|
May 2014
|
|
GeneHunter Adds Evolution Strategy, 64-bit, and Excel 2013
|
|
Ward Systems Group has announced Release 4 of GeneHunter, a powerful software solution for many types of optimization problems. GeneHunter is actually two products in one: an Excel Add-in and a library of genetic optimization functions that you can call from your programs. New Optimization Method Release 4 of GeneHunter adds the Evolution Strategy optimization method to the Excel interface . The classic GeneHunter optimization method remains in the program. Evolution Strategy is a variant of genetic algorithms that uses real numbers instead of integers in chromosomes, and therefore does not cross segments of a chromosome, but instead crosses whole chromosomes. The individuals represent potential solutions to a problem. The individuals are tested by a fitness function and the results are used to determine if the individual will be included in the next generation of potential solutions. The Evolution Strategy method is the default optimization method used in NeuroShell Trader. Excel Version Can Run as 64-Bit and Works with Excel 2013 The Excel Add-in is updated to work in both 64-bit as well as 32-bit versions of Excel and Windows, up to and including Excel 2013. The GeneHunter Excel Add-in may now be run as either a 64-bit or 32-bit application in Excel 2013 and 2010. The GeneHunter Add-in continues to run as a 32-bit application in Excel 2003 and Excel 2007. GeneHunter runs under Windows XP, 7, 8, and 8.1 operating systems. New .NET Programming Examples Release 4 adds examples for Microsoft .NET versions of C* and VB. The GeneHunter programming toolkit also includes examples written in VB6, MSVC6, and Delphi6. Optimization with Multiple Constraints The GeneHunter Excel Add-in can evolve a solution that meets the requirements of a primary fitness function and at the same time satisfy other constraints that affect the problem. For example, if you’re trying to design a manufacturing process that minimizes cost or time, but at the same time maximizes product quality GeneHunter is the tool to use. Neural Nets as Fitness Functions If you don’t know the formula for the fitness function, as long as you have relevant data about the problem you can use the NeuroShell Predictor to create the prediction model and then call that model as a fitness function in Excel by using the NeuroShell Run-Time Server. The AI Trilogy software program bundles the NeuroShell Predictor, Classifier, and Run-Time Server as well as GeneHunter. |
|
|
|
|
| Neural Indicators Change with Market Conditions
Commentary by Marge Sherald, CEO |
| Neural Indicators are another unsupervised neural net type that trains in a moving window so the model keeps up with current market conditions. Neural Indicators (NI) provide signals that range from -1 to 1 which may be interpreted as probabilities for buy/sell decisions. NI learn how to give their signals based on evolutionary pressure. The genetic algorithm in the NeuroShell Trader “evolves” NI that gives better and better signals. Survival of the fittest controls the evolutionary process, where fitness is determined by how much money the Neural Indicators make.
The neural nets used in the USG Corporation example are trained to give buy/sell signals rather than making predictions. This chart displays trading signals for both the in and out-of-sample periods. Neural Indicators may simply be used as indicators on a chart or special conditional versions may be substituted for true/false rules in Trading Strategies. If you prefer to see the probabilities for each trading signal, use them in a rule such as Ward4 > 0 where Ward4 is a neural net with four inputs. There are several other types of Neural Indicators. Recurrent Nets feed in data from the current bar and also retain information from previous bars in the same training pattern. Sparse Nets have less connections by design, but they may prove to be a good choice for generalizing on data not included in the training set. If all of this neural net talk seems intimidating, remember that the genetic optimizer in the Trader can make most of the decisions for you.
|
