Clustering Students Based on Academic Performance and Social Factors: An Unsupervised Learning Approach to Identify Student Patterns

Abstract

This study explores the application of K-Means clustering, an unsupervised learning method, to group students based on academic performance and social factors. The primary objective is to uncover hidden patterns among students by analyzing academic scores in mathematics, reading, and writing, as well as demographic attributes including gender, ethnicity, parental education level, and lunch type. Data preprocessing steps, such as normalization and one-hot encoding, were conducted to prepare the dataset for clustering. The optimal number of clusters was determined using the Elbow Method and Silhouette Score, with K=3 selected for its balance between cluster quality and interpretability. The clustering results revealed three distinct groups of students: low performers, average performers, and high performers. These clusters were visualized using PCA and t-SNE, which showed clear separation and internal consistency. Interpretation of the clusters suggests that social factors may influence academic outcomes, with students from disadvantaged backgrounds more likely to fall into the lower-performing group. The study highlights the importance of data-driven approaches in understanding student diversity and designing targeted interventions. Furthermore, this research underlines the potential of clustering techniques to inform educational strategies by identifying students' needs more precisely. However, limitations include reliance on academic and basic demographic variables, and sensitivity of the K-Means algorithm to outliers and the predefined number of clusters. Future research should incorporate additional factors such as emotional well-being and learning preferences to develop more comprehensive educational models. Overall, the study demonstrates that clustering can serve as a valuable tool for enhancing the effectiveness and equity of educational programs