diff --git a/docs/learning.ipynb b/docs/learning.ipynb index 96aaa69..4469b47 100644 --- a/docs/learning.ipynb +++ b/docs/learning.ipynb @@ -2,19 +2,10 @@ "cells": [ { "cell_type": "code", - "execution_count": 31, + "execution_count": 1, "id": "faafb9a0", "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "The autoreload extension is already loaded. To reload it, use:\n", - " %reload_ext autoreload\n" - ] - } - ], + "outputs": [], "source": [ "%matplotlib inline\n", "%load_ext autoreload\n", @@ -110,7 +101,7 @@ }, { "cell_type": "code", - "execution_count": 32, + "execution_count": 2, "id": "8342340f", "metadata": {}, "outputs": [], @@ -135,7 +126,7 @@ }, { "cell_type": "code", - "execution_count": 33, + "execution_count": 3, "id": "9dfdc01f", "metadata": {}, "outputs": [], @@ -184,7 +175,7 @@ }, { "cell_type": "code", - "execution_count": 34, + "execution_count": 4, "id": "99de3ed7", "metadata": {}, "outputs": [ @@ -217,7 +208,7 @@ }, { "cell_type": "code", - "execution_count": 35, + "execution_count": 5, "id": "09eca16d", "metadata": {}, "outputs": [ @@ -255,7 +246,7 @@ }, { "cell_type": "code", - "execution_count": 36, + "execution_count": 6, "id": "b94a89f2", "metadata": {}, "outputs": [ @@ -332,7 +323,7 @@ }, { "cell_type": "code", - "execution_count": 37, + "execution_count": 7, "id": "4f1c169f", "metadata": {}, "outputs": [ @@ -369,7 +360,7 @@ }, { "cell_type": "code", - "execution_count": 38, + "execution_count": 8, "id": "91cedffb", "metadata": {}, "outputs": [ @@ -466,7 +457,7 @@ }, { "cell_type": "code", - "execution_count": 39, + "execution_count": 9, "id": "4c21cd56", "metadata": {}, "outputs": [ @@ -528,9 +519,9 @@ "text/markdown": [ "| Méthode | R² |\n", "| :---: | :---: |\n", - "|AD|0.511089|\n", - "|Normalisation + AD|0.508518|\n", - "|Standardisation + AD|0.512649|\n" + "|AD|0.510519|\n", + "|Normalisation + AD|0.508927|\n", + "|Standardisation + AD|0.506827|\n" ], "text/plain": [ "" @@ -543,7 +534,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "best score= 0.512649496806262 depth= 5 method= Standardisation + AD\n" + "best score= 0.5105192880780061 depth= 5 method= AD\n" ] } ], @@ -614,6 +605,14 @@ "Refaites l’expérience avec un paramètre n_neigbors=5. L’augmentation du nombre de voisins considérés permet-elle de meilleurs résultats ? Si cette amélioration n’est pas significative, nous conserverons n_neighbors = 4 pour la suite, par souci de simplicité." ] }, + { + "cell_type": "markdown", + "id": "88185c62", + "metadata": {}, + "source": [ + "Aucune amélioration signification par rapport au pré et post traitement" + ] + }, { "cell_type": "markdown", "id": "72869cf6", @@ -625,7 +624,7 @@ }, { "cell_type": "code", - "execution_count": 40, + "execution_count": 10, "id": "f72f499f", "metadata": {}, "outputs": [ @@ -641,9 +640,9 @@ "text/markdown": [ "| Méthode | R² |\n", "| :---: | :---: |\n", - "|KNN|0.485764|\n", - "|Normalisation + KNN|0.497729|\n", - "|Standardisation + KNN|0.489906|\n" + "|KNN|0.463459|\n", + "|Normalisation + KNN|0.465584|\n", + "|Standardisation + KNN|0.480575|\n" ], "text/plain": [ "" @@ -664,9 +663,9 @@ "text/markdown": [ "| Méthode | R² |\n", "| :---: | :---: |\n", - "|KNN|0.504888|\n", - "|Normalisation + KNN|0.500504|\n", - "|Standardisation + KNN|0.493475|\n" + "|KNN|0.492485|\n", + "|Normalisation + KNN|0.485298|\n", + "|Standardisation + KNN|0.495728|\n" ], "text/plain": [ "" @@ -679,7 +678,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "best score= 0.5048884037669835 neighbor= 5 method= KNN\n" + "best score= 0.49572822708745434 neighbor= 5 method= Standardisation + KNN\n" ] } ], @@ -707,7 +706,7 @@ " else make_pipeline(scaler(), KNeighborsRegressor(n_neighbors=n))\n", " )\n", " model.fit(X_train, y_train)\n", - " score: float = cross_val_score(model, X_train, y_train, cv=5).mean()\n", + " score: float = model.score(X_test, y_test)\n", " knn_table.ajoutligne(f\"{name}\", score)\n", "\n", " if score > best_score_knn:\n", @@ -749,7 +748,7 @@ }, { "cell_type": "code", - "execution_count": 41, + "execution_count": 11, "id": "9f764f3a", "metadata": {}, "outputs": [ @@ -759,8 +758,8 @@ "| Méthode | R² |\n", "| :---: | :---: |\n", "|LR|0.452908|\n", - "|AD|0.512649|\n", - "|KNN|0.504888|\n" + "|AD|0.510519|\n", + "|KNN|0.495728|\n" ], "text/plain": [ "" @@ -773,7 +772,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "best_model= AD best_scaler= StandardScaler\n" + "best_model= AD best_scaler= DecisionTreeRegressor\n" ] } ], @@ -813,7 +812,7 @@ }, { "cell_type": "code", - "execution_count": 42, + "execution_count": 12, "id": "9084e87e", "metadata": {}, "outputs": [ @@ -850,17 +849,17 @@ }, { "cell_type": "code", - "execution_count": 43, + "execution_count": 13, "id": "fdcdfb17", "metadata": {}, "outputs": [ { "data": { "text/plain": [ - "0.3938363073944231" + "0.39383630739442343" ] }, - "execution_count": 43, + "execution_count": 13, "metadata": {}, "output_type": "execute_result" } @@ -902,7 +901,7 @@ }, { "cell_type": "code", - "execution_count": 44, + "execution_count": 14, "id": "c4f6c27f", "metadata": {}, "outputs": [ @@ -912,7 +911,7 @@ "" ] }, - "execution_count": 44, + "execution_count": 14, "metadata": {}, "output_type": "execute_result" }, @@ -947,7 +946,7 @@ }, { "cell_type": "code", - "execution_count": 45, + "execution_count": 15, "id": "2e512052", "metadata": {}, "outputs": [ @@ -1134,7 +1133,7 @@ "App_Haut-Médoc -0.070383" ] }, - "execution_count": 45, + "execution_count": 15, "metadata": {}, "output_type": "execute_result" } @@ -1155,7 +1154,7 @@ }, { "cell_type": "code", - "execution_count": 46, + "execution_count": 16, "id": "4010aa12", "metadata": {}, "outputs": [ @@ -1165,7 +1164,7 @@ "0.4489472637581393" ] }, - "execution_count": 46, + "execution_count": 16, "metadata": {}, "output_type": "execute_result" } @@ -1213,7 +1212,7 @@ }, { "cell_type": "code", - "execution_count": 47, + "execution_count": 17, "id": "1510f763", "metadata": {}, "outputs": [ @@ -1222,9 +1221,9 @@ "text/markdown": [ "| Méthode | R² |\n", "| :---: | :---: |\n", - "|Random Forest|0.496806|\n", - "|Normalisation + RF|0.494283|\n", - "|Standardisation + RF|0.486890|\n" + "|Random Forest|0.492518|\n", + "|Normalisation + RF|0.491729|\n", + "|Standardisation + RF|0.483625|\n" ], "text/plain": [ ""