An Artificial Neural Network Decision Support Model For University Students Progression

Abstract

This study is motivated by the recent developments in the Kenyan education sector. The government has introduced tough measures to curb cheating of K.C.S.E exams thus resulting to decreasing number of students who qualify for university placement. This means that the number of students being admitted to the university has drastically declined. The number of students achieving the minimum entry points to the universities has steadily declined. This is evident from the fact that previously the entry point to JAB programs was B plus and above but currently this has changed to C plus and above. The number of students dropping out of campus has also increased as well as the number of students deferring their studies. It is important to predict the progression rate of students in order to target potential students for early intervention. The main objective of the study was to develop an artificial neural network model for progression rate of university students. The specific objectives of the study were to determine the enrolment rate, dropout rate and deferment rate of students, to develop an appropriate artificial neural network model that uses the identified factors for predicting progression rate and to validate the developed model. Data was obtained from the Technical University of Kenya database system. The data contained information on students enrolled for the 2015 to 2018 period of study. A total of 2976 students were used for the study. The data was split into training and test set and then the artificial neural network model validated using the sigmoid activation function. The progression rate was found to be 78.5%. The study recommends that universities should have intervention programs for students who are at risk of deferment or dropping out of the university.