BrainMaker Professional for Windows 95 or Windows NT 4.0

BrainMaker Professional slashes development time, reduces trial-and-error tedium, and creates networks that recognize patterns and predict better than any other neural network system. Professional has all the features and ease of use of standard BrainMaker plus the most advanced features available. Professional is the best choice for quickly solving problems, especially complex ones such as financial forecasting. Professional includes a Runtime License and source code to build trained networks into your own program for resale. BrainMaker Professional v3 is the Advanced Neural Network Application Development System for people who demand the most highly powered development tools, most analysis functions, and greatest ease of use.

Professional includes proprietary algorithms offered by no other system, including Data Correlator, Cyclic Analysis, Sensitivity Analysis, Global Network Analysis, Contour Analysis, and Hypersonic Training. You can experiment with leading technology features such as recurrence, heavy weights, and other learning algorithm variations. Professional also supports new optional Genetic Training algorithms. Some of Professional's latest advances in neural network technology include:

1. Integration with Windows including Windows Dynamic Data Exchange (DDE). With DDE you can build neural network capability into your DDE-supporting spreadsheet, database, graphics program, etc.
2. Stop training on any number of criteria based on training performance or testing performance. Allows you to easily find a network that tests well and lets you stop at a predefined point.
3. You get the latest development in back-propagation enhancement: recurrent operation. Recurrence builds feedback into your neural network and automates the design of historical inputs for time-series data.
4. Financial users can read MetaStock, CompuTrack and CSI files into NetMaker and can create four financial indicators: RSI, stochastic, MACD, and on-balance volume.
5. Batch mode operations and file directives let you do nearly anything without interaction, even while you're away.
   1. Tune the error tolerance, number of hidden neurons and learning rate as needed while training.
   2. Save training and testing results as often as you'd like, automatically stopping with the best network.

Five types of analysis help you understand your problem better:

1. Cyclic Analysis finds periodic behavior in your data. Knowing that a commodity price tends to oscillate on a four day price cycle helps you predict future oscillations.
2. Data Correlator tells you how strong the relationship is between any two types of data (such as between rainfall and corn production), and what the time delay is between them. This helps you determine which data which will be most predictive.
3. Sensitivity Analysis shows you how each piece of data affects the network's answer. You can display this information as bar graphs which depict the relative impact, or print the results. These graphs help you to understand the network and answer questions such as "How can I best increase the value of my investment?"
4. Global Analysis answers the question "How does the network solve my problem overall?" Reports statistics for entire data sets.
5. Contour Analysis helps answer the question "How did the network arrive at this particular answer?" Color contour plots show you graphically how pairs of inputs affect the answer.