python支持向量机回归_scikit-learn代码实现SVM分类与SVR回归以及调参

分类

二分类：

from sklearn.model_selection import train_test_split

from sklearn.svm import SVC

import mglearn.datasets

import matplotlib.pyplot as plt

#forge数据集是一个二维二分类数据集

X,y=mglearn.tools.make_handcrafted_dataset()

X_train,X_test,y_train,y_test=train_test_split(X,y,test_size=0.2, random_state=33)

svm=SVC(kernel='rbf',C=10,gamma=0.1,probability=True).fit(X_train,y_train)

print(svm.predict(X_test))

#输出分类概率

print(svm.predict_proba(X_test))

print(svm.score(X_test,y_test))

[0 0 1 1 1 0]

[[0.91919503 0.08080497]

[0.94703815 0.05296185]

[0.04718756 0.95281244]

[0.08991918 0.91008082]

[0.18789225 0.81210775]

[0.83350967 0.16649033]]

0.6666666666666666

多分类：

用的是鸢尾花数据集，其实代码和分类差不多

from sklearn.model_selection import train_test_split

from sklearn.svm import SVC

import mglearn.datasets

from sklearn.datasets import load_iris

lr=load_iris()

X_train,X_test,y_train,y_test=train_test_split(lr.data,lr.target,test_size=0.2, random_state=33)

svm=SVC(kernel='rbf',C=10,gamma=0.1,probability=True).fit(X_train,y_train)

print(svm.predict(X_test))

#输出分类概率

print(svm.predict_proba(X_test))

print(svm.score(X_test,y_test))

[1 1 0 1 2 2 0 0 2 2 2 0 2 1 2 1 1 0 1 2 0 0 2 0 2 1 1 1 2 2]

[[0.00854955 0.97805528 0.01339518]

[0.02046015 0.96924699 0.01029286]

[0.95821345 0.02958702 0.01219953]

[0.01262418 0.9436578 0.04371802]

[0.01150194 0.21819026 0.7703078 ]

[0.01106185 0.0025207 0.98641745]

[0.93603376 0.04933543 0.01463081]

[0.97333858 0.01668576 0.00997566]

[0.01176501 0.10464775 0.88358724]

[0.01189284 0.01007565 0.97803151]

[0.01141951 0.0055538 0.98302669]

[0.96745325 0.01936203 0.01318472]

[0.01085658 0.0074639 0.98167952]

[0.00991103 0.934642 0.05544697]

[0.01539224 0.07365073 0.91095703]

[0.01314571 0.9801558 0.00669849]

[0.01167736 0.56576966 0.42255298]

[0.97030247 0.01935948 0.01033805]

[0.01612402 0.9036916 0.08018438]

[0.01584532 0.03144821 0.95270647]

[0.96485254 0.02414839 0.01099907]

[0.95276015 0.03287629 0.01436357]

[0.00943625 0.00639159 0.98417216]

[0.96755483 0.02137253 0.01107264]

[0.00979122 0.52403131 0.46617748]

[0.01150947 0.57310996 0.41538057]

[0.01065238 0.95469906 0.03464856]

[0.01084855 0.98409152 0.00505992]

[0.0121375 0.12769585 0.86016666]

[0.01086478 0.25277848 0.73635675]]

0.9333333333333333

回归

回归的调参和分类是一样的。

# 导入库

import numpy as np # numpy库

from sklearn.linear_model import BayesianRidge, LinearRegression, ElasticNet # 批量导入要实现的回归算法

from sklearn.svm import SVR # SVM中的回归算法

from sklearn.ensemble.gradient_boosting import GradientBoostingRegressor # 集成算法

from sklearn.model_selection import cross_val_score # 交叉检验

from sklearn.metrics import explained_variance_score, mean_absolute_error, mean_squared_error, r2_score # 批量导入指标算法

import pandas as pd # 导入pandas

import matplotlib.pyplot as plt # 导入图形展示库

# 数据准备

raw_data = np.loadtxt('regression.txt') # 读取数据文件

X = raw_data[:, :-1] # 分割自变量

y = raw_data[:, -1] # 分割因变量

# 训练回归模型

n_folds = 6 # 设置交叉检验的次数

model_br = BayesianRidge() # 建立贝叶斯岭回归模型对象

model_lr = LinearRegression() # 建立普通线性回归模型对象

model_etc = ElasticNet() # 建立弹性网络回归模型对象

model_svr = SVR() # 建立支持向量机回归模型对象

model_gbr = GradientBoostingRegressor() # 建立梯度增强回归模型对象

model_names = ['BayesianRidge', 'LinearRegression', 'ElasticNet', 'SVR', '