Drop Out Students Identification Using Knowledge Base

The purpose of this study is modeling for the initial identification of college dropout students. Samples were taken from drop out student data for the past 4 years. Information from this sample will be acquired as a knowledge base in system modeling. The research aims to provide a knowledge-based system approach using Dempster Shafer for the management of student drop outs at universities especially in Yogyakarta. The symptoms of DO students are obtained from knowledge about DO that appears on campus in Yogyakarta. The system output is in the form of 3 groups of classification namely initial potential DO, enough potential and once potential. The results of the study produced a system that could help university managers deal with drop-out problems early.