Neural Network Modeling for Windows 95/98/2000/XP

prev home next

Qnet - ZD Shareware Awards Finalistshareill.gif (2964 bytes)

Qnet 2000 for Windows
  • Excellent for Large models.
  • Blazing speeds - over 20,000,000 CPS possible.
  • Technical support via email for registered users.

General Features

  • Simple, powerful model design and setup.
  • Interactive training analysis and optimization tools.
  • Multi-model training mode to train and maintain large numbers of models automatically.
  • QnetTool, DLL and C library options for easy model access.
  • Complete online help and online User's Guide (PDF format).
  • Sample financial, scientific and AI problems included.

Qnet Interactive Training Features

  • Interactive graphing analysis with NetGraph, the industry's best training graphing tool.
  • Error, correlation & tolerance overtraining analysis and recovery.
  • Learn rate control feature.
  • Full, interactive control of all training parameters.
  • Backprop and FAST-Prop training algorithms.
  • Unique threshold checks for financial forecasting.
  • Input relevance rankings.
  • Hidden node analyzer for sizing hidden layers.
  • Color contour plots.

Network Design Options

  • Configure up to 10 network layers.
  • Unlimited nodes per layer.
  • Layer selectable transfer functions.
  • Connection editor to customize network designs.
  • NetView for fast setup.

Training Data Options

  • Unlimited number of training cases.
  • Copy/Paste training data directly to Qnet via DataPro.
  • Import popular ASCII file formats (CSV, TXT, PRN).
  • Automatically incorporate test sets (random, sequential).
  • Automatically limit data outliers.
  • Apply case weight factors to unequally weight training cases in a model.

Qnet Recall Options

  • Qnet recall mode.
  • QnetTool for integrating trained models into popular Windows applications.
  • C source library.
  • Windows DLL access. Compatible with virtually all common programming platforms including VB and VB for Applications.

System Requirements

  • Windows 95, 98, 2000 or XP operating systems.
  • 32MB RAM (or higher depending on hardware and modeling requirements).