sklearn模块总结

sklearn module

持续更新中…
5.17 update:linear regression&support vector machine

linear_regression

数据集导入

• datasets.load_diabetes(*, return_X_y = False, as_frame = False)

  糖尿病数据集（回归任务）

  • return_X_y : Flase则返回bunch对象（类字典），True则返回二元组（data,target）

  • as_frame : True则数据为pandas DataFrame

核心函数

• reg = linear_module.LinearRegression()

  创建linear regression对象

• reg.fit(X, y)

  核心步骤，拟合直线

  X为train_data

  y为train_target

• y = reg.predict(X)

  在测试集上预测结果

  X为test_data

  y为test_pred

评价指标

• MSE(mean square error)

$$MSE=\frac{1}{n}{\Sigma }_{i=0}^{n-1}(y_i-{\hat{y}}_i{)}^2$$

越小表示预测值与真实值越接近，即拟合效果好

• MAE(mean absolute error)

$$MAE=\frac{1}{n}{\Sigma }_{i=0}^{n-1}|y_i-{\hat{y}}_i|$$

越小表示预测值与真实值越接近，即拟合效果好

• $R^2$决定系数

$$R^2=1-\frac{{\Sigma }_{i=0}^{n-1}(y_i-{\hat{y}}_i{)}^2}{{\Sigma }_{i=0}^{n-1}(y_i-{\overline{y}}_i{)}^2}$$

$R^2\in (0,1)$，且越接近1效果越好。

sklearn实现

from sklearn import linear_model, datasets
import numpy as np
import matplotlib.pyplot as plt
from sklearn.metrics import mean_squared_error, r2_score


if __name__ == '__main__':
    #load data
    diabetes_X, diabetes_y = datasets.load_diabetes(return_X_y=True)

    # Use only one feature
    diabetes_X = diabetes_X[:, np.newaxis, 2]

    # Split the data into training/testing sets
    diabetes_X_train = diabetes_X[:-20]
    diabetes_X_test = diabetes_X[-20:]

    # Split the targets into training/testing sets
    diabetes_y_train = diabetes_y[:-20]
    diabetes_y_test = diabetes_y[-20:]

    # Create linear regression object
    regr = linear_model.LinearRegression()

    # Train the model using the training sets
    regr.fit(diabetes_X_train, diabetes_y_train)

    # Make predictions using the testing set
    diabetes_y_pred = regr.predict(diabetes_X_test)

    # The coefficients
    print('Coefficients: \n', regr.coef_)
    # The mean squared error
    print('Mean squared error: %.2f'
        % mean_squared_error(diabetes_y_test, diabetes_y_pred))
    # The coefficient of determination: 1 is perfect prediction
    print('Coefficient of determination: %.2f'
        % r2_score(diabetes_y_test, diabetes_y_pred))

    # Plot outputs
    plt.scatter(diabetes_X_test, diabetes_y_test,  color='black')
    plt.plot(diabetes_X_test, diabetes_y_pred, color='blue', linewidth=3)

    plt.xticks(())
    plt.yticks(())

    plt.show()

support_vector_machine

• 核心库

#分类任务
from sklearn.svm import SVC

数据集导入

纸币真伪数据集

from sklearn.model_selection import train_test_split	
#load data
filename = ''
bankdata = pd.read_csv(filename)
#print(bankdata.head)
#properties
# Variance:图像的方差
# Skewness:偏度
# Kurtosis:峰度
# Entropy:熵
# Class:类别
    
#pretreatment
X = bankdata.drop('Class', axis=1)
y = bankdata['Class']
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)

核心函数

• svclassifier = SVC(kernel=‘linear’)

  使用线性核，创建svc对象

• svclassifier.fit(X, y)

  拟合超平面

  X:train_data

  y:train_target

• svclassifier.predict(X)

  预测测试集的target

  X:test_data

评价指标

• 混淆矩阵(confusion matrix)

又称可能性表格和错误矩阵。下图很清晰地给出了含义：

可以发现对角线上即为正确分类的数据。

from sklearn.metrics import confusion_matrix
print(confusion_matrix(y_test, y_pred))

• classification_report

  包括精度、召回率、F1值等

  • 精度(precision)

  关注被预测为阳性的样本中有多少是阳性的

  • 召回率(recall)

  关注所有标签为阳性的样本有多少被准确预测出来了

  • F1值(F1_score)

  $$F1=\frac{2\ast precision\ast recall}{precision+recall}$$

sklearn实现

# -*- coding:utf-8 -*-

import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.svm import SVC
from sklearn.metrics import classification_report, confusion_matrix

if __name__ == '__main__':
    #load data
    filename = r'D:\VS-Code-python\ML_algorithm\support_vector_machine\bill_authentication.csv'
    bankdata = pd.read_csv(filename)
    #print(bankdata.head)
    #properties
    # Variance:图像的方差
    # Skewness:偏度
    # Kurtosis:峰度
    # Entropy:熵
    # Class:类别
    
    #pretreatment
    X = bankdata.drop('Class', axis=1)
    y = bankdata['Class']
    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.20)

    #training
    svclassifier = SVC(kernel='linear')
    svclassifier.fit(X_train, y_train)

    #prediction
    y_pred = svclassifier.predict(X_test)

    #assessment
    print('confusion_matrix\n',confusion_matrix(y_test, y_pred))
    print('classification_report\n',classification_report(y_test, y_pred))

• out:

confusion_matrix:
[[170   0]
 [  2 103]]
classification_report:
              precision    recall  f1-score   support

           0       0.99      1.00      0.99       170
           1       1.00      0.98      0.99       105

   micro avg       0.99      0.99      0.99       275
   macro avg       0.99      0.99      0.99       275
weighted avg       0.99      0.99      0.99       275

reference：

scikit-learn (sklearn) 官方文档中文版