<![CDATA[This study aims to analyze the competitiveness of study programs in the 2023 SNBT dataset using the Knowledge Discovery in Databases (KDD) method and the K-Means Clustering algorithm. The competitiveness of study programs is measured by the ratio between the number of applicants and available slots, reflecting the level of competition and popularity of the programs. Two main issues are addressed: the urgency of data-driven decision-making for formulating effective student admission policies and the lengthy execution time on large datasets such as the 2023 SNBT data, which includes thousands of study programs with complex variables. The number of clusters was determined using the elbow method, dividing the data into three categories: low, medium, and high. Clustering evaluation was conducted using the silhouette score metric, revealing that Cluster 0 (low) demonstrated the best quality with the highest silhouette score. To accelerate the analysis process, parallel computing was implemented using the joblib, scikit learn and multiprocessing library, significantly reducing execution time compared to conventional methods. With an average silhouette score of 0.684816, the results indicate good clustering quality. These findings provide valuable insights for universities in understanding the competitiveness patterns of study programs and support the development of more effective and efficient data-driven student admission policies.]]>
Copyrights © 2025
