当前位置:网站首页>特征选择:最大信息系数(MIC;Maximal Information Coefficient)【用于衡量两个变量X和Y之间的关联程度,线性或非线性的强度,常用于机器学习的特征选择】
特征选择:最大信息系数(MIC;Maximal Information Coefficient)【用于衡量两个变量X和Y之间的关联程度,线性或非线性的强度,常用于机器学习的特征选择】
2022-06-29 07:04:00 【u013250861】
一、概念
1.1 MIC
MIC,即(Maximal Information Coefficient)最大信息系数,属于Maximal Information-based Nonparametric Exploration (MINE) 最大的基于信息的非参数性探索,用于衡量两个变量X和Y之间的关联程度,线性或非线性的强度,常用于机器学习的特征选择。
MIC相较于Mutual Information(MI)互信息而言有更高的准确度,是一种优秀的数据关联性的计算方式。
1.2 互信息
互信息(Mutual Information)是信息论里一种有用的信息度量,它可以看成是一个随机变量中包含的关于另一个随机变量的信息量,或者说是一个随机变量由于已知另一个随机变量而减少的不肯定性。这个已经是机器学习中老生常谈的内容了,如果想不起来,请参考百度百科-互信息。
参考资料:
MIC:最大信息系数
边栏推荐
- [6G] collation of white paper on computing network technology
- Common MySQL errors and solutions summarized painstakingly (II)
- AC自动机
- NOR flash application layer operation
- 笔记本电脑快速连接手机热点的方法
- C编译器 - 隐式函数声明
- Robotframework learning notes: introduction to robot framework and browserlibrary (playwright)
- Mosaic notes
- Sonic communication - streaming data processing - window alignment
- 常见的七大排序
猜你喜欢

Hands on deep learning (I) -- linear neural network

MySQL中有哪些约束?(实例验证)

SQL Server 开启cdc

Robotframework learning notes: introduction to robot framework and browserlibrary (playwright)

Notice on organizing the second round of the Northwest Division (Shaanxi) of the 2021-2022 National Youth electronic information intelligent innovation competition

线段树以及使用

在colaboratory上云端使用GPU训练(以YOLOv5举例)

STM32基于HAL库的USART+DMA使用

After crossing, she said that the multiverse really exists
![[eye of depth wuenda machine learning homework class phase IV] regularization regularization summary](/img/24/3d0b892c0eaa330f0c69764de5da13.png)
[eye of depth wuenda machine learning homework class phase IV] regularization regularization summary
随机推荐
SizeBalanceTree
Django - installing mysqlclient error: mysqlclient 1.4.0 or newer is required; you have 0.9.3
Codeforces Round #799 (Div. 4)
RobotFramework学习笔记:Robot Framework和BrowserLibrary(PlayWright)简介
AI and the meta universe sparked a spark: human beings lost only shackles and gained all-round liberation
友元,静态关键字,静态方法以及对象间的关系
How to share the virtual environment of pycharm to jupyter Lab
SQL Server 2008 publish and subscribe to SQL Server 2017 pit avoidance Guide
Indextree and Application
VSLAM特征之线特征&面特征
802.11--802.11n协议 PHY
aws elastic beanstalk入门之简单使用
蓝图基础
mysql 开启日志功能
Robotframework learning notes: introduction to robot framework and browserlibrary (playwright)
MySQL系统关键字汇总(官网)
Development trend of garment industry | supply chain | intelligent manufacturing
jsp学习部分
C # import CSV into MySQL database
华为云的AI深潜之旅