def Em(name1,feature,n1,n2):
data = pd.read_csv('E:\\feature.csv', encoding='gbk')
datafeature = data[feature]
data2values = datafeature[feature].values
emval = EM()
datacompete = emval.solve(data2values)
return datacompete
def EMmake(name1, name2):
trafficflow, b, ind = share.Csvopreation(name1, name2)
datavalue = trafficflow.values
datavalue = share.Standardize(datavalue, np.nanmean(datavalue, axis=0),
np.nanstd(datavalue, axis=0))
#print(trafficflow.values.dtype)
em = EM() # EM算法
model = em.solve(datavalue)
return model
def test_em():
X_filled = EM().complete(missing_data)
complete_data_, _, _ = min_max_scale(complete_data)
X_filled, _, _ = min_max_scale(X_filled)
score = RMSE(complete_data_[missing_mask], X_filled[missing_mask])
print(score)
def evaluate(self, X_mis, X_full):
missing_index = evaluate.get_missing_index(np.isnan(X_mis))
original_arr = X_full[missing_index]
em_X_filled = EM().complete(copy.copy(X_mis))
em_filled_arr = em_X_filled[missing_index]
rmse_em_score = evaluate.RMSE(original_arr, em_filled_arr)
return rmse_em_score
def em_missing_value(self, **params) -> pd.DataFrame:
"""
EM基于高斯分布能处理混合数据,但在连续型数据上表现的相对较好。在多种数
据缺失机制下,EM相对于其他方法有着较好的表现。
:param params:
:return:
"""
before = self.df[self.columns].values
after = EM(max_iter=1000).complete(before)
self.df[self.columns] = after
return self.df
#pip install ycimpute
#EM Algoritma
import seaborn as sns
import missingno as msno
import numpy as np
import pandas as pd
from ycimpute.imputer import EM
df = sns.load_dataset("titanic")
df = df.select_dtypes(include =["float64","int64"])
print(df)
print(df.isnull().sum())
var_name = list(df)
print(var_name)
arr_df = np.array(df)
print(arr_df)
print(arr_df.shape)
arr_dffill = EM().complete(arr_df)
df_fill = pd.DataFrame(arr_dffill,columns=var_name)
print(df_fill.isnull().sum())
try:
imputedData = knnimput.KNN(k=3).complete(missData)
score = evaluate.RMSE(originData, imputedData)
knn_rmse.append(score)
logger.info("knn missing rate:{},RMSE:{}".format(i, score))
except:
knn_rmse.append(np.nan)
try:
imputedData = mice.MICE().complete(missData)
score = evaluate.RMSE(originData, imputedData)
mice_rmse.append(score)
logger.info("MICE missing rate:{},RMSE:{}".format(i, score))
except:
mice_rmse.append(np.nan)
try:
imputedData = EM().complete(missData)
score = evaluate.RMSE(originData, imputedData)
em_rmse.append(score)
logger.info("EM missing rate:{},RMSE:{}".format(i, score))
except:
em_rmse.append(np.nan)
try:
imputedData = BiScaler().fit_transform(missData)
imputedData = SoftImpute().fit_transform(imputedData)
score = evaluate.RMSE(originData, imputedData)
fi_bs_rmse.append(score)
logger.info("fi BiScaler missing rate:{},RMSE:{}".format(
i, score))
except:
fi_bs_rmse.append(np.nan)
try:
var_names=list(df)
import numpy as np
n_df=np.array(df)
from ycimpute.imputer import iterforest
dff=iterforest.IterImput().complete(n_df)
import pandas as pd
dff=pd.DataFrame(dff, columns=var_names)
dff.isnull().sum()
###EM
import seaborn as sns
import missingno as msno
df=sns.load_dataset('titanic')
df=df.select_dtypes(include=['float64','int64'])
from ycimpute.imputer import EM
var_names=list(df)
import numpy as np
n_df=np.array(df)
dff=EM().complete(n_df)
dff=pd.DataFrame(dff, columns=var_names)
dff.isnull().sum() #yani eksiklerin hepsi dolmuş oldu.