at-tan/Forecasting_Air_Pollution

Stacking a machine learning ensemble for multivariate time series forecasting, with the goal of predicting the one-period ahead PM 2.5 air pollution level, as published in Towards Data Science on Medium.com

Archived
34
/ 100
Emerging

Implements a diverse base-learner stack (linear, tree-based, SVM, and neural network models) with OLS meta-learner on the Beijing air pollution dataset, addressing missing values and complex seasonalities through careful temporal splitting and forward-chain cross-validation. Achieves 5-6% improvement over persistence baseline on MAE/RMSE metrics, with the ensemble consistently outperforming individual base models across multivariate PM 2.5 prediction tasks.

Archived No License No Package No Dependents
Maintenance 0 / 25
Adoption 8 / 25
Maturity 8 / 25
Community 18 / 25

How are scores calculated?

Stars

43

Forks

12

Language

Jupyter Notebook

License

Last pushed

Oct 30, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/at-tan/Forecasting_Air_Pollution"

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