Mário R. G. Meireles Filho Meireles Filho, M. R. G.

Aldir S. Sousa Sousa, A. S.

Resumo

Changing climatic factors such as wind, temperature and humidity are one of the factors that directly affect honey production. In this work, some combinations of neural network models (CNN-Dense; GRU-Dense and LSTM-Dense) were tested with the objective of predicting honey production through daily climatic data in the region. All studied networks were successful in reaching predictions with RMSE and standard deviation indices of less than 5% of production for the period studied, with emphasis on the LSTM-Dense model, where the best indices were obtained, thus demonstrating the effectiveness of use of Artificial Neural Networks in the recognition of weather patterns for the prediction of honey.