{"cells": [{"cell_type": "markdown", "metadata": {}, "source": ["# 2A.ml - Arbres de d\u00e9cision / Random Forest\n", "\n", "Classification, r\u00e9gression, visualisation avec des m\u00e9thodes ensemblistes (arbres, for\u00eats, ...)."]}, {"cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": ["%matplotlib inline"]}, {"cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": ["import matplotlib.pyplot as plt"]}, {"cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [{"data": {"text/html": ["
run previous cell, wait for 2 seconds
\n", ""], "text/plain": [""]}, "execution_count": 4, "metadata": {}, "output_type": "execute_result"}], "source": ["from jyquickhelper import add_notebook_menu\n", "add_notebook_menu()"]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Description du probl\u00e8me\n", "\n", "Le code suivant t\u00e9l\u00e9charge les donn\u00e9es n\u00e9cessaires [salaries2010.zip](http://www.xavierdupre.fr/enseignement/complements/salaries2010.zip)."]}, {"cell_type": "markdown", "metadata": {}, "source": ["Le machine learning peut se r\u00e9sumer \u00e0 la construction d'une fonction de pr\u00e9diction $Y=f(X) + \\epsilon$. $f$ est le plus souvent le r\u00e9sultat d'une minimisation de l'erreur $\\sum_i E(Y_i,f(X_i))$ o\u00f9 $(X_i,Y_i)$ est une liste de couples (features, cible). Les [arbres de d\u00e9cision](http://fr.wikipedia.org/wiki/Arbre_de_d%C3%A9cision) sont des mod\u00e8les assez faciles \u00e0 apprendre et ils ont l'avantage d'accepter des [features](http://en.wikipedia.org/wiki/Feature_%28machine_learning%29) continues et discr\u00e8tes. Pour cet exercice, on reprend la base des salari\u00e9s vu dans un pr\u00e9c\u00e9dent notebook et on va essayer de pr\u00e9dire le salaire en fonction de plus de variables que l'\u00e2ge ou le sexe :"]}, {"cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [{"data": {"text/html": ["
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
TRNETTOTAGESEXEDEPTDEPRTYP_EMPLOIPCSCSCONT_TRAVCONV_COLLVARIABLEMODALITEMODLIBELLEmontant
01450.01972O628G62ZZZTRNETTOT1418 000 \u00e0 19 999 euros18999.5
11441.017575O354C35CDD1734TRNETTOT1418 000 \u00e0 19 999 euros18999.5
21429.017575O373C37CDD0014TRNETTOT1418 000 \u00e0 19 999 euros18999.5
31430.017575O651A65CDD9999TRNETTOT1418 000 \u00e0 19 999 euros18999.5
41455.017892O623E62ZZZTRNETTOT1418 000 \u00e0 19 999 euros18999.5
\n", "
"], "text/plain": [" TRNETTOT AGE SEXE DEPT DEPR TYP_EMPLOI PCS CS CONT_TRAV CONV_COLL \\\n", "0 14 50.0 1 972 O 628G 62 ZZZ \n", "1 14 41.0 1 75 75 O 354C 35 CDD 1734 \n", "2 14 29.0 1 75 75 O 373C 37 CDD 0014 \n", "3 14 30.0 1 75 75 O 651A 65 CDD 9999 \n", "4 14 55.0 1 78 92 O 623E 62 ZZZ \n", "\n", " VARIABLE MODALITE MODLIBELLE montant \n", "0 TRNETTOT 14 18 000 \u00e0 19 999 euros 18999.5 \n", "1 TRNETTOT 14 18 000 \u00e0 19 999 euros 18999.5 \n", "2 TRNETTOT 14 18 000 \u00e0 19 999 euros 18999.5 \n", "3 TRNETTOT 14 18 000 \u00e0 19 999 euros 18999.5 \n", "4 TRNETTOT 14 18 000 \u00e0 19 999 euros 18999.5 "]}, "execution_count": 5, "metadata": {}, "output_type": "execute_result"}], "source": ["import os\n", "if not os.path.exists(\"salaries2010.db3\"):\n", " import pyensae.datasource\n", " db3 = pyensae.datasource.download_data(\"salaries2010.zip\")\n", "\n", "import sqlite3, pandas\n", "con = sqlite3.connect(\"salaries2010.db3\")\n", "df = pandas.io.sql.read_sql(\"select * from varmod\", con)\n", "con.close()\n", "\n", "values = df[ df.VARIABLE == \"TRNETTOT\"].copy()\n", "\n", "def process_intervalle(s):\n", " if \"euros et plus\" in s : \n", " return float ( s.replace(\"euros et plus\", \"\").replace(\" \",\"\") )\n", " spl = s.split(\"\u00e0\")\n", " if len(spl) == 2 : \n", " s1 = spl[0].replace(\"Moins de\",\"\").replace(\"euros\",\"\").replace(\" \",\"\")\n", " s2 = spl[1].replace(\"Moins de\",\"\").replace(\"euros\",\"\").replace(\" \",\"\")\n", " return (float(s1)+float(s2))/2\n", " else : \n", " s = spl[0].replace(\"Moins de\",\"\").replace(\"euros\",\"\").replace(\" \",\"\")\n", " return float(s)/2\n", "\n", "values[\"montant\"] = values.apply(lambda r : process_intervalle(r [\"MODLIBELLE\"]), axis = 1)\n", "\n", "con = sqlite3.connect(\"salaries2010.db3\")\n", "data = pandas.io.sql.read_sql(\"select TRNETTOT,AGE,SEXE,DEPT,DEPR,TYP_EMPLOI,PCS,CS,CONT_TRAV,CONV_COLL from salaries\", con)\n", "con.close()\n", "\n", "salaires = data.merge ( values, left_on = \"TRNETTOT\", right_on=\"MODALITE\" )\n", "salaires.dropna(inplace=True)\n", "salaires.head()"]}, {"cell_type": "markdown", "metadata": {}, "source": ["Le module scikit-learn n'accepte pas les features sous forme de cha\u00eenes de caract\u00e8res :\n", "* [Encoding categorical features](http://scikit-learn.org/stable/modules/feature_extraction.html#dict-feature-extraction)\n", "* [Loading features from dicts](http://scikit-learn.org/stable/modules/feature_extraction.html#loading-features-from-dicts)\n", "* [Vectorizing a Pandas dataframe for Scikit-Learn](http://stackoverflow.com/questions/20024584/vectorizing-a-pandas-dataframe-for-scikit-learn)\n", "\n", "Il faut transformer les variables qui ne sont pas num\u00e9riques (et non ordonn\u00e9es) en variables bool\u00e9ennes (on fait cela sur un \u00e9chantillon d'abord) :"]}, {"cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [], "source": ["import random\n", "salaires[\"rnd\"] = salaires.apply (lambda r : random.randint(0,50),axis=1)\n", "ech = salaires [ salaires.rnd == 0 ]"]}, {"cell_type": "markdown", "metadata": {}, "source": ["La taille de l'\u00e9chantillon doit \u00eatre ajust\u00e9e en fonction de la m\u00e9moire de l'ordinateur et il est aussi pr\u00e9f\u00e9rable de commencer avec un \u00e9chantillon petit. Le d\u00e9veloppement du mod\u00e8le prend moins de temps. On agrandit la taille de l'\u00e9chantillon quand tout fonctionne bien (on perd souvent pas mal de temps parce que le type d'une variable n'est pas celui attendu, qu'on s'est tromp\u00e9 de nom, qu'une valeur est manquante...)."]}, {"cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [{"data": {"text/plain": ["(43111, 4)"]}, "execution_count": 7, "metadata": {}, "output_type": "execute_result"}], "source": ["X,Y = ech[[\"AGE\",\"SEXE\",\"TYP_EMPLOI\",\"CONT_TRAV\"]], ech[[\"montant\"]]\n", "Xd = X.T.to_dict().values()\n", "X.shape"]}, {"cell_type": "markdown", "metadata": {}, "source": ["On transforme les variables sous forme de cha\u00eenes de caract\u00e8res en variables binaires :"]}, {"cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [], "source": ["from sklearn.feature_extraction import DictVectorizer\n", "prep = DictVectorizer()\n", "Xt = prep.fit_transform(Xd).toarray()"]}, {"cell_type": "markdown", "metadata": {}, "source": ["``Xt`` est un [numpy.ndarray](http://docs.scipy.org/doc/numpy/reference/generated/numpy.ndarray.html) mais la variable ``prep`` a conserv\u00e9 le nom des features."]}, {"cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [{"data": {"text/plain": ["['AGE',\n", " 'CONT_TRAV=APP',\n", " 'CONT_TRAV=AUT',\n", " 'CONT_TRAV=CDD',\n", " 'CONT_TRAV=CDI',\n", " 'CONT_TRAV=TTP',\n", " 'CONT_TRAV=ZZZ',\n", " 'SEXE=',\n", " 'SEXE=1',\n", " 'SEXE=2',\n", " 'TYP_EMPLOI=A',\n", " 'TYP_EMPLOI=O',\n", " 'TYP_EMPLOI=X']"]}, "execution_count": 9, "metadata": {}, "output_type": "execute_result"}], "source": ["prep.feature_names_"]}, {"cell_type": "markdown", "metadata": {}, "source": ["**Remarque :** On transforme une variable cat\u00e9gorielle en une s\u00e9rie de variables bool\u00e9ennes mais lorsque les cat\u00e9gories sont exclusives, une observation est n\u00e9cessairement dans l'une d'elles. La somme des variables bool\u00e9ennes qui en d\u00e9coulent est \u00e9gale \u00e0 1. Cela revient \u00e0 cr\u00e9er une s\u00e9ries de variables dont la somme est corr\u00e9l\u00e9e \u00e0 une constante : ce cas est \u00e0 \u00e9viter lors d'un mod\u00e8le lin\u00e9aire comme la r\u00e9gression. Il faut enlever une variable. Comme on cale un arbre de d\u00e9cision par la suite, ce n'est pas indispensable."]}, {"cell_type": "markdown", "metadata": {}, "source": ["On entra\u00eene l'arbre, on limite la profondeur \u00e0 3 histoire de pouvoir visualiser l'arbre r\u00e9sultant. Ce n'est certainement pas assez puisque $2^3=8$ = le nombre de feuilles est un nombre inf\u00e9rieure au nombres de tranches de salaires possibles."]}, {"cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [{"data": {"text/plain": ["0.21069611975253544"]}, "execution_count": 10, "metadata": {}, "output_type": "execute_result"}], "source": ["from sklearn.tree import DecisionTreeRegressor\n", "clf = DecisionTreeRegressor(min_samples_leaf=10, max_depth=3)\n", "clf = clf.fit(Xt,Y)\n", "clf.score(Xt,Y)"]}, {"cell_type": "markdown", "metadata": {}, "source": ["On repr\u00e9sente l'arbre de d\u00e9cision (et \u00e7a devient un peu complexe) :"]}, {"cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [], "source": ["from sklearn.tree import export_graphviz\n", "export_graphviz(clf, out_file=\"arbre.dot\") "]}, {"cell_type": "markdown", "metadata": {}, "source": ["Pour visualiser l'arbre, il faut installer [graphviz](http://www.graphviz.org/) et lancer la commande (il faudra sans doute remplacer le chemin vers votre installation de Graphviz)."]}, {"cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [], "source": ["import sys\n", "cwd = os.getcwd()\n", "if sys.platform.startswith(\"win\"):\n", " exe = 'C:\\\\Program Files (x86)\\\\Graphviz2.38\\\\bin\\\\dot.exe'\n", " if not os.path.exists(exe):\n", " raise FileNotFoundError(exe)\n", " exe = '\"{0}\"'.format(exe)\n", "else:\n", " exe = \"dot\"\n", "cmd = '\"{0}\" -Tpng {1}\\\\arbre.dot -o {1}\\\\arbre.png'.format(exe, cwd)"]}, {"cell_type": "code", "execution_count": 12, "metadata": {}, "outputs": [{"data": {"text/plain": ["0"]}, "execution_count": 13, "metadata": {}, "output_type": "execute_result"}], "source": ["os.system(cmd)"]}, {"cell_type": "code", "execution_count": 13, "metadata": {}, "outputs": [{"data": {"image/png": "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\n", "text/plain": [""]}, "execution_count": 14, "metadata": {}, "output_type": "execute_result"}], "source": ["from IPython.core.display import Image\n", "Image(\"arbre.png\")"]}, {"cell_type": "markdown", "metadata": {}, "source": ["Dans un notebook, le javascript peut \u00eatre utilis\u00e9 pour tracer de le graphe (voir [Visualiser un arbre de d\u00e9cision](http://www.xavierdupre.fr/app/papierstat/helpsphinx/notebooks/decision_tree_visualization.html))."]}, {"cell_type": "code", "execution_count": 14, "metadata": {}, "outputs": [{"data": {"text/html": ["
\n", ""], "text/plain": [""]}, "execution_count": 15, "metadata": {}, "output_type": "execute_result"}], "source": ["from jyquickhelper import RenderJsVis\n", "dot = export_graphviz(clf, out_file=None, feature_names=prep.feature_names_)\n", "RenderJsVis(dot=dot, height=\"400px\", layout='hierarchical')"]}, {"cell_type": "markdown", "metadata": {}, "source": ["Oups... j'ai oubli\u00e9 de s\u00e9parer base d'apprentissage et base de test. Il ne restera plus qu'\u00e0 tracer la courbe ROC : [Receiver operating characteristic (ROC)](http://scikit-learn.org/0.11/auto_examples/plot_roc.html)."]}, {"cell_type": "markdown", "metadata": {}, "source": ["### Exercice 1 : Bases d'apprentissage, test, courbes\n", "\n", "A vous de jouer. Quelques id\u00e9es :\n", "\n", "* [train_test_split](http://scikit-learn.org/stable/modules/generated/sklearn.model_selection.train_test_split.html)\n", "* [random forest](http://blog.yhathq.com/posts/random-forests-in-python.html)"]}, {"cell_type": "code", "execution_count": 15, "metadata": {}, "outputs": [], "source": []}, {"cell_type": "markdown", "metadata": {}, "source": ["### Exercice 2 : Courbes ROC\n", " \n", "On retourne le probl\u00e8me, on essaye de pr\u00e9voir le sexe en fonction des autres variables dont le salaire.\n", "\n", "* [RandomForestClassifier](http://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestClassifier.html)"]}, {"cell_type": "code", "execution_count": 16, "metadata": {}, "outputs": [], "source": []}], "metadata": {"kernelspec": {"display_name": "Python 3", "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": "3.6.4"}}, "nbformat": 4, "nbformat_minor": 2}