Dengue Vector Population Forecasting Using Multisource Earth Observation Products and Recurrent Neural Networks

This article introduces a technique for using recurrent neural networks to forecast Ae. aegypti mosquito (Dengue transmission vector) counts at neighborhood-level, using Earth Observation data inputs as proxies to environmental variables. The model is validated using in situ data in two Brazilian cities, and compared with state-of-the-art multioutput random forest and k-nearest neighbor models. The approach exploits a clustering step performed before the model definition, which simplifies the task by aggregating mosquito count sequences with similar temporal patterns.

Automated summary

This article presents a technique using recurrent neural networks to forecast Ae. aegypti mosquito counts, with Earth Observation data inputs as proxies. The method involves a clustering step before model definition, aggregating mosquito count sequences with similar temporal patterns to simplify the task. The model is validated and compared with other state-of-the-art models using in situ data in Brazilian cities.