Data Mining Modeling of the Alignment Between Graduate Capabilities and Stakeholder Needs Using the Apriori Algorithm

Abstract

Service speed, precision, data accuracy, ease of information delivery, and accountability are critical justifications for implementing information systems. Universitas Semarang (USM) is a private higher education institution in Semarang with the second-largest student population in Central Java. As a rapidly growing institution, USM bears a significant responsibility to ensure its students are well-prepared for the workforce, meeting the specific demands of the business and industrial sectors. Data from USM’s 2019 tracer study—which assesses "horizontal alignment," or the degree of correlation between a graduate's field of study and their subsequent employment—reveals a degree of mismatch between the competencies of graduates and the expectations of stakeholders (1.6% reported "no alignment," 19.2% "low alignment," and 27.5% "moderate alignment"). This issue warrants special attention, prompting the university to refine its strategies to reduce these figures. The Apriori algorithm is a widely recognized method for identifying high-frequency patterns, which are then used to formulate association rules and support various other data mining techniques. Rules that define associations between specific attributes are often referred to as "affinity analysis" or "market basket analysis." In this application, the Apriori algorithm was used to analyze data from the USM tracer study, with a minimum support threshold of 50% and a minimum confidence threshold of 100%, resulting in the generation of four distinct rules. These findings demonstrate that the Apriori algorithm can produce various rule configurations, providing the university with a basis for evaluation and the development of targeted strategic initiatives an approach made possible by the fact that each rule reflects the unique nature of the relationship between a graduate’s field of study and the requirements of their respective stakeholders.