{ "cells": [ { "cell_type": "code", "execution_count": 1, "id": "465ef7c3", "metadata": { "ExecuteTime": { "end_time": "2023-01-10T06:48:15.893603Z", "start_time": "2023-01-10T06:48:14.076493Z" } }, "outputs": [], "source": [ "#import the packages\n", "\n", "import pandas as pd\n", "import numpy as np\n", "\n", "from sklearn.tree import DecisionTreeClassifier\n", "from sklearn.metrics import classification_report\n", "from sklearn.metrics import confusion_matrix\n", "\n", "from warnings import simplefilter\n", "simplefilter(action='ignore', category=FutureWarning)" ] }, { "cell_type": "code", "execution_count": 2, "id": "ef8db77c", "metadata": { "ExecuteTime": { "end_time": "2023-01-10T06:48:15.910189Z", "start_time": "2023-01-10T06:48:15.894567Z" } }, "outputs": [], "source": [ "#Data Pre-process \n", "#limit the data range and normalize the data\n", "def Dataprocess(Data):\n", " P_Data = np.clip(Data,0,255)\n", " M_Data = P_Data/255\n", " return M_Data" ] }, { "cell_type": "code", "execution_count": 3, "id": "d35b249e", "metadata": { "ExecuteTime": { "end_time": "2023-01-10T06:50:38.762856Z", "start_time": "2023-01-10T06:48:21.394739Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "best_score= 0.8082\n", "best_criterion entropy\n", "best_max_depth 15\n", "best_min_samples_leaf 2\n", "best_min_samples_split 2\n" ] } ], "source": [ "#This cell is to train and tune decision tree model to get the best performance\n", "#import the original data\n", "#import train data\n", "M_Train = pd.read_csv('fashion-mnist_train.csv')\n", "M_Train_array = np.array(M_Train)\n", "TrainMat = M_Train_array[:,1:785]\n", "TrainMat = Dataprocess(TrainMat)\n", "TrainLabel = M_Train_array[:,0]\n", "#import test data\n", "M_Test = pd.read_csv('fashion-mnist_test.csv')\n", "M_Test_array = np.array(M_Test)\n", "TestMat = M_Test_array[:,1:785]\n", "TestMat = Dataprocess(TestMat)\n", "TestLabel = M_Test_array[:,0]\n", "\n", "best_score = 0\n", "#A nested for-loop is defined to traverse four important parameters to get the best decision tree model\n", "#criterion\n", "for c in ['gini','entropy']:\n", " #max_depth\n", " for j in range(5,20):\n", " #min_samples_leaf\n", " for m in (2,10):\n", " #min_samples_split\n", " for n in (2,10):\n", " #create decision tree model\n", " classifier = DecisionTreeClassifier(criterion=c,max_depth=j,min_samples_leaf=m,min_samples_split=n,max_features='auto',random_state=0)\n", " #train the model\n", " classifier.fit(TrainMat, TrainLabel)\n", " #Test accuracy\n", " score = classifier.score(TestMat,TestLabel)\n", " #If the accuracy rate this time is greater than the previous maximum accuracy rate, the parameter values of this time will be stored\n", " if score >best_score:\n", " best_criterion = c\n", " best_score = score\n", " best_max_depth = j\n", " best_min_samples_leaf =m\n", " best_min_samples_split =n\n", "#print the best parameter values\n", "print('best_score=',best_score)\n", "print('best_criterion',best_criterion)\n", "print('best_max_depth',best_max_depth)\n", "print('best_min_samples_leaf',best_min_samples_leaf)\n", "print('best_min_samples_split',best_min_samples_split)" ] }, { "cell_type": "code", "execution_count": 4, "id": "fe3a7aa2", "metadata": { "ExecuteTime": { "end_time": "2023-01-10T06:52:06.501049Z", "start_time": "2023-01-10T06:52:04.703580Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Confusion Matrix: \n", "[[782 5 27 57 10 1 108 0 10 0]\n", " [ 14 942 7 24 3 2 5 0 3 0]\n", " [ 20 3 720 17 135 0 91 0 14 0]\n", " [ 47 33 13 832 53 0 19 0 3 0]\n", " [ 5 4 116 33 751 5 82 0 4 0]\n", " [ 1 1 0 2 1 864 1 89 9 32]\n", " [180 7 105 52 121 0 518 0 17 0]\n", " [ 0 0 1 0 0 64 0 866 5 64]\n", " [ 12 5 19 6 10 10 27 1 907 3]\n", " [ 0 0 0 1 0 22 1 69 7 900]]\n", "\n", "\n", " precision recall f1-score support\n", "\n", " 0 0.74 0.78 0.76 1000\n", " 1 0.94 0.94 0.94 1000\n", " 2 0.71 0.72 0.72 1000\n", " 3 0.81 0.83 0.82 1000\n", " 4 0.69 0.75 0.72 1000\n", " 5 0.89 0.86 0.88 1000\n", " 6 0.61 0.52 0.56 1000\n", " 7 0.84 0.87 0.86 1000\n", " 8 0.93 0.91 0.92 1000\n", " 9 0.90 0.90 0.90 1000\n", "\n", " accuracy 0.81 10000\n", " macro avg 0.81 0.81 0.81 10000\n", "weighted avg 0.81 0.81 0.81 10000\n", "\n", "mean accuracy= 0.8082\n" ] } ], "source": [ "#This cell is to evaluate the best Decision tree model\n", "\n", "#use the best parameter values to create the Decision tree model\n", "DT = DecisionTreeClassifier(criterion='entropy',max_depth=15,max_features='auto',min_samples_leaf=2,min_samples_split=2,random_state=0)\n", "#train\n", "DT.fit(TrainMat,TrainLabel)\n", "#Calculate the confusion matrix\n", "pred=DT.predict(TestMat)\n", "cm = confusion_matrix(y_true=TestLabel, y_pred=pred)\n", "\n", "# Print confusion matrix\n", "print(\"Confusion Matrix: \")\n", "print(cm)\n", "print('\\n')\n", "# Print precision ,recall and accuracy\n", "print(classification_report(TestLabel,pred))\n", "print('mean accuracy=', DT.score(TestMat,TestLabel))\n", " " ] } ], "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": "3.10.8" } }, "nbformat": 4, "nbformat_minor": 5 }