To search, Click below search items.

 

All Published Papers Search Service

Title

Optimizing Training Data Selection for Decision Trees using Genetic Algorithms

Author

Ahmad Alhindi

Citation

Vol. 20  No. 4  pp. 84-90

Abstract

Genetic algorithms (GAs) have recently been used as a search method for training set selection in supervised machine learning. The assumption is made that not all the data are equally useful in training supervised algorithms. In this paper, we empirically study the performance of classical GA for selecting a ‘good’ training set for decision tree classifiers. We also discuss different fitness functions and their influence on the results. A set of widely used classification datasets from Kaggle and UCI machine learning repository are used. Empirical results show that improved generalization can indeed be obtained using this approach.

Keywords

Genetic Algorithm (GA) Machine Learning (ML) Supervised Learning Optimization Training Set Selection Data Mining Pattern Recognition.

URL

http://paper.ijcsns.org/07_book/202004/20200410.pdf