pelt

Time series changepoint detection

Fast Change Point Detection (R Mirror)

A library testing for python, this package allows us to test algorithms and generate test data.

Changepoint detection with Pruned Exact Linear Time

Data Mining Project - STIB network quality assessment code

Auto-ARIMA timeseries forecasting in combination with PELT changepoint detection to predict social media viewership performance and identify major changes in performance trends. Models are deployed into a streamlit webapp for analytical functionality.

Changepoint detection toolkit for offline and online in Rust with Python bindings

6. Vreme / lag ACF, PACF, cross-correlation, lagged MI, Granger causality, transfer entropy, spectral density, FFT/periodogram, wavelet, change-point (PELT, binary segmentation, CUSUM), STL/decomposition, seasonal strength

Event-based ML framework analyzing structural morphology of rainfall events in Pakistan (1981–2024). Segments 47K+ dekadal records into 4,563 events, discovers 7 rainfall typologies via GMM clustering, detects a 1996 regime shift with PELT, and classifies typologies with 96.6% accuracy (Random Forest).

Code, data, figures, and tables for the paper introducing ρ a changepoint robustness metric and a causal regime-shift augmentation for reliable demand forecasting under structural change. Submitted to Knowledge-Based Systems (Elsevier).

Fast change point detection in Python using PELT and SeGD algorithms. Supports multiple model families including mean/variance, GLM, regression, ARMA, and GARCH.

Bayesian changepoint detection for insurance pricing — BOCPD and PELT, exposure-weighted Poisson-Gamma, UK regulatory event priors, Consumer Duty reports