A Neural Network Prediction Model For Diploma And Certificate Student’s Progression In Universities

Abstract

The most important priority of a private academic university is financial stability, which is determined by 100% student progression to the next level of study. Poor student advancement will result in the university's demise, as student progression is the primary source of revenue for a private university. The institution can plan adequately for the next semester and determine the number of workers required without undue stress as a result of the students' progress. Students in Kenyan colleges make predictions using linear forecasting models, which presume that data is linear. As a result, the progression of students based on linear models may be erroneous. Nonlinear models have been used to predict student outcomes with great effectiveness. As evidenced by the literature study, the artificial neural network stands out. The suitability of several models of student advancement to predict student progression in Kenya will be investigated. A literature review will be used to investigate the viability and performance of various models. The study's particular goals were to evaluate students' dropout and deferment rates, create an appropriate artificial neural network model that employs the identified elements to forecast progression rate, and validate the model. The data for this project will be gathered via the Zetech University database system. The report included information on students who were enrolled from 2007 to 2019. The study included a total of 5000 pupils. The artificial neural network model was validated using the sigmoid activation function after the data was separated into training and test sets. The rate of advancement was discovered to be 78.5 percent. Universities should establish intervention programs for students who are on the verge of deferral or dropping out, according to the report.