Machine-Learning Model for Program Selection in Technical and Vocational Educational Training (TVET) in Kenya

Abstract

This study presents a novel machine learning model to improve program selection processes within Kenya’s Technical and Vocational Educational Training (TVET) institutions. With the increasing demand for vocational education driven by Kenya’s Vision 2030, there is a critical need for systems that align students’ skills and career aspirations with suitable educational programs. We used data from 1,500 current students and alumni students in Nairobi, Kisumu, and Mombasa cities to develop the model. The model uses a Random Forest algorithm to analyze inputs such as Kenya Certificate of Secondary Education grades, personality traits based on the Holland Code, and gender. The proposed model achieved an accuracy of 87 % with alumni data and 83% with current students' data, proving its potential to significantly affect educational outcomes by providing data-driven, personalized program recommendations in Technical and Vocational Educational Training (TVET) institutions in Kenya. The developed method can be extended and used for different educational institutions in Kenya.