Scholarlink

This study explores the relationship between student satisfaction and academic performance using advanced clustering techniques on a large dataset from the National University of Altiplano in Puno (UNAP), which includes over 18,000 students. A representative sample of 12,156 students was selected. The analysis focused on 32 items related to student satisfaction, covering teaching quality and university services, which were combined with students’ semester averages and admission scores. The data were managed using a preprocessing approach that included normalization and data cleaning, followed by clustering techniques based on the K-Means algorithm. Dimensionality reduction was performed through Principal Component Analysis (PCA) to enhance data interpretation and facilitate the identification of relevant patterns. The elbow method and the silhouette coefficient indicated that the optimal number of clusters was three (K=3). The results revealed three student profiles: Cluster 0 (C0), which includes students with low performance (12.96) and high satisfaction (teaching: 1.87; services: 1.69); Cluster 1 (C1), which groups students with high performance (17.05) and low satisfaction (teaching: 0.94; services: 0.78); and Cluster 2 (C2), composed of students with average performance (15.09) and high satisfaction (teaching: 1.88; services: 1.68). Additionally, a small group of 146 students with low performance and low satisfaction was found within Cluster 1. The validation of the findings through Analysis of Variance (ANOVA) revealed statistically significant differences between the groups (p<0.001), reinforcing the validity of the identified profiles. This work provides valuable tools for educational segmentation that can be used to design differentiated and personalized academic policies.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Clustering Analysis of Student Satisfaction–Performance Relationships in Higher Education Using Large-Scale Educational Data

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Michael Newton Cutipa-Santi,
  • Daisy Crishyely Quelca-Velasquez,
  • Yeni Cutipa-Santi,
  • Renzo Apaza-Cutipa,
  • Vladimiro Ibañez-Quispe

摘要

This study explores the relationship between student satisfaction and academic performance using advanced clustering techniques on a large dataset from the National University of Altiplano in Puno (UNAP), which includes over 18,000 students. A representative sample of 12,156 students was selected. The analysis focused on 32 items related to student satisfaction, covering teaching quality and university services, which were combined with students’ semester averages and admission scores. The data were managed using a preprocessing approach that included normalization and data cleaning, followed by clustering techniques based on the K-Means algorithm. Dimensionality reduction was performed through Principal Component Analysis (PCA) to enhance data interpretation and facilitate the identification of relevant patterns. The elbow method and the silhouette coefficient indicated that the optimal number of clusters was three (K=3). The results revealed three student profiles: Cluster 0 (C0), which includes students with low performance (12.96) and high satisfaction (teaching: 1.87; services: 1.69); Cluster 1 (C1), which groups students with high performance (17.05) and low satisfaction (teaching: 0.94; services: 0.78); and Cluster 2 (C2), composed of students with average performance (15.09) and high satisfaction (teaching: 1.88; services: 1.68). Additionally, a small group of 146 students with low performance and low satisfaction was found within Cluster 1. The validation of the findings through Analysis of Variance (ANOVA) revealed statistically significant differences between the groups (p<0.001), reinforcing the validity of the identified profiles. This work provides valuable tools for educational segmentation that can be used to design differentiated and personalized academic policies.