change-point-detection

What problem does it solve?

Gaussian Process Change-Point Detection (GP-CPD) enables segmentation of financial time-series into regimes by identifying change-points where statistical properties shift, supporting regime identification and improved decision-making.

Core Features & Use Cases

• Gaussian Process change-point detection that compares a Matérn GP with a CP kernel to reveal regime shifts.
• Regime segmentation using lookback windows and min/max length constraints to produce meaningful market states.
• Context-set construction for few-shot learning and adaptive trading strategies based on detected regimes.
• Use cases include momentum crash detection, adaptive strategy selection, and volatility regime analysis.

Quick Start

Run GP-CPD on a time-series to identify change-points and generate regime segments for modeling.