吉安做网站优化百度排名优化软件

## 区分DataFrame里面的数值变量和离散变量
## DataFrame_data：待处理的DataFrame类型的变量
## O_index:数值型变量列名
## C_index：离散型变量的列名
def distinguish_Char_Num(DataFrame_data):import copy m, n = DataFrame_data.shape## 存放数值型变量所在的列O = []## 存放离散型变量所在的列C = []data = copy.deepcopy(DataFrame_data) for i in range(n):try:if isinstance(data.iloc[0, i], int) or isinstance(data.iloc[0, i], float) or isinstance(data.iloc[0, i], np.int64 ) or isinstance(data.iloc[0, i], np.int32):O.append(i)elif isinstance(data.iloc[0, i], str):C.append(i)else:raise ValueError("the %d column of data is not a number or a string column" % i)except TypeError as e:print(e)# 数值型变量O_data = copy.deepcopy(data.iloc[:, O])# 分类型变量C_data = copy.deepcopy(data.iloc[:, C])##  数值型变量的列名O_index = O_data.columns.tolist()## 分类型变量的列名C_index = C_data.columns.tolist()return   O_index , C_index    

## 对离散变量进行标签化处理，对数值型变量进行归一化处理
## sample_X: 返回整合好的DataFrame类型的变量        
def DataIntegration(DataFrame_data):import copyfrom sklearn.preprocessing import LabelEncodersample_X = copy.deepcopy(DataFrame_data)O_index , C_index = distinguish_Char_Num(DataFrame_data)## 数值型变量标准化sample_X[O_index] = 1.0*(sample_X[O_index] - sample_X[O_index].min())/(sample_X[O_index].max() - sample_X[O_index].min()) ## 离散变量标签化for i in range(len(C_index)):sample_X[C_index[i]] = LabelEncoder().fit_transform(sample_X[C_index[i]]) return sample_X