Numap
| FREE |
|
Description by Publisher
Numap7.0 was developed by the Neural Networks and Image Processing Lab of Univ. of Texas at Arlington. It has training and testing algorithms for Multilayer Perceptron, Functional Link Networks, Piecewise Linear Networks, Self Organization Mapping and K-Mean. The algorithm written for MLP is called Output Weight Optimization - Hidden Weight Optimization (OWO-HWO). The Self Organization Map and K-Means algorithms are assosiated with fast VB Graphics that show the cluster formation and plots the mean sqare error simultaneously. It has Utilities and Tools for controlling the working directory, Counting the number of patterns in a given training file, Deleting columns, combining files, splitting files, to calculate mean and standard deviation of a given column of a file and to plot the histogram of each column of the training file.
It includes Data Preprocessing wherein the training data can be compressed using the discrete Karhunen-Loeve' transform (KLT). It also includes a Feature Selection Algorithm which specifies the importance of each input.
Extensive help files are provided in the software.
This is version of Numap7 is freeware which limits the MLP to 10 hidden units, and limts the PLN to 10 clusters. Interested parties can contact Prof. Manry concerning the commercial version which lacks these limitations.
The classification (decision making) version of this software, called Nuclass7, will be available in late Spring 2003.
