PredictEd: COMPARATIVE ANALYSIS OF ALGORITHMS IN PREDICTING LICENSURE EXAMINATION PERFORMANCE FOR BACHELOR OF ELEMENTARY EDUCATION

Abstract

This study focuses on the development of the PredictEd tool, designed to predict the Licensure Examination for Teachers (LET) performance of Bachelor of Elementary Education (BEED) students, using their GPAs from Year Levels 1 to 4 in major subjects. The study utilized academic records from BEED graduates of Notre Dame of Midsayap College between 2018 and 2023 and gathered usability evaluations from 36 fourth-year BEED students. Three machine learning algorithms—Decision Tree (ID3), Random Forest, and Naïve Bayes—were implemented using Node.js and evaluated to determine their predictive performance. Among the three, Naïve Bayesachieves the highest overall accuracy (100%), particularly excelling in correctly identifying failing cases. This demonstrates the algorithm’s strength in handling predictors with varying levels of independence. The PredictEd tool was further assessed using the ISO 9126 software quality standard, receiving consistently high ratings in functionality, usability, and reliability. The results affirm that PredictEd is an accurate and effective forecasting tool for LEPT performance based on GPA data. It is recommended that future studies incorporate additional variables such as mock exam scores, study habits, and support systems, expand the dataset for broader generalizability,and consider web or cloud deployment for increased accessibility.