zoukankan      html  css  js  c++  java
  • sklearn-数据预处理scale

    python机器学习-乳腺癌细胞挖掘(博主亲自录制视频,包含数据预处理scale)

    https://study.163.com/course/introduction.htm?courseId=1005269003&utm_campaign=commission&utm_source=cp-400000000398149&utm_medium=share

    数据预处理方法包括scale,normalization,Binarizer

    # -*- coding: utf-8 -*-
    """
    Created on Sat Apr 14 09:09:41 2018
    
    @author:Toby 
    standardScaler==features with a mean=0 and variance=1
    minMaxScaler==features in a 0 to 1 range
    normalizer==feature vector to a euclidean length=1
    
    normalization
    bring the values of each feature vector on a common scale
    L1-least absolute deviations-sum of absolute values(on each row)=1;it is insensitive to outliers
    L2-Least squares-sum of squares(on each row)=1;takes outliers in consideration during traing
    
    """
    
    from sklearn import preprocessing
    import numpy as np
    
    data=np.array([[2.2,5.9,-1.8],[5.4,-3.2,-5.1],[-1.9,4.2,3.2]])
    bindata=preprocessing.Binarizer(threshold=1.5).transform(data)
    print('Binarized data:',bindata)
    
    #mean removal
    print('Mean(before)=',data.mean(axis=0))
    print('standard deviation(before)=',data.std(axis=0))
    
    #features with a mean=0 and variance=1
    scaled_data=preprocessing.scale(data)
    print('Mean(before)=',scaled_data.mean(axis=0))
    print('standard deviation(before)=',scaled_data.std(axis=0))
    print('scaled_data:',scaled_data)
    '''
    scaled_data: [[ 0.10040991  0.91127074 -0.16607709]
     [ 1.171449   -1.39221918 -1.1332319 ]
     [-1.27185891  0.48094844  1.29930899]]
    '''
    
    #features in a 0 to 1 range
    minmax_scaler=preprocessing.MinMaxScaler(feature_range=(0,1))
    data_minmax=minmax_scaler.fit_transform(data)
    print('MinMaxScaler applied on the data:',data_minmax)
    '''
    MinMaxScaler applied on the data: [[ 0.56164384  1.          0.39759036]
     [ 1.          0.          0.        ]
     [ 0.          0.81318681  1.        ]]
    '''
    
    data_l1=preprocessing.normalize(data,norm='l1')
    data_l2=preprocessing.normalize(data,norm='l2')
    print('l1-normalized data:',data_l1)
    '''
    [[ 0.22222222  0.5959596  -0.18181818]
     [ 0.39416058 -0.23357664 -0.37226277]
     [-0.20430108  0.4516129   0.34408602]]
    '''
    print('l2-normalized data:',data_l2)
    '''
    [[ 0.3359268   0.90089461 -0.2748492 ]
     [ 0.6676851  -0.39566524 -0.63059148]
     [-0.33858465  0.74845029  0.57024784]]
    '''

    数据处理——One-Hot Encoding

    一、One-Hot Encoding

        One-Hot编码,又称为一位有效编码,主要是采用位状态寄存器来对个状态进行编码,每个状态都由他独立的寄存器位,并且在任意时候只有一位有效。
        在实际的机器学习的应用任务中,特征有时候并不总是连续值,有可能是一些分类值,如性别可分为“male”和“female”。在机器学习任务中,对于这样的特征,通常我们需要对其进行特征数字化,如下面的例子:
    有如下三个特征属性:
    • 性别:["male","female"]
    • 地区:["Europe","US","Asia"]
    • 浏览器:["Firefox","Chrome","Safari","Internet Explorer"]
    对于某一个样本,如["male","US","Internet Explorer"],我们需要将这个分类值的特征数字化,最直接的方法,我们可以采用序列化的方式:[0,1,3]。但是这样的特征处理并不能直接放入机器学习算法中。

    二、One-Hot Encoding的处理方法

        对于上述的问题,性别的属性是二维的,同理,地区是三维的,浏览器则是思维的,这样,我们可以采用One-Hot编码的方式对上述的样本“["male","US","Internet Explorer"]”编码,“male”则对应着[1,0],同理“US”对应着[0,1,0],“Internet Explorer”对应着[0,0,0,1]。则完整的特征数字化的结果为:[1,0,0,1,0,0,0,0,1]。这样导致的一个结果就是数据会变得非常的稀疏。
     
     

     https://study.163.com/provider/400000000398149/index.htm?share=2&shareId=400000000398149( 欢迎关注博主主页,学习python视频资源,还有大量免费python经典文章)


    机器学习项目合作QQ:231469242

     
     
  • 相关阅读:
    灰哥的二叉树
    BZOJ1029: [JSOI2007]建筑抢修[模拟 贪心 优先队列]
    POJ1155TELE[树形背包]
    HDU4003Find Metal Mineral[树形DP 分组背包]
    POJ1837 Balance[分组背包]
    HDU2639Bone Collector II[01背包第k优值]
    POJ3666Making the Grade[DP 离散化 LIS相关]
    HDU2955 Robberies[01背包]
    HDU2191悼念512汶川大地震遇难同胞——珍惜现在,感恩生活[多重背包]
    POJ1014Dividing[多重背包可行性]
  • 原文地址:https://www.cnblogs.com/webRobot/p/8830816.html
Copyright © 2011-2022 走看看