使用SVM进行癫痫检测

背景：

这是一片论文的简单复现，只是还原思想，不知道是不是我看的不够仔细，我觉得论文说的也就是这样简单的操作

论文名称：Seizure prediction with spectral power of EEG using cost-sensitive support vector machines

有需要的同学可以好好看一下，论文中不仅仅只是检测，还有预测功能。

过程：

SVM Procces

代码：

一个单纯的SVM例子：
 

import numpy as np
import matplotlib.pyplot
from sklearn import svm
from numpy.fft import fft
import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import GridSearchCV

from sklearn.metrics import classification_report
from sklearn.metrics import confusion_matrix

#数据预处理
def datapro():
    data = pd.read_csv(x_src, error_bad_lines=False,header=None)
    label = pd.read_csv(y_src, error_bad_lines=False,header=None)
    x = np.array(data.values)
    print(x)
    y = label.values
    return x,y

if __name__ == '__main__':
    x_src = 'training_data/C3_P3_CZ_PZ.csv'
    y_src = 'training_data/data_y.csv'
    x,y = datapro(x_src,y_src)
    #train_test_split是交叉验证中常用的函数，功能是从样本中随机的按比例选取train data和testdata
    X_train, X_test, y_train, y_test = train_test_split(x, y, test_size=1/3.) # 分割训练集和测试集

     # 标准化
    scaler = StandardScaler()
    X_train_std = scaler.fit_transform(X_train)
    X_test_std = scaler.transform(X_test)

    # 交叉验证，调整参数

    param_grid = {'C':[1e1,1e2,1e3, 5e3,1e4,5e4],
                  'gamma':[0.0001,0.0008,0.0005,0.008,0.005,]}
    print(X_train_std.shape)
    print(y_train.shape)
    #cv 几叉验证
    clf = GridSearchCV(svm.SVC(kernel= 'rbf',class_weight='balanced',max_iter=6000,tol=0.01,probability=True),param_grid,cv=10)
    clf = clf.fit(X_train_std,y_train)
    print(clf.best_estimator_)

    print(clf.predict_proba(X_test_std))

    #打分

    clf.score(X_test_std,y_test)

    y_pred = clf.predict(X_test_std)

    #测试准确率

    print(classification_report(y_test,y_pred))
    print(confusion_matrix(y_test,y_pred))
    """
    https://blog.csdn.net/akadiao/article/details/78788864
    precision: 识别为此类中正确的样本所占比例
    recall: 一类样本中被正确识别的概率
    """