news
Published: 06 February 2015

Function approximation and feature selection tool v1.0

Varun Ojha (ESR9) developed a function approximation and features selection tool (a preliminary version: Version 1.0).

This tool is able to provide facility of performing function approximation task that finds underlying relation between a dependent variable and one or more independent variables. The tool is able to process data for the prediction and classification problems. It offers an adaptive way to select the significant feature among the available features in the problem that it processes.  The basic idea of the tool is based on Flexible Neural Tree, an adaptive data structure, performs automatic feature selection and function approximation [Yuehui Chen, Bo Yang, Jiwen Dong, Ajith Abraham, Time-series forecasting using flexible neural tree model, Information Sciences, Volume 174, Issues 3–4, 11 August 2005, Pages 219-235, ISSN 0020-0255].

The current version is shared among the IPROCPM group. The feedback gathered from the group will be visible in the upgraded version of the tool.

The tool is platform independent. Therefore it can run on Mac, Linux, and Windows system. The IPROCOM fellows are advised to use this tool for their own research purpose and write to varun.kumar.ojha@vsb.cz for their feedback and queries.

Share what you've read?