機器學習——決策樹

決策樹是一種用於分類和回歸的非參數監督學習方法。目標是創建一個模型，通過從數據特性中推導出簡單的決策規則來預測目標變量的值

導入類庫

import numpy as np
import pandas as pd
from sklearn.feature_extraction import DictVectorizer
from sklearn.tree import DecisionTreeClassifier
from sklearn.model_selection import train_test_split

簡單版

def decide_play1():
    df = pd.read_csv('dtree.csv')
    dict_train = df.to_dict(orient='record')

    dv = DictVectorizer(sparse=False)
    dv_train = dv.fit_transform(dict_train)
    # print(dv_train)
    # dv_train1 = np.append(dv_train, dv_train[:, 5].reshape(-1, 1), axis=1)
    # dv_train2 = np.delete(dv_train1, 5, axis=1)
    # print('*' * 50)
    # print(dv_train2)

    # print(dv_train[:,:5])
    # print(dv_train[:,6:])
    # print(dv_train[:,5])
    y = dv_train[:, 5]
    x = np.delete(dv_train, 5, axis=1)
    print(x)
    print(y)
    dtc = DecisionTreeClassifier()
    dtc.fit(x, y.reshape(-1, 1))
    print(dtc.predict(np.array([x[3]])))

正式版

def decide_play():
    # ID3
    df = pd.read_csv('dtree.csv')
    # 將數據轉換為字典格式，orient='record'參數指定數據格式為{column:value,column:value}的形式
    dict_train = df.loc[:, ['Outlook', 'Temperatur', 'Humidity', 'Windy']].to_dict(orient='record')
    dict_target = pd.DataFrame(df['PlayGolf'], columns=['PlayGolf']).to_dict(orient='record')


    # 訓練數據字典向量化
    dv_train = DictVectorizer(sparse=False)
    x_train = dv_train.fit_transform(dict_train)

    # 目標數據字典向量化
    dv_target = DictVectorizer(sparse=False)
    y_target = dv_target.fit_transform(dict_target)

    # 創建訓練模型並訓練
    d_tree = DecisionTreeClassifier()
    d_tree.fit(x_train, y_target)

    data_predict = {
        'Humidity': 85,
        'Outlook': 'sunny',
        'Temperatur': 85,
        'Windy': False
    }

    x_data = dv_train.transform(data_predict)
    print(dv_target.inverse_transform(d_tree.predict(x_data)))


if __name__ == '__main__':
    decide_play()

泰坦尼克生存率決策

import numpy as np
import pandas as pd
from sklearn.feature_extraction import DictVectorizer
from sklearn.model_selection import train_test_split
from sklearn.tree import DecisionTreeClassifier
from sklearn.metrics import r2_score


def titanic_tree():
    # 獲取數據
    df = pd.read_csv('Titanic.csv')
    # df = df.fillna(0)
    # dict_train = df.loc[:, ['Pclass', 'Age', 'Sex']].to_dict(orient='record')
    # dict_target = pd.DataFrame(df['Survived'], columns=['Survived']).to_dict(orient='record')
    # x_train, x_test, y_train, y_test = train_test_split(dict_train, dict_target, test_size=0.25)

    # 處理數據，找出特征值和目標值
    x = df.loc[:, ['Pclass', 'Age', 'Sex']]
    y = df.loc[:, ['Survived']]
    # 缺失值處理
    x['Age'].fillna(x['Age'].mean(), inplace=True)
    # 分割數據集到訓練集和測試集
    x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.25)
    # print(y_test)
    dv_train = DictVectorizer(sparse=False)
    x_train = dv_train.fit_transform(x_train.to_dict(orient='record'))
    x_test = dv_train.transform(x_test.to_dict(orient='record'))

    dv_target = DictVectorizer(sparse=False)
    y_target = dv_target.fit_transform(y_train.to_dict(orient='record'))
    y_test = dv_target.transform(y_test.to_dict(orient='record'))
    # print(y_test)
    # 用決策樹進行預測
    d_tree = DecisionTreeClassifier()
    d_tree.fit(x_train, y_train)

    data_predict = {
        'Pclass': 1,
        'Age': 38,
        'Sex': 'female'

    }

    x_data = dv_train.transform(data_predict)
    print(dv_target.inverse_transform(d_tree.predict(x_data).reshape(-1,1)))
    # print(d_tree.predict(x_test))
    # print(y_test)
    # 預測准確率
    # print(d_tree.score(x_test, y_test))


if __name__ == '__main__':
    titanic_tree()

 (Decision Tree)及其變種是另一類將輸入空間分成不同的區域，每個區域有獨立參數的算法。

決策樹分類算法是一種基於實例的歸納學習方法，它能從給定的無序的訓練樣本中，提煉出樹型的分類模型。樹中的每個非葉子節點記錄了使用哪個特征來進行類別的判斷，每個葉子節點則代表了最后判斷的類別。根節點到每個葉子節點均形成一條分類的路徑規則。而對新的樣本進行測試時，只需要從根節點開始，在每個分支節點進行測試，沿着相應的分支遞歸地進入子樹再測試，一直到達葉子節點，該葉子節點所代表的類別即是當前測試樣本的預測類別