| |
Downloads
Computational Intelligence Software
| Name |
File (kb) |
Version |
Description |
D2CSVM
(1/9/06) |
D2CSVM.zip
|
b 1.3.
|
This C++ program trains the SVM classifier using an adaptive heuristic framework for selecting working sets. Supports several common kernels, crossvalidation and testing. |
| D2C-Matlab |
D2CMatlabv1.zip
Manual (pdf) |
v 1.0. |
This is a MATLAB GUI interface programmed from MATLAB GUIDE. It allows the user to train and test a SVM classifier and perform basic analysis on the results such as crossvalidation, ROC curve plotting, ROC area calculation and probabilistic computation of SVM outputs. The core SVM program uses D2CSVM and is included in this package. |
The following is a Matlab tool package written for use with SVM analysis and the D2CSVM package.
The zip file contains the following functions.
D2CMatlabTools
D2CMtools.zip |
AR Model Tool
(31/10/06) |
b 1.0. |
A small GUI driven program which was written to enable ease of converting coefficients of the autoregressive model (AR) into training files for the SVM. The current version supports only Excel input files with digital signals in columns. The first row of the Excel file can be appended with class labels. The program outputs a sparse dataset consisting of features which are the AR coefficients. Supports the major AR estimation methods in Matlab.
Run ARtool.m in Matlab 6.1 or higher.
|
|
(1/9/06) |
b 1.0. |
Accepts Excel files of features and plots a 2D scatter plot. This tool is useful for viewing the seperability of 2 data features. Utility also allows specific graphs to be saved and automatic generation of sparse training data files for selected 2 features.
Run scatter2D.m in Matlab 6.1 or higher.
|
Convert Sparse
(18/9/06) |
b 1.0 |
Reads a sparse data file and parses it into several output files. The function takes 3 arguments; Input File name ; Vector containing ratio of data for each fold; Random Distribution to draw the examples from the main data file.
Run ConvSparse.m in Matlab 6.1 or higher |
| Convert Excel to Sparse |
b 1.0 |
Reads Excel File containing the training examples in the following format
<label> <attr1> <attr2> ....
Function accepts a vector, V containing the attributes numbers which need to be written to the output file in sparse format.
Run convExcelSp.m in Matlab 6.1 or higher |
BibTeX citation:
@misc{D2CSVM,
author = {Lai, D.},
title = {{D2C-SVM}: A heuristic algorithm for training {S}upport {V}ector {M}achines},
year = {2005},
note = {Software available at {\tt http://www.ee.unimelb.edu.au/people/dlai/}}
Disclaimer: The code is freely available for public use and can be downloaded and modified. Please cite this site when using the code for your research purposes. The author is not responsible for any untoward errors that may occur due to bugs in the code. Please send all questions and bug reports to the author.
|
|
