Improving Penalized-Based Clustering Model in Big Fusion Data by Hybrid Black Hole Algorithm

Sarah G. M. Al- Kababchee¹, Zakariya Y. Algamal², Omar S. Qasim³

¹Department of Mathematics, Education College, University of AL-Hamdaniya

²Department of Statistics and Informatics, University of Mosul, 41002 Mosul, Iraq; College of

Engineering, University of Warith Al-Anbiyaa, 56001 Karbala, Iraq

³Department of Mathematics, University of Mosul, Mosul, Iraq

Email: sarahghanim@uohamdaniya.edu.iq; zakariya.algamal@uomosul.edu.iq; omar.saber@uomosul.edu.iq

Abstract

This paper presents an improved penalized regression-based clustering algorithm using a nature-inspired approach. Clustering is an unsupervised learning method widely used in data fusion mining, including gene analysis, to group unclassified fusion data based on their features. The proposed algorithm is an extension of the "Sum of Norms" model and aims to better estimate the data by fusing information from various sources. The performance of the proposed algorithm is evaluated on gene expression data. Results show that our approach outperforms other methods, indicating its potential impact on clustering research with data fusion.

Keywords: Black hole algorithm; Data fusion mining; Clustering fusion data, Bat algorithm; K-means.