sketch_kmeans1d.nx

sketch_kmeans1d.nx -- streaming 1D k-means clustering. Online single-pass clustering via Lloyd's-style centroid updates. For each new value x: 1. Find nearest centroid i = argmin |x - centroid[j]| 2. count[i] += 1 3. centroid[i] += (x - centroid[i]) / count[i] (running mean) Centroids initialized via simple priming: first k distinct values observed seed the centroids; subsequent values cluster against them. USE CASES: - log-message latency clustering (fast / slow / outlier) - sensor value bucketing - online categorical discretization - simple anomaly via distance-to-nearest-centroid COMPLEMENTS: - sketch_histogram: pre-defined uniform bins - sketch_kmeans1d: adaptive bins driven by data density LOSSLESS-LANGUAGE DISCIPLINE: nx_kmeans_query_centroid returns the running-mean centroid value with NX_ENV_REL_STDDEV ~ 1/sqrt(count). Production tier (exact integer running means, no probabilistic error).