Forecasting COVID-19 new cases using NBEATS deep learning and mobility data
Amril Nazir, Mohammad Shorfuzzaman, Muhammad Lujaini Lotfi, Firuz Kamalov, Sufian Badawi, Maen Takruri, Abdul-Halim Jallad · 2026
Read Forecasting COVID-19 new cases using NBEATS deep learning and mobility data by Amril Nazir, Mohammad Shorfuzzaman, Muhammad Lujaini Lotfi, Firuz Kamalov, Sufian Badawi, Maen Takruri, Abdul-Halim Jallad (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/plos-pone-0350264.txt.
Description
Amril Nazir and six colleagues forecast COVID-19 case counts with a neural network and phone movement data. Their 2026 study tests a model called N-BEATS in four large countries: Brazil, Russia, Britain, and the United States. The virus spreads mainly through human contact. How people move should therefore foreshadow case curves. Mobility records from Google and Apple supplied that signal. Case data came from a public global dashboard.
N-BEATS beat the benchmark model everywhere. Its errors came out lower on both standard measures, across all datasets and countries. Adding mobility data improved it further. The same model without movement inputs performed worse, confirming the data's real value. Interpretability improved alongside accuracy, which matters when officials must explain forecasts. Epidemic planners gain a tool that reads tomorrow's case pressure partly from today's journeys, buying hospitals and governments precious days of warning.
Catalog record
| CREATOR | Amril Nazir, Mohammad Shorfuzzaman, Muhammad Lujaini Lotfi, Firuz Kamalov, Sufian Badawi, Maen Takruri, Abdul-Halim Jallad |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 78905 |
| LICENSE | CC_BY |
| RIGHTS | CC BY |
| SOURCE | PLOS |