Where I started — a UCL undergrad project on K-means initialisation

This is where my publication record actually starts. UCL Data Science, STAT0041 (Computational Statistics). The course was taught by Professor Yvo Pokern, and the experience reshaped how I think about three things at once — statistics, algorithms, and programming. His teaching had a particular German rigour to it: an algorithm is not just a recipe to memorise but an object you derive, dissect, and rebuild from first principles. After that module I no longer experienced “stats” and “code” as two separate subjects. They were the same craft. This paper was the first thing I wrote with that perspective in mind.

The annoyance that drove it: random and k-means++ initialisation can place initial centres in low-density regions, then converge to clusters that don’t match anything a marketer would recognise. The fix was almost embarrassingly simple — pick initial centres from the centroids of the densest grid cells in feature space. The interesting bit was the data representation. Treating customer features as interval-valued data (each feature an interval [min, max] capturing per-customer variability) rather than as point estimates preserved a layer of information that simpler approaches squashed.

On simulation experiments: silhouette +0.1249 over k-means++, +0.4903 over vanilla K-means.

Looking back, the methodological reflex that runs through everything I’ve done since started here. What is the data actually telling you, beyond the point estimate? That question carries from clustering all the way to TikTok exposure measurement and clinical LLM evaluation.

Conference paper at ICIICS 2023 (IEEE): link.