Projects‎ > ‎

SOMTree

Abstract

The proposed Self-Organizing Map Tree (SOMT) utilizes Self-Organizing Maps as nodes of a tree to make association among different domain data and to observe the prediction processes. Nonlinear data relationships and possible prediction outcomes are inspected through the processes of the SOMT that shows a good predictability of the target output for the given inputs.

People

Aim

To visualize the prediction processes for inspecting data causal relationships

Background

Environmental problems and solutions - Ecological data analyses for management decision makings

Idea

Interactive and visualized method for Data relationships - Data predictions - Data Causalities

Progress

The method design and two experimental evaluations...

Publications

  • Younjin Chung and Masahiro Takatsuka, The Self-Organizing Map Tree (SOMT) for Nonlinear Data Causality Prediction, In the Proceedings of 18th International Conference on Neural Information Processing (ICONIP 2011, Part II, LNCS 7063), pp. 133–142, Springer-Verlag Berlin Heidelberg, 2011

Repository

We strongly suggest you to use git rather than subversion.  If you don't have a repository for this newly created project, please contact ViSLAB admin to create one.

git clone ssh://project.vislab.usyd.edu.au/Groups/groupname/projectname.git
or

svn co svn+ssh://project.vislab.usyd.edu.au/Groups/groupname/projectname

Resources

You should create a "Resources" page as a "File Cabinet" and put the link to "Resources" title.  You should release various files such as released vernon binary in this page.

Links

  • links to related sites go here


You should Insert "Recent Posts" pointing to your Progress here.