Hybrid Clustering Using N-Soft Set and Artificial Bee Colony for Digital Literacy

Abstract

Introduction: Open and distance education requires a good level of digital literacy. Students in open and distance educatiom come from various ages and backgrounds, resulting in differing levels of literacy. Clustering is needed to identify the digital literacy skills of new students so that digital literacy improvement can be tailored to their respective clusters. Traditional clustering methods like K-Means and Agglomerative Clustering often struggle with uncertainty in digital literacy data. This study introduces the NSS-ABC clustering method, which combines N-Soft Set (NSS) theory with the Artificial Bee Colony (ABC) algorithm, to improve the accuracy of clustering digital literacy profiles in open and distance education students.

Objectives: The primary objective of this research is to develop and evaluate the performance of the NSS-ABC clustering method for digital literacy data. The purpose of the study is to ascertain whether NSS-ABC performs better at managing uncertainty and enhancing cluster separation than traditional clustering techniques.

Methods: The population of this study was students of Universitas Terbuka, with a sample size of 3088. The samples were taken randomly with a research instrument in the form of a questionnaire distributed online. Based on data processing using the NSS-ABC algorithm, UT students' grouping, and digital behavior patterns were obtained based on generation profiles. The collected data was preprocessed before being subjected to clustering analysis using the NSS-ABC algorithm. The NSS-ABC method was implemented by integrating N-Soft Set decision making principles with the optimization capabilities of the Artificial Bee Colony algorithm to enhance cluster performance.

Results: The NSS-ABC, K-Means, and Agglomerative Clustering algorithms are used to analyze primary data on distance learning students’ digital literacy. The results show that the NSS-ABC hybrid approach regularly outperforms K-Means and Agglomerative Clustering, particularly when three to six clusters are used. This shows that the combination of N-Soft Set and ABC can optimize clustering by handling data uncertainty better than conventional methods. The division of clusters in digital literacy can facilitate various parties in determining programs that suit the needs of each cluster. 

Conclusions: The NSS-ABC method improves clustering by addressing data uncertainty and oprimizing group formations. The results suggest that the method can be applied to digital transformation policies in open and distance education.