机器学习二（二分类问题）

最新推荐文章于 2024-07-03 11:24:00 发布

wangxihe2012

最新推荐文章于 2024-07-03 11:24:00 发布

阅读量3.7k

点赞数

分类专栏： Python 机器学习

本文链接：https://blog.csdn.net/wangxihe2012/article/details/81480915

版权

本文通过分析二分类问题的数据集，利用传统统计方法进行特征选择，然后运用逻辑回归、LDA、KNN、CART、NB、SVM等机器学习算法进行模型建立，并通过交叉验证评估模型性能。进一步，通过调整参数和集成学习方法提升模型预测能力，最终选用GBM并调优，得到较好的预测效果。

摘要由CSDN通过智能技术生成

# -*- coding: utf-8 -*-
"""
Created on Mon Aug 6 20:37:19 2018

@author: wangxihe
"""
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from scipy import stats
from statsmodels.formula.api import ols
import statsmodels.api as sm
import statsmodels.formula.api as smf
import os
os.chdir(r'E:\spyderwork\wxh\数据科学\二分类问题')
columns=['A0','A1','A2','A3','A4','A5','A6','A7','A8','A9','A10','A11','A12','A13','A14','A15',
         'A16','A17','A18','A19','A20','A21','A22','A23','A24','A25','A26','A27','A28','A29',
         'A30','A31','A32','A33','A34','A35','A36','A37','A38','A39','A40','A41','A42','A43',
         'A44','A45','A46','A47','A48','A49','A50','A51','A52','A53','A54','A55',
         'A56','A57','A58','A59','Y']
sonar=pd.read_csv('sonar.all-data.csv',names=columns,header=None)
sonar.shape
sonar.dtypes
sonar['Y'].value_counts()#大致均衡
#R=0,M=1
sonar['Y'].replace({'R':0},inplace=True)
sonar['Y'].replace({'M':1},inplace=True)

# 数据的分类分布
sonar['Y'].value_counts().plot(kind='bar')
#%%
      使用传统的方法
#%%
#由于自变量都是连续的，使用双样本T检验
columned=[]
X=sonar.copy()
for ct in X.columns:
    if ct!='Y':
      TT0=X[X['Y']==0][ct]
      TT1=X[X['Y']==1][ct]
      #方差齐性检验
      leveneTest=stats.levene(TT0,TT1,center='median')
#      print('w-value=%6.4f, p-value=%6.4f' %leveneTest)
      _,fp_value=leveneTest
      if fp_value<0.05:
          Flag=False
      else:
          Flag=True

      _,p_value=stats.stats.ttest_ind(TT0,TT1,equal_var=Flag)
      if p_value<0.05:
         columned.append(ct)
         print('p-value=%6.4f' %p_value)
len(columned) #建议保留 34个变量
##使用共性判断
#%%共线性
def vif(df, col_i):
    cols = list(df.columns)
    cols.remove(col_i)
    cols_noti = cols
    formula = col_i + '~' + '+'.join(cols_noti)
    r2 = ols(formula, df).fit().rsquared
    return 1. / (1. - r2)

#%%
exog = X[columned].copy()

for i in exog.columns:
    print(i, '\t', vif(df=exog, col_i=i))

exog.drop(['A19'],axis=1,inplace=True)
exog.drop(['A45'],axis=1,inplace=True)
exog.drop(['A35'],axis=1,inplace=True)
exog.drop(['A10'],axis=1,inplace=True)
exog.drop(['A47'],axis=1,inplace=True)
#%%
# 向前法
def forward_select(data, response):
    remaining = set(data.columns)
    remaining.remove(response)
    selected = []
    current_score, best_new_score = float('inf'), float('inf')
    while remaining:
        aic_with_candidates=[]
        for candidate in remaining:
            formula = "{} ~ {}".format(
                response,' + '.join(selected + [candidate]))
            aic = smf.glm(
                formula=formula, data=data,
                family=sm.families.Binomial(sm.families.links.logit)
            ).fit().aic
            aic_with_candidates.append((aic, candidate))
        aic_with_candidates.sort(reverse=True)
        best_new_score, best_candidate=aic_with_candidates.pop()
        if current_score > best_new_score:
            remaining.remove(best_candidate)
            selected.append(best_candidate)
            current_score = best_new_score