A preliminary study on the reuse of subtrees within decision trees in a genetic programming context for data classification

摘要: Genetic programming (GP) has been successful in creating models for data classification which obtain high accuracies. In a context functions is common practice as this serves way to isolate part of code can be reused. The encapsulation genetic operator capable promoting modularization the sense that encapsulate subtrees reused by GP trees during execution algorithm. Models created problems tend large and certain complexity, thus rendering need modular acquisition methods promote reuse existing order solve problems. effect when solving not previously investigated. Two approaches were proposed, first incorporated with no limitations on how use encapsulated subtrees. second approach made maintained list two proposed tested eight sets results show improved training accuracy nearly every set.