Duo attention with deep learning on tomato yield prediction and factor interpretation

© 2019, Springer Nature Switzerland AG. Although many smart farming related approaches have been proposed to support farmers, crop modeling in smart farming and most effective factors for the yield remains an open problem. In this paper, we introduce Long Short Term Memory (LSTM) and Attention score mechanism, which gives the most effective factors to tomato yield using tomato growing under smart farm condition data set. Our finding shows that plant factors are more important as well as environmental factors. Next, we proposed DA-LSTM model for tomato yield prediction and best time frame for harvest based on a deep learning algorithm. This model shows high accuracy when compared with LSTM, XGBR and Support Vector Regression (SVR).