![](\"img/uci-logo.png\"){kind=logo}
\n",
"\n",
"## Example problem description:\n",
"\n",
"We'll be using data on blood donations from the UC Irvine Machine Learning Repository. Given some data on some blood donors' previous donations, we'll be trying to predict a binary outcome of whether the donor gave blood in a certain time period.\n",
"\n",
"Here's the official description:\n",
"\n",
"Note: in this example we're not putting much thought into which model to use and we're sticking with the default model parameters. The point is not that we would *want* to do this. The point is that all of these very different models have such similar structure and conventions that we can iterate through a bunch of them, fit, and evaluate them while changing none of the syntax.
"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"from sklearn.ensemble import AdaBoostClassifier\n",
"from sklearn.ensemble import RandomForestClassifier\n",
"from sklearn.naive_bayes import MultinomialNB\n",
"from sklearn.neighbors import KNeighborsClassifier\n",
"from sklearn.svm import LinearSVC\n",
"\n",
"other_clfs = {\n",
" 'Random forest': RandomForestClassifier(),\n",
" 'AdaBoost': AdaBoostClassifier(),\n",
" 'Naive Bayes': MultinomialNB(),\n",
" 'Linear SVC': LinearSVC(),\n",
" 'KNN': KNeighborsClassifier(5),\n",
"}\n",
"\n",
"# iterating through all of these models we want to fit ...\n",
"for name, other_clf in other_clfs.iteritems():\n",
" \n",
" # fit the model with the training data\n",
" print other_clf.fit(X_train, y_train)\n",
" \n",
" # cross validation score\n",
" print '---\\nScore:', other_clf.score(X_test, y_test)\n",
" print '\\n'"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',\n",
" metric_params=None, n_neighbors=5, p=2, weights='uniform')\n",
"---\n",
"Score: 0.756684491979\n",
"\n",
"\n",
"LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,\n",
" intercept_scaling=1, loss='l2', multi_class='ovr', penalty='l2',\n",
" random_state=None, tol=0.0001, verbose=0)\n",
"---\n",
"Score: 0.299465240642\n",
"\n",
"\n",
"MultinomialNB(alpha=1.0, class_prior=None, fit_prior=True)\n",
"---\n",
"Score: 0.729946524064\n",
"\n",
"\n",
"RandomForestClassifier(bootstrap=True, compute_importances=None,\n",
" criterion='gini', max_depth=None, max_features='auto',\n",
" max_leaf_nodes=None, min_density=None, min_samples_leaf=1,\n",
" min_samples_split=2, n_estimators=10, n_jobs=1,\n",
" oob_score=False, random_state=None, verbose=0)\n",
"---\n",
"Score: 0.756684491979\n",
"\n",
"\n",
"AdaBoostClassifier(algorithm='SAMME.R', base_estimator=None,\n",
" learning_rate=1.0, n_estimators=50, random_state=None)"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n",
"---\n",
"Score: 0.775401069519\n",
"\n",
"\n"
]
}
],
"prompt_number": 20
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Anatomy of a `scikit-learn` model"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The object itself:"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"clf"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 21,
"text": [
"LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n",
" intercept_scaling=1, penalty='l2', random_state=None, tol=0.0001)"
]
}
],
"prompt_number": 21
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Coefficients (if fit):"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"clf.coef_"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 22,
"text": [
"array([[ -9.74764830e-02, 1.49194232e-06, 3.72985587e-04,\n",
" -1.17108498e-02]])"
]
}
],
"prompt_number": 22
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Intercept (if fit):"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"clf.intercept_"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 23,
"text": [
"array([-0.53074433])"
]
}
],
"prompt_number": 23
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Model parameters:"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"clf.get_params()"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 24,
"text": [
"{'C': 1.0,\n",
" 'class_weight': None,\n",
" 'dual': False,\n",
" 'fit_intercept': True,\n",
" 'intercept_scaling': 1,\n",
" 'penalty': 'l2',\n",
" 'random_state': None,\n",
" 'tol': 0.0001}"
]
}
],
"prompt_number": 24
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Making predictions for $\\hat y$:"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"clf.predict(X_test)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 25,
"text": [
"array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0,\n",
" 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0,\n",
" 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0,\n",
" 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
" 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
" 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
" 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
" 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
" 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
" 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1,\n",
" 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
" 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
" 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
" 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
" 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0,\n",
" 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
" 0, 0, 0, 0, 0, 0])"
]
}
],
"prompt_number": 25
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Predicting probabilities, $p(y_i = 1)$"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"pd.DataFrame(clf.predict_proba(X_test))\\\n",
" .head(10)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"html": [
"\n",
"\n",
" \n",
"
\n",
"
"
],
"metadata": {},
"output_type": "pyout",
"prompt_number": 26,
"text": [
" 0 1\n",
"0 0.691294 0.308706\n",
"1 0.952584 0.047416\n",
"2 0.726055 0.273945\n",
"3 0.770998 0.229002\n",
"4 0.680134 0.319866\n",
"5 0.705628 0.294372\n",
"6 0.637364 0.362636\n",
"7 0.888941 0.111059\n",
"8 0.559386 0.440614\n",
"9 0.936608 0.063392"
]
}
],
"prompt_number": 26
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"| \n", " | 0 | \n", "1 | \n", "
|---|---|---|
| 0 | \n", "0.691294 | \n", "0.308706 | \n", "
| 1 | \n", "0.952584 | \n", "0.047416 | \n", "
| 2 | \n", "0.726055 | \n", "0.273945 | \n", "
| 3 | \n", "0.770998 | \n", "0.229002 | \n", "
| 4 | \n", "0.680134 | \n", "0.319866 | \n", "
| 5 | \n", "0.705628 | \n", "0.294372 | \n", "
| 6 | \n", "0.637364 | \n", "0.362636 | \n", "
| 7 | \n", "0.888941 | \n", "0.111059 | \n", "
| 8 | \n", "0.559386 | \n", "0.440614 | \n", "
| 9 | \n", "0.936608 | \n", "0.063392 | \n", "
![](\"img/question-mark.png\")
\n",
"\n",
"### Challenge: evaluating model performance\n",
"\n",
"The default metric for logistic regression is mean accuracy:\n",
"\n",
"$$\\mathrm{accuracy}(y, \\hat y) = \\frac{1}{n} \\sum_{i=1}^n \\mathbb{I}(\\hat y_i = y_i)$$"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"clf.score(X_test, y_test)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 27,
"text": [
"0.77272727272727271"
]
}
],
"prompt_number": 27
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"But what if we want a more holistic view of how well this model works?\n",
"\n",
"![](\"img/lego-brick.png\")
\n",
"\n",
"### Common solution: use cross validation tools and other metrics\n",
"\n",
"#### Cross validation\n",
"\n",
"Our data set here is fairly small $(N=748)$, so what if we happened to randomly pick a non-representative chunk for testing?\n",
"\n",
"Instead of summarizing the performance just on the test data we want to use $k$-fold cross-validation, splitting the data randomly into chunks — \"folds\" — and evaluating on each to get an idea of the variation in performance."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"from sklearn import cross_validation\n",
"\n",
"# come up with random folds of the data\n",
"kf = cross_validation.KFold(len(X), n_folds=5, shuffle=True)\n",
"\n",
"def plot_scores(scores):\n",
" N = len(scores)\n",
" plt.bar(np.arange(1, N + 1) - 0.4, scores, color=set2[2])\n",
" plt.title('{}-fold cross-validation scores'.format(N), fontsize=18)\n",
" plt.xlabel('fold', fontsize=14)\n",
" plt.ylabel('score', fontsize=14)\n",
" plt.xlim(0.5, N + 0.5)\n",
" plt.ylim(0, 1)\n",
" plt.show()\n",
"\n",
"# evaluate the fitted model on each fold in turn, returns a score for each fold\n",
"scores = cross_validation.cross_val_score(clf, X, y, cv=kf, n_jobs=1)\n",
"\n",
"print 'scores:', scores\n",
"print 'average score:', np.mean(scores)\n",
"plot_scores(scores)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"scores: [ 0.78 0.76666667 0.81333333 0.75167785 0.73154362]\n",
"average score: 0.768644295302\n"
]
},
{
"metadata": {},
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAmMAAAICCAYAAACDYMqCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABJRU5ErkJggg==\n",
"text": [
"\n",
"\n",
" \n",
"
\n",
"
"
],
"metadata": {},
"output_type": "pyout",
"prompt_number": 31,
"text": [
" pred 0 pred 1\n",
"actual 0 283 5\n",
"actual 1 80 6"
]
}
],
"prompt_number": 31
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"there are **many** available ...\n",
"\n",
"```\n",
"metrics.accuracy_score, metrics.auc, metrics.average_precision_score, metrics.classification_report, metrics.confusion_matrix, metrics.f1_score, metrics.fbeta_score, metrics.hamming_loss, metrics.hinge_loss, metrics.jaccard_similarity_score, metrics.log_loss, metrics.matthews_corrcoef, metrics.precision_recall_curve, metrics.precision_recall_fscore_support, metrics.precision_score, metrics.recall_score, metrics.roc_auc_score, metrics.roc_curve, metrics.zero_one_loss, metrics.explained_variance_score, metrics.mean_absolute_error, metrics.mean_squared_error, metrics.r2_score, metrics.adjusted_mutual_info_score, metrics.adjusted_rand_score, metrics.completeness_score, metrics.homogeneity_completeness_v_measure, metrics.homogeneity_score, metrics.mutual_info_score, metrics.normalized_mutual_info_score, metrics.silhouette_score, metrics.silhouette_samples, metrics.v_measure_score, metrics.consensus_score, metrics.pairwise.additive_chi2_kernel, metrics.pairwise.chi2_kernel, metrics.pairwise.distance_metrics, metrics.pairwise.euclidean_distances, metrics.pairwise.kernel_metrics, metrics.pairwise.linear_kernel, metrics.pairwise.manhattan_distances, metrics.pairwise.pairwise_distances, metrics.pairwise.pairwise_kernels, metrics.pairwise.polynomial_kernel, metrics.pairwise.rbf_kernel, metrics.pairwise_distances, metrics.pairwise_distances_argmin, metrics.pairwise_distances_argmin_min\n",
"```\n",
"\n",
"... and they **all use the same basic syntax**."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"| \n", " | pred 0 | \n", "pred 1 | \n", "
|---|---|---|
| actual 0 | \n", "283 | \n", "5 | \n", "
| actual 1 | \n", "80 | \n", "6 | \n", "
\n", "\n", "## Refining the model and preparing the data" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "
\n",
"\n",
"### Challenge: finding reasonable model parameters\n",
"\n",
"The regularization constant, $C$, is unknown. Let's say we want to try some different orders of magnitude as well experimenting with L1 regularization (default for `sklearn` logistic regression is L2):\n",
"\n",
"$$C \\in \\{0.01, 0.1, 1, 10, 100\\}$$\n",
"$$\\textrm{penalty} \\in \\{\\textrm{L1}, \\textrm{L2}\\}$$\n",
"\n",
"We know that the outcome can be different depending on what settings we choose."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"from IPython.html.widgets import interact\n",
"\n",
"from sklearn.metrics import roc_curve, auc\n",
"\n",
"def plot_roc_curve(y_test, probas):\n",
" # Compute ROC curve and area the curve\n",
" fpr, tpr, thresholds = roc_curve(y_test, probas[:, 1])\n",
" roc_auc = auc(fpr, tpr)\n",
" \n",
" # Plot ROC curve\n",
" plt.clf()\n",
" plt.plot(fpr, tpr, label='ROC curve (AUC = %0.2f)' % roc_auc)\n",
" plt.plot([0, 1], [0, 1], 'k--')\n",
" plt.xlim([0.0, 1.0])\n",
" plt.ylim([0.0, 1.0])\n",
" plt.xlabel('False Positive Rate', fontsize=14)\n",
" plt.ylabel('True Positive Rate', fontsize=14)\n",
" plt.legend(loc=\"lower right\", fontsize=20)\n",
" plt.show()\n",
"\n",
"def fit_model(penalty, C):\n",
" clf = LogisticRegression(penalty=penalty, C=C)\n",
" clf.fit(X_train, y_train)\n",
" plot_roc_curve(y_test, clf._predict_proba_lr(X_test))\n",
" \n",
"interact(fit_model, penalty=('l2','l1'), C=(0.01, 100, 1))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAmkAAAH0CAYAAABmYbrSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSJI0IYQQQggvJEmaEEIIIYQX+j/qgjzfD2PTSwAAAABJRU5ErkJggg==\n",
"text": [
"\n",
"\n",
"### Common solution: grid search\n",
"\n",
"From `sklearn`, we get built in grid search for exploring the space. Using grid search cross validation, we can try all combinations, and automatically keep the most performant."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"from sklearn.grid_search import GridSearchCV\n",
"\n",
"params_to_try = {\n",
" 'C': [0.01, 0.1, 1, 10, 100, 100],\n",
" 'penalty': ['l1', 'l2']\n",
"}\n",
"\n",
"gs = GridSearchCV(clf, param_grid=params_to_try, cv=5)\n",
"gs.fit(X, y)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 33,
"text": [
"GridSearchCV(cv=5,\n",
" estimator=LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n",
" intercept_scaling=1, penalty='l2', random_state=None, tol=0.0001),\n",
" fit_params={}, iid=True, loss_func=None, n_jobs=1,\n",
" param_grid={'penalty': ['l1', 'l2'], 'C': [0.01, 0.1, 1, 10, 100, 100]},\n",
" pre_dispatch='2*n_jobs', refit=True, score_func=None, scoring=None,\n",
" verbose=0)"
]
}
],
"prompt_number": 33
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"print \"Best parameters:\", gs.best_params_\n",
"print \"Best score:\", gs.best_score_"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"Best parameters: {'penalty': 'l1', 'C': 0.01}\n",
"Best score: 0.791443850267\n"
]
}
],
"prompt_number": 34
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"\n",
"\n",
"### Challenge: data transformations often needed\n",
"\n",
"One thing we might want to do is scale all of the data before fitting any models. As the [sklearn docs](http://scikit-learn.org/stable/modules/preprocessing.html#standardization-or-mean-removal-and-variance-scaling) point out:\n",
"\n",
"Standardization of datasets is a common requirement for many machine learning estimators implemented in the scikit: they might behave badly if the individual feature do not more or less look like standard normally distributed data: Gaussian with zero mean and unit variance.\n", "\n", "
\n",
"\n",
"### Common solution: use plug-and-play data prep tools\n",
"\n",
"It's painful and error prone to wing it every time we want to do common transformations. The `sklearn` toolbox comes with many useful tools for doing this in a consistent way.\n",
"\n",
"#### Scaling"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"from sklearn.preprocessing import StandardScaler\n",
"\n",
"scaler = StandardScaler()\n",
"\n",
"X_train_standardized = scaler.fit_transform(X_train.astype(np.float))\n",
"X_train_standardized"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 35,
"text": [
"array([[ 1.15376999, 0.35655147, 0.35655147, 2.38280942],\n",
" [-0.56644764, 6.41792654, 6.41792654, 2.50245107],\n",
" [ 0.22749896, -0.60050775, -0.60050775, -0.9671566 ],\n",
" ..., \n",
" [-0.03714991, -0.12197814, -0.12197814, -0.64811222],\n",
" [ 0.62447226, -0.44099788, -0.44099788, -0.01002345],\n",
" [-0.83109651, 0.03753173, 0.03753173, -0.56835112]])"
]
}
],
"prompt_number": 35
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"print 'column means:', np.round(X_train_standardized.mean(axis=0))\n",
"print 'column variances:', np.round(X_train_standardized.var(axis=0))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"column means: [-0. 0. 0. 0.]\n",
"column variances: [ 1. 1. 1. 1.]\n"
]
}
],
"prompt_number": 36
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Now that the scaler is *fit* on the training data, we can use it to *transform* new data:"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"X_new = np.array([25., 35., 9200., 90.])\n",
"\n",
"scaler.transform(X_new)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "pyout",
"prompt_number": 37,
"text": [
"array([ 2.08004102, 4.66331797, 4.95043573, 2.18340669])"
]
}
],
"prompt_number": 37
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Dimensionality reduction\n",
"\n",
"We could also reduce the dimensionality with [PCA](http://en.wikipedia.org/wiki/Principal_component_analysis) and [whiten](http://en.wikipedia.org/wiki/Whitening_transformation) the data:"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"from sklearn.decomposition import PCA\n",
"\n",
"# instantiate the PCA transformation object\n",
"pca = PCA(n_components=2, whiten=True)\n",
"\n",
"# fit the PCA object on and transform the training data\n",
"X_train_pca = pca.fit_transform(X_train_standardized)\n",
"\n",
"# create a 3d figure\n",
"fig = plt.figure()\n",
"ax = fig.add_subplot(111)\n",
"\n",
"# scatterplot the PCA points\n",
"ax.scatter(*np.hsplit(X_train_pca, 2), c=y_train, s=40,\n",
" cmap='coolwarm')\n",
"\n",
"# annotate and show the figure\n",
"ax.set_xlabel('component 1')\n",
"ax.set_ylabel('component 2')\n",
"\n",
"plt.show()"
],
"language": "python",
"metadata": {},
"outputs": [
{
"metadata": {},
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAmQAAAH1CAYAAABYw+LHAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nMXUTAAAAAElFTkSuQmCC\n",
"text": [
"\n",
"\n",
"### Challenge: all the different steps add up into a big jumbled mess\n",
"\n",
"We have all seen the data science version of this:\n",
"\n",
"\n",
"\n",
"But there's no reason for **data_aug_13_v2_scaled_pca_3_dims_WHITENED_plus_some_tweaks_final_V4_FINAL.csv**, right?\n",
"\n",
"We shouldn't have to do that."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"\n",
"\n",
"### Common solution: putting all the steps together with a `Pipeline`\n",
"\n",
"One big takeaway: **it's all about documented, repeatable data-to-model pipelines**.\n",
"\n",
"At some point, especially for tasks we will be doing more than once, we might want to codify all of the exploratory work we've done. The standard way to express a beginning-to-end data transformation and model fitting workflow in `sklearn` is the [Pipeline](http://scikit-learn.org/stable/modules/generated/sklearn.pipeline.Pipeline.html#sklearn.pipeline.Pipeline). "
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"from sklearn.pipeline import Pipeline\n",
"\n",
"# define a pipeline with some transforms and a simple classifier\n",
"pipeline = Pipeline([\n",
" ('scale', StandardScaler()),\n",
" ('reduce_dim', PCA()),\n",
" ('clf', LogisticRegression()),\n",
"])\n",
"\n",
"# enumerate all of the different settings we wish to try out\n",
"parameters = {\n",
" 'reduce_dim__n_components': (1, 2, 3),\n",
" 'reduce_dim__whiten': (True, False),\n",
" 'clf__penalty': ('l1', 'l2'),\n",
" 'clf__C': (1e-3, 1e-2, 1, 1e1, 1e2, 1e3),\n",
"}\n",
"\n",
"# grid search the parameter space\n",
"grid_search = GridSearchCV(pipeline, parameters, n_jobs=1, verbose=1)\n",
"\n",
"print(\"Performing grid search...\\n\")\n",
"print(\"pipeline:\", [name for name, _ in pipeline.steps])\n",
"print(\"parameters:\")\n",
"print(parameters)\n",
"grid_search.fit(X, y)\n",
"\n",
"print(\"Best score: %0.3f\" % grid_search.best_score_)\n",
"print(\"Best parameters set:\")\n",
"best_parameters = grid_search.best_estimator_.get_params()\n",
"for param_name in sorted(parameters.keys()):\n",
" print(\"\\t%s: %r\" % (param_name, best_parameters[param_name]))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stderr",
"text": [
"[Parallel(n_jobs=1)]: Done 1 jobs | elapsed: 0.0s\n",
"[Parallel(n_jobs=1)]: Done 50 jobs | elapsed: 0.2s\n"
]
},
{
"output_type": "stream",
"stream": "stderr",
"text": [
"[Parallel(n_jobs=1)]: Done 200 jobs | elapsed: 0.7s\n",
"[Parallel(n_jobs=1)]: Done 216 out of 216 | elapsed: 0.8s finished\n"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"Performing grid search...\n",
"\n",
"('pipeline:', ['scale', 'reduce_dim', 'clf'])\n",
"parameters:\n",
"{'clf__penalty': ('l1', 'l2'), 'clf__C': (0.001, 0.01, 1, 10.0, 100.0, 1000.0), 'reduce_dim__whiten': (True, False), 'reduce_dim__n_components': (1, 2, 3)}\n",
"Fitting 3 folds for each of 72 candidates, totalling 216 fits\n",
"Best score: 0.777"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n",
"Best parameters set:\n",
"\tclf__C: 0.01\n",
"\tclf__penalty: 'l2'\n",
"\treduce_dim__n_components: 2\n",
"\treduce_dim__whiten: True\n"
]
}
],
"prompt_number": 39
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"\n",
"\n",
"## Takeaways for using `sklearn` in the real world\n",
"\n",
"* API consistency is the \"killer feature\" of `sklearn`\n",
"* Treat the original data as immutable, everything later is a view into or a transformation of the original and **all the steps are obvious**\n",
"* To that end, building transparent and repeatable dataflows using Pipelines cuts down on black magic"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"\n", "\n", "## Public service announcement: using other tools\n", "\n", "Why choose? Use whichever one is best for the task at hand." ] }, { "cell_type": "code", "collapsed": false, "input": [ "# load the %R cell magic extension\n", "%load_ext rmagic\n", "\n", "# send the dataframe over to the R instance\n", "%Rpush df" ], "language": "python", "metadata": {}, "outputs": [], "prompt_number": 40 }, { "cell_type": "code", "collapsed": false, "input": [ "%%R\n", "library(ggplot2)\n", "qplot(log(time), log(cc), data=df, color=donated)" ], "language": "python", "metadata": {}, "outputs": [ { "metadata": {}, "output_type": "display_data", "png": "iVBORw0KGgoAAAANSUhEUgAAAeAAAAHgCAIAAADytinCAAAgAElEQVR4nO3de5gU1YH38VPV09Nc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\nAElFTkSuQmCC\n" } ], "prompt_number": 41 }, { "cell_type": "code", "collapsed": false, "input": [ "%%R\n", "blood.glm <- glm(donated ~ log(cc) + log(time), data=df, family=\"binomial\")\n", "print(summary(blood.glm))\n", "\n", "par(mfrow=c(2, 2))\n", "plot(blood.glm)" ], "language": "python", "metadata": {}, "outputs": [ { "metadata": {}, "output_type": "display_data", "text": [ "\n", "Call:\n", "glm(formula = donated ~ log(cc) + log(time), family = \"binomial\", \n", " data = df)\n", "\n", "Deviance Residuals: \n", " Min 1Q Median 3Q Max \n", "-1.5214 -0.7665 -0.5352 -0.2782 2.5447 \n", "\n", "Coefficients:\n", " Estimate Std. Error z value Pr(>|z|) \n", "(Intercept) -7.9523 0.8749 -9.089 < 2e-16 ***\n", "log(cc) 1.4448 0.1693 8.535 < 2e-16 ***\n", "log(time) -1.0278 0.1454 -7.069 1.56e-12 ***\n", "---\n", "Signif. codes: 0 \u2018***\u2019 0.001 \u2018**\u2019 0.01 \u2018*\u2019 0.05 \u2018.\u2019 0.1 \u2018 \u2019 1\n", "\n", "(Dispersion parameter for binomial family taken to be 1)\n", "\n", " Null deviance: 820.89 on 747 degrees of freedom\n", "Residual deviance: 728.63 on 745 degrees of freedom\n", "AIC: 734.63\n", "\n", "Number of Fisher Scoring iterations: 5\n", "\n" ] }, { "metadata": {}, "output_type": "display_data", "png": "iVBORw0KGgoAAAANSUhEUgAAAeAAAAHgCAIAAADytinCAAAgAElEQVR4nOzdd1wUx/s48GevUo7e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\nlnNl6wIAAAAASUVORK5CYII=\n" } ], "prompt_number": 42 }, { "cell_type": "markdown", "metadata": {}, "source": [ "... and we can pull data back from R to Python:" ] }, { "cell_type": "code", "collapsed": false, "input": [ "r_coeffs = %R coef(blood.glm)\n", "r_coeffs" ], "language": "python", "metadata": {}, "outputs": [ { "metadata": {}, "output_type": "pyout", "prompt_number": 43, "text": [ "array([-7.95228363, 1.44481591, -1.02784654])" ] } ], "prompt_number": 43 }, { "cell_type": "markdown", "metadata": {}, "source": [ "Even better for those in the Julia community ... you can replicate this whole experience when working in an IJulia notebook:" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "In fact - IPython notebooks are not just for Python anymore. The core functionality has been generalized as Project Jupyter, and language kernels are available or under active development for Julia, Haskell, and R." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "\n", "\n", "
" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Project plug: data science competitions for social good!\n", "\n", "
\n",
"\n",
"
\n",
"\n",
"Currently in beta — launching our first for-prize competitions in the next few weeks.\n",
"\n",
"* Twitter: [@drivendataorg](https://twitter.com/drivendataorg)\n",
"* E-mail: [isaac@drivendata.org](mailto:isaac@drivendata.org)\n",
"\n",
""
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"[www.drivendata.org](http://www.drivendata.org)\n", "
\n", "\n", "# Questions?\n", "\n", "" ] } ], "metadata": {} } ] }