annotate README.TXT @ 0:e9a9cd732c1e tip

first hg version after svn
author wolffd
date Tue, 10 Feb 2015 15:05:51 +0000
parents
children
rev   line source
wolffd@0 1 This page allows you to reproduce the experiments from our paper "Feature Preprocessing using RBM for Music Similarity Learning". This includes the newest release of the CAMIR system for modelling music similarity.
wolffd@0 2
wolffd@0 3 MLR and SVMLIGHT were tested using a framework developed by Daniel Wolff (daniel.wolff.1@city.ac.uk) at the MIRG group.
wolffd@0 4 This version of CAMIR contains the RBM toolbox of Son Tran (son.tran.1@city.ac.uk).
wolffd@0 5 You can download the code using subversion from the following repository. It is licensed under the GNU GPL v3.
wolffd@0 6
wolffd@0 7 user: anonymous
wolffd@0 8 pass: citymirg
wolffd@0 9 http://chivm.soi.city.ac.uk/svn/camir/branches/code_publications/AES2013/
wolffd@0 10
wolffd@0 11 For reproducing the results in the paper using Matlab on a Windows machine follow these steps:
wolffd@0 12
wolffd@0 13 -Download the code from above using a subversion client such as collabnet svn and the above password
wolffd@0 14 -Set the working directories in the editme_startup.m file of the downloaded code to the location of the downloaded code
wolffd@0 15 -Include the code directory with all subdirectories in your matlab path
wolffd@0 16 -Start Matlab, change the working directory to the location of the downloaded code
wolffd@0 17 -Edit and execute the editme_startup file, this initialises the dataset and could take a few minutes.
wolffd@0 18 -Run the experiments using the scripts in the folder reproduce_AES53rd.
wolffd@0 19 -Most of the results are output as figures. Otherwise values are printed in the Matlab console
wolffd@0 20
wolffd@0 21 If you use this code please cite our paper. We are more than welcoming your additional bits of functionality to be included in the package.