Authors:
Jouma Ali Al-Mohamad, Ghefar Alaa Aldeen Alrefai, Mykhailo Paslavskyi, Rahul Chauhan, M. Rehena Sulthana, Prasanna Ranjith Christodoss
Addresses:
Department of Computer and Mobile Communication Engineering, Faculty of Information Engineering, Al-Shahbaa Private University, Aleppo, Syria. Department of Informatics and Communications Engineering, Faculty of Engineering, EBLA Private University, Heraitan, Aleppo, Syria. Department of Computer Science, Ukrainian National Forestry University, Lviv, Lviv Oblast, Ukraine. Department of Fitch Ratings, The Fitch Group, New York, United States of America. School of Information Technology and Engineering, Melbourne Institute of Technology, Melbourne, Victoria, Australia. Department of Computing, Mathematics and Physics, Messiah University, Mechanicsburg, Pennsylvania, United States of America.
Abstract:
Unsupervised learning (UL) is the cornerstone of exploratory data analysis, enabling the discovery of hidden structures, clusters, latent representations, and anomalies without labeled examples. In an era of exponentially growing unlabeled data – from IoT sensor streams to medical images and customer transactions – UL provides the essential toolkit for pattern discovery, dimensionality reduction, and generative modeling. This paper delivers a rigorous, technically deep, and application-driven survey of unsupervised learning. Researchers cover: foundational concepts (probability density estimation, latent variable models, information theory), clustering algorithms (k means, DBSCAN, hierarchical, spectral, Gaussian mixture models), dimensionality reduction (PCA, t SNE, UMAP, autoencoders), association rule mining (Apriori, FP Growth, Eclat), anomaly detection (statistical, proximity based, isolation forests, autoencoder reconstruction error), evaluation metrics (silhouette score, Davies–Bouldin index, adjusted Rand index, reconstruction loss, domain specific validity), modern advances (contrastive learning, self-supervised learning, variational autoencoders, generative adversarial networks for unsupervised tasks), and real world case studies (customer segmentation, gene expression clustering, network intrusion detection, recommender systems). Researchers provide a decision framework for selecting UL algorithms based on data characteristics (size, dimensionality, noise, cluster shapes) and propose a benchmarking methodology for fair comparison.
Keywords: Unsupervised Learning; Clustering and Dimensionality Reduction; Anomaly Detection; Pattern Discovery; Association Rule Mining; Evaluation Metrics; Self-Supervised Learning.
Received on: 08/01/2025, Revised on: 01/03/2025, Accepted on: 20/05/2025, Published on: 05/06/2026
DOI: 10.64091/ATICL.2026.000326
AVE Trends in Intelligent Computer Letters, 2026 Vol. 2 No. 2 , Pages: 67-81