ztxtech/cep_ts

This is an official PyTorch implementation of "Continuous Evolution Pool: Taming Recurring Concept Drift in Online Time Series Forecasting"

22
/ 100
Experimental

This project helps businesses and researchers make more accurate predictions when analyzing time-series data that changes its patterns over time, especially when old patterns reappear. You input your historical time-series data, and it outputs improved forecasts that adapt to these shifting trends without forgetting past knowledge. This is useful for data scientists, financial analysts, operations managers, or anyone relying on predictions from real-time data streams.

No commits in the last 6 months.

Use this if your online time-series forecasts are often inaccurate because the underlying data patterns (concepts) frequently change and reappear, leading to outdated models.

Not ideal if your time-series data has very stable patterns without significant, recurring concept shifts, or if you need to optimize for extremely low latency over prediction accuracy.

time-series-forecasting financial-prediction energy-management traffic-forecasting environmental-monitoring
Stale 6m No Package No Dependents
Maintenance 2 / 25
Adoption 5 / 25
Maturity 15 / 25
Community 0 / 25

How are scores calculated?

Stars

11

Forks

Language

Python

License

MIT

Last pushed

Jun 19, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/ztxtech/cep_ts"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.