Article: Cluster analysis of innovation and entrepreneurship education quality based on mutual information features Journal: International Journal of Information and Communication Technology (IJICT) 2026 Vol.27 No.88 pp.93 - 113 Abstract: In the quality evaluation of innovation and entrepreneurship education, redundancy of the original indicator system often leads to ambiguous clustering boundaries, failing to distinguish substantial differences among universities. To address this, this paper adopts mutual information screening to select the feature subset most correlated with educational quality, then conducts cluster analysis on universities. Experiments on 2,100 student entrepreneurial projects show that after mutual information screening, normalised mutual information increases from 0.623 to 0.847, the Davies-Bouldin index decreases from 1.052 to 0.738, and inter-group differences of two domain-specific indicators (entrepreneurial intention retention rate and industryacademia integration tightness) expand by 38.6% and 42.3%, respectively. This framework reveals that innovation score, funding amount, and team experience are core factors driving educational quality differentiation, providing a quantifiable classification basis for educational management departments. Inderscience Publishers - linking academia, business and industry through research

Title: Cluster analysis of innovation and entrepreneurship education quality based on mutual information features

Authors: Xiujuan Fan

Addresses: School of Culture and Media, Zhanjiang University of Science and Technology, Zhanjiang, 524000, China

Abstract: In the quality evaluation of innovation and entrepreneurship education, redundancy of the original indicator system often leads to ambiguous clustering boundaries, failing to distinguish substantial differences among universities. To address this, this paper adopts mutual information screening to select the feature subset most correlated with educational quality, then conducts cluster analysis on universities. Experiments on 2,100 student entrepreneurial projects show that after mutual information screening, normalised mutual information increases from 0.623 to 0.847, the Davies-Bouldin index decreases from 1.052 to 0.738, and inter-group differences of two domain-specific indicators (entrepreneurial intention retention rate and industryacademia integration tightness) expand by 38.6% and 42.3%, respectively. This framework reveals that innovation score, funding amount, and team experience are core factors driving educational quality differentiation, providing a quantifiable classification basis for educational management departments.

Keywords: mutual information; feature selection; innovation and entrepreneurship education; clustering evaluation.

DOI: 10.1504/IJICT.2026.155568

International Journal of Information and Communication Technology, 2026 Vol.27 No.88, pp.93 - 113

Received: 25 Apr 2026
Accepted: 01 Jun 2026
Published online: 05 Aug 2026 *

Title: Cluster analysis of innovation and entrepreneurship education quality based on mutual information features

Keep up-to-date