svm+voting

# encoding:utf-8
import getopt
from sklearn.preprocessing import MinMaxScaler
import os,time
from multiprocessing import Process, Manager
import pandas as pd
import numpy as np
import itertools
from sklearn.model_selection import KFold  
from sklearn import svm
# from sklearn.cross_validation import train_test_split
import math
from sklearn.model_selection import *
import sklearn.ensemble
from sklearn import metrics
from sklearn.metrics import roc_curve, auc
import sys
from sklearn.model_selection import GridSearchCV
import warnings 
whole_result=[]
input_files=""
whole_dimension=[]
default_l = 1
cross_validation_value = 10
CPU_value = 1
opts, args = getopt.getopt(sys.argv[1:], "hi:l:c:n:", )
final_out_to_excel=[]
row0 = [u'特征集', u'样本个数', u'分类器', u'Accuracy', u'Precision', u'Recall', u'SN', u'SP',
                u'Gm', u'F_measure', u'F_score', u'MCC', u'ROC曲线面积', u'tp', u'fn', u'fp', u'tn']
final_out_to_excel.append(row0) #above was used to generate xlsx format Excel file
for op, value in opts:
    if op == "-i":
        input_files = str(value)
        input_files = input_files.replace(" ", "").split(',')
        for input_file in input_files:
            if input_file == "":
                print("Warning: please insure no blank in your input files !")
                sys.exit()
    elif op == "-l":
        if int(value) == 1:
            default_l = 1
        else:
            default_l = -1
    elif op == "-c":
        cross_validation_value = int(value)
    
    elif op == "-n":
        CPU_value = int(value)

def performance(labelArr, predictArr):
    #labelArr[i] is actual value,predictArr[i] is predict value
    TP = 0.; TN = 0.; FP = 0.; FN = 0.
    for i in range(len(labelArr)):
        if labelArr[i] == 1 and predictArr[i] == 1:
            TP += 1.
        if labelArr[i] == 1 and predictArr[i] == 0:
            FN += 1.
        if labelArr[i] == 0 and predictArr[i] == 1:
            FP += 1.
        if labelArr[i] == 0 and predictArr[i] == 0:
            TN += 1.
    if (TP + FN)==0:
        SN=0
    else:
        SN = TP/(TP + FN) #Sensitivity = TP/P  and P = TP + FN
    if (FP+TN)==0:
        SP=0
    else:
        SP = TN/(FP + TN) #Specificity = TN/N  and N = TN + FP
    if (TP+FP)==0:
        precision=0
    else:
        precision=TP/(TP+FP)
    if (TP+FN)==0:
        recall=0
    else:
        recall=TP/(TP+FN)
    GM=math.sqrt(recall*SP)
    #MCC = (TP*TN-FP*FN)/math.sqrt((TP+FP)*(TP+FN)*(TN+FP)*(TN+FN))
    return precision,recall,SN,SP,GM,TP,TN,FP,FN

def worker(X_train, y_train, cross_validation_value, CPU_value, input_file, share_y_predict_dict, share_y_predict_proba_dict):
    print("子进程执行中>>> pid={0},ppid={1}".format(os.getpid(),os.getppid()))
    svc = svm.SVC(probability=True)
    parameters = {'kernel': ['rbf'], 'C':map(lambda x:2**x,np.linspace(-2,5,7)), 'gamma':map(lambda x:2**x,np.linspace(-5,2,7))}
    clf = GridSearchCV(svc, parameters, cv=cross_validation_value, n_jobs=CPU_value, scoring='accuracy')
    clf.fit(X_train, y_train)
    C=clf.best_params_['C']
    gamma=clf.best_params_['gamma']
    print('c:',C,'gamma:',gamma)

    
    y_predict=cross_val_predict(svm.SVC(kernel='rbf',C=C,gamma=gamma,),X_train,y_train,cv=cross_validation_value,n_jobs=CPU_value)
    y_predict_prob=cross_val_predict(svm.SVC(kernel='rbf',C=C,gamma=gamma,probability=True),X_train,y_train,cv=cross_validation_value,n_jobs=CPU_value,method='predict_proba')
    input_file = input_file.replace(".csv","")
    y_predict_path = input_file + "_predict.csv"
    y_predict_proba_path = input_file + "_predict_proba.csv"
    share_y_predict_dict[input_file] = y_predict
    share_y_predict_proba_dict[input_file] = y_predict_prob[:,1]
    pd.DataFrame(y_predict).to_csv(y_predict_path, header = None, index = False)
    pd.DataFrame(y_predict_prob[:,1]).to_csv(y_predict_proba_path, header = None, index = False)
    print("子进程终止>>> pid={0}".format(os.getpid()))
        
if __name__=="__main__":
    print("主进程执行中>>> pid={0}".format(os.getpid()))
    manager = Manager()
    share_y_predict_dict = manager.dict()
    share_y_predict_proba_dict = manager.dict()
    ps=[]
    if default_l == 1:
        data = ""
        x_len = 1000
        y_len = 1000
        file_len = len(input_files)
        threshold = file_len/2
        for index, input_file in enumerate(input_files):
            data = pd.read_csv(input_file,header=None)
            (x_len,y_len) = data.shape

            X_train = data.iloc[:,0:y_len-1]
            y_train = data.iloc[:,[y_len-1]]
            X_train = X_train.values
            y_train = y_train.values
            y_train = y_train.reshape(-1)
            p=Process(target=worker,name="worker"+str(index),args=(X_train, y_train, cross_validation_value, CPU_value,input_file,share_y_predict_dict,share_y_predict_proba_dict))
            ps.append(p)
        # 开启进程
        for index, input_file in enumerate(input_files):
            ps[index].start()

        # 阻塞进程
        for index, input_file in enumerate(input_files):
            ps[index].join()
        ensembling_prediction = 0
        ensembling_prediction_proba = 0
        for key, value in share_y_predict_dict.items():
            ensembling_prediction = ensembling_prediction + value
        ensembling_prediction = [1 if e > threshold else 0 for e in ensembling_prediction]
        print(ensembling_prediction)
        for key, value in share_y_predict_proba_dict.items():
            ensembling_prediction_proba = ensembling_prediction_proba + value
        ensembling_prediction_proba = ensembling_prediction_proba/3.0
        print(ensembling_prediction_proba/3.0)
        ACC=metrics.accuracy_score(y_train,ensembling_prediction)
        print("ACC",ACC)
        precision, recall, SN, SP, GM, TP, TN, FP, FN = performance(y_train, ensembling_prediction) 
        F1_Score=metrics.f1_score(y_train, ensembling_prediction)
        F_measure=F1_Score
        MCC=metrics.matthews_corrcoef(y_train, ensembling_prediction)
        auc = metrics.roc_auc_score(y_train, ensembling_prediction_proba)
        pos=TP+FN
        neg=FP+TN
        savedata=[str(input_files),"正："+str(len(y_train[y_train == 1]))+'负：'+str(len(y_train[y_train == 1])),'svm',ACC,precision, recall,SN,SP, GM,F_measure,F1_Score,MCC,auc,TP,FN,FP,TN]
        final_out_to_excel.append(savedata)
        print("final_out_to_excel",final_out_to_excel)
        pd.DataFrame(ensembling_prediction).to_csv("voting_prediction_label.csv", header = None, index = False)
        pd.DataFrame(ensembling_prediction_proba).to_csv("voting_prediction_proba_label.csv", header = None, index = False)
        pd.DataFrame(final_out_to_excel).to_excel('output'+'.xlsx',sheet_name="results",index=False,header=False)
        print("主进程终止")
    else:
        data = ""
        x_len = 1000
        y_len = 1000
        file_len = len(input_files)
        threshold = file_len/2
        for index, input_file in enumerate(input_files):
            data = pd.read_csv(input_file,header=None)
            (x_len,y_len) = data.shape
            X_train = data.values
            half_sequence_number = x_len / 2
            y_train = np.array([1 if e < half_sequence_number else 0 for (e,value) in enumerate(X_train)])
            y_train = y_train.reshape(-1)
            print("default y_train: ", y_train)
            p=Process(target=worker,name="worker"+str(index),args=(X_train, y_train, cross_validation_value, CPU_value,input_file,share_y_predict_dict,share_y_predict_proba_dict))
            ps.append(p)
        # 开启进程
        for index, input_file in enumerate(input_files):
            ps[index].start()

        # 阻塞进程
        for index, input_file in enumerate(input_files):
            ps[index].join()
        ensembling_prediction = 0
        ensembling_prediction_proba = 0
        for key, value in share_y_predict_dict.items():
            ensembling_prediction = ensembling_prediction + value
        ensembling_prediction = [1 if e > threshold else 0 for e in ensembling_prediction]
        print(ensembling_prediction)
        for key, value in share_y_predict_proba_dict.items():
            ensembling_prediction_proba = ensembling_prediction_proba + value
        ensembling_prediction_proba = ensembling_prediction_proba/3.0
        print(ensembling_prediction_proba/3.0)
        ACC=metrics.accuracy_score(y_train,ensembling_prediction)
        print("ACC",ACC)
        precision, recall, SN, SP, GM, TP, TN, FP, FN = performance(y_train, ensembling_prediction) 
        F1_Score=metrics.f1_score(y_train, ensembling_prediction)
        F_measure=F1_Score
        MCC=metrics.matthews_corrcoef(y_train, ensembling_prediction)
        auc = metrics.roc_auc_score(y_train, ensembling_prediction_proba)
        pos=TP+FN
        neg=FP+TN
        savedata=[str(input_files),"正："+str(len(y_train[y_train == 1]))+'负：'+str(len(y_train[y_train == 1])),'svm',ACC,precision, recall,SN,SP, GM,F_measure,F1_Score,MCC,auc,TP,FN,FP,TN]
        final_out_to_excel.append(savedata)
        print("final_out_to_excel",final_out_to_excel)
        pd.DataFrame(ensembling_prediction).to_csv("voting_prediction_label.csv", header = None, index = False)
        pd.DataFrame(ensembling_prediction_proba).to_csv("voting_prediction_proba_label.csv", header = None, index = False)
        pd.DataFrame(final_out_to_excel).to_excel('output'+'.xlsx',sheet_name="results",index=False,header=False)
        print("主进程终止")

 

 

该代码用于实现多个特征数据的集合输入输入同一个程序中实现程序的投票，并输入结果

最终会生成

• 文件名_predict.csv  对应文件的预测标签
• 文件名_predict_prob.csv 对应文件的预测分数
• output.xlsx  最终的评估结果

example 

python simple_voting.py -l 1 -c 5 -n 1 -i 1.csv,2.csv,3.csv

• -i ：表示输入的特征文件，以逗号分隔多个特征文件
• -l : 表示是否默认csv格式特征文件尾有标签，默认为1（因此需要保证你的csv文件中尾部带有标签（1,0））,若csv默认为前一半标签为1，后一半为0，则将-l设为0
• -c ：几折交叉验证，5 代表五折交叉验证
• -n : 是否开多进程在单个数据集训练的时候，因为是多个数据集，所以已经实现了多进程，这边设置为1较为稳妥，如果cpu核数不是很多请不要轻易增加这个值，否则可能出现不知名bug

 

github链接