Data Reduction Algorithm for Machine Learning and Data Mining

摘要

The paper proposes an approach to data reduction. The data reduction procedures are of vital importance to machine learning and data mining. To solve the data reduction problems the agent-based population learning algorithm was used. The proposed approach has been used to reduce the original dataset in two dimensions including selection of reference instances and removal of irrelevant attributes. To validate the approach the computational experiment has been carried out. Presentation and discussion of experiment results conclude the paper.