Python实现遗传算法（GA）+支持向量回归机（SVR）

本实验使用环境为Anaconda3 Jupyter，调用Sklearn包，请提前准备好。

1.引入一些常见包

主要包含pandas、numpy、绘图包、SVR、标准化、验证函数等包。

from sklearn.metrics import accuracy_score
from sklearn.metrics import confusion_matrix
from sklearn.metrics import classification_report
from sklearn.metrics import explained_variance_score
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split
import pandas as pd
import numpy as np
import time
from sklearn import metrics
import csv
from sklearn.svm import SVR
import matplotlib.pyplot as plt

2.引入数据

将准备好的CSV文件导入，将数据分为目标集和特征集，比如本文预测’土壤水分’，使用’土壤温度’,‘空气湿度’,‘空气温度’,'光照强度’作为特征。使用pands进行格式化显示。

#全部数据
data=[]
#特征集
traffic_feature=[]
#目标集
traffic_target=[]
#打开数据文件
csv_file = csv.reader(open('turang.csv'))
#遍历 格式设置为float
for content in csv_file:content=list(map(float,content))if len(content)!=0:data.append(content)traffic_feature.append(content[0:4])traffic_target.append(content[-1])
#将数据变成np.array的格式
data=np.array(data)
traffic_feature=np.array(traffic_feature)
traffic_target=np.array(traffic_target)
#更直观的观察数据，本文中没有太多用途
df=pd.DataFrame(data=data,columns = ['土壤温度','空气湿度','空气温度','光照强度','土壤水分'])

数据最终样子输入df即可查看

目标值图：
(噪音有点大…先将就用吧 - -!!，以后我会写一篇数据降噪的文章)

plt.plot(traffic_target)#测试数组
fig = plt.gcf()
fig.set_size_inches(18.5, 10.5)
plt.show()

3.数据标准化

使用StandardScaler()方法将数据标准化归一化。

scaler = StandardScaler() # 标准化转换
scaler.fit(traffic_feature)  # 训练标准化对象
traffic_feature= scaler.transform(traffic_feature)   # 转换数据集

标准化前：

标准化后：

4.使用支持向量回归机（SVR）

先将数据随机90%做测试集，剩下10%当验证集，随机种子任意设置。本文采取随机抽取形式，而不是时间序列。如使用时间序列预测，请将特征集和目标集的后10%留出来，做为验证集。同理如单步时间序列与从，则保留最后一行作为验证集。
如：

feature_test=traffic_feature[int(len(traffic_feature)*0.9):int(len(traffic_feature))]
target_test=traffic_target[int(len(traffic_target)*0.9):int(len(traffic_target))]

使用SVR（），参数默认，使用R方、EVS、时间作为评价指标。

feature_train,feature_test,target_train, target_test = train_test_split(traffic_feature,traffic_target,test_size=0.1,random_state=10)start1=time.time()
model_svr = SVR()
model_svr.fit(feature_train,target_train)
predict_results1=model_svr.predict(feature_test)
end1=time.time()plt.plot(target_test)#测试数组
plt.plot(predict_results1)#测试数组
plt.legend(['True','SVR'])
fig = plt.gcf()
fig.set_size_inches(18.5, 10.5)
plt.title("SVR")  # 标题
plt.show()
print("EVS:",explained_variance_score(target_test,predict_results1))
print("R2:",metrics.r2_score(target_test,predict_results1))
print("Time:",end1-start1)

结果：
看的出来，随机抽取的精度不高，R方只有区区0.54。(废话，噪音那么大，精度能高吗…)

时间序列预测结果更离谱，我就不放图了。

5.使用GA算法对SVR进行调参

对SVR参数的惩罚参数C、损失函数epsilon、核系数gamma进行调参，设置其范围为[0,10]、[0,2]、[0,100]，可以自行设置。一共20代，每代10人，可以自行设置。
GA算法简单来说，就是每一代人都会有人基因突变，基因突变如果效果很好，适应度很高，那么其他人也会向其进化。
上代码：

#设置适应度，这里设置为R2
def msefunc(predictval,realval):print("R2 = ",metrics.r2_score(realval,predictval)) # R2return metrics.r2_score(realval,predictval)#设置优化函数，这里为SVR，参数在此绑定，使用验证集输入验证得出适应度
def SVMResult(vardim, x, bound):X = feature_train.tolist()y = target_train.tolist()c=x[0]e=x[1]g=x[2]clf = SVR(C=c,epsilon=e,gamma=g)clf.fit(X, y)predictval=clf.predict(feature_test.tolist())return msefunc(predictval,target_test.tolist())
class GAIndividual:'''individual of genetic algorithm'''def __init__(self,  vardim, bound):'''vardim: dimension of variablesbound: boundaries of variables'''self.vardim = vardimself.bound = boundself.fitness = 0.def generate(self):'''generate a random chromsome for genetic algorithm'''len = self.vardimrnd = np.random.random(size=len)self.chrom = np.zeros(len)for i in range(0, len):self.chrom[i] = self.bound[0, i] + \(self.bound[1, i] - self.bound[0, i]) * rnd[i]def calculateFitness(self):'''calculate the fitness of the chromsome'''self.fitness = SVMResult(self.vardim, self.chrom, self.bound)import random
import copyclass GeneticAlgorithm:'''The class for genetic algorithm'''def __init__(self, sizepop, vardim, bound, MAXGEN, params):'''sizepop: population sizepop人口规模vardim: dimension of variables变量维数bound: boundaries of variables变量边界MAXGEN: termination condition终止条件param: algorithm required parameters, it is a list which is consisting of crossover rate, mutation rate, alpha'''self.sizepop = sizepopself.MAXGEN = MAXGENself.vardim = vardimself.bound = boundself.population = []self.fitness = np.zeros((self.sizepop, 1))self.trace = np.zeros((self.MAXGEN, 3))self.params = paramsdef initialize(self):'''initialize the population'''for i in range(0, self.sizepop):ind = GAIndividual(self.vardim, self.bound)ind.generate()self.population.append(ind)def evaluate(self):'''evaluation of the population fitnesses'''for i in range(0, self.sizepop):self.population[i].calculateFitness()self.fitness[i] = self.population[i].fitnessdef solve(self):'''evolution process of genetic algorithm'''self.t = 0self.initialize()self.evaluate()best = np.max(self.fitness)bestIndex = np.argmax(self.fitness)self.best = copy.deepcopy(self.population[bestIndex])self.avefitness = np.mean(self.fitness)self.maxfitness = np.max(self.fitness)self.trace[self.t, 0] =  self.best.fitnessself.trace[self.t, 1] =  self.avefitnessself.trace[self.t, 2] =  self.maxfitnessprint("Generation %d: optimal function value is: %f; average function value is %f;max function value is %f"% (self.t, self.trace[self.t, 0], self.trace[self.t, 1],self.trace[self.t, 2]))while (self.t < self.MAXGEN - 1):self.t += 1self.selectionOperation()self.crossoverOperation()self.mutationOperation()self.evaluate()best = np.max(self.fitness)bestIndex = np.argmax(self.fitness)if best > self.best.fitness:self.best = copy.deepcopy(self.population[bestIndex])self.avefitness = np.mean(self.fitness)self.maxfitness = np.max(self.fitness)self.trace[self.t, 0] =  self.best.fitnessself.trace[self.t, 1] = self.avefitnessself.trace[self.t, 2] =  self.maxfitnessprint("Generation %d: optimal function value is: %f; average function value is %f;max function value is %f"% (self.t, self.trace[self.t, 0], self.trace[self.t, 1],self.trace[self.t, 2]))print("Optimal function value is: %f; " %self.trace[self.t, 0])print ("Optimal solution is:")print (self.best.chrom)self.printResult()def selectionOperation(self):'''selection operation for Genetic Algorithm'''newpop = []totalFitness = np.sum(self.fitness)accuFitness = np.zeros((self.sizepop, 1))sum1 = 0.for i in range(0, self.sizepop):accuFitness[i] = sum1 + self.fitness[i] / totalFitnesssum1 = accuFitness[i]for i in range(0, self.sizepop):r = random.random()idx = 0for j in range(0, self.sizepop - 1):if j == 0 and r < accuFitness[j]:idx = 0breakelif r >= accuFitness[j] and r < accuFitness[j + 1]:idx = j + 1breaknewpop.append(self.population[idx])self.population = newpopdef crossoverOperation(self):'''crossover operation for genetic algorithm'''newpop = []for i in range(0, self.sizepop, 2):idx1 = random.randint(0, self.sizepop - 1)idx2 = random.randint(0, self.sizepop - 1)while idx2 == idx1:idx2 = random.randint(0, self.sizepop - 1)newpop.append(copy.deepcopy(self.population[idx1]))newpop.append(copy.deepcopy(self.population[idx2]))r = random.random()if r < self.params[0]:crossPos = random.randint(1, self.vardim - 1)for j in range(crossPos, self.vardim):newpop[i].chrom[j] = newpop[i].chrom[j] * self.params[2] + (1 - self.params[2]) * newpop[i + 1].chrom[j]newpop[i + 1].chrom[j] = newpop[i + 1].chrom[j] * self.params[2] + \(1 - self.params[2]) * newpop[i].chrom[j]self.population = newpopdef mutationOperation(self):'''mutation operation for genetic algorithm'''newpop = []for i in range(0, self.sizepop):newpop.append(copy.deepcopy(self.population[i]))r = random.random()if r < self.params[1]:mutatePos = random.randint(0, self.vardim - 1)theta = random.random()if theta > 0.5:newpop[i].chrom[mutatePos] = newpop[i].chrom[mutatePos] - (newpop[i].chrom[mutatePos] - self.bound[0, mutatePos]) * (1 - random.random() ** (1 - self.t / self.MAXGEN))else:newpop[i].chrom[mutatePos] = newpop[i].chrom[mutatePos] + (self.bound[1, mutatePos] - newpop[i].chrom[mutatePos]) * (1 - random.random() ** (1 - self.t / self.MAXGEN))self.population = newpopdef printResult(self):'''plot the result of the genetic algorithm'''x = np.arange(0, self.MAXGEN)y1 = self.trace[:, 0]y2 = self.trace[:, 1]y3 = self.trace[:, 2]#plt.plot(x, y1, 'r', label='optimal value')plt.plot(x, y2, 'g', label='average value')plt.plot(x, y3, 'b', label='max value')fig = plt.gcf()fig.set_size_inches(18.5, 10.5)plt.xlabel("GENS")plt.ylabel("R2")plt.title("GA")plt.legend()plt.show()

开跑开跑!

if __name__ == "__main__":bound = np.array([[0,0,0],[10,2,100]])ga = GeneticAlgorithm(10, 3, bound, 20, [0.9, 0.1, 0.5])ga.solve()

结果：
图不是很好看，GA确实容易陷入局部极值。
c、e、g等与[ 9.93055626 0.28102442 24.58580654]

6.融合GA-SVR模型

from sklearn.svm import SVR
import matplotlib.pyplot as plt
start1=time.time()
model_svr = SVR(C=9.93055626,epsilon=0.28102442,gamma=24.58580654)
model_svr.fit(feature_train,target_train)
predict_results1=model_svr.predict(feature_test)
end1=time.time()plt.plot(target_test)#测试数组
plt.plot(predict_results1)#测试数组
plt.legend(['True','SVR'])
fig = plt.gcf()
fig.set_size_inches(18.5, 10.5)
plt.title("SVR")  # 标题
plt.show()
print("EVS:",explained_variance_score(target_test,predict_results1))
print("R2:",metrics.r2_score(target_test,predict_results1))
print("Time:",end1-start1)

结果如下：

时间虽然增大，但R方和EVS均有较大提升。