|
|
|
|
|
|
|
|
SouperPrediktor
|
|
|
|
|
|
|
|
|
|
|
|
2005 Wolfram Technology Conference
|
|
|
|
|
|
Champaign IL
|
|
|
|
|
|
SouperPrediktor is a Mathematica package that searches function space for functional fits to arbitrary data using genetic algorithm techniques. The program provides a large variety of control parameters to allow the user to refine his search for the optimal functional form. SouperPrediktor outputs detailed and insightful graphics that summarize the information content and evolution of the population of organisms, providing the user with the feedback necessary to guide his search. SouperPrediktor introduces a number of important innovations: - a new complexity function
- an efficiency frontier fitness function
- non-deterministic functional forms
- arbitrary data inputs and outputs
- user-defined functional search space
- new measures of functional homogeneity
- subpopulation evolution
- and a number of new visualization techniques
SouperPrediktor outputs Mathematica expressions that can be used directly by the user in any other context. This also allows the user to interpret the optimized function according to applicable theoretical models. In this presentation I will demonstrate SouperPrediktor’s interface and ability to search an infinite functional space for optimal functional fits to arbitrary functions.
|
|
|
|
|
|
|
|
|
|
|
|
| BGress - SouperPrediktor.E.nb (2.7 MB) - Mathematica Notebook |
|
|