糖尿病预测--KNN算法

    import numpy as py
    import pandas as pd
    import matplotlib.pyplot as plt
    import seaborn as sns
     
    diabetes_data = pd.read_csv('diabetes.csv')
    print(diabetes_data.head())
     
    # 查看数据信息
    print(diabetes_data.info(verbose=True))
    # 设置参数verbose为True，允许冗长信息
     
    # 数据描述
    print(diabetes_data.describe())
    # 通过describe可以观察到数据的数量，平均值，标准差，最小值，最大值等数据
     
    #数据形状
    print("dimension of diabetes data: {}".format(diabetes_data.shape))
     
    #查看标签分布 
    print(diabetes_data.Outcome.value_counts()) 
     
    #使用柱状图的方式画出标签个数统计
    plt.figure()
    diabetes_data.Outcome.value_counts().plot(kind="bar")
     
    plt.figure()
    sns.countplot(diabetes_data['Outcome'], label="Count")
    plt.savefig("0_1_graph")
     
     
    # 可视化数据分布
    # 对角线上是各个属性的直方图（分布图）
    # 而非对角线上是两个不同属性之间的相关图
    plt.figure()
    sns.pairplot(diabetes_data)
    plt.savefig("0_2_graph")
    '''
    sns.pairplot(diabetes_data, hue="Outcome");
    plt.savefig("0_3_grap")
    '''
    # 画热力图，数值为两个变量之间的相关系数
    # annot: 默认为False，为True,在格子上显示数字
    plt.figure()
    sns.heatmap(diabetes_data.corr(), annot=True)
    plt.savefig("0_4_grap")
     
    # KNN算法
    from sklearn.model_selection import train_test_split
     
    X_train,X_test,y_train,y_test=train_test_split(\
    diabetes_data.loc[:,diabetes_data.columns !='Outcome'],\
    diabetes_data['Outcome'],stratify=diabetes_data['Outcome'],random_state=66)
     
    # 把数据切分为特征X和标签y
    # 切分数据集，test_size=0.3表示30%为测试集。
    # stratify=y表示切分后训练集和测试集中的数据类型的比例跟切分前y中的比例一致
    # 如切分前y中0和1的比例为1:2，切分后y_train和y_test中0和1的比例也都是1:2
    #X = diabetes_data.drop("Outcome",axis = 1)
    #y = diabetes_data.Outcome
    #X_train,X_test,y_train,y_test = train_test_split(X,y,test_size=0.3, stratify=y)
     
    from sklearn.neighbors import KNeighborsClassifier
    # 保存不同k值测试集准确率
    training_accuracy = []
    # 保存不同k值训练集准确率
    test_accuracy = []
    # try n_neighbors from 1 to 10
    neighbors_settings = range(1,11)
     
    for n_neighbors in neighbors_settings:
    # build the model
    knn = KNeighborsClassifier(n_neighbors = n_neighbors)
    knn.fit(X_train,y_train)
    #record training set accuracy 保存训练集准确率
    training_accuracy.append(knn.score(X_train,y_train))
    #record test set accuracy  保存测试集准确率
    test_accuracy.append(knn.score(X_test,y_test))
    plt.figure()
    plt.plot(neighbors_settings,training_accuracy,label="training accuracy")
    plt.plot(neighbors_settings,test_accuracy,label="test accuracy")
    plt.ylabel("Accuracy")
    plt.xlabel("n_neighbors")
    plt.legend()  # 给图像加上图例
    plt.savefig('knn_compare_model')
     
    knn = KNeighborsClassifier(n_neighbors=9)
    knn.fit(X_train,y_train)
     
    print('Accuracy of K-NN classifier on training set:{:.2f}'.format(\
    knn.score(X_train,y_train)))
    print('Accuracy of K-NN classifier on training set:{:.2f}'.format(\
    knn.score(X_test,y_test)))
     
    from sklearn.metrics import confusion_matrix
    from sklearn.metrics import classification_report
    y_pred = knn.predict(X_test)
    print(classification_report(y_pred, y_test))
     
    confusion = confusion_matrix(y_pred, y_test)
    df_cm = pd.DataFrame(confusion)
    plt.figure()
    sns.heatmap(df_cm, annot=True)
    plt.savefig('confusion_matrix')
    plt.show()