"
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {
"ExecuteTime": {
"end_time": "2020-09-09T04:09:59.269223Z",
"start_time": "2020-09-09T04:09:58.955734Z"
}
},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"\n",
""
],
"text/plain": [
""
]
},
"execution_count": 1,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"import os\n",
"\n",
"# path : store the current path to convert back to it later\n",
"path = os.getcwd()\n",
"os.chdir(os.path.join('..', '..', 'notebook_format'))\n",
"\n",
"from formats import load_style\n",
"load_style(css_style='custom2.css', plot_style=False)"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {
"ExecuteTime": {
"end_time": "2020-09-09T04:10:01.056271Z",
"start_time": "2020-09-09T04:09:59.272407Z"
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Ethen 2020-09-08 21:10:01 \n",
"\n",
"CPython 3.6.4\n",
"IPython 7.15.0\n",
"\n",
"numpy 1.19.1\n",
"pandas 1.0.5\n",
"sklearn 0.23.1\n",
"matplotlib 3.1.0\n",
"xgboost 1.0.0\n"
]
}
],
"source": [
"os.chdir(path)\n",
"\n",
"# 1. magic for inline plot\n",
"# 2. magic to print version\n",
"# 3. magic so that the notebook will reload external python modules\n",
"# 4. magic to enable retina (high resolution) plots\n",
"# https://gist.github.com/minrk/3301035\n",
"%matplotlib inline\n",
"%load_ext watermark\n",
"%load_ext autoreload\n",
"%autoreload 2\n",
"%config InlineBackend.figure_format='retina'\n",
"\n",
"import os\n",
"import time\n",
"import numpy as np\n",
"import pandas as pd\n",
"import matplotlib.pyplot as plt\n",
"from xgboost import XGBClassifier\n",
"from sklearn.model_selection import train_test_split\n",
"\n",
"# prevent scientific notations\n",
"pd.set_option('display.float_format', lambda x: '%.4f' % x)\n",
"\n",
"%watermark -a 'Ethen' -d -t -v -p numpy,pandas,sklearn,matplotlib,xgboost"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Probability Calibration"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Well calibrated classifiers are classifiers for which the output probability can be directly interpreted as a confidence level. The definition of a well calibrated (binary) classifier should classify samples such that among the samples which the model gave a predicted probability value close to 0.8, approximately 80% of them actually belong to the positive class. For example, when looking up the weather forecast, we usually get a precipitation probability. e.g. If the weather forecast says there's a 80% chance of raining, then how trustworthy is this probability? In other words, if we take 100 days of data that were claimed to have a 80% chance of raining, how many rainy days were there? If the number of rainy days were around 80, then that means that particular rain forecast is indeed well calibrated.\n",
"\n",
"As it turns out, a lot of the classifiers/models that we used on a day to day basis might not be calibrated right out of the box, either due to the objective function of the model or simply when working with highly imbalanced datasets, our model's probability estimates can be skewed towards the majority class. Another way to put it is: After training a classifier, the output we get might just be a ranking score instead of well calibrated probability. A ranking score is essentially evaluating how well does the model score positive examples above negative ones, whereas a calibrated probability is evaluating how closely the scores generated by our model resembles an actual probability.\n",
"\n",
"Obtaining a well calibrated probability becomes important when:\n",
"\n",
"- We wish to use the probability threshold to inform some action. e.g. We'll reject the loan approval if the default rate is higher than 50% or we'll defer the judgment to humans if the probability is lower than some threshold.\n",
"- If our ranking formula is not solely based on the original model's score. In some cases, we may wish to use the score/probability along with some additional factors for ranking purpose. e.g. In the advertising cost per click model, we're going to rank the ads by its expected value (the probability of clicking on the ad multiplied by the ad fee for the click)."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Data Preprocessing"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We'll be using the credit card default dataset from UCI, we can download this dataset from [Kaggle](https://www.kaggle.com/uciml/default-of-credit-card-clients-dataset) as well."
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {
"ExecuteTime": {
"end_time": "2020-09-09T04:10:01.225819Z",
"start_time": "2020-09-09T04:10:01.061287Z"
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"(30000, 25)\n"
]
},
{
"data": {
"text/html": [
"
"
],
"text/plain": [
" ID LIMIT_BAL SEX EDUCATION MARRIAGE AGE PAY_0 PAY_2 PAY_3 \\\n",
"9256 9257 20000.0000 2 3 1 23 1 2 2 \n",
"23220 23221 150000.0000 2 3 2 35 -1 2 -1 \n",
"11074 11075 260000.0000 2 2 1 43 2 2 2 \n",
"1583 1584 50000.0000 2 1 2 70 2 2 0 \n",
"8623 8624 390000.0000 2 2 1 45 1 -2 -2 \n",
"\n",
" PAY_4 ... BILL_AMT4 BILL_AMT5 BILL_AMT6 PAY_AMT1 PAY_AMT2 \\\n",
"9256 -2 ... 0.0000 0.0000 0.0000 480.0000 0.0000 \n",
"23220 2 ... 1143.0000 163.0000 2036.0000 0.0000 2264.0000 \n",
"11074 2 ... 2500.0000 2500.0000 2500.0000 0.0000 0.0000 \n",
"1583 0 ... 17793.0000 18224.0000 18612.0000 0.0000 2200.0000 \n",
"8623 -2 ... 0.0000 0.0000 0.0000 0.0000 0.0000 \n",
"\n",
" PAY_AMT3 PAY_AMT4 PAY_AMT5 PAY_AMT6 default.payment.next.month \n",
"9256 0.0000 0.0000 0.0000 0.0000 1 \n",
"23220 0.0000 163.0000 2036.0000 0.0000 0 \n",
"11074 0.0000 0.0000 0.0000 0.0000 1 \n",
"1583 700.0000 700.0000 674.0000 608.0000 0 \n",
"8623 0.0000 0.0000 0.0000 3971.0000 1 \n",
"\n",
"[5 rows x 25 columns]"
]
},
"execution_count": 6,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"test_size = 0.1\n",
"val_size = 0.3\n",
"random_state = 1234\n",
"\n",
"df_train, df_test = train_test_split(\n",
" df,\n",
" test_size=test_size,\n",
" random_state=random_state,\n",
" stratify=df[label_col])\n",
"\n",
"df_train, df_val = train_test_split(\n",
" df_train,\n",
" test_size=val_size,\n",
" random_state=random_state,\n",
" stratify=df_train[label_col])\n",
"\n",
"print('train shape: ', df_train.shape)\n",
"print('validation shape: ', df_val.shape)\n",
"print('test shape: ', df_test.shape)\n",
"\n",
"df_train.head()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Model Training"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We'll train a binary classifier to predict default payment, and evaluate the model using some common evaluation metrics. In our example, we'll only focus on the widely used boosted tree open sourced library xgboost, though the calibration process and technique introduced in later section is applicable for any arbitrary model."
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {
"ExecuteTime": {
"end_time": "2020-09-09T04:10:02.525423Z",
"start_time": "2020-09-09T04:10:01.422864Z"
}
},
"outputs": [
{
"data": {
"text/plain": [
"XGBClassifier(base_score=0.5, booster=None, colsample_bylevel=1,\n",
" colsample_bynode=1, colsample_bytree=1, gamma=0, gpu_id=-1,\n",
" importance_type='gain', interaction_constraints=None,\n",
" learning_rate=0.1, max_delta_step=0, max_depth=6,\n",
" min_child_weight=1, missing=nan, monotone_constraints=None,\n",
" n_estimators=30, n_jobs=0, num_parallel_tree=1, random_state=0,\n",
" reg_alpha=0, reg_lambda=1, scale_pos_weight=1, subsample=1,\n",
" tree_method=None, validate_parameters=False, verbosity=None)"
]
},
"execution_count": 7,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# parameters chosen in an adhoc manner\n",
"xgb_params = {\n",
" 'learning_rate': 0.1,\n",
" 'max_depth': 6,\n",
" 'n_estimators': 30\n",
"}\n",
"\n",
"xgb = XGBClassifier(**xgb_params)\n",
"xgb.fit(df_train[input_cols].values, df_train[label_col].values)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"A lot of the helper functions/class are organized as under the `calibration_module`, which can be found under the same folder as this notebook. [link](https://github.com/ethen8181/machine-learning/tree/master/model_selection/prob_calibration)"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {
"ExecuteTime": {
"end_time": "2020-09-09T04:10:02.755766Z",
"start_time": "2020-09-09T04:10:02.532486Z"
}
},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"\n",
"
\n",
" \n",
"
\n",
"
\n",
"
auc
\n",
"
precision
\n",
"
recall
\n",
"
f1
\n",
"
log_loss
\n",
"
brier
\n",
"
name
\n",
"
\n",
" \n",
" \n",
"
\n",
"
0
\n",
"
0.8293
\n",
"
0.5667
\n",
"
0.6163
\n",
"
0.5904
\n",
"
0.3958
\n",
"
0.1213
\n",
"
xgb_train
\n",
"
\n",
"
\n",
"
1
\n",
"
0.7872
\n",
"
0.5177
\n",
"
0.5954
\n",
"
0.5539
\n",
"
0.4257
\n",
"
0.1330
\n",
"
xgb_val
\n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" auc precision recall f1 log_loss brier name\n",
"0 0.8293 0.5667 0.6163 0.5904 0.3958 0.1213 xgb_train\n",
"1 0.7872 0.5177 0.5954 0.5539 0.4257 0.1330 xgb_val"
]
},
"execution_count": 8,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"from calibration_module.utils import compute_binary_score\n",
"\n",
"# evaluate the metrics for training and validation set\n",
"estimators = {\n",
" 'xgb': xgb\n",
"}\n",
"df_groups = {\n",
" 'train': df_train,\n",
" 'val': df_val\n",
"}\n",
"\n",
"estimator_metrics = []\n",
"for name, estimator in estimators.items():\n",
" for df_name, df_group in df_groups.items():\n",
" y_prob = estimator.predict_proba(df_group[input_cols].values)[:, 1]\n",
" # compute various binary classification metrics\n",
" metric_dict = compute_binary_score(df_group[label_col], y_prob)\n",
" metric_dict['name'] = name + '_' + df_name\n",
" estimator_metrics.append(metric_dict)\n",
"\n",
"df_metrics = pd.DataFrame(estimator_metrics)\n",
"df_metrics"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Measuring Calibration"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We'll first discuss how do we measure whether a model is well-calibrated or not. The main idea here is to first discretize our model predictions into $M$ interval bins, and calculate the average fraction of positives and predicted probability of each bin. Here, the number of bin is configurable, and samples that have similar predicted score will fall into the same bin.\n",
"\n",
"Let $B_m$ be the set of samples whose predicted probability falls into interval $I_m = \\big( \\frac{m - 1}{M}, \\frac{m}{M}\\big]$. The fraction of positives for $B_m$ can be computed by:\n",
"\n",
"\\begin{align}\n",
"pos(B_m) = \\frac{1}{|B_m|} \\sum_{i \\in B_m} y_i\n",
"\\end{align}\n",
"\n",
"Where $y_i$ is the true class label for sample $i$ (assuming in the binary classification setting 1 denotes a positive class and 0 otherwise). On the other hand, the predicted probability within bin $B_m$ is defined as:\n",
"\n",
"\\begin{align}\n",
"prob(B_m) = \\frac{1}{|B_m|} \\sum_{i \\in B_m} \\hat{p_i}\n",
"\\end{align}\n",
"\n",
"Where $\\hat{p_i}$ is the predicted probability for sample $i$. Given the two terms, fraction of positives and predicted probability within each bin, we can either build a calibration curve to visualize the amount of miscalibration or directly compute a summary statistics.\n",
"\n",
"**Calibration Curve** or also known as a **Reliability Diagram**. For each bin, the mean predicted probability, $prob(B_m)$, is plotted against the fraction of positive cases for that bin, $pos(B_m)$. If the model is well-calibrated, then the points will fall near the diagonal line, and any deviation from that diagonal line in the visualization depicts some level of miscalibration with our model.\n",
"\n",
"**Expected Calibrator Error (ECE)** is one commonly used summary statistic that measures the difference between the expected probability and fraction of positives.\n",
"\n",
"\\begin{align}\n",
"ECE = \\sqrt{ \\sum_{m=1}^M \\frac{|B_m|}{n} \\big(prob(B_m) - pos(B_m)\\big)^2 }\n",
"\\end{align}\n",
"\n",
"Where $n$ is the total number of samples. Here the expected calibration error is measured by the RMSE (Root Meas Squared Error) between $prob(B_m)$ and $pos(B_m)$. If we wish to have a metric that is less sensitive to outliers, we could also switch to MAE (Mean Absolute Error).\n",
"\n",
"\n",
"We'll now take a look at these concepts in action."
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {
"ExecuteTime": {
"end_time": "2020-09-09T04:10:02.833104Z",
"start_time": "2020-09-09T04:10:02.758932Z"
}
},
"outputs": [],
"source": [
"# extract the validation and test true label and predicted probability,\n",
"# as we are working with binary classification in this use case, we can\n",
"# extract the predicted probability for the positive class\n",
"labels_val = df_val[label_col].values\n",
"xgb_pred_val = xgb.predict_proba(df_val[input_cols].values)[:, 1]\n",
"\n",
"labels_test = df_test[label_col].values\n",
"xgb_pred_test = xgb.predict_proba(df_test[input_cols].values)[:, 1]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We implement a `compute_calibration_summary` that builds on top of scikit-learn's calibration module [[5]](https://scikit-learn.org/stable/modules/calibration.html). Instead of only plotting the calibration curve, we also return a table that contains summary statistics on the model performance, and calibration error."
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {
"ExecuteTime": {
"end_time": "2020-09-09T04:10:04.693632Z",
"start_time": "2020-09-09T04:10:02.836751Z"
}
},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"
"
],
"text/plain": [
" auc precision recall f1 log_loss brier calibration_error \\\n",
"0 0.7872 0.5177 0.5954 0.5539 0.4257 0.1330 0.0240 \n",
"1 0.7739 0.5423 0.5602 0.5511 0.4355 0.1369 0.0296 \n",
"\n",
" name \n",
"0 xgb_val \n",
"1 xgb_test "
]
},
"execution_count": 10,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"from calibration_module.utils import compute_calibration_summary\n",
"\n",
"# link the label and probability into a dataframe\n",
"score_col = 'score'\n",
"df_xgb_eval_val = pd.DataFrame({\n",
" label_col: labels_val,\n",
" score_col: xgb_pred_val\n",
"})\n",
"df_xgb_eval_test = pd.DataFrame({\n",
" label_col: labels_test,\n",
" score_col: xgb_pred_test\n",
"})\n",
"\n",
"# key to the dictionary is for giving the result\n",
"# a descriptive name\n",
"eval_dict = {\n",
" 'xgb_val': df_xgb_eval_val,\n",
" 'xgb_test': df_xgb_eval_test\n",
"}\n",
"\n",
"# change default style figure and font size\n",
"plt.rcParams['figure.figsize'] = 12, 8\n",
"plt.rcParams['font.size'] = 12\n",
"\n",
"n_bins = 15\n",
"df_result = compute_calibration_summary(eval_dict, label_col, score_col, n_bins=n_bins)\n",
"df_result"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Judging from the calibration plot, we can see there are some points that fall above and below the diagonal line.\n",
"\n",
"- Below the diagonal: The model has over-forecast; the probabilities are too large.\n",
"- Above the diagonal: The model has under-forecast; the probabilities are too small.\n",
"\n",
"But from the looks of it, it seems like the predicted score is pretty well calibrated as the dots fall closely to the diagonal line."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Calibration Model"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The calibration technique that we'll be introducing here are all rescaling operation that is applied after the predictions have been made by a predictive mode, i.e. this assumes we already have a model, and we would only like to perform some post-processing steps to calibration our original model's prediction. As mentioned in the data preprocessing step, When training/learning the calibration function, we should ensure the data that is used to fit the original model and the one that is used for calibration does not overlap. e.g.\n",
"\n",
"- We can split the data into training / validation set, After our base model is trained on the training\n",
"set, the predictions on the validation set are used to fit the calibration model.\n",
"- Or do it in a cross validation way, where the data is split into $C$ folds. For each fold, one part is held aside for use as an validation set while the training is performed using the other $C-1$ fold. After repeating the process for all $C$ folds, we can compute final probability by doing an arithmetic mean of the calibrated classifier's predictions.\n",
"- Whether we're using train/validation split, or cross validation. It boils down to using the predicted probability as the single input feature, and the hold set's label as the target.\n",
"- To evaluate whether we successfully calibrated our model, we can/should check various evaluation metrics. e.g. Our ranking metrics such as AUC should remain somewhat the same, whereas our probability-related metrics such as calibration error should improve."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We'll introduce some notations for discussing the calibration models itself. Assuming a binary classification setting, where given a sample $x_i$ and its corresponding label $y_i$, our original model will produce a predicted probability of the positive class $\\hat{p_i}$. Given that most models are not calibrated out of the box, the calibration model's goal is to post processed $\\hat{p_i}$ and produce a well calibrated probability $\\hat{q_i}$.\n",
"\n",
"Two popular choices are **Platt Scaling** and **Isotonic Regression** [[6]](https://arxiv.org/abs/1207.1403).\n",
"\n",
"**Platt Scaling:** Is a parametric approach. At a high level, Platt Scaling amounts to training a logistic regression on of the original classifier's output with respect to the true class labels.\n",
"\n",
"**Isotonic Regression:** A non-parametric approach. With this approach, the idea is to fit a piecewise constant non-decreasing function, where we would merge similar scores into bins such that each bin becomes monotonically increasing. e.g. The first bin may have the range [0, 0.2] and probability 0.15, meaning that any instance with a score between 0 and 0.2 should be assigned a probability estimate of 0.15. More formally,\n",
"\n",
"\\begin{align}\n",
"& \\underset{\\mathbf{\\theta, a}}{\\text{min}}\n",
"&& \\sum_{m=1}^M \\sum_{i=1}^n \\mathbb{1} (a_m \\leq \\hat{p_i} < a_{m+1}) (\\theta_m - y_i)^2 \\\\ \\nonumber\n",
"& \\text{subject to}\n",
"&& 0 = a_1 \\leq a_2 \\leq ... \\leq a_{M+1} = 1, \\theta_1 \\leq \\theta_2 \\leq ... \\theta_M\n",
"\\end{align}\n",
"\n",
"Where $M$ is the number of bins, $a_1, ..., a_{M+1}$ are the interval boundaries that defines each mutually exclusive bins, $B_1, ..., B_M$. $\\theta_1, ..., \\theta_M$ are the corresponding calibrated score for $\\hat{p_i}$ that fall under the bin's boundaries. During the optimization process, the bin boundaries and prediction values are jointly optimized.\n",
"\n",
"In general, Platt Scaling is preferable if the calibration curve has a sigmoid shape and when there is few calibration data. Whereas, Isotonic Regression, being a non-parametric method, is preferable for non-sigmoid calibration curves and in situations where many additional data can be used for calibration. But again, it doesn't hurt to try both approaches on our data and see which one leads to a lower calibration error on the holdout test set."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Apart from Platt Scaling and Isotonic Regression that we'll often come across in online materials, here we'll also introduce two additional methods. Namely, **Histogram Binning** and **Platt Scaling Binning**\n",
"\n",
"**Histogram Binning** is a stricter version of Isotonic Regression, where we would directly define the bin boundaries either by choosing it to be of equal length interval or equal sample size interval. As for the prediction values for each bin, we would set it equal to the $pos(B_m)$, the average number of positive samples in that bin.\n",
"\n",
"**Platt Scaling Binning** is a blend of Platt Scaling and Histogram Binning [[9]](https://arxiv.org/abs/1909.10155). We first fit a Platt Scaling function, $f$, then just like Histogram Binning, we would bin the input samples. The main difference here is that we would bin the samples with the output from Platt Scaling instead of the original predicted probability. And for each bin, the calibrated prediction is the average of the scaling function, $f$, instead of the average number of positive samples. The motivation behind this method is the output from our scaling function lies within a narrower range than the original label values, hence it should introduce a lower calibration error."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"One important thing to note is (not commonly mentioned) while applying Platt Scaling related calibration method is that logistic regression assumes a linear relationship between the input and log odds class probability output. Hence in theory, it should be beneficial to first transform the class probability $p$ into a log odds scale, $z$ before passing it to Platt Scaling [[7]](https://arxiv.org/abs/1808.00111).\n",
"\n",
"\\begin{align}\n",
"z = \\log \\left({p\\over 1-p}\\right)\n",
"\\end{align}"
]
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {
"ExecuteTime": {
"end_time": "2020-09-09T04:10:04.744591Z",
"start_time": "2020-09-09T04:10:04.699088Z"
}
},
"outputs": [],
"source": [
"from sklearn.calibration import IsotonicRegression\n",
"from calibration_module.calibrator import (\n",
" HistogramCalibrator,\n",
" PlattCalibrator,\n",
" PlattHistogramCalibrator\n",
")"
]
},
{
"cell_type": "code",
"execution_count": 12,
"metadata": {
"ExecuteTime": {
"end_time": "2020-09-09T04:10:04.787865Z",
"start_time": "2020-09-09T04:10:04.749236Z"
}
},
"outputs": [
{
"data": {
"text/plain": [
"array([0.22972973, 0.14263076, 0.05276734, ..., 0.13581395, 0.11445783,\n",
" 0.36842105], dtype=float32)"
]
},
"execution_count": 12,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"isotonic = IsotonicRegression(out_of_bounds='clip',\n",
" y_min=xgb_pred_val.min(),\n",
" y_max=xgb_pred_val.max())\n",
"isotonic.fit(xgb_pred_val, labels_val)\n",
"isotonic_probs = isotonic.predict(xgb_pred_test)\n",
"isotonic_probs"
]
},
{
"cell_type": "code",
"execution_count": 13,
"metadata": {
"ExecuteTime": {
"end_time": "2020-09-09T04:10:04.837559Z",
"start_time": "2020-09-09T04:10:04.790170Z"
}
},
"outputs": [
{
"data": {
"text/plain": [
"array([0.22407407, 0.14074074, 0.02962963, ..., 0.12592593, 0.10555556,\n",
" 0.39814815])"
]
},
"execution_count": 13,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"histogram = HistogramCalibrator(n_bins=n_bins)\n",
"histogram.fit(xgb_pred_val, labels_val)\n",
"histogram_probs = histogram.predict(xgb_pred_test)\n",
"histogram_probs"
]
},
{
"cell_type": "code",
"execution_count": 14,
"metadata": {
"ExecuteTime": {
"end_time": "2020-09-09T04:10:04.891131Z",
"start_time": "2020-09-09T04:10:04.840045Z"
}
},
"outputs": [
{
"data": {
"text/plain": [
"array([0.2116992 , 0.16321564, 0.05074317, ..., 0.12652918, 0.10286036,\n",
" 0.35680279])"
]
},
"execution_count": 14,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"platt = PlattCalibrator(log_odds=True)\n",
"platt.fit(xgb_pred_val, labels_val)\n",
"platt_probs = platt.predict(xgb_pred_test)\n",
"platt_probs"
]
},
{
"cell_type": "code",
"execution_count": 15,
"metadata": {
"ExecuteTime": {
"end_time": "2020-09-09T04:10:04.940493Z",
"start_time": "2020-09-09T04:10:04.893898Z"
}
},
"outputs": [
{
"data": {
"text/plain": [
"array([0.21283885, 0.15907912, 0.05334451, ..., 0.12737967, 0.10104891,\n",
" 0.3687527 ])"
]
},
"execution_count": 15,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"platt_histogram = PlattHistogramCalibrator(n_bins=n_bins, log_odds=True)\n",
"platt_histogram.fit(xgb_pred_val, labels_val)\n",
"platt_histogram_probs = platt_histogram.predict(xgb_pred_test)\n",
"platt_histogram_probs"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Calibration Model Evaluation"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"In this section, we compare the calibration error of various calibration models."
]
},
{
"cell_type": "code",
"execution_count": 16,
"metadata": {
"ExecuteTime": {
"end_time": "2020-09-09T04:10:06.857700Z",
"start_time": "2020-09-09T04:10:04.943956Z"
}
},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"
"
],
"text/plain": [
" auc precision recall f1 log_loss brier calibration_error \\\n",
"0 0.7739 0.5423 0.5602 0.5511 0.4355 0.1369 0.0296 \n",
"2 0.7706 0.5525 0.5151 0.5331 0.4408 0.1388 0.0363 \n",
"1 0.7739 0.5423 0.5602 0.5511 0.4385 0.1379 0.0425 \n",
"\n",
" name \n",
"0 xgb \n",
"2 xgb platt histogram \n",
"1 xgb platt "
]
},
"execution_count": 19,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df_xgb_platt_eval_test = pd.DataFrame({\n",
" label_col: labels_test,\n",
" score_col: platt_probs\n",
"})\n",
"df_xgb_platt_histogram_eval_test = pd.DataFrame({\n",
" label_col: labels_test,\n",
" score_col: platt_histogram_probs\n",
"})\n",
"\n",
"eval_dict = {\n",
" 'xgb': df_xgb_eval_test,\n",
" 'xgb platt': df_xgb_platt_eval_test,\n",
" 'xgb platt histogram': df_xgb_platt_histogram_eval_test\n",
"}\n",
"\n",
"df_result = compute_calibration_summary(eval_dict, label_col, score_col, n_bins=n_bins)\n",
"df_result.sort_values('calibration_error')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Final Notes"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Although our primary focus was on calibrating binary classification models, we can extend the concepts and notations to multi-class setting by treating the problem as $K$ one versus all problems, where $K$ is the number of distinct classes. For $k = 1, ..., K$, we would create a binary classification where the label is $\\mathbb{1}(y_i = k)$, giving us $K$ calibration model, one for each class.\n",
"\n",
"Other than the techniques introduced here, there are many other methods that can be used to calibrate our model. e.g. for ease of production, some work resort to using a piecewise linear function:\n",
"\n",
"\\begin{align}\n",
"\\hat{q_i}=\n",
"\\begin{cases}\n",
"\\hat{p_i} & \\hat{p_i} < t_c \\\\\n",
"t_c \\cdot \\big( 1 + log( \\frac{\\hat{p_i}}{t_c} ) \\big) & \\hat{p_i} \\geq t_c\n",
"\\end{cases}\n",
"\\end{align}\n",
"\n",
"In this case, the calibration function is saying for any predicted probability higher than a user-defined calibration threshold $t_c$, we will scale the prediction using the function specified above. The piecewise linear function can be of any arbitrary function, and unlike the other estimators that we can directly plug and play, this requires us to have a much better understanding of our data's distribution.\n",
"\n",
"Given all the rage with deep learning models lately, there are even ones that are tailored for them [[8]](https://arxiv.org/abs/1706.04599). Calibration also becomes an important topic there, as modern neural networks often times optimizes for negative log likelihood. Upon being able to correctly classify the majority of the training samples, that measure can be further minimized by increasing the probability of its prediction, which will ultimately result in over/under confident predicted score.\n",
"\n",
"One caveat to note about measuring calibration error is that the number of bins do matter, play with the parameter and we might find surprising results. As we are measuring the calibration error of a continuous output (probability output from the model) by grouping samples into finite set of bins, the measure that we've obtained will only be an approximation of the true calibration error. The intuition behind this is that averaging a continuous number within a bin allows errors at different regions of a bin to cancel out with each other."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Reference"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"- [[1]](https://gdmarmerola.github.io/probability-calibration/) Blog: Calibration of probabilities for tree-based models\n",
"- [[2]](https://changhsinlee.com/python-calibration-plot/) Blog: A Guide to Calibration Plots in Python\n",
"- [[3]](https://machinelearningmastery.com/calibrated-classification-model-in-scikit-learn/) Blog: How and When to Use a Calibrated Classification Model with scikit-learn\n",
"- [[4]](https://www.youtube.com/watch?v=FkfDlOnQVvQ) Youtube: Model Calibration - is your model ready for the real world?\n",
"- [[5]](https://scikit-learn.org/stable/modules/calibration.html) Sklearn Documentation: Probability calibration\n",
"- [[6]](https://arxiv.org/abs/1207.1403) Alexandru Niculescu-Mizil, Richard A. Caruana - Obtaining Calibrated Probabilities from Boosting (2012)\n",
"- [[7]](https://arxiv.org/abs/1808.00111) Tim Leathart, Eibe Frank, Geoffrey Holmes, Bernhard Pfahringer - Probability Calibration Trees (2017)\n",
"- [[8]](https://arxiv.org/abs/1706.04599) Chuan Guo, Geoff Pleiss, Yu Sun, et al. - On Calibration of Modern Neural Networks (2017)\n",
"- [[9]](https://arxiv.org/abs/1909.10155) Ananya Kumar, Percy Liang, Tengyu Ma - Verified Uncertainty Calibration (2020)"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3 (ipykernel)",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.9.12"
},
"toc": {
"base_numbering": 1,
"nav_menu": {},
"number_sections": true,
"sideBar": true,
"skip_h1_title": false,
"title_cell": "Table of Contents",
"title_sidebar": "Contents",
"toc_cell": true,
"toc_position": {
"height": "calc(100% - 180px)",
"left": "10px",
"top": "150px",
"width": "235.594px"
},
"toc_section_display": true,
"toc_window_display": true
},
"varInspector": {
"cols": {
"lenName": 16,
"lenType": 16,
"lenVar": 40
},
"kernels_config": {
"python": {
"delete_cmd_postfix": "",
"delete_cmd_prefix": "del ",
"library": "var_list.py",
"varRefreshCmd": "print(var_dic_list())"
},
"r": {
"delete_cmd_postfix": ") ",
"delete_cmd_prefix": "rm(",
"library": "var_list.r",
"varRefreshCmd": "cat(var_dic_list()) "
}
},
"types_to_exclude": [
"module",
"function",
"builtin_function_or_method",
"instance",
"_Feature"
],
"window_display": false
}
},
"nbformat": 4,
"nbformat_minor": 4
}