24/8/8算法条记 差别分类算法的差异

2. import numpy as np
3. from sklearn.tree import DecisionTreeClassifier
4. from sklearn.linear_model import LogisticRegression
5. from sklearn.svm import SVC
7. from sklearn import datasets

复制代码

加载数据

我们加载的是啤酒的数据

2. wine = datasets.load_wine()
3. wine

复制代码

LR逻辑斯蒂回归模子应用

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

复制代码

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

复制代码

SVC支持向量机模子应用

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

复制代码

决议树模子应用

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

复制代码

差别算法总结对比

递归树对数据是否归一化不敏感

逻辑回归，假如不举行归一化，精确率低落，运行时间会增长

svc支持向量机，假如不举行归一化，精确率，大大低落

2. model = DecisionTreeClassifier()
3. model.fit(X_train,y_train)
4. model.feature_importances_

复制代码

回归模子中，就线性回归可以体现告急性的巨细