Home of CPC-X Software

  

  

  Copyright © CPC-X Software, 1997-2005  

(Last Update: May 22, 2005)  

Auto2Fit

  Home

  Introduction

  Nonlinear Regression

  Example Problems

  Screen Shots

  Download

  Order

 

NeuralPower

  Home

  Screen Shots

  Download

  Order

  

Introduction of NeuralPower

1: What is NeuralPower?

      NeuralPower is an easiest-to-use yet powerful Artificial Neural Network (ANN) program. It is designed especially for common users who want to employ ANN in their studies or projects effortlessly. With its friendly user-interface, simple operations and high efficient performance, most of general ANN problems can be handed with great results

Return to top

2: What can be done with NeuralPower and who needs it?

NeuralPower can be used in almost all research fields such as multi-nonlinear regression, forecasting, curve fit, pattern recognition, classification, decision making, problem optimization, time series analysis and market performance.

NeuralPower is suitable for:

  • People who have to work with or analyses data in any form.
  • People with problems are complex, laborious, 'fuzzy' or simply un-resolvable with present methods.
  • People who simply want to improve on their current techniques and gain competitive advantage.         

Return to top

3: Features of NeuralPower

  • Unique! Multi-learning data files supported: almost all other similar products support only one data file for learning process. Sometimes, this is a not enough for real problem (see gRainfall-runoff" example).
  • Unique! Normal and grid data types supported: no any other products support gGridh data type learning.
  • Visual Control of Learning! Visual and real time monitoring of learning process. Almost all aspects of network could be controlled in graphical mode, i.e. any connection weight, variations of RMSE.
  • Visual design of network structures.
  • Unique! Steps-by-Steps learning.
  • Interrupting/suspending and resuming of learning process are supported. Learning can start from a set of random weights or weights specified by customer.
  • Five learning algorithms, including Genetic Algorithms and Levnberg-Marquardt Algorithm. Fast converge of learning can be easily achieved.
  • No limitation on hidden layer numbers and nodes number of each layer. Each layer can have different transfer functions.
  • Besides six types of most-often used transfer functions, customers can define functions of their self.
  • One-click settlement for time series analysis.
  • Abilities for optimal study and 2D-3D graphical analysis.
  • Excel-like, multi-windows data file editor enhanced by VB, Java and Delphi/Pascal script languages. Input/export Excel File (.xls), DBase File (.dbf), Access File (.mdb) and Comma-delimited Text File (.csv) directly.
  • Chart Function: create, edit, and store multi-charts within a single data file.
  • Linear, multi-linear regression with over 30 pre-defined functions.
  • Matrix operation, data arrange tools, and much more...

 Return to top

4: System Requirements

     Running the 32-bit version of NeuralPower requires a minimum 486 processor and Windows 95/98/2000/ME/XP or NT 4.0. A complete installation requires less than 10 MB of hard disk space.

We welcome your input on how we can improve NeuralPower to make it more useful to the users.

 

Return to top

 

 

Hosted by www.Geocities.ws

1