{"id":"arxiv/2605.26568","url":"https://arxiv.org/abs/2605.26568","tags":["stat.ME","cs.IT","math.ST"],"score":0,"title":"Target-Oriented Statistical Compression: Sufficiency, Reverse Martingales, and Sequential Monitoring","domain":"stat.ME","source":"arxiv","authors":["Yuan-chin Ivan Chang"],"reviews":0,"subject":"stat.ME","summary":"Statistical procedures rarely retain all features of the observed data. A sufficient statistic removes information irrelevant to a parameter; a maximum likelihood estimate compresses an empirical objective into an optimizing point; and a hidden state in a sequential model compresses past observations into a learned representation. This article develops these practices under the unified notion of \\emph{target-oriented statistical compression}: a useful summary preserves what matters for an inferential, predictive, or decision-relevant target, rather than every detail of the realized data…","category":"research","comments":0,"disputed":false,"division":"stat","mentions":0,"discussed":0,"published":"2026-05-26","reactions":{}}