Self-Organizing Map Plugin
Welcome to the homepage of the Self-Organizing Map Plugin!
Click on these links to get to the corresponding sections.
[ About ]
[ Applicability ]
[ Requirements ]
[ Download ]
About
During the last few years, the channels of acquiring musical data have increased rapidly. Due to the growing importance of broad-band network connections, home computer users are no longer limited to radio or television broadcasting services or other non-computer aided methods to acquire copies of musical pieces. More or less legal peer-to-peer networks, online music stores and personalized internet radio stations provide the user with a constant and neverending flow of musical information. The choice seems to be endless. On the one hand this surplus supply makes it possible to get any piece of musical information any time with a minimal effort of time and costs. On the other hand, this gives rise to a new, challenging problem: Retrieving musical information in large corpora. This plugin for the Machine Learning Environment Yale provides one operator to classify digital song files with regard to extracted musical features as supplied before the classification.
Applicability
The plugin classifies aidio data into similiarity clusters by making use of a Self-Organizing Map. This allows for a variety of different tasks. As an example, meta information for unnamed song files can be retrieved. Song files that are not named or named not according to the actual song name can easily be identified in a cluster containing much fewer songs than the original corpus (which might be excessively large). In addition, applying this operator on a set of files that are most often played by a user can help automating user preference analysis to idiviualize the offer of online music stores and other personalized music services.
Requirements
Software Requirements
In order to use this plugin, you need the following software.
- Yale – Yet Another Learning Environment, Version 3.3
- Value Series Processing Plugin for Yale, Version 3.3
- Java (TM) 2 Runtime Environment, Standard Edition, build 1.5.0_06-b05 or later
- any operating system, the Java Runtime Environment is available for (this software was developed on Microsoft Windows XP and Debian Linux, support for other operating systems can not yet be provided)
Hardware Requirements
Minimal tested requirements (No support for evolutionary algorithm):
- IBM compatible processor with 900MHz
- 224MB SD-Ram
- 32MB Shared Memory Grafics Adaptor
- 80MB of free hard disk space
Recommended system set up (evolutionary algorithm possible):
- IBM compatible processor with at least 2,7GHz or any 64Bit or Multi-Core Processor
- 1536MB DDR-Ram or more
- 64MB DDR Memory Grafics Adaptor
- 1GB free hard disk space (for swap files)
Download
In this section, you can find all downloads to the plugin. Choose the distribution of your preference.
- SOM-Plugin full distribution
- SOM-Plugin binary
- SOM-Plugin source
- sample experiments
- SOM-Plugin Manual
- SOM-Plugin javadoc download or view online