基于机器学习的光纤传感器扰动模式识别系统.zip

""" main 临近种类支持向量机-交叉验证器 求超参数 说明：目前为学习版本的代码 """ import numpy as np from sklearn.model_selection import cross_val_score #交叉验证函数 from preprocesser import dataset #数据集类 from sklearn.svm import SVC #SVM分类器 class iterator:#迭代器 c=0 min=0 max=0 a=0 def __init__(self,min,max): self.min=min; self.a=min self.max=max; return None def __iter__(self): self.c=0 return self def __next__(self): if(self.c>self.max): return self.max self.c=self.c+self.a if(self.c>=self.a*10 and self.a<self.min*100): self.a*=10 return self.c def cross_validator(trainData):#遍历迭代器中所有超参数 iteratorC=iter(iterator(1,32000)) C=0 bestScore=0 bestC=0 bestGamma=0 while(C<32000): C=next(iteratorC) iteratorGamma=iter(iterator(0.01,10)) Gamma=0 print(f"当前进度：C-{C/320:.4f}% ") while(Gamma<1): Gamma=next(iteratorGamma) Classifier=SVC(kernel='rbf',gamma=Gamma,C=C) Score=cross_val_score(Classifier, trainData.data, trainData.target, cv=5,scoring='accuracy').mean() if(Score>bestScore): bestScore=Score bestC=C bestGamma=Gamma print(f"当前最高准确率：{bestScore*100:.2f}% \n C：{C} \n gamma：{Gamma}") print(f"最高准确率：{bestScore*100:.2f}%") return bestC,bestGamma #导入数据 file=open('D:\\大创\\兵王.txt') file1=file.readlines() file.close() #由基本的列表生成数据 totalData=dataset().build_from_outside(file1) #随机分割出训练集 trainData,testData=totalData.split(0.2) svmC,svmGamma=cross_validator(trainData) print(f"交叉验证成功 C:{svmC}, gamma:{svmGamma}")