24/8/8算法笔记 不同分类算法的差异

import numpy as np
from sklearn.tree import DecisionTreeClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.svm import SVCfrom sklearn import datasets

加载数据

我们加载的是啤酒的数据

wine = datasets.load_wine()
wine

LR逻辑斯蒂回归模型应用

import warnings
warnings.filterwarnings('ignore')#隐藏ignore报错

%%time
score = 0
for i in range(100):X_train,X_test,y_train,y_test = train_test_split(X,y,test_size=0.2)lr = LogisticRegression()lr.fit(X_train,y_train)s=lr.score(X_test,y_test)score +=s/100
print('LR逻辑斯蒂回归算法多次运算平均是',score)

SVC支持向量机模型应用

%%time
score = 0
for i in range(1000):X_train,X_test,y_train,y_test = train_test_split(X,y,test_size=0.2)model = SVC()model.fit(X_train,y_train)s=model.score(X_test,y_test)score +=s/1000
print('SVC算法多次运算平均是',score)

决策树模型应用

%%time
score = 0
for i in range(1000):X_train,X_test,y_train,y_test = train_test_split(X,y,test_size=0.2)model = DecisionTreeClassifier()model.fit(X_train,y_train)s=model.score(X_test,y_test)score +=s/1000
print('决策树算法多次运算平均是',score)

不同算法总结对比

递归树对数据是否归一化不敏感

逻辑回归，如果不进行归一化，准确率降低，运行时间会增加

svc支持向量机，如果不进行归一化，准确率，大大降低

model = DecisionTreeClassifier()
model.fit(X_train,y_train)
model.feature_importances_

回归模型中，就线性回归可以表示重要性的大小