feature-selection-master
所属分类：matlab编程
开发工具：matlab
文件大小：11KB
下载次数：49
上传日期：2018-04-26 20:09:58
上 传 者：
smilingcost
说明： 最小冗余最大相关性（MRMR）（MRMR.M）
需要外部库。详情请见MRMR。下载一个更新版本的互信息工具箱
偏最小二乘（PLS）回归系数（ReGCOEF.m）
使用MATLAB统计工具箱中的PLSReress
ReliefF（分类）和RReliefF（回归）（ReleFracePr.M.）
从Matlab STATS工具箱中包装Releff.m。这是Matlab R2010B以后提供的。
ReliefF的另一个选择是使用ASU特征选择工具箱中的代码。这使用WEKA工具箱的ReleFEF，因此需要额外的库。请参阅相应的文档。
费雪评分（Fisher评分）
围绕ASFS特征选择工具箱围绕FSFisher。M
(Minimum Redundancy Maximum Relevance (mRMR) (mRMR.m)
Needs external library. See mRMR.m for details.
Download a newer version of the mutual information toolbox
Partial Least Squares (PLS) regression coefficients (regCoef.m)
Uses plsregress.m from MATLAB statistics toolbox
ReliefF (classification) and RReliefF (regression) (relieffWrapper.m)
Wraps around relieff.m from the MATLAB stats toolbox. This is available MATLAB r2010b onwards.
Another option for ReliefF is to use the code from ASU Feature Selection toolbox. This uses ReliefF from weka toolbox and hence needs additional libraries. Please see the corresponding documentation.
Fisher Score (fisherScore.m)
Wraps around fsFisher.m from the ASU Feature Selection toolbox)
文件列表:
feature-selection-master (0, 2018-04-26)
feature-selection-master\compareFeatSelAlgos.m (8691, 2016-07-18)
feature-selection-master\featSelOptions.m (1870, 2016-07-18)
feature-selection-master\fisherScore.m (228, 2016-07-18)
feature-selection-master\mRMR.m (534, 2016-07-18)
feature-selection-master\mycandexch.m (947, 2016-07-18)
feature-selection-master\optimalLoadings.m (926, 2016-07-18)
feature-selection-master\plots.m (6300, 2016-07-18)
feature-selection-master\regCoef.m (399, 2016-07-18)
feature-selection-master\relieffWrapper.m (106, 2016-07-18)
feature-selection-master\setup_feat_sel.m (478, 2016-07-18)
Feature Selection using Partial Least Squares Regression and Optimal Experiment Design
============================================================
This repository contains code for the _Optimal Loadings_ feature selection technique proposed in the following paper [pdf](http://www.umiacs.umd.edu/~varun/files/optimal-loadings-IJCNN15.pdf).
```
@inproceedings{NagarajaPLS15,
author = {Varun K. Nagaraja and
Wael Abd{-}Almageed},
title = {Feature Selection using Partial Least Squares Regression and Optimal
Experiment Design},
booktitle = {International Joint Conference on Neural Networks, {IJCNN}},
year = {2015}
}
```
The determinant maximization is performed using a modified version of the candidate exchange function present in the MATLAB Statistics Toolbox. Since I cannot share the original source of the MATLAB function, I have created a protected file. Contact me if you want to know the edits.
Other techniques compared in the paper
--------------------------------------
Minimum Redundancy Maximum Relevance (mRMR) (`mRMR.m`)
* Needs external library. See `mRMR.m` for details.
* Download a newer version of the [mutual information toolbox](http://www.mathworks.com/matlabcentral/fileexchange/14888)
Partial Least Squares (PLS) regression coefficients (`regCoef.m`)
* Uses `plsregress.m` from MATLAB statistics toolbox
ReliefF (classification) and RReliefF (regression) (`relieffWrapper.m`)
* Wraps around `relieff.m` from the MATLAB stats toolbox. This is available MATLAB r2010b onwards.
* Another option for ReliefF is to use the [code](http://featureselection.asu.edu/old/algorithms/fs_sup_relieff.zip) from ASU Feature Selection toolbox. This uses ReliefF from weka toolbox and hence needs additional libraries. Please see the corresponding documentation.
Fisher Score (`fisherScore.m`)
* Wraps around [`fsFisher.m`](http://featureselection.asu.edu/old/algorithms/fs_sup_fisher_score.zip) from the ASU Feature Selection toolbox
Usage
-----
* Load the data
* Create an options structure using `featSelOptions.m`
* Perform experiments using `compareFeatSelAlgos.m`
近期下载者：
相关文件：
收藏者：