GaTree - Breeding Decision Trees Using Genetic Algorithms


	GATree Breeding Decision Trees Using Genetic Algorithms	:
Downloads : Screenshots : Usage Tutorial : Buy the program : Presentation


This is the main evolution screen. You can select the train and test data and by using the evolve button you can see how the decision tree evolves in real time. This screen also has several info fields about the current size and accuracy of the decision tree.	The program includes memo boxes with the training and testing data to help you better understand the evolution process.

Here is the statistics tab where we present -in real time- info about the current decision tree size, accuracy and fitness. We also present info about the average fitness of the population. You can control what type of fields will be presented here.	This is the settings tab where you control every aspect of the evolution. The settings include the number of generations and the initial population, the crossover and mutation rate as well as the crossover/mutation heuristic, the percent of replacement between generations, an acceptable error rate which can avoid classification of instances if the decision tree classification error exceeds a limit, as well as a way to vague control the characteristics of the produced decision tree.