<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[小方哥哥的博客]]></title><description><![CDATA[]]></description><link>https://blog.youkuaiyun.com/OnTheWayGoGoing</link><language>zh-cn</language><generator>https://blog.youkuaiyun.com/</generator><copyright><![CDATA[Copyright &copy; OnTheWayGoGoing]]></copyright><item><title><![CDATA[数据预处理之定量特征二值化与定性特征哑变量编码]]></title><link>https://blog.youkuaiyun.com/OnTheWayGoGoing/article/details/79887973</link><guid>https://blog.youkuaiyun.com/OnTheWayGoGoing/article/details/79887973</guid><author>OnTheWayGoGoing</author><pubDate>Tue, 10 Apr 2018 21:39:38 +0800</pubDate><description><![CDATA[   1. 定量特征二值化      在数据挖掘领域，定量特征二值化的目的是为了对定量的特征进行“好与坏”的划分，以剔除冗余信息。举个例子，银行对5名客户的征信进行打分，分别为50，60，70，80，90。现在，我们不在乎一个人的征信多少分，只在乎他的征信好与坏(如大于90为好，低于90就不好)；再比如学生成绩，大于60及格，小于60就不及格。这种“好与坏”、“及格与不及格”的关系可以转化为0-1...]]></description><category></category></item><item><title><![CDATA[数据预处理之数据无量纲化(标准化/归一化)]]></title><link>https://blog.youkuaiyun.com/OnTheWayGoGoing/article/details/79871559</link><guid>https://blog.youkuaiyun.com/OnTheWayGoGoing/article/details/79871559</guid><author>OnTheWayGoGoing</author><pubDate>Tue, 10 Apr 2018 14:21:37 +0800</pubDate><description><![CDATA[   在进行特征选择之前，一般会先进行数据无量纲化处理，这样，表征不同属性（单位不同）的各特征之间才有可比性，如1cm 与 0.1kg 你怎么比？无量纲处理方法很多，使用不同的方法，对最终的机器学习模型会产生不同的影响。本文将对常用的无量纲化技术进行总结，试图指出他们的适用场景，并给出在Python中的相应调用方式。正文中每列代表一个 属性/特征，每行表示一个/条 样本。   1. min-max...]]></description><category></category></item><item><title><![CDATA[Windows环境下配置python spark（windows7+python+spark）]]></title><link>https://blog.youkuaiyun.com/OnTheWayGoGoing/article/details/78236828</link><guid>https://blog.youkuaiyun.com/OnTheWayGoGoing/article/details/78236828</guid><author>OnTheWayGoGoing</author><pubDate>Sat, 14 Oct 2017 21:01:24 +0800</pubDate><description><![CDATA[@@注意，配置环境变量确保路径全部正确，比如我下载的安装包spark-1.6.0-bin-hadoop2.6/spark-1.6.0-bin-hadoop2.6/bin    文件路径这种，要去掉外层文件夹，额能理解吧！
@@环境变量配置网上有，注意，系统环境变量 path后面要加 ;分号


@@要安装java环境]]></description><category></category></item><item><title><![CDATA[用对数几率回归实现周志华《机器学习》习题3.3西瓜分类，python编程]]></title><link>https://blog.youkuaiyun.com/OnTheWayGoGoing/article/details/68485682</link><guid>https://blog.youkuaiyun.com/OnTheWayGoGoing/article/details/68485682</guid><author>OnTheWayGoGoing</author><pubDate>Thu, 30 Mar 2017 13:04:23 +0800</pubDate><description><![CDATA[数据集如下，要求根据西瓜的两个属性x1(密度)，x2(含糖率)实现对西瓜好瓜（1）还是坏瓜（0）的分类代码如下：# -*- coding: utf-8 -*-
#对率回归分类
import numpy as np
from numpy import linalg
import pandas as pd
#读取数据集
inputfile = 'd:/pythonwork/data/xigua.xls...]]></description><category></category></item><item><title><![CDATA[LDA 线性判别分析]]></title><link>https://blog.youkuaiyun.com/OnTheWayGoGoing/article/details/64906144</link><guid>https://blog.youkuaiyun.com/OnTheWayGoGoing/article/details/64906144</guid><author>OnTheWayGoGoing</author><pubDate>Wed, 22 Mar 2017 11:00:22 +0800</pubDate><description><![CDATA[线性判别分析]]></description><category></category></item><item><title><![CDATA[向量、矩阵求导]]></title><link>https://blog.youkuaiyun.com/OnTheWayGoGoing/article/details/64121430</link><guid>https://blog.youkuaiyun.com/OnTheWayGoGoing/article/details/64121430</guid><author>OnTheWayGoGoing</author><pubDate>Mon, 20 Mar 2017 09:02:43 +0800</pubDate><description><![CDATA[向量、矩阵求导]]></description><category></category></item><item><title><![CDATA[对称矩阵的一些性质]]></title><link>https://blog.youkuaiyun.com/OnTheWayGoGoing/article/details/63682959</link><guid>https://blog.youkuaiyun.com/OnTheWayGoGoing/article/details/63682959</guid><author>OnTheWayGoGoing</author><pubDate>Sun, 19 Mar 2017 11:06:07 +0800</pubDate><description><![CDATA[对称矩阵的一些性质和定理]]></description><category></category></item><item><title><![CDATA[python 安装库 出错合集及专业答疑]]></title><link>https://blog.youkuaiyun.com/OnTheWayGoGoing/article/details/61615620</link><guid>https://blog.youkuaiyun.com/OnTheWayGoGoing/article/details/61615620</guid><author>OnTheWayGoGoing</author><pubDate>Sun, 12 Mar 2017 11:38:59 +0800</pubDate><description><![CDATA[想安装最新版的python库，请在python库大全及下载地址里下载whl版本的安装文件，当然这种安装容易出错，但是成功安装后能比exe文件安装在后期出错的机会小，特别是scipy，如果跟numpy不配套，就很容易各种出错。本人能够帮你解决各种whl格式库安装的问题，有问题请在博客下留言并贴上报错，通过大量的工作本人发现，绝大所数exe格式的python库安装文件都不是最新版，这在以后的应用中会出]]></description><category></category></item></channel></rss>