A Recommendation System for College Majors for High School Students Using a Hybrid Random Forest and K-Nearest Neighbors Method Based on Alumni Profiles

Abstract

Choosing a higher education field of study remains a challenge for senior high school students because decisions are often influenced by personal assumptions, social environment, and study trends without measurable academic and interest-based mapping. This study developed a field-of-study recommendation model based on alumni profiles using a Hybrid Random Forest and K-Nearest Neighbors approach. The dataset consisted of 2,440 records, 16 variables, and eight field-of-study categories. The research stages included data collection, data curation, preprocessing, 80:20 data splitting, Random Forest and K-Nearest Neighbors model training, probability fusion with weights of 0.6 and 0.4, and evaluation using confusion matrix, accuracy, precision, recall, and F1-score. The results showed that the hybrid model achieved the best performance with an accuracy of 98.77%, precision of 0.99, recall of 0.99, and F1-score of 0.99. This result was higher than Random Forest with an accuracy of 98.36% and K-Nearest Neighbors with an accuracy of 96.31%. Feature importance analysis indicated that interest-related variables contributed the most to the recommendation process. These findings show that the hybrid model can be used as a basis for developing a web-based field-of-study recommendation system.