|
Overview
What's new in v2000
The Qnet v2000 Software Package
Overview
Vesta Services, Inc. is now releasing Qnet v2000 for Windows, the complete solution for
backpropagation neural network modeling under Windows. Qnet v2000 combines exceptional
value with unmatched performance and functionality. The Qnet neural network development
system is designed for use with all current Windows OS's. Qnet's performance soars with
highly optimized floating point operations. Qnet offers virtually unrestricted model sizes
and now makes it easier than ever to train, maintain and implement large numbers of Qnet
models.
What's new in v2000
AutoTrain/AutoProcess your Qnet models. Ideal for training and
maintaining a large number of Qnet neural networks. This feature allows complete
unattended training and optimizing of any number of Qnet models. AutoProcessing lets you
return and review your AutoTraining results if desired. This Qnet 2000 feature is a must
for anyone who must periodically update or retrain on a regular basis.
Case weighting factors. Rank/weight your training cases for training
importance. A great feature for models where you wish to attach greater or lesser
significance to your training cases.
Additional updates:
Halt training at RMS error
Extend runs after the run has reached the max iteration.
Improved Learn Rate Control
Input Interrogator and hidden node analyzer results are now available
from the info menu (text mode)
Splash screen turnoff
Improved Windows HTML help
The Qnet v2000
Software package:
Qnet v2000 - This is the training/analysis application for Qnet's neural
network modeling system. You design your neural network, configure the training data and
options, and train and analyze your model. Check out Qnet's features and visit our online documentation to see how easy and simple it is to
develop Qnet neural network models.
DataPro - DataPro is designed to simplify training data transfer and
setup from spreadsheet applications like Microsoft's Excel to Qnet. Simple copy/paste
steps will complete the entire process. Easily setup data labels or apply bounding limits
to raw data. A click of the button saves the data for Qnet and configures Qnet with the
proper training data information. See the DataPro
manual for detailed information.
QnetTool - QnetTool is designed to give basic access to your Qnet
models from common Windows data application (spreadsheet or database applications).
Non-programmers can easily configure QnetTool to automatically work with common Windows
applications. Qnet includes a sample Excel spreadsheet that shows how a simple button can
automatically retrieve your results and place them in your spreadsheet! Automating access
to your models is one example of how Qnet helps to make neural network technology part of
your everyday functions. Visit the QnetTool manual for
detailed operating information.
C-Source and Dll's - More sophisticated applications require programming
access to Qnet developed models. The C source library
included with Qnet provides access and can easily be used on virtually any hardware
platform or operating environment. For Windows based access, Qnet now comes with a 32-bit Dll that can be called directly from popular programming
environments like Microsoft's Visual Basic, VC++, etc.
Help and Online Manuals - Complete online documentation. Windows HTML
help gives you run time support for virtually every input and option. With Qnet's
extensive help system, you can be up and running with a short learning curve. Our PDF
manual (requires Acrobat Reader)
gives you the full printed documentation online or print to your printer! For those who
prefer HTML, you can get help from your browser too. Make sure you have the latest HTML help
system (v 1.2 or higher is recommended) for Windows!
Sample Problems - Qnet's sample
problems cover building financial forecasting models, data analysis, pattern
recognition and more. Qnet's potential application
areas are virtually endless (popular neural network applications).
(top)
|
