实现功能：

python实现KNN建模，选择最佳K值，对数据样本进行分类预测，并验证评估。

实现代码：

# 导入需要的库

from warnings import simplefilter

simplefilter(action='ignore', category=FutureWarning)

import pandas as pd

from sklearn.model_selection import train_test_split

import seaborn as sns

import matplotlib.pyplot as plt

from sklearn import metrics

from sklearn.metrics import roc_curve, auc

from sklearn.neighbors import KNeighborsClassifier

def Read_data(file):

    dt = pd.read_csv(file)

    dt.columns = ['age', 'sex', 'chest_pain_type', 'resting_blood_pressure', 'cholesterol',

                  'fasting_blood_sugar', 'rest_ecg', 'max_heart_rate_achieved','exercise_induced_angina',

                  'st_depression', 'st_slope', 'num_major_vessels', 'thalassemia', 'target']

    data =dt

    print(data.head())

    return data

    # ===================数据清洗======================

def data_clean(data):

    # 重复值处理

    print('存在' if any(data.duplicated()) else '不存在', '重复观测值')

    data.drop_duplicates()

    print('不存在' if any(data.isnull()) else '存在', '缺失值')

    data.dropna()  # 直接删除记录

    data.fillna(method='ffill')  # 前向填充

    data.fillna(method='bfill')  # 后向填充

    data.fillna(value=2)  # 值填充

    data.fillna(value={'resting_blood_pressure': data['resting_blood_pressure'].mean()})  # 统计值填充

    # 异常值处理

    data1 = data['resting_blood_pressure']

    # 标准差监测

    xmean = data1.mean()

    xstd = data1.std()

    print('存在' if any(data1 > xmean + 2 * xstd) else '不存在', '上限异常值')

    print('存在' if any(data1 < xmean - 2 * xstd) else '不存在', '下限异常值')

    # 箱线图监测

    q1 = data1.quantile(0.25)

    q3 = data1.quantile(0.75)

    up = q3 + 1.5 * (q3 - q1)

    dw = q1 - 1.5 * (q3 - q1)

    print('存在' if any(data1 > up) else '不存在', '上限异常值')

    print('存在' if any(data1 < dw) else '不存在', '下限异常值')

    data1[data1 > up] = data1[data1 < up].max()

    data1[data1 < dw] = data1[data1 > dw].min()

    return data

    #========================数据编码===========================

def data_encoding(data):

    data = data[["age", 'sex', "chest_pain_type", "resting_blood_pressure", "cholesterol",

                "fasting_blood_sugar", "rest_ecg","max_heart_rate_achieved", "exercise_induced_angina",

                "st_depression", "st_slope", "num_major_vessels","thalassemia","target"]]

    Discretefeature=['sex',"chest_pain_type", "fasting_blood_sugar", "rest_ecg",

          "exercise_induced_angina",  "st_slope", "thalassemia"]

    Continuousfeature=["age", "resting_blood_pressure", "cholesterol",

                      "max_heart_rate_achieved","st_depression","num_major_vessels"]

    df = pd.get_dummies(data,columns=Discretefeature)

    df[Continuousfeature]=(df[Continuousfeature]-df[Continuousfeature].mean())/(df[Continuousfeature].std())

    df["target"]=data[["target"]]

    return df

def data_partition(data):

    #======================数据集划分==========================

    # 1.4查看样本是否平衡

    print(data["target"].value_counts())

    # X提取变量特征；Y提取目标变量

    X = data.drop('target', axis=1)

    y = data['target']

    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=.2,random_state=10)

    return X_train, y_train, X_test, y_test

def Draw_ROC(list1,list2):

    fpr_model,tpr_model,thresholds=roc_curve(list1,list2,pos_label=1)

    roc_auc_model=auc(fpr_model,tpr_model)

    font = {'family': 'Times New Roman',

            'size': 12,

            }

    sns.set(font_scale=1.2)

    plt.rc('font',family='Times New Roman')

    plt.plot(fpr_model,tpr_model,'blue',label='AUC = %0.2f'% roc_auc_model)

    plt.legend(loc='lower right',fontsize = 12)

    plt.plot([0,1],[0,1],'r--')

    plt.ylabel('True Positive Rate',fontsize = 14)

    plt.xlabel('Flase Positive Rate',fontsize = 14)

    plt.show()

    return

    # =========================================KNN====================================

def KNN(X_train, y_train, X_test, y_test):

    training_accuracy = []

    test_accuracy = []

    neighbors_settings = range(1, 15)

    for n_neighbors in neighbors_settings:

        knn = KNeighborsClassifier(n_neighbors=n_neighbors)

        knn.fit(X_train, y_train)

        training_accuracy.append(knn.score(X_train, y_train))

        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()

    plt.show()

    knn = KNeighborsClassifier(n_neighbors=11)

    knn.fit(X_train, y_train)

    print("Accuracy of K-NN classifier on training set: {:.3f}".format(knn.score(X_train, y_train)))

    print("Accuracy of K-NN classifier on test set: {:.3f}".format(knn.score(X_test, y_test)))

    predict_target = knn.predict(X_test)

    predict_target_prob = knn.predict_proba(X_test)

    predict_target_prob_knn = predict_target_prob[:, 1]

    print('预测正确总数：')

    print(sum(predict_target == y_test))

    print('KNN测试集：')

    print(metrics.classification_report(y_test, predict_target))

    print(metrics.confusion_matrix(y_test, predict_target))

    print('KNN训练集：')

    predict_Target = knn.predict(X_train)

    print(metrics.classification_report(y_train, predict_Target))

    print(metrics.confusion_matrix(y_train, predict_Target))

    return y_test,predict_target_prob_knn

if __name__=="__main__":

    data1=Read_data("F:\数据杂坛\\0504\heartdisease\Heart-Disease-Data-Set-main\\UCI Heart Disease Dataset.csv")

    data1=data_clean(data1)

    data2=data_encoding(data1)

    X_train, y_train, X_test, y_test= data_partition(data2)

    y_test,predict_target_prob_knn=KNN(X_train, y_train, X_test, y_test)

    Draw_ROC(y_test,predict_target_prob_knn)

实现效果：

喜欢记得点赞，在看，收藏，

关注V订阅号：数据杂坛，获取完整代码和效果，将持续更新！