<![CDATA[This research aims to produce a classification system for students who have the potential to drop out. This classification system is expected to help identify students who have the potential to drop out early on in prevention efforts. This research uses academic data in the form of Semester Grade Point Average (IPS) 1-7, Cumulative Grade Point Average Semester 7 (IPKS7), and Cumulative SKS 7, as well as non-academic data including Study Program and Entry Path as classification parameters. The method used is the Long Short-Term Memory (LSTM) algorithm with system development using the CRISP-DM approach. System testing is done using black box testing method and performance evaluation using confusion matrix. The results showed that the classification system developed achieved an accuracy rate of 93% based on confusion matrix evaluation, and all system functionality runs as expected based on the results of black box testing.]]>
Copyrights © 2025
