{"nbformat_minor": 0, "cells": [{"source": "# Hyper Parameter Tuning with sklearn (GridSearch)\n\u200b\nThe objective of any learning algorithm $A$ is to find a function $f$ that minimizes some expected loss. A learning algorithm produces $f$ through the optimization of a training criterion with respect to a set of parameters $\u03b8$. Most of the time the actual learning algorithm itself will be having some hyper-parameters. The actual learning algorithm is the one which obtained after choosing lambda. The process is described the below given algorithm. ", "cell_type": "markdown", "metadata": {}}, {"execution_count": 1, "cell_type": "code", "source": "from IPython.display import Image\nImage(\"http://topepo.github.io/caret/TrainAlgo.png\")", "outputs": [{"execution_count": 1, "output_type": "execute_result", "data": {"image/png": "iVBORw0KGgoAAAANSUhEUgAAAnUAAADqCAYAAADEWLPrAAAACXBIWXMAAAsTAAALEwEAmpwYAABA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\nB5YIz3gWgfoeAAAAAElFTkSuQmCC\n", "text/plain": ""}, "metadata": {}}], "metadata": {"collapsed": false, "trusted": true}}, {"source": "In this notebook we will discuss how to apply the hyper-parameter tuning with scikit-learn.Let's play with this using the IRIS data-set.\n\nLet's explore.", "cell_type": "markdown", "metadata": {}}, {"execution_count": 43, "cell_type": "code", "source": "from sklearn.datasets import load_iris\nfrom sklearn.svm import SVC\nfrom sklearn.grid_search import GridSearchCV\nfrom sklearn.metrics import f1_score\nfrom sklearn.metrics import classification_report\n\n\niris = load_iris()\n\nsvc = SVC(kernel=\"poly\")\nsvo = SVC(kernel=\"poly\").fit(iris.data, iris.target)\n\nparameters = {\"C\": [1,2,4,8],\"kernel\": [\"poly\",\"rbf\"],\"degree\":[1, 2, 3, 4]}\n\nclf = GridSearchCV(svc, param_grid=parameters,score_func=f1_score,refit=True)\n\nclf.fit(iris.data, iris.target)\n\n\n\nprint model_tunning.best_score_\nprint model_tunning.best_params_\nprint classification_report(iris.target,clf.predict(iris.data))\nprint classification_report(iris.target,svo.predict(iris.data))\n", "outputs": [{"output_type": "stream", "name": "stdout", "text": "0.986666666667\n{'kernel': 'rbf', 'C': 2, 'degree': 1}\n precision recall f1-score support\n\n 0 1.00 1.00 1.00 50\n 1 1.00 0.98 0.99 50\n 2 0.98 1.00 0.99 50\n\navg / total 0.99 0.99 0.99 150\n\n precision recall f1-score support\n\n 0 1.00 1.00 1.00 50\n 1 0.98 0.96 0.97 50\n 2 0.96 0.98 0.97 50\n\navg / total 0.98 0.98 0.98 150\n\n"}], "metadata": {"collapsed": false, "trusted": true}}, {"source": "### What just happened ?\nIn the above code we are trying to build a classier to classify the IRIS data with Support Vector Machine. We have created two svm objects one is for fitting the entire data without hyper parameter tuning and other for tuning. Then we created a set of parameters. The svm object and parameter is passed to $GridSearchCV$ class for fitting the classifier. After fitting the model we observed that the best estimator give a score of $0.986666666667$ and the best parameters as ${'kernel': 'rbf', 'C': 2, 'degree': 1}$. The classification report says that the tuned classifier gives and F-1 score of $0.99$ and the other produces $0.98$\n\n### Applying pre-processing to the grid\nNow lets see how we can apply some pre-processing to the tuning. The one easy way is to pre-process the entire training set and then pass it to the grid. The second option is to pass the pre-processor to the grid. We will try to use the Standard Scalar with IRIS data.", "cell_type": "markdown", "metadata": {}}, {"execution_count": 56, "cell_type": "code", "source": "from sklearn.preprocessing import StandardScaler\nfrom sklearn.pipeline import make_pipeline\n\npipe_line = make_pipeline(StandardScaler(),SVC())\n\nparam_grid = {\"standardscaler__copy\":[True,False],\"standardscaler__with_mean\":[True,False],\\\n\"standardscaler__with_std\":[True,False],\"svc__C\": [1,2,4,8],\"svc__kernel\": [\"poly\",\"rbf\"],\\\n\"svc__degree\":[1, 2, 3, 4]}\n\n\nclf_p = GridSearchCV(pipe_line, param_grid=param_grid,score_func=f1_score,refit=True)\n\nclf_p.fit(iris.data, iris.target)\n\nprint clf_p.best_score_\nprint clf_p.best_params_\nprint classification_report(iris.target,clf_p.predict(iris.data))\n", "outputs": [{"output_type": "stream", "name": "stdout", "text": "0.986666666667\n{'standardscaler__with_mean': True, 'svc__C': 2, 'standardscaler__copy': True, 'svc__degree': 1, 'svc__kernel': 'poly', 'standardscaler__with_std': True}\n precision recall f1-score support\n\n 0 1.00 1.00 1.00 50\n 1 0.98 0.96 0.97 50\n 2 0.96 0.98 0.97 50\n\navg / total 0.98 0.98 0.98 150\n\n"}], "metadata": {"collapsed": false, "trusted": true}}, {"source": "### What just happened ?\n\nIn the above example we have we have applied the StandardScalar to the data set and applied to the training data. While the cross validated and tuned training in progress, each train and test split will be scaled with the specified parameters. For the grid search we have created a pipe-line. For the pipe line we have created the parameters. Parameter for each item in the pipe-line is specified after name of element name and doube under score like \"svc__C\". \n\nThe example was all about classifiers. In the same way we can use the GridSearch with any Regressor class available in scikit-learn.\n\n### Clustering with GridSearch\n\nThe GridSearch in sklearn supports clustering too. But it won't be having the \".labels_\" attribute available. Let's see an example with IRIS data.", "cell_type": "markdown", "metadata": {}}, {"execution_count": 62, "cell_type": "code", "source": "from sklearn.cluster import KMeans\n\nkmeans_params = {'n_clusters':[3,6],'init':['k-means++','random'],\\\n'precompute_distances':[True,False],'random_state':[0,1],'n_init':[5,10]}\n\nkmean_grid = GridSearchCV(KMeans(),kmeans_params,refit=True)\n\nkmean_grid.fit(iris.data)\n\nprint kmean_grid.best_score_\nprint kmean_grid.best_estimator_", "outputs": [{"output_type": "stream", "name": "stdout", "text": "-143.81758994\nKMeans(copy_x=True, init='k-means++', max_iter=300, n_clusters=6, n_init=5,\n n_jobs=1, precompute_distances=True, random_state=1, tol=0.0001,\n verbose=0)\n"}], "metadata": {"collapsed": false, "trusted": true}}, {"source": "### Conclusion \nIn this note we had a short walk though of how to perform hyper parameter tuning in scikit-learn. Try it and enjoy.\n\nHappy Hacking !!!", "cell_type": "markdown", "metadata": {}}], "nbformat": 4, "metadata": {"kernelspec": {"display_name": "Python 2", "name": "python2", "language": "python"}, "language_info": {"mimetype": "text/x-python", "nbconvert_exporter": "python", "version": "2.7.9", "name": "python", "file_extension": ".py", "pygments_lexer": "ipython2", "codemirror_mode": {"version": 2, "name": "ipython"}}}}