机器学习-SVM支持向量机

自由随风飘2025-07-26 22:31

支持向量机是一类监督学习算法，实现二分类，其决策边界是对学习样本求解的最大边距超平面。

课程代码：

import numpy as np

my_seed = 2017

np.random.seed(my_seed)

import random

random.seed(my_seed)import matplotlib

import matplotlib.pyplot as plt

matplotlib.rcParams['font.family'] = 'SimHei'

data_set_name = "ec_data"

all_data = np.fromfile(data_set_name)/(1024 * 1024)

sequence_len = 10

X = []

Y = []

for i in range(len(all_data)-sequence_len):

X.append(all_data[i:i+sequence_len])

Y.append(all_data[i+sequence_len])

X = np.array(X)

Y = np.array(Y)

from sklearn.model_selection import train_test_split

默认shufft

X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.05)

from sklearn import svm

SVR = svm.SVR()

SVR.fit(X_train, Y_train)

Y_predict = SVR.predict(X_test)

预测的值

Y_test_predict = Y_predict

Y_test_real = Y_test

fig = plt.figure(figsize=(15, 6))

fig.suptitle(data_set_name)

x = np.arange(100, 150, 1)

plt.subplots_adjust(wspace =0, hspace =0.5)#调整子图间距

plt.subplot(2,1,1)

plt.plot(x,Y_test_predict[100:150],label="预测值")

plt.plot(x,Y_test_real[100:150],label="真实值")

plt.ylabel('流量大小(Mb)')

plt.legend(loc=1,ncol=1)

plt.subplot(2,1,2)

plt.ylabel('流量大小(Mb)')

plt.plot(Y_test_predict,label="预测值")

plt.plot(Y_test_real,label="真实值")

#指定图例位置，1右上角，2左上角，3右下角，4左下角，0自动适应图像

plt.legend(loc=1,ncol=1)

plt.savefig(data_set_name+"_svm_predict.png",dpi=400)