Prediction of Rice Yield via Stacked LSTM

Xiangyan Meng,Muyan Liu,Qiufeng Wu

doi:10.4018/ijaeis.2020010105

Abstract

In order to guarantee the rice yield more effectively, the prediction of rice yield should be taken into account. Because the rice yield every year can be seen as a sequence of time series, many methods applied in prediction of time series can be considered. Long Short-Term Memory recurrent neural network (LSTM) is one of the most popular methods of time series prediction. In consideration of its own characteristics and the popularity of deep learning, an improved LSTM architecture called Stacked LSTM which has multiple layers is proposed in this article. It is based on the idea of increasing the depth of LSTM. The comparison among the Stacked LSTM architectures which have different numbers of LSTM layers and other methods including ARIMA, GRU, and ANN has been carried out on the data of rice yield in Heilongjiang Province, China, from 1980 to 2017. The results showed the superior performance of Stacked LSTM and the effectiveness of increasing the depth of LSTM.

Highlights

Rice is one of the most important food sources for more than half of the world’s population (Jeon et al, 2011), it is the second most widely grown cereal crop worldwide (Hirooka et al, 2018) and the demand for rice is expected to grow because of the increasing population on earth (Daniela et al, 2018)
The results showed the superior performance of Stacked Long Short-Term Memory recurrent neural network (LSTM) and the effectiveness of increasing the depth of LSTM
Three methods of time series prediction including Auto regressive Integrated Moving Average Model (ARIMA), Gated Recurrent Unit (GRU), Artificial neural networks (ANN) (Artificial Neural Network) are used here to compare with Stacked LSTM in order to prove the state of art performance of it

Summary

Introduction

Rice is one of the most important food sources for more than half of the world’s population (Jeon et al, 2011), it is the second most widely grown cereal crop worldwide (Hirooka et al, 2018) and the demand for rice is expected to grow because of the increasing population on earth (Daniela et al, 2018). Crop yield prediction is a representative measure which is vital for food security (Hutchinson, 1991). It can obtain the result whether the future crop yield can achieve the demand of population, it plays a key role in government’s policy making and preparing production plan for following year. Crop yield prediction can provide a reference for farmers and enterprise, helping them increase outcome (Na-Udom & Rungrattanaubol, 2015), so rice yield prediction is a matter of importance

Methods

Results

Conclusion