from sklearn.neural_network import MLPClassifier # import the classifier
from imblearn.under_sampling import RandomUnderSampler
from sklearn.model_selection import StratifiedKFold
import matplotlib.pyplot as plt
pandas_data = pd.read_csv('sql_eigen.csv')
data = pandas_data.fillna(np.mean(pandas_data)) #缺失数据插补
data['age'][data['age'] > 200] = 91.4 #年龄大于200岁的用中位数91.4代替
data2 = data.drop([
'hr_cov', 'bpsys_cov', 'bpdia_cov', 'bpmean_cov', 'pulse_cov', 'resp_cov',
'spo2_cov', 'height'
],
axis=1) #去掉没用的列
dataSet = np.array(data2)
dataSet[:, 0:78] = ann.preprocess(dataSet[:, 0:78]) #归一化到[0,1]
dataSet[:, 0:78] = ann.preprocess1(dataSet[:, 0:78]) #归一化到[-1,1]
dataMat = dataSet[:, 0:78]
labelMat = dataSet[:, 78]
# dataMat=np.array(dataMat)
# labelMat=np.array(labelMat)
'''
计算各个特征值的平均影响值
'''
clf = MLPClassifier(hidden_layer_sizes=(78, ),
activation='tanh',
shuffle=True,
solver='sgd',
alpha=1e-6,
batch_size=5,
# @File : global_new.py
#提取特征值，对哑变量性别进行处理，然后进行归一化
#0——79列是特征值，最后一列是标签
import pandas as pd
import numpy as np
import ann
global dataMat
global labelMat
global colnames
global framefile
#建立全局变量，提取全部特征值
###read the data###
data = pd.read_csv('final_eigen.csv')
dummy_sex = data['sex'] #哑变量
dummies = pd.get_dummies(dummy_sex, prefix='sex') #哑变量变换
data = data.drop(['sex'], axis=1) #去掉原来的列
datawithdummy = dummies.join(data) #矩阵进行拼接
colnames = datawithdummy.keys()
dataSet = np.array(datawithdummy)
dataSet[:, 0:80] = ann.preprocess(dataSet[:, 0:80])
# dataSet[:,0:80]=ann.preprocess1(dataSet[:,0:80])
dataMat = dataSet[:, 0:80]
labelMat = np.array(datawithdummy['Label'])
framefile = pd.DataFrame(dataSet, columns=colnames)
#writefile.to_csv("F:/testformean.csv")#归一化之后的结果
#print("test")
"""
#------------calculate the FS score with scikit-feature package--------------#
from skfeature.function.similarity_based import fisher_score
from skfeature.function.information_theoretical_based import MRMR
from skfeature.function.similarity_based import reliefF
from skfeature.function.statistical_based import gini_index
Relief = reliefF.reliefF(datamat, labelmat)
Fisher= fisher_score.fisher_score(datamat, labelmat)
# mRMR,J,M,=MRMR.mrmr(datamat,labelmat,n_selected_features=80)
# mRMR=-mRMR
gini= gini_index.gini_index(datamat,labelmat)
gini=-gini
FSscore=np.column_stack((Relief,Fisher,gini))#合并三个分数
FSscore=ann.preprocess(FSscore)
FinalScore=np.sum(FSscore,axis=1)
FS=np.column_stack((FSscore,FinalScore))
FS_nor=ann.preprocess(FS)#将最后一列联合得分归一化
FS=pd.DataFrame(FS_nor,columns=["Relief", "Fisher","gini","FinalScore"],index=featurenames)
# FS.to_csv("F:\Githubcode\AdaBoost\myown\FSscore.csv")
sorteigen=FS.sort_values(by='FinalScore',ascending=False,axis=0)
sorteigen.to_csv('FSsort.csv')
#------------crossalidation with ann--------------#
meanfit=[]#用来存储逐渐增加特征值过程中，不同数目特征值对应的BER平均值
stdfit=[]#用来存储逐渐增加特征值过程中，不同数目特征值对应的BER标准差
names=sorteigen.index#排序之后的特征值
sortfeatures=dataFrame[names]
import pandas as pd
import numpy as np
import ann
from sklearn.neural_network import MLPClassifier # import the classifier
from imblearn.under_sampling import RandomUnderSampler
from sklearn.model_selection import StratifiedKFold
import matplotlib.pyplot as plt
pandas_data = pd.read_csv('sql_eigen.csv')
data = pandas_data.fillna(np.mean(pandas_data))
data['age'][data['age'] > 200] = 91
data2 = data.drop(['hr_cov', 'bpsys_cov', 'bpdia_cov', 'bpmean_cov', 'pulse_cov', 'resp_cov', 'spo2_cov','height'], axis=1)
dataSet=np.array(data2)
dataSet[:,0:78]=ann.preprocess(dataSet[:,0:78])
dataSet[:,0:78]=ann.preprocess1(dataSet[:,0:78])
dataMat=dataSet[:,0:78]
labelMat=dataSet[:,78]
# dataMat=np.array(dataMat)
# labelMat=np.array(labelMat)
clf = MLPClassifier(hidden_layer_sizes=(78,), activation='tanh',
shuffle=True, solver='sgd', alpha=1e-6, batch_size=5,
learning_rate='adaptive')
clf.fit(dataMat,labelMat)
IV=[]
for i in range(78):
tmpdata=dataMat.copy()
tmpdata[:, i]=tmpdata[:,i]*0.8
evaluate_test：各折测试集的各个评价指标，最后两行分别为平均值及标准差
prenum_train：各折训练集混淆矩阵的结果
prenum_test：各折测试集混淆矩阵的结果
'''
import ann
import pandas as pd
import numpy as np
from imblearn.over_sampling import RandomOverSampler
from sklearn.model_selection import StratifiedKFold
import matplotlib.pyplot as plt
data = pd.read_csv('sortedFeature.csv')
labelMat = data['classlabel']
# dataMat=data.ix[:,0:80]
dataMat = data.ix[:, 0:11]
dataMat = ann.preprocess(dataMat)
# dataMat = ann.preprocess1(dataMat)
evaluate_train = []
evaluate_test = []
prenum_train = []
prenum_test = []
skf = StratifiedKFold(n_splits=10)
for train, test in skf.split(dataMat, labelMat):
print("%s %s" % (train, test))
train_in = dataMat[train]
test_in = dataMat[test]
train_out = labelMat[train]
test_out = labelMat[test]
train_in, train_out = RandomOverSampler().fit_sample(train_in, train_out)
from sklearn.model_selection import StratifiedKFold
from imblearn.over_sampling import RandomOverSampler
import pandas as pd #python data analysis
import matplotlib.pyplot as plt
data = pd.read_csv('sortedFeature.csv')
labelMat = data['classlabel']
dataMat = data.ix[:, 0:7]
evaluate_train = []
evaluate_test = []
prenum_train = []
prenum_test = []
data01 = ann.preprocess(dataMat)
dataMat1 = ann.preprocess1(data01)
addones = np.ones((1293, 1))
dataMat = np.c_[addones, dataMat1]
evaluate_train = []
evaluate_test = []
prenum_train = []
prenum_test = []
skf = StratifiedKFold(n_splits=10)
for train, test in skf.split(dataMat, labelMat):
#==============================================================================
# skf=StratifiedShuffleSplit(n_splits=10)
# for train,test in skf.split(dataMat,labelMat):
import pandas as pd
import numpy as np
import ann
global dataMat
global labelMat
###read the data###
pandas_data = pd.read_csv('eigen62.csv')
data = pandas_data.fillna(np.mean(pandas_data))
data['age'][data['age'] > 200] = 91.4
#data2 = data.drop(['hr_cov', 'bpsys_cov', 'bpdia_cov', 'bpmean_cov', 'pulse_cov', 'resp_cov', 'spo2_cov','height'], axis=1)
data2 = data
dataSet = np.array(data2)
dataSet[:, 0:62] = ann.preprocess(dataSet[:, 0:62])
dataSet[:, 0:62] = ann.preprocess1(dataSet[:, 0:62])
# dataSet=np.array(dataSet)
# print("test"）
prenum_test：各折测试集混淆矩阵的结果
"""
import ann
import pandas as pd
import numpy as np
from imblearn.over_sampling import RandomOverSampler
from sklearn.model_selection import StratifiedKFold
import matplotlib.pyplot as plt
data = pd.read_csv('sortedFeature.csv') #排序后的特征值
labelMat = data['classlabel']
# dataMat=data.ix[:,0:80]#全部特征值
# dataMat=data.ix[:,0:67]#BER最小对应的特征值子集
dataMat = data.ix[:, 0:8] #特征值个数最少的特征值子集
dataMat = ann.preprocess(dataMat) #归一化到[0,1]
dataMat = ann.preprocess1(dataMat) #归一化到[-1,1]
neuo = np.shape(dataMat)[1] #隐含层中神经元的数目和特征值个数一致
evaluate_train = []
evaluate_test = []
prenum_train = []
prenum_test = []
skf = StratifiedKFold(n_splits=10) #十折交叉验证
kfold = 1
for train, test in skf.split(dataMat, labelMat):
print("第%s 次交叉验证：" % kfold)
train_in = dataMat[train]
test_in = dataMat[test]
train_out = labelMat[train]