Investigating The Role of Predictive Analytics and Machine Learning in Optimizing Student Support Services Resource Allocation in Universities

Abstract

The research paper examined the application of predictive analytics and machine learning in helping to optimize student support services resource allocation in Pakistani universities. It was a case study using a mixed-method design that focused on three large public sector universities in Punjab, which had a student population of 42,000. The study collected the necessary data within five years of study, between the years 2018 and 2023 when the institutional data bases were used, the structured questionnaires were given to 350 students and 60 members of staff, and the semi-structured interviews were conducted with 12 administrators. Random Forest, Support Vector Machines and Logistic Regression were machine learning algorithms that used to build predictive models in identifying at-risk students and predicting service demand patterns. The quantitative analysis indicated that predictive model models had 84 percent accuracy in enabling students who needed support services and the qualitative analysis indicated that the limitations were observed in the form of low technological infrastructures and budget shortages. The research found that there was a considerable resource distribution disparage among various categories of students and support services. Findings revealed that machine learning-enabled systems had the potential of optimizing resource allocation by 37 percent more as opposed to the conventional allocation procedures. The study found that introducing predictive analytics into student support services was much more efficient and provided better interventions on a timely basis, but its application had contextual issues unique to Pakistani institutions of higher learning.

Keywords: Application, predictive analytics, machine learning, student support, resource allocation, Pakistani universities.