Keywords: data analytics, optimization, predictive modelling, statistical analysis, student admissions, student behaviour
Every year universities globally receive numerous applications to their undergraduate programs. The admissions cycle is an operation which consists of multiple steps requiring vigorous review of application data to determine program eligibility. Program eligibility is based on a number of factors: grades, courses taken, program choice, supplemental application, etc. Eligible applicants are sent an offer which they may accept or refuse. Each educational institution is responsible for meeting a specific target enrolment capacity; under-enrolment is financially devastating and over-enrolment can be extremely costly.
We created a mathematical model using historical data in conjunction with current application data to predict the number of acceptances. The model utilizes correlation analysis to determine key indicators which dynamically categorize the applicants, providing the foundation of the projection. Historical data is then analyzed to determine past yields and optimized coefficients are applied. Finally, grade inflation is integrated into the model to provide a more accurate estimate.
Additionally, this work resulted in the development of a software application which university personnel use to forecast current year enrolment numbers.