Unsupervised Clustering Methods Comparison

This question requires a comprehensive comparison of several unsupervised clustering approaches, detailing their suitability for different scenarios.

At a minimum, you must discuss and compare the following methods:

• Centroid-based clustering (e.g., K-Means)
• Hierarchical clustering
• Density-based clustering (e.g., DBSCAN)
• Gaussian Mixture Models (GMM)
• Spectral clustering

For each method, your discussion should cover:

• Its core objective or underlying intuition.
• Assumptions it makes regarding cluster shape and scale.
• Important hyperparameters, including strategies for their selection.
• Its computational complexity and scalability characteristics.
• Its primary strengths and common failure modes.
• Techniques for evaluating clustering quality in the absence of ground truth labels.
• Strategies for preprocessing high-dimensional embeddings or sparse features prior to applying the clustering method.