Air Quality Prediction System in Lima Using Long Short Term Memory

Cadenas G.; Alvarez M.; Montero R. · Proceedings - 2025 3rd International Conference on Intelligent Control and Computing, IC and C 2025 (2025) · DOI: 10.1109/ic-c68228.2025.00024

Resumen

The city of Lima has a high level of air pollution due to the increase in the number of vehicles using fossil fuels and the lack of natural aspects that can mitigate it. This is detrimental to people's quality of life. Authorities lack tools that allow them to make decisions on environmental aspects. This research has aimed to create a system that allows measuring air quality in different environments of the city in order to provide quantitative aspects that facilitate taking actions in favor of decontamination. The proposed system is based on LSTM neural networks to model a real-time prediction of air quality that allows the metropolitan municipality of Lima to take action on environmental policies. The results have shown that LSTM outperforms other models such as DTR and SVR by better capturing temporal dependencies, so the validations of the system have been very satisfactory and based on them it has been possible to generate policies on vehicle access in places with a high degree of pollution, which benefits citizens.

Instituciones: Universidad Peruana de Ciencias Aplicadas