数据预处理--归一化

数据预处理–归一化
一.分类：
1.归一化
2.标准化
二.归一化：将数据变换映射到[0,1]之间
计算：mx代表的是想要放缩的区间，如[0,1]，mx = 1,mi = 0
X’= (x-min)/(max-min)
X’’ =X’ *(mx - mi) +mi
代码：sklearn.preproccessing.MinMaxScaler(feature_range = (0,1))
~ MinMaxScaler.fit_transform(x)
X_numpy array格式的数据[n_sample,n_features]
~ 返回值 ：转换后的形状相同的array
三。实例
from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split
import numpy as np
from sklearn.preprocessing import MinMaxScaler
import pandas as pd
def maxmin_demo():
#归一化
#1.获取数据
#2.实例化一个转换器类
#3.调用fit_transfor
data = pd.read_csv(“simlization.txt”)

data.iloc[:,:3]
print(data)
transfor = MinMaxScaler(feature_range=[0,1])
data_new = transfor.fit_transform(data)
# print("data_new:\n",data_new)
return data_new

if name == ‘main’:
datas = maxmin_demo()
# print(type(datas))
s = np.shape(datas)
for i in range(s[0]):
print(’-------’)
for n in datas[i]:
print(round(n, 6))
# for data in datas:
# print(data)
# print(’------’)