The Çebi Method Anomaly detection and correction in a univariate number series — a parameter-free, scale-invariant, two-sided method

A closed-form, parameter-free method for detecting and correcting anomalies — extreme values that break away from the body of a univariate number series. Both tails are pulled inward toward a band derived entirely from the series’ own structure; the method needs no threshold, training, or labels and is scale-invariant, order-invariant, and deterministic. Keywords anomaly detection · outlier correction · robust statistics · Winsorization · expected shortfall · tail estimation · parameter-free · scale-invariant