Change Point Detection in Irregularly Sampled Time Series: AGN Light Curves Toy Example

This repository contains the materials of the talk 'Change Point Detection in Irregularly Sampled Time Series: AGN Light Curves Toy Example' presented at the XXI Escuela de Verano en Matemáticas Discretas (Centro de Modelamiento Matemático (CMM) 2026), Universidad Adolfo Ibáñez, Viña del Mar, Chile. The implementation uses PELT (via the `ruptures` package) for segmenting the light curve, detecting changes in mean, variance, or both, and classical Binary Segmentation for the piecewise model. Each segment is modeled with quadratic + harmonic regression. Methods are standard and widely used in time series analysis.