Low-Cost Sensor Calibration for Indoor Air Quality Monitoring: A Dataset, Evaluation Scenarios, and a Lightweight Model

The paper presents a six‑month dataset of indoor air‑quality measurements from low‑cost and reference sensors, collected at five locations with contextual metadata. It defines four evaluation scenarios—reference‑efficient, location‑transfer, long‑term drift, and event‑conditioned—to test spatial generalization and robustness to distribution shifts. A lightweight temporal model that compresses input windows and fuses residual temporal and feature information achieves strong calibration performance across all scenarios while keeping edge‑inference cost low.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.