Statistical learning: classification and cross-validation

MACS 30500 University of Chicago

Should I Have a Cookie?

Interpreting a decision tree

A more complex tree

A more complexier tree

Benefits/drawbacks to decision trees

  • Easy to explain
  • Easy to interpret/visualize
  • Good for qualitative predictors
  • Lower accuracy rates
  • Non-robust

Random forests

Sampling with replacement

(numbers <- seq(from = 1, to = 10))
##  [1]  1  2  3  4  5  6  7  8  9 10
# sample without replacement
rerun(5, sample(numbers, replace = FALSE))
## [[1]]
##  [1]  6  4  1 10  9  7  5  2  3  8
## 
## [[2]]
##  [1]  9  8  7  1  4  5  3 10  6  2
## 
## [[3]]
##  [1]  2  4  7  1 10  8  3  5  6  9
## 
## [[4]]
##  [1]  4  6  3  1  7 10  5  8  2  9
## 
## [[5]]
##  [1]  8  6  7  5  9  3 10  1  4  2
# sample with replacement
rerun(5, sample(numbers, replace = TRUE))
## [[1]]
##  [1]  5  4  2  3 10  1  5  3  5  8
## 
## [[2]]
##  [1]  8 10  2  9  9 10  9  1  5  3
## 
## [[3]]
##  [1]  6  3  1 10  4  3  7  8  1  7
## 
## [[4]]
##  [1]  2  3  7  2  4 10  8  5  4  8
## 
## [[5]]
##  [1]  3  3  4  9  9  1 10  2  1  7

Random forests

  • Bootstrapping
  • Reduces variance
  • Bagging
  • Random forest
    • Reliability

Estimating statistical models using caret

  • Not part of tidyverse (yet)
  • Aggregator of hundreds of statistical learning algorithms
  • Provides a single unified interface to disparate range of functions
    • Similar to scikit-learn for Python

train()

library(caret)

titanic_clean <- titanic %>%
  filter(!is.na(Survived), !is.na(Age))

caret_glm <- train(Survived ~ Age, data = titanic_clean,
                   method = "glm",
                   family = binomial,
                   trControl = trainControl(method = "none"))
summary(caret_glm)
## 
## Call:
## NULL
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -1.1488  -1.0361  -0.9544   1.3159   1.5908  
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)  
## (Intercept) -0.05672    0.17358  -0.327   0.7438  
## Age         -0.01096    0.00533  -2.057   0.0397 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 964.52  on 713  degrees of freedom
## Residual deviance: 960.23  on 712  degrees of freedom
## AIC: 964.23
## 
## Number of Fisher Scoring iterations: 4

Estimating a random forest

age_sex_rf <- train(Survived ~ Age + Sex, data = titanic_rf_data,
                   method = "rf",
                   ntree = 200,
                   trControl = trainControl(method = "oob"))
## note: only 1 unique complexity parameters in default grid. Truncating the grid to 1 .
age_sex_rf
## Random Forest 
## 
## 714 samples
##   2 predictor
##   2 classes: 'Died', 'Survived' 
## 
## No pre-processing
## Resampling results:
## 
##   Accuracy   Kappa    
##   0.7492997  0.4725343
## 
## Tuning parameter 'mtry' was held constant at a value of 2

Structure of train() object

## List of 24
##  $ method      : chr "rf"
##  $ modelInfo   :List of 15
##  $ modelType   : chr "Classification"
##  $ results     :'data.frame':    1 obs. of  3 variables:
##  $ pred        : NULL
##  $ bestTune    :'data.frame':    1 obs. of  1 variable:
##  $ call        : language train.formula(form = Survived ~ Age + Sex, data = titanic_rf_data,      method = "rf", ntree = 200, trControl = t| __truncated__
##  $ dots        :List of 1
##  $ metric      : chr "Accuracy"
##  $ control     :List of 26
##  $ finalModel  :List of 23
##   ..- attr(*, "class")= chr "randomForest"
##  $ preProcess  : NULL
##  $ trainingData:Classes 'tbl_df', 'tbl' and 'data.frame':    714 obs. of  3 variables:
##  $ resample    : NULL
##  $ resampledCM : NULL
##  $ perfNames   : chr [1:2] "Accuracy" "Kappa"
##  $ maximize    : logi TRUE
##  $ yLimits     : NULL
##  $ times       :List of 3
##  $ levels      : atomic [1:2] Died Survived
##   ..- attr(*, "ordered")= logi FALSE
##  $ terms       :Classes 'terms', 'formula'  language Survived ~ Age + Sex
##   .. ..- attr(*, "variables")= language list(Survived, Age, Sex)
##   .. ..- attr(*, "factors")= int [1:3, 1:2] 0 1 0 0 0 1
##   .. .. ..- attr(*, "dimnames")=List of 2
##   .. ..- attr(*, "term.labels")= chr [1:2] "Age" "Sex"
##   .. ..- attr(*, "order")= int [1:2] 1 1
##   .. ..- attr(*, "intercept")= int 1
##   .. ..- attr(*, "response")= int 1
##   .. ..- attr(*, ".Environment")=<environment: R_GlobalEnv> 
##   .. ..- attr(*, "predvars")= language list(Survived, Age, Sex)
##   .. ..- attr(*, "dataClasses")= Named chr [1:3] "factor" "numeric" "factor"
##   .. .. ..- attr(*, "names")= chr [1:3] "Survived" "Age" "Sex"
##  $ coefnames   : chr [1:2] "Age" "Sexmale"
##  $ contrasts   :List of 1
##  $ xlevels     :List of 1
##  - attr(*, "class")= chr [1:2] "train" "train.formula"

Model statistics

## 
## Call:
##  randomForest(x = x, y = y, ntree = 200, mtry = param$mtry) 
##                Type of random forest: classification
##                      Number of trees: 200
## No. of variables tried at each split: 2
## 
##         OOB estimate of  error rate: 24.79%
## Confusion matrix:
##          Died Survived class.error
## Died      352       72   0.1698113
## Survived  105      185   0.3620690

Results of a single tree

##     left daughter right daughter split var split point status prediction
## 1               2              3   Sexmale       0.500      1       <NA>
## 2               4              5       Age       3.500      1       <NA>
## 3               6              7       Age       5.500      1       <NA>
## 4               8              9       Age       1.375      1       <NA>
## 5              10             11       Age      32.250      1       <NA>
## 6              12             13       Age       0.960      1       <NA>
## 7              14             15       Age      77.000      1       <NA>
## 8               0              0      <NA>       0.000     -1   Survived
## 9              16             17       Age       2.500      1       <NA>
## 10             18             19       Age      24.500      1       <NA>
## 11             20             21       Age      38.500      1       <NA>
## 12              0              0      <NA>       0.000     -1   Survived
## 13             22             23       Age       2.000      1       <NA>
## 14             24             25       Age      50.500      1       <NA>
## 15              0              0      <NA>       0.000     -1   Survived
## 16              0              0      <NA>       0.000     -1       Died
## 17              0              0      <NA>       0.000     -1       Died
## 18             26             27       Age       5.500      1       <NA>
## 19             28             29       Age      25.500      1       <NA>
## 20              0              0      <NA>       0.000     -1   Survived
## 21             30             31       Age      50.500      1       <NA>
## 22              0              0      <NA>       0.000     -1   Survived
## 23             32             33       Age       3.500      1       <NA>
## 24             34             35       Age      47.500      1       <NA>
## 25             36             37       Age      56.500      1       <NA>
## 26              0              0      <NA>       0.000     -1   Survived
## 27             38             39       Age      12.000      1       <NA>
## 28              0              0      <NA>       0.000     -1       Died
## 29             40             41       Age      30.250      1       <NA>
## 30             42             43       Age      49.500      1       <NA>
## 31              0              0      <NA>       0.000     -1   Survived
## 32              0              0      <NA>       0.000     -1   Survived
## 33              0              0      <NA>       0.000     -1   Survived
## 34             44             45       Age      45.250      1       <NA>
## 35             46             47       Age      48.500      1       <NA>
## 36             48             49       Age      55.500      1       <NA>
## 37              0              0      <NA>       0.000     -1       Died
## 38             50             51       Age       8.500      1       <NA>
## 39             52             53       Age      19.500      1       <NA>
## 40             54             55       Age      28.500      1       <NA>
## 41             56             57       Age      31.500      1       <NA>
## 42             58             59       Age      41.500      1       <NA>
## 43              0              0      <NA>       0.000     -1       Died
## 44             60             61       Age      43.500      1       <NA>
## 45              0              0      <NA>       0.000     -1       Died
## 46              0              0      <NA>       0.000     -1   Survived
## 47             62             63       Age      49.500      1       <NA>
## 48              0              0      <NA>       0.000     -1       Died
## 49              0              0      <NA>       0.000     -1       Died
## 50             64             65       Age       6.500      1       <NA>
## 51              0              0      <NA>       0.000     -1       Died
## 52             66             67       Age      18.500      1       <NA>
## 53             68             69       Age      20.500      1       <NA>
## 54             70             71       Age      27.500      1       <NA>
## 55             72             73       Age      29.500      1       <NA>
## 56              0              0      <NA>       0.000     -1       Died
## 57              0              0      <NA>       0.000     -1   Survived
## 58             74             75       Age      39.500      1       <NA>
## 59             76             77       Age      46.000      1       <NA>
## 60             78             79       Age      32.250      1       <NA>
## 61             80             81       Age      44.500      1       <NA>
## 62              0              0      <NA>       0.000     -1       Died
## 63              0              0      <NA>       0.000     -1       Died
## 64              0              0      <NA>       0.000     -1       Died
## 65             82             83       Age       7.500      1       <NA>
## 66             84             85       Age      13.500      1       <NA>
## 67              0              0      <NA>       0.000     -1   Survived
## 68              0              0      <NA>       0.000     -1       Died
## 69             86             87       Age      22.500      1       <NA>
## 70             88             89       Age      26.500      1       <NA>
## 71              0              0      <NA>       0.000     -1       Died
## 72              0              0      <NA>       0.000     -1   Survived
## 73              0              0      <NA>       0.000     -1   Survived
## 74              0              0      <NA>       0.000     -1   Survived
## 75             90             91       Age      40.500      1       <NA>
## 76              0              0      <NA>       0.000     -1   Survived
## 77             92             93       Age      47.500      1       <NA>
## 78             94             95       Age      30.500      1       <NA>
## 79             96             97       Age      35.500      1       <NA>
## 80              0              0      <NA>       0.000     -1       Died
## 81              0              0      <NA>       0.000     -1   Survived
## 82              0              0      <NA>       0.000     -1   Survived
## 83              0              0      <NA>       0.000     -1   Survived
## 84              0              0      <NA>       0.000     -1   Survived
## 85             98             99       Age      14.500      1       <NA>
## 86            100            101       Age      21.500      1       <NA>
## 87            102            103       Age      23.500      1       <NA>
## 88              0              0      <NA>       0.000     -1       Died
## 89              0              0      <NA>       0.000     -1   Survived
## 90              0              0      <NA>       0.000     -1   Survived
## 91              0              0      <NA>       0.000     -1   Survived
## 92              0              0      <NA>       0.000     -1       Died
## 93              0              0      <NA>       0.000     -1   Survived
## 94            104            105       Age      24.500      1       <NA>
## 95            106            107       Age      31.500      1       <NA>
## 96              0              0      <NA>       0.000     -1       Died
## 97            108            109       Age      41.500      1       <NA>
## 98              0              0      <NA>       0.000     -1   Survived
## 99            110            111       Age      16.500      1       <NA>
## 100             0              0      <NA>       0.000     -1   Survived
## 101             0              0      <NA>       0.000     -1   Survived
## 102             0              0      <NA>       0.000     -1   Survived
## 103             0              0      <NA>       0.000     -1   Survived
## 104           112            113       Age      21.500      1       <NA>
## 105           114            115       Age      27.500      1       <NA>
## 106             0              0      <NA>       0.000     -1       Died
## 107             0              0      <NA>       0.000     -1       Died
## 108           116            117       Age      40.250      1       <NA>
## 109           118            119       Age      42.500      1       <NA>
## 110           120            121       Age      15.500      1       <NA>
## 111           122            123       Age      17.500      1       <NA>
## 112           124            125       Age       9.500      1       <NA>
## 113           126            127       Age      22.500      1       <NA>
## 114           128            129       Age      26.500      1       <NA>
## 115           130            131       Age      28.750      1       <NA>
## 116           132            133       Age      38.500      1       <NA>
## 117             0              0      <NA>       0.000     -1       Died
## 118             0              0      <NA>       0.000     -1       Died
## 119             0              0      <NA>       0.000     -1       Died
## 120             0              0      <NA>       0.000     -1   Survived
## 121             0              0      <NA>       0.000     -1   Survived
## 122             0              0      <NA>       0.000     -1   Survived
## 123             0              0      <NA>       0.000     -1   Survived
## 124           134            135       Age       7.500      1       <NA>
## 125           136            137       Age      15.500      1       <NA>
## 126             0              0      <NA>       0.000     -1       Died
## 127           138            139       Age      23.250      1       <NA>
## 128           140            141       Age      25.500      1       <NA>
## 129             0              0      <NA>       0.000     -1       Died
## 130           142            143       Age      28.250      1       <NA>
## 131           144            145       Age      29.500      1       <NA>
## 132           146            147       Age      36.250      1       <NA>
## 133           148            149       Age      39.500      1       <NA>
## 134             0              0      <NA>       0.000     -1       Died
## 135           150            151       Age       8.500      1       <NA>
## 136             0              0      <NA>       0.000     -1       Died
## 137           152            153       Age      20.250      1       <NA>
## 138             0              0      <NA>       0.000     -1       Died
## 139           154            155       Age      23.750      1       <NA>
## 140             0              0      <NA>       0.000     -1       Died
## 141             0              0      <NA>       0.000     -1       Died
## 142             0              0      <NA>       0.000     -1       Died
## 143             0              0      <NA>       0.000     -1       Died
## 144             0              0      <NA>       0.000     -1       Died
## 145             0              0      <NA>       0.000     -1       Died
## 146             0              0      <NA>       0.000     -1       Died
## 147           156            157       Age      37.500      1       <NA>
## 148             0              0      <NA>       0.000     -1       Died
## 149             0              0      <NA>       0.000     -1       Died
## 150             0              0      <NA>       0.000     -1       Died
## 151             0              0      <NA>       0.000     -1   Survived
## 152           158            159       Age      18.500      1       <NA>
## 153           160            161       Age      20.750      1       <NA>
## 154             0              0      <NA>       0.000     -1       Died
## 155             0              0      <NA>       0.000     -1       Died
## 156             0              0      <NA>       0.000     -1       Died
## 157             0              0      <NA>       0.000     -1       Died
## 158           162            163       Age      17.500      1       <NA>
## 159           164            165       Age      19.500      1       <NA>
## 160             0              0      <NA>       0.000     -1       Died
## 161             0              0      <NA>       0.000     -1       Died
## 162           166            167       Age      16.500      1       <NA>
## 163             0              0      <NA>       0.000     -1       Died
## 164             0              0      <NA>       0.000     -1       Died
## 165             0              0      <NA>       0.000     -1       Died
## 166             0              0      <NA>       0.000     -1       Died
## 167             0              0      <NA>       0.000     -1       Died

Variable importance

Exercise: depression and voting

Resampling methods

  • Evaluating model fit/predictive power
  • How to avoid overfitting the data

Validation set

  • Randomly split data into two distinct sets
    • Training set
    • Test set
  • Train model on training set
  • Evaluate fit on test set

Regression

Mean squared error

\[MSE = \frac{1}{n} \sum_{i = 1}^{n}{(y_i - \hat{f}(x_i))^2}\]

  • \(y_i =\) the observed response value for the \(i\)th observation
  • \(\hat{f}(x_i) =\) the predicted response value for the \(i\)th observation given by \(\hat{f}\)
  • \(n =\) the total number of observations

Split data

set.seed(1234)

auto_split <- resample_partition(Auto, c(test = 0.5, train = 0.5))
auto_train <- as_tibble(auto_split$train)
auto_test <- as_tibble(auto_split$test)

Train model

auto_lm <- glm(mpg ~ horsepower, data = auto_train)
summary(auto_lm)
## 
## Call:
## glm(formula = mpg ~ horsepower, data = auto_train)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -12.892   -2.864   -0.545    2.793   13.298  
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)    
## (Intercept) 38.005404   0.921129   41.26   <2e-16 ***
## horsepower  -0.140459   0.007968  -17.63   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for gaussian family taken to be 20.48452)
## 
##     Null deviance: 10359.4  on 196  degrees of freedom
## Residual deviance:  3994.5  on 195  degrees of freedom
## AIC: 1157.9
## 
## Number of Fisher Scoring iterations: 2

Test model

mse(auto_lm, auto_test)
## [1] 28.57255

Compare models

Classification

survive_age_woman_x <- glm(Survived ~ Age * Sex, data = titanic,
                           family = binomial)
summary(survive_age_woman_x)
## 
## Call:
## glm(formula = Survived ~ Age * Sex, family = binomial, data = titanic)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -1.9401  -0.7136  -0.5883   0.7626   2.2455  
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)   
## (Intercept)  0.59380    0.31032   1.913  0.05569 . 
## Age          0.01970    0.01057   1.863  0.06240 . 
## Sexmale     -1.31775    0.40842  -3.226  0.00125 **
## Age:Sexmale -0.04112    0.01355  -3.034  0.00241 **
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 964.52  on 713  degrees of freedom
## Residual deviance: 740.40  on 710  degrees of freedom
##   (177 observations deleted due to missingness)
## AIC: 748.4
## 
## Number of Fisher Scoring iterations: 4

Test error rate

titanic_split <- resample_partition(titanic, c(test = 0.3, train = 0.7))
map(titanic_split, dim)
## $test
## [1] 267  12
## 
## $train
## [1] 624  12
train_model <- glm(Survived ~ Age * Sex, data = titanic_split$train,
                   family = binomial)
summary(train_model)
## 
## Call:
## glm(formula = Survived ~ Age * Sex, family = binomial, data = titanic_split$train)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -2.0901  -0.6520  -0.6507   0.7352   1.8223  
## 
## Coefficients:
##             Estimate Std. Error z value Pr(>|z|)    
## (Intercept)  0.37929    0.35668   1.063 0.287607    
## Age          0.02957    0.01287   2.297 0.021612 *  
## Sexmale     -1.82897    0.49728  -3.678 0.000235 ***
## Age:Sexmale -0.02931    0.01642  -1.785 0.074193 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 667.64  on 491  degrees of freedom
## Residual deviance: 500.14  on 488  degrees of freedom
##   (132 observations deleted due to missingness)
## AIC: 508.14
## 
## Number of Fisher Scoring iterations: 4
x_test_accuracy <- titanic_split$test %>%
  as_tibble() %>%
  add_predictions(train_model) %>%
  mutate(pred = logit2prob(pred),
         pred = as.numeric(pred > .5))

mean(x_test_accuracy$Survived != x_test_accuracy$pred, na.rm = TRUE)
## [1] 0.2387387

Drawbacks to validation sets

Leave-one-out cross-validation

\[CV_{(n)} = \frac{1}{n} \sum_{i = 1}^{n}{MSE_i}\]

  • Extension of validation set to repeatedly split data and average results
  • Minimizes bias of estimated error rate
  • Low variance
  • Highly computationally intensive

LOOCV in linear regression

(loocv_data <- crossv_kfold(Auto, k = nrow(Auto)))
## # A tibble: 392 x 3
##             train           test   .id
##            <list>         <list> <chr>
##  1 <S3: resample> <S3: resample>   001
##  2 <S3: resample> <S3: resample>   002
##  3 <S3: resample> <S3: resample>   003
##  4 <S3: resample> <S3: resample>   004
##  5 <S3: resample> <S3: resample>   005
##  6 <S3: resample> <S3: resample>   006
##  7 <S3: resample> <S3: resample>   007
##  8 <S3: resample> <S3: resample>   008
##  9 <S3: resample> <S3: resample>   009
## 10 <S3: resample> <S3: resample>   010
## # ... with 382 more rows
(loocv_models <- map(loocv_data$train, ~ lm(mpg ~ horsepower, data = .)))
## $`1`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8759      -0.1574  
## 
## 
## $`2`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9329      -0.1578  
## 
## 
## $`3`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9261      -0.1578  
## 
## 
## $`4`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8891      -0.1575  
## 
## 
## $`5`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8551      -0.1573  
## 
## 
## $`6`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9366      -0.1578  
## 
## 
## $`7`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.7980      -0.1568  
## 
## 
## $`8`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9336      -0.1578  
## 
## 
## $`9`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8643      -0.1574  
## 
## 
## $`10`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9345      -0.1578  
## 
## 
## $`11`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9170      -0.1577  
## 
## 
## $`12`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9351      -0.1578  
## 
## 
## $`13`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9322      -0.1578  
## 
## 
## $`14`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9400      -0.1578  
## 
## 
## $`15`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9383      -0.1579  
## 
## 
## $`16`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8884      -0.1575  
## 
## 
## $`17`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9319      -0.1579  
## 
## 
## $`18`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8823      -0.1575  
## 
## 
## $`19`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9226      -0.1576  
## 
## 
## $`20`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8868      -0.1575  
## 
## 
## $`21`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.968       -0.158  
## 
## 
## $`22`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.963       -0.158  
## 
## 
## $`23`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9532      -0.1581  
## 
## 
## $`24`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9523      -0.1579  
## 
## 
## $`25`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.957       -0.158  
## 
## 
## $`26`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.946       -0.158  
## 
## 
## $`27`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.971       -0.158  
## 
## 
## $`28`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8616      -0.1573  
## 
## 
## $`29`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9137      -0.1577  
## 
## 
## $`30`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9230      -0.1577  
## 
## 
## $`31`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9307      -0.1578  
## 
## 
## $`32`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9429      -0.1578  
## 
## 
## $`33`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9234      -0.1578  
## 
## 
## $`34`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     40.0854      -0.1595  
## 
## 
## $`35`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9453      -0.1579  
## 
## 
## $`36`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9447      -0.1579  
## 
## 
## $`37`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.962       -0.158  
## 
## 
## $`38`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9247      -0.1578  
## 
## 
## $`39`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9445      -0.1578  
## 
## 
## $`40`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9505      -0.1579  
## 
## 
## $`41`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9362      -0.1579  
## 
## 
## $`42`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.976       -0.158  
## 
## 
## $`43`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9084      -0.1577  
## 
## 
## $`44`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8647      -0.1573  
## 
## 
## $`45`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9215      -0.1576  
## 
## 
## $`46`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9680      -0.1581  
## 
## 
## $`47`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9682      -0.1581  
## 
## 
## $`48`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9220      -0.1578  
## 
## 
## $`49`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     40.0504      -0.1586  
## 
## 
## $`50`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9148      -0.1577  
## 
## 
## $`51`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9330      -0.1578  
## 
## 
## $`52`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9320      -0.1578  
## 
## 
## $`53`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9257      -0.1578  
## 
## 
## $`54`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9532      -0.1579  
## 
## 
## $`55`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9477      -0.1578  
## 
## 
## $`56`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9431      -0.1579  
## 
## 
## $`57`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     40.0019      -0.1586  
## 
## 
## $`58`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     40.0279      -0.1589  
## 
## 
## $`59`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9603      -0.1581  
## 
## 
## $`60`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.949       -0.158  
## 
## 
## $`61`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9241      -0.1578  
## 
## 
## $`62`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9464      -0.1579  
## 
## 
## $`63`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9430      -0.1578  
## 
## 
## $`64`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9093      -0.1577  
## 
## 
## $`65`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9205      -0.1577  
## 
## 
## $`66`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9035      -0.1576  
## 
## 
## $`67`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8583      -0.1573  
## 
## 
## $`68`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9293      -0.1578  
## 
## 
## $`69`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8952      -0.1576  
## 
## 
## $`70`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9520      -0.1579  
## 
## 
## $`71`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9566      -0.1579  
## 
## 
## $`72`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9347      -0.1578  
## 
## 
## $`73`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9375      -0.1578  
## 
## 
## $`74`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9536      -0.1581  
## 
## 
## $`75`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8909      -0.1575  
## 
## 
## $`76`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     40.0066      -0.1583  
## 
## 
## $`77`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.966       -0.158  
## 
## 
## $`78`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9433      -0.1579  
## 
## 
## $`79`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9326      -0.1578  
## 
## 
## $`80`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8918      -0.1576  
## 
## 
## $`81`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9076      -0.1577  
## 
## 
## $`82`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9346      -0.1578  
## 
## 
## $`83`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9150      -0.1577  
## 
## 
## $`84`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9387      -0.1579  
## 
## 
## $`85`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8433      -0.1572  
## 
## 
## $`86`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9680      -0.1582  
## 
## 
## $`87`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8732      -0.1574  
## 
## 
## $`88`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9817      -0.1581  
## 
## 
## $`89`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9216      -0.1577  
## 
## 
## $`90`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9346      -0.1578  
## 
## 
## $`91`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9402      -0.1578  
## 
## 
## $`92`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9188      -0.1577  
## 
## 
## $`93`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9219      -0.1578  
## 
## 
## $`94`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9361      -0.1578  
## 
## 
## $`95`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9283      -0.1578  
## 
## 
## $`96`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     40.0504      -0.1586  
## 
## 
## $`97`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9682      -0.1581  
## 
## 
## $`98`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9583      -0.1581  
## 
## 
## $`99`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8956      -0.1577  
## 
## 
## $`100`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9165      -0.1576  
## 
## 
## $`101`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9230      -0.1577  
## 
## 
## $`102`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8700      -0.1574  
## 
## 
## $`103`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8627      -0.1573  
## 
## 
## $`104`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     40.0605      -0.1592  
## 
## 
## $`105`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9532      -0.1579  
## 
## 
## $`106`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9035      -0.1576  
## 
## 
## $`107`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9287      -0.1577  
## 
## 
## $`108`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9306      -0.1578  
## 
## 
## $`109`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9444      -0.1579  
## 
## 
## $`110`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9461      -0.1579  
## 
## 
## $`111`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9019      -0.1576  
## 
## 
## $`112`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9112      -0.1577  
## 
## 
## $`113`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9429      -0.1579  
## 
## 
## $`114`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9321      -0.1579  
## 
## 
## $`115`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9438      -0.1578  
## 
## 
## $`116`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8735      -0.1574  
## 
## 
## $`117`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9527      -0.1579  
## 
## 
## $`118`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9419      -0.1579  
## 
## 
## $`119`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9226      -0.1576  
## 
## 
## $`120`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9271      -0.1577  
## 
## 
## $`121`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8916      -0.1576  
## 
## 
## $`122`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     40.0000      -0.1586  
## 
## 
## $`123`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     40.0085      -0.1587  
## 
## 
## $`124`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9540      -0.1578  
## 
## 
## $`125`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9411      -0.1579  
## 
## 
## $`126`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9194      -0.1577  
## 
## 
## $`127`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     40.1919      -0.1606  
## 
## 
## $`128`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     40.0240      -0.1585  
## 
## 
## $`129`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9389      -0.1578  
## 
## 
## $`130`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.945       -0.158  
## 
## 
## $`131`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9360      -0.1578  
## 
## 
## $`132`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9524      -0.1579  
## 
## 
## $`133`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8775      -0.1575  
## 
## 
## $`134`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9461      -0.1578  
## 
## 
## $`135`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9088      -0.1577  
## 
## 
## $`136`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9310      -0.1578  
## 
## 
## $`137`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9157      -0.1576  
## 
## 
## $`138`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9188      -0.1577  
## 
## 
## $`139`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9429      -0.1579  
## 
## 
## $`140`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9804      -0.1581  
## 
## 
## $`141`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9491      -0.1578  
## 
## 
## $`142`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8801      -0.1576  
## 
## 
## $`143`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.970       -0.158  
## 
## 
## $`144`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9363      -0.1578  
## 
## 
## $`145`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9229      -0.1577  
## 
## 
## $`146`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9329      -0.1578  
## 
## 
## $`147`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8550      -0.1573  
## 
## 
## $`148`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9285      -0.1578  
## 
## 
## $`149`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9482      -0.1579  
## 
## 
## $`150`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.952       -0.158  
## 
## 
## $`151`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9287      -0.1577  
## 
## 
## $`152`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9235      -0.1576  
## 
## 
## $`153`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9422      -0.1578  
## 
## 
## $`154`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.971       -0.158  
## 
## 
## $`155`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9220      -0.1577  
## 
## 
## $`156`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9699      -0.1581  
## 
## 
## $`157`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9321      -0.1578  
## 
## 
## $`158`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9612      -0.1579  
## 
## 
## $`159`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9998      -0.1583  
## 
## 
## $`160`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9356      -0.1578  
## 
## 
## $`161`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9420      -0.1578  
## 
## 
## $`162`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.969       -0.158  
## 
## 
## $`163`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9566      -0.1579  
## 
## 
## $`164`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9441      -0.1579  
## 
## 
## $`165`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9441      -0.1579  
## 
## 
## $`166`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9367      -0.1579  
## 
## 
## $`167`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9348      -0.1578  
## 
## 
## $`168`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9618      -0.1579  
## 
## 
## $`169`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9133      -0.1577  
## 
## 
## $`170`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8731      -0.1574  
## 
## 
## $`171`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.966       -0.158  
## 
## 
## $`172`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     40.1249      -0.1599  
## 
## 
## $`173`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9173      -0.1576  
## 
## 
## $`174`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     40.1249      -0.1599  
## 
## 
## $`175`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9120      -0.1577  
## 
## 
## $`176`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9484      -0.1579  
## 
## 
## $`177`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.7873      -0.1567  
## 
## 
## $`178`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9427      -0.1579  
## 
## 
## $`179`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8929      -0.1576  
## 
## 
## $`180`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9268      -0.1576  
## 
## 
## $`181`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8772      -0.1575  
## 
## 
## $`182`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.968       -0.158  
## 
## 
## $`183`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.975       -0.158  
## 
## 
## $`184`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9362      -0.1578  
## 
## 
## $`185`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9374      -0.1578  
## 
## 
## $`186`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9230      -0.1577  
## 
## 
## $`187`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.7826      -0.1567  
## 
## 
## $`188`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     40.0346      -0.1586  
## 
## 
## $`189`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9445      -0.1578  
## 
## 
## $`190`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     40.0783      -0.1594  
## 
## 
## $`191`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9213      -0.1578  
## 
## 
## $`192`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.964       -0.158  
## 
## 
## $`193`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9210      -0.1576  
## 
## 
## $`194`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9399      -0.1579  
## 
## 
## $`195`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9085      -0.1577  
## 
## 
## $`196`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9287      -0.1577  
## 
## 
## $`197`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9300      -0.1578  
## 
## 
## $`198`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9497      -0.1579  
## 
## 
## $`199`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9349      -0.1578  
## 
## 
## $`200`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.975       -0.158  
## 
## 
## $`201`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9375      -0.1578  
## 
## 
## $`202`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9324      -0.1578  
## 
## 
## $`203`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9241      -0.1578  
## 
## 
## $`204`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8866      -0.1575  
## 
## 
## $`205`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9309      -0.1577  
## 
## 
## $`206`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9490      -0.1579  
## 
## 
## $`207`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9354      -0.1578  
## 
## 
## $`208`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9618      -0.1579  
## 
## 
## $`209`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9412      -0.1579  
## 
## 
## $`210`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9566      -0.1579  
## 
## 
## $`211`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9399      -0.1578  
## 
## 
## $`212`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9673      -0.1581  
## 
## 
## $`213`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.960       -0.158  
## 
## 
## $`214`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9548      -0.1579  
## 
## 
## $`215`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9816      -0.1581  
## 
## 
## $`216`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9299      -0.1578  
## 
## 
## $`217`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9147      -0.1577  
## 
## 
## $`218`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8975      -0.1576  
## 
## 
## $`219`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9279      -0.1578  
## 
## 
## $`220`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9578      -0.1579  
## 
## 
## $`221`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9379      -0.1579  
## 
## 
## $`222`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9450      -0.1579  
## 
## 
## $`223`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     40.0071      -0.1586  
## 
## 
## $`224`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9262      -0.1578  
## 
## 
## $`225`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9341      -0.1579  
## 
## 
## $`226`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.969       -0.158  
## 
## 
## $`227`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9326      -0.1579  
## 
## 
## $`228`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.7714      -0.1567  
## 
## 
## $`229`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9347      -0.1577  
## 
## 
## $`230`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.967       -0.158  
## 
## 
## $`231`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9106      -0.1577  
## 
## 
## $`232`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9427      -0.1579  
## 
## 
## $`233`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9165      -0.1576  
## 
## 
## $`234`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9344      -0.1578  
## 
## 
## $`235`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9842      -0.1584  
## 
## 
## $`236`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     40.0240      -0.1585  
## 
## 
## $`237`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9379      -0.1579  
## 
## 
## $`238`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9247      -0.1577  
## 
## 
## $`239`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8277      -0.1571  
## 
## 
## $`240`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9344      -0.1578  
## 
## 
## $`241`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9086      -0.1577  
## 
## 
## $`242`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9633      -0.1579  
## 
## 
## $`243`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9339      -0.1578  
## 
## 
## $`244`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.931       -0.158  
## 
## 
## $`245`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9356      -0.1578  
## 
## 
## $`246`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.954       -0.158  
## 
## 
## $`247`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9261      -0.1577  
## 
## 
## $`248`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.956       -0.158  
## 
## 
## $`249`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9349      -0.1578  
## 
## 
## $`250`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9427      -0.1578  
## 
## 
## $`251`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9491      -0.1578  
## 
## 
## $`252`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9412      -0.1579  
## 
## 
## $`253`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.946       -0.158  
## 
## 
## $`254`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9366      -0.1578  
## 
## 
## $`255`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8784      -0.1574  
## 
## 
## $`256`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9084      -0.1577  
## 
## 
## $`257`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9474      -0.1579  
## 
## 
## $`258`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9630      -0.1582  
## 
## 
## $`259`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9287      -0.1577  
## 
## 
## $`260`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8733      -0.1574  
## 
## 
## $`261`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9148      -0.1577  
## 
## 
## $`262`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     40.1008      -0.1596  
## 
## 
## $`263`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9261      -0.1577  
## 
## 
## $`264`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.957       -0.158  
## 
## 
## $`265`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9230      -0.1578  
## 
## 
## $`266`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9448      -0.1579  
## 
## 
## $`267`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9540      -0.1578  
## 
## 
## $`268`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9340      -0.1578  
## 
## 
## $`269`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8547      -0.1573  
## 
## 
## $`270`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8864      -0.1575  
## 
## 
## $`271`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9360      -0.1578  
## 
## 
## $`272`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9478      -0.1579  
## 
## 
## $`273`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9294      -0.1578  
## 
## 
## $`274`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9532      -0.1579  
## 
## 
## $`275`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9392      -0.1578  
## 
## 
## $`276`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9171      -0.1575  
## 
## 
## $`277`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.953       -0.158  
## 
## 
## $`278`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9352      -0.1578  
## 
## 
## $`279`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8312      -0.1572  
## 
## 
## $`280`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9491      -0.1578  
## 
## 
## $`281`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8943      -0.1576  
## 
## 
## $`282`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9433      -0.1578  
## 
## 
## $`283`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9368      -0.1579  
## 
## 
## $`284`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9599      -0.1579  
## 
## 
## $`285`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.945       -0.158  
## 
## 
## $`286`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9064      -0.1578  
## 
## 
## $`287`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9688      -0.1585  
## 
## 
## $`288`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9338      -0.1579  
## 
## 
## $`289`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9804      -0.1581  
## 
## 
## $`290`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9758      -0.1583  
## 
## 
## $`291`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9743      -0.1581  
## 
## 
## $`292`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9247      -0.1577  
## 
## 
## $`293`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     40.0068      -0.1586  
## 
## 
## $`294`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8807      -0.1576  
## 
## 
## $`295`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9532      -0.1579  
## 
## 
## $`296`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.965       -0.158  
## 
## 
## $`297`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9205      -0.1577  
## 
## 
## $`298`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.969       -0.158  
## 
## 
## $`299`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8550      -0.1573  
## 
## 
## $`300`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.7941      -0.1568  
## 
## 
## $`301`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9241      -0.1578  
## 
## 
## $`302`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9345      -0.1577  
## 
## 
## $`303`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9498      -0.1579  
## 
## 
## $`304`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9346      -0.1578  
## 
## 
## $`305`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9383      -0.1579  
## 
## 
## $`306`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9498      -0.1579  
## 
## 
## $`307`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9391      -0.1579  
## 
## 
## $`308`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9532      -0.1581  
## 
## 
## $`309`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9454      -0.1579  
## 
## 
## $`310`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9422      -0.1579  
## 
## 
## $`311`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9025      -0.1577  
## 
## 
## $`312`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9182      -0.1577  
## 
## 
## $`313`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9570      -0.1579  
## 
## 
## $`314`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9287      -0.1578  
## 
## 
## $`315`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9410      -0.1578  
## 
## 
## $`316`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9306      -0.1578  
## 
## 
## $`317`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9299      -0.1578  
## 
## 
## $`318`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9173      -0.1576  
## 
## 
## $`319`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9993      -0.1583  
## 
## 
## $`320`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9273      -0.1577  
## 
## 
## $`321`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9401      -0.1579  
## 
## 
## $`322`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.944       -0.158  
## 
## 
## $`323`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9689      -0.1581  
## 
## 
## $`324`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9146      -0.1577  
## 
## 
## $`325`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8451      -0.1572  
## 
## 
## $`326`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9249      -0.1578  
## 
## 
## $`327`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9063      -0.1577  
## 
## 
## $`328`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9230      -0.1577  
## 
## 
## $`329`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9173      -0.1576  
## 
## 
## $`330`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9173      -0.1576  
## 
## 
## $`331`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9764      -0.1582  
## 
## 
## $`332`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9391      -0.1579  
## 
## 
## $`333`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9913      -0.1582  
## 
## 
## $`334`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9363      -0.1578  
## 
## 
## $`335`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9322      -0.1577  
## 
## 
## $`336`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9285      -0.1577  
## 
## 
## $`337`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9901      -0.1581  
## 
## 
## $`338`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9399      -0.1579  
## 
## 
## $`339`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.961       -0.158  
## 
## 
## $`340`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9287      -0.1577  
## 
## 
## $`341`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9331      -0.1578  
## 
## 
## $`342`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9727      -0.1581  
## 
## 
## $`343`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9293      -0.1578  
## 
## 
## $`344`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9394      -0.1579  
## 
## 
## $`345`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9336      -0.1579  
## 
## 
## $`346`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9349      -0.1578  
## 
## 
## $`347`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9633      -0.1579  
## 
## 
## $`348`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9344      -0.1578  
## 
## 
## $`349`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9380      -0.1579  
## 
## 
## $`350`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9200      -0.1577  
## 
## 
## $`351`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8908      -0.1575  
## 
## 
## $`352`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.964       -0.158  
## 
## 
## $`353`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9916      -0.1585  
## 
## 
## $`354`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9301      -0.1577  
## 
## 
## $`355`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9292      -0.1578  
## 
## 
## $`356`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9767      -0.1581  
## 
## 
## $`357`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.961       -0.158  
## 
## 
## $`358`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9349      -0.1578  
## 
## 
## $`359`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9152      -0.1577  
## 
## 
## $`360`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9250      -0.1578  
## 
## 
## $`361`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8863      -0.1576  
## 
## 
## $`362`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9287      -0.1577  
## 
## 
## $`363`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9258      -0.1578  
## 
## 
## $`364`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.7933      -0.1569  
## 
## 
## $`365`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9287      -0.1577  
## 
## 
## $`366`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9337      -0.1578  
## 
## 
## $`367`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9461      -0.1578  
## 
## 
## $`368`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9059      -0.1574  
## 
## 
## $`369`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9398      -0.1578  
## 
## 
## $`370`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9306      -0.1578  
## 
## 
## $`371`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9195      -0.1576  
## 
## 
## $`372`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9326      -0.1578  
## 
## 
## $`373`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9303      -0.1578  
## 
## 
## $`374`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9429      -0.1579  
## 
## 
## $`375`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9230      -0.1577  
## 
## 
## $`376`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8660      -0.1575  
## 
## 
## $`377`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.960       -0.158  
## 
## 
## $`378`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##      39.952       -0.158  
## 
## 
## $`379`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9221      -0.1577  
## 
## 
## $`380`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9230      -0.1577  
## 
## 
## $`381`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9581      -0.1579  
## 
## 
## $`382`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8929      -0.1575  
## 
## 
## $`383`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9526      -0.1579  
## 
## 
## $`384`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9306      -0.1578  
## 
## 
## $`385`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9451      -0.1579  
## 
## 
## $`386`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9383      -0.1578  
## 
## 
## $`387`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8899      -0.1575  
## 
## 
## $`388`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9307      -0.1578  
## 
## 
## $`389`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9344      -0.1579  
## 
## 
## $`390`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.8443      -0.1572  
## 
## 
## $`391`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9667      -0.1579  
## 
## 
## $`392`
## 
## Call:
## lm(formula = mpg ~ horsepower, data = .)
## 
## Coefficients:
## (Intercept)   horsepower  
##     39.9860      -0.1584
(loocv_mse <- map2_dbl(loocv_models, loocv_data$test, mse))
##            1            2            3            4            5 
## 3.213342e+01 2.020010e+00 1.250857e+00 2.849248e+01 6.484768e+01 
##            6            7            8            9           10 
## 9.584908e-03 1.172437e+02 1.759818e-01 6.139599e+01 7.341613e-02 
##           11           12           13           14           15 
## 7.297568e+00 8.428034e+00 3.924368e+00 7.181963e+00 3.926033e-01 
##           16           17           18           19           20 
## 2.241539e+01 1.529641e+01 3.580449e+01 4.146999e+01 4.490646e+01 
##           21           22           23           24           25 
## 3.437743e+01 9.357612e+00 2.910330e+00 1.483044e+01 9.995138e+00 
##           26           27           28           29           30 
## 4.679773e+00 3.291832e+01 6.880037e+01 7.917981e+00 5.166911e+00 
##           31           32           33           34           35 
## 2.613037e-01 8.233947e+00 6.724508e+00 6.075519e+01 3.367885e+00 
##           36           37           38           39           40 
## 5.377148e-01 1.914019e+01 2.069744e+01 3.122004e+01 4.656941e+00 
##           41           42           43           44           45 
## 1.103363e-02 6.009991e+01 9.782282e+00 5.420830e+01 1.874967e+01 
##           46           47           48           49           50 
## 1.403014e+01 1.692639e+01 7.171493e+00 1.857938e+02 5.330425e+00 
##           51           52           53           54           55 
## 3.451690e-01 2.871244e-01 1.907775e+00 2.667455e+01 5.764925e+01 
##           56           57           58           59           60 
## 4.652496e+00 2.634304e+01 4.135667e+01 2.767354e+00 5.083533e+00 
##           61           62           63           64           65 
## 5.183800e+00 2.165397e+00 2.102111e+01 1.616981e+01 2.715686e+00 
##           66           67           68           69           70 
## 1.478994e+01 5.766164e+01 1.622758e+00 2.147951e+01 5.366573e+00 
##           71           72           73           74           75 
## 3.803600e+01 1.017226e+00 1.157220e+00 3.674287e+00 2.422052e+01 
##           76           77           78           79           80 
## 9.332138e+01 2.724249e+01 2.056389e+00 7.082556e-01 6.365750e+01 
##           81           82           83           84           85 
## 1.042080e+01 1.837023e-01 1.111564e+01 5.283052e-01 1.115532e+02 
##           86           87           88           89           90 
## 1.476578e+01 5.106756e+01 3.555996e+01 4.046029e+00 1.150586e-01 
##           91           92           93           94           95 
## 4.209601e+00 2.074089e+00 4.719855e+00 3.006473e-02 8.090347e-01 
##           96           97           98           99          100 
## 1.857938e+02 1.692639e+01 5.865479e+00 6.072868e+01 1.656735e+01 
##          101          102          103          104          105 
## 5.166911e+00 5.005465e+01 5.720928e+01 5.141271e+01 2.667455e+01 
##          106          107          108          109          110 
## 1.696762e+01 1.605078e+00 9.165497e-01 2.403135e+00 6.926842e+00 
##          111          112          113          114          115 
## 1.866232e+01 7.038055e+00 4.826153e-01 5.921415e+00 2.586927e+01 
##          116          117          118          119          120 
## 6.295836e+01 9.331482e+00 2.464358e+00 4.146999e+01 2.110725e+00 
##          121          122          123          124          125 
## 3.164374e+01 1.583746e+01 3.181246e+01 5.447911e+01 8.994396e-01 
##          126          127          128          129          130 
## 3.262850e+00 1.625041e+02 4.531854e+01 3.912844e+00 1.250924e+00 
##          131          132          133          134          135 
## 8.093849e-02 7.440905e+00 8.749214e+01 4.342942e+01 1.128019e+01 
##          136          137          138          139          140 
## 4.819221e-01 7.300693e+00 2.074089e+00 4.826153e-01 6.512312e+01 
##          141          142          143          144          145 
## 2.890009e+01 1.349922e+02 4.421187e+01 2.643188e-02 1.166297e+00 
##          146          147          148          149          150 
## 5.171868e-01 7.539327e+01 6.423470e-01 1.340184e+01 4.435387e+00 
##          151          152          153          154          155 
## 1.605078e+00 4.352799e+01 1.667557e+01 4.555565e+01 2.529958e+00 
##          156          157          158          159          160 
## 2.154283e+01 4.106222e-01 3.548262e+01 5.782504e+01 3.546176e-03 
##          161          162          163          164          165 
## 1.586678e+01 4.030122e+01 3.803600e+01 3.786435e+00 3.786435e+00 
##          166          167          168          169          170 
## 1.033653e+01 2.749188e-02 4.412290e+01 5.458248e+00 4.376707e+01 
##          171          172          173          174          175 
## 3.302308e+01 9.704988e+01 1.075349e+01 9.704988e+01 7.748422e+00 
##          176          177          178          179          180 
## 8.694181e+00 1.528278e+02 2.011132e+00 4.472444e+01 1.964564e+01 
##          181          182          183          184          185 
## 7.092852e+01 5.566352e+01 4.994053e+01 6.653002e-02 2.048085e-01 
##          186          187          188          189          190 
## 5.166911e+00 1.449052e+02 7.175958e+01 3.122004e+01 6.714948e+01 
##          191          192          193          194          195 
## 5.432293e+00 3.122569e+01 1.486748e+01 8.895572e-01 1.014678e+01 
##          196          197          198          199          200 
## 7.581236e+00 2.813085e-01 9.653262e+00 1.066789e-02 4.994053e+01 
##          201          202          203          204          205 
## 1.157220e+00 5.561744e-01 5.183800e+00 2.174016e+01 5.880433e+00 
##          206          207          208          209          210 
## 2.358597e+00 4.636272e-01 4.412290e+01 2.324873e-01 3.803600e+01 
##          211          212          213          214          215 
## 6.654632e+00 1.140456e+01 2.250222e+01 1.982618e+01 5.690043e+01 
##          216          217          218          219          220 
## 4.133647e-01 8.334101e+00 2.269988e+01 1.053729e+01 3.180794e+01 
##          221          222          223          224          225 
## 1.511906e-01 1.723940e+00 1.679054e+01 2.702564e+00 1.432359e+01 
##          226          227          228          229          230 
## 3.677007e+01 2.530169e+01 2.894490e+02 2.032089e+01 2.892397e+01 
##          231          232          233          234          235 
## 1.033955e+01 1.457053e+00 1.656735e+01 6.799020e-02 1.140418e+01 
##          236          237          238          239          240 
## 4.531854e+01 1.724878e-01 3.231044e+00 1.250516e+02 6.799020e-02 
##          241          242          243          244          245 
## 2.020818e+01 6.679005e+01 1.264151e-01 4.949936e+01 3.546176e-03 
##          246          247          248          249          250 
## 5.717027e+00 4.241411e+00 1.136904e+01 1.066789e-02 7.668633e+00 
##          251          252          253          254          255 
## 2.890009e+01 2.324873e-01 3.068052e+00 9.584908e-03 2.758466e+01 
##          256          257          258          259          260 
## 8.728675e+00 1.478825e+01 8.566786e+00 1.605078e+00 6.048098e+01 
##          261          262          263          264          265 
## 1.666316e+01 8.137554e+01 4.241411e+00 1.411405e+01 1.771618e+00 
##          266          267          268          269          270 
## 3.009836e+00 5.447911e+01 1.497268e-01 5.899216e+01 3.180083e+01 
##          271          272          273          274          275 
## 3.393771e-02 2.657232e+00 2.299575e-01 2.667455e+01 1.864996e+00 
##          276          277          278          279          280 
## 2.362534e+01 4.326951e+00 8.472761e-03 1.853442e+02 2.890009e+01 
##          281          282          283          284          285 
## 1.990884e+01 1.519954e+01 1.198846e-02 5.140783e+01 2.073870e+00 
##          286          287          288          289          290 
## 7.693554e+01 1.863862e+02 1.040145e+01 6.512312e+01 9.620669e+00 
##          291          292          293          294          295 
## 4.018772e+01 3.231044e+00 1.853546e+01 9.969267e+01 2.667455e+01 
##          296          297          298          299          300 
## 4.843384e+01 2.715686e+00 3.065622e+01 7.539327e+01 1.238847e+02 
##          301          302          303          304          305 
## 5.183800e+00 3.036942e+01 1.732348e+01 1.150586e-01 3.926033e-01 
##          306          307          308          309          310 
## 1.732348e+01 7.861988e+00 2.910330e+00 1.146842e+01 2.654407e+00 
##          311          312          313          314          315 
## 2.852183e+01 2.898726e+00 2.454227e+01 8.211334e-01 1.822059e+01 
##          316          317          318          319          320 
## 9.165497e-01 4.133647e-01 1.075349e+01 3.601165e+01 4.129160e+00 
##          321          322          323          324          325 
## 5.556481e-01 4.905186e+00 7.551530e+00 7.680024e+00 8.493035e+01 
##          326          327          328          329          330 
## 6.586063e+00 5.236824e+01 5.166911e+00 1.075349e+01 1.075349e+01 
##          331          332          333          334          335 
## 1.041250e+01 6.160147e-01 4.355852e+01 5.953056e-03 1.034145e+01 
##          336          337          338          339          340 
## 2.423504e+00 8.006204e+01 8.895572e-01 8.410440e+00 1.605078e+00 
##          341          342          343          344          345 
## 8.941908e-02 1.215199e+01 1.622758e+00 1.235372e-01 2.047146e+00 
##          346          347          348          349          350 
## 1.294907e-02 6.679005e+01 6.799020e-02 2.061807e-01 4.278732e+00 
##          351          352          353          354          355 
## 2.638785e+01 3.075798e+01 2.538815e+01 2.758083e+00 5.999879e-01 
##          356          357          358          359          360 
## 4.543656e+01 1.480422e+01 1.294907e-02 4.559232e+00 3.843146e+00 
##          361          362          363          364          365 
## 8.066902e+01 3.408942e+00 6.666276e+00 2.345776e+02 1.605078e+00 
##          366          367          368          369          370 
## 5.072715e-02 1.741235e+01 2.800088e+01 1.066612e+01 9.165497e-01 
##          371          372          373          374          375 
## 6.186339e+00 2.754308e-01 6.313349e-01 4.826153e-01 5.166911e+00 
##          376          377          378          379          380 
## 1.326611e+02 9.672737e+00 4.435387e+00 6.707682e+00 5.166911e+00 
##          381          382          383          384          385 
## 3.579356e+01 1.977894e+01 1.049286e+01 9.165497e-01 6.118529e+00 
##          386          387          388          389          390 
## 2.487076e+00 2.255236e+01 4.707417e-01 8.639328e-01 7.008530e+01 
##          391          392 
## 8.418264e+01 2.054254e+01
mean(loocv_mse)
## [1] 24.23151

LOOCV in linear regression

LOOCV in classification

titanic_loocv <- crossv_kfold(titanic, k = nrow(titanic))
titanic_models <- map(titanic_loocv$train, ~ glm(Survived ~ Age * Sex,
                                                 data = .,
                                                 family = binomial))
titanic_mse <- map2_dbl(titanic_models, titanic_loocv$test, mse.glm)
mean(titanic_mse, na.rm = TRUE)
## [1] 0.1703518

Exercise: LOOCV in linear regression

\(k\)-fold cross-validation

\[CV_{(k)} = \frac{1}{k} \sum_{i = 1}^{k}{MSE_i}\]

  • Split data into \(k\) folds
  • Repeat training/test process for each fold
  • LOOCV: \(k=n\)

k-fold CV in linear regression

cv10_data <- crossv_kfold(Auto, k = 10)

Computational speed of LOOCV

Computational speed of 10-fold CV

k-fold CV in logistic regression

titanic_kfold <- crossv_kfold(titanic, k = 10)
titanic_models <- map(titanic_kfold$train, ~ glm(Survived ~ Age * Sex,
                                                 data = .,
                                                 family = binomial))
titanic_mse <- map2_dbl(titanic_models, titanic_kfold$test, mse.glm)
mean(titanic_mse, na.rm = TRUE)
## [1] 0.1709727

Exercise: k-fold CV