Abstract
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.
Highlights
Z OONOTIC diseases are one of the most widespread threats to human lives in many parts of the world
This study looks to tackle the question of obtaining a qualitative time series prediction of the population of female Ae. aegypti at neighborhood-level based on Earth Observation (EO) data inputs to recurrent neural networks (RNNs)
The model was trained for one-week-ahead female Ae. aegypti population prediction starting from T training populations and environmental condition features
Summary
Z OONOTIC diseases are one of the most widespread threats to human lives in many parts of the world. Dengue is a very prevalent one of such This disease is transmitted by the female Ae. aegypti mosquito species. The environmental variables, which have shown empirical relationship with the development and population of female Ae. aegypti are precipitation, humidity, vegetation condition, and land surface temperature (LST) [1]– [4]. There is significant evidence of the effects of temperature, humidity, precipitation and surface vegetation on the life-cycle, development and density of female Ae. aegypti mosquito species in urban environments [3]. XT ) with xt ∈ Ru as input independent covariate features, a simple RNN can be expressed as follows: ht = f (ht−1, xt) (1). The access to st is controlled by three sigmoid gates: forget gate ft, input gate it, and output gate ot.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
