IDTS5101 - Data Clustering

This course provides a comprehensive introduction to data clustering within the context of data mining and unsupervised machine learning algorithms. It focuses on understanding how modern intelligence systems discover underlying structural patterns, segment populations, and partition unlabeled datasets into homogenous groups without prior training labels.

Students will develop essential knowledge of clustering principles, including similarity metrics, vector space representations, and proximity criteria, while exploring core algorithmic taxonomies such as partitioning (K-Means/K-Medoids), hierarchical linkage systems (Agglomerative/Divisive), and density-based frameworks (DBSCAN). The course also examines the role of internal and external validation metrics (Silhouette Index, Davies-Bouldin) in ensuring model stability and scalability across high-dimensional features.

In addition, the curriculum integrates ethical, legal, and regulatory considerations, including profiling bias mitigation, data privacy compliance, and responsibilities of institutions in protecting sensitive demographic variables during automated segmentation. Emerging trends such as big data scaling configurations and advanced visualization pipelines are also explored. Through real-world case studies and practical software implementation, students will gain the ability to deploy robust unsupervised clustering models and propose strategic, data-driven solutions to industry optimization challenges.