Introduction

There is currently enormous interest in neural network modeling systems and what can be accomplished by employing them to solve everyday data modeling problems. Qnet has been designed to provide both the expert and novice with a powerful tool for creating and implementing backpropagation style neural networks into everyday problem solving.

A neural network is best defined as a set of simple, highly interconnected processing elements that are capable of learning information presented to them. The foundation of neural network theory is based on studies of the biological activities of the brain. A neural network’s ability to learn and process information classifies it as a form of artificial intelligence (AI).

The most exciting feature of this new technology is that it can be effectively applied to a vast array of problems, many of which have been thought to be too complex or lacking in sophisticated theoretical models. Neural networks are responsible for making significant advances in the traditional AI fields of speech and visual recognition. Investment managers are creating investment models to better manage money and improve profits. Scientists and engineers use them to model and predict complex phenomena. Marketing professionals are employing neural networks to accurately target products to potential customers. Geologists can increase their probability of finding oil. Lenders use neural networks to determine the credit risk of loan applicants. Complete neurocomputing hardware systems are being built for use in everything from automobiles to manufacturing systems. The variety of problems that can be solved effectively by neural networks is virtually endless.

Qnet is a backpropagation neural modeling system that is designed to exploit the ever increasing power of PC hardware and operating systems. Qnet utilizes the full power of   today's 32-bit operating system environments and hardware (Windows 95, 98, 2000 and NT, along with Pentium class and higher CPU's). Problem sizes that can be tackled by Qnet are virtually unlimited (a PC's  speed and memory will provide practical limits). Qnet offers advanced network design features for creating complex networks that learn using highly optimized backpropagation training algorithms. Qnet features include:

• Blazing speed. Speeds over 6,000,000 connection updates per second on current generation PC’s are possible with Qnet. Accuracy and stability are retained by Qnet with full 32-bit floating point representation of all training data.

• Multiple training modes. Train and maintain hundreds of network models with ease using Qnet's new AutoTrain option. Set up single runs that will train and optimize any number of models in a batch style, automated mode.

• On-line help. Help is available for all menus, dialog screens and input items.

• Fast, easy network design. Select transfer functions on a layer-by-layer basis to give networks unique characteristics. Qnet’s connection editor can be used to customize network connections and control logic flow through the network.

• Easy data interfacing. Cut and paste data directly to Qnet through DataPro. Also, training data is easily imported to Qnet via the universally compatible ASCII file format with support for space, comma and tab delimited formats (these formats are supported by virtually all spreadsheets, word processors, text editors and database programs).

• Automated test set inclusion for overtraining and model integrity analysis. This powerful feature takes the guess work out of determining when a network is properly trained and optimized for use.

• Complete interactive analysis of the training process using NetGraph™ and its powerful AutoZoom feature. NetGraph instantly creates graphs of all key network and training information. AutoZoom can be used to interrogate plotted information to any level of detail required.

• Sophisticated network analysis tools. Interrogate your network to find relative importance of network inputs. Analyze node contributions to network performance. Perform tolerance checks to quickly determine accuracy. All tools are on-line and can be accessed interactively during the training process.

• AutoSave to automatically store the network model during training. The AutoSave feature protects you from overtraining and network divergence by allowing you to retrieve the network state from any prescribed point of the training run.

• Learn Rate Control to automate network training. Qnet takes the drudgery out of adjusting learn rates during training. The Learn Rate Control feature improves convergence times and promotes stable network training with minimal user interaction.

• Multiple training algorithms. Highly optimized backpropagation and FAST-Prop methods can be employed interchangeably during network training.

• Full-featured recall mode. recall mode can be used for analyzing trained networks with new sets of data. The same powerful network analysis features used for training are available in Qnet’s recall mode.

• Easy integration of Qnet developed neural network models into your everyday work environment. Use QnetTool to automate the integration of your Qnet developed neural networks into Windows spreadsheets or database applications. Programmers will appreciate the ANSI C source code or 32-bit DLL that allows you to easily interface your own applications with Qnet neural models royalty free.

• Example problems. The example problems included with Qnet will get you started learning and using Qnet immediately.

All these features combine to make Qnet the most powerful and easy to use neural network model generator available. It is strongly recommended that you read through Qnet’s help file or user’s guide to become familiar with all of Qnet’s features and options. Doing so can greatly improve your productivity with this software.