Machine Learning Paradigm for Internet of Things Applications

2An Empirical Study on Paddy Harvest and Rice Demand Prediction for an Optimal Distribution Plan

W. H. Rankothge

Sri Lanka Institute of Information Technology, Colombo, Sri Lanka

Abstract

Rice is one of the main types of nutrient supplying foods in Asian communities and has a direct impact on the socioeconomic development of the communities. Generally, paddy harvest, price, and demand for rice depend on several factors: rainfall, humidity, paddy cultivation area, temperature, etc. Therefore, precisely predicting the future harvest of paddy and consumption demand of rice is complicated. It creates a necessary requirement of a management platform to predict future paddy harvest and rice demands considering important parameters. Such platform will help the countries in the processes of national strategy development, to maintain a sustainable approach between paddy cultivation and rice demand.

We have proposed a centralized management platform, where we would analyze numerous factors affecting on the amount of crop yield on the next harvesting season and the factors affecting on future demand for rice. The proposed platform follows the smart agri/farm approach for paddy related processes, with three functional features: (1) paddy harvest prediction module, (2) rice demand prediction module, and (3) an optimal planning module for crop distribution. Two popular machine learning algorithms, Recurrent Neural Network (RNN) and Long Short-Term Memory (LSTM), are used to develop the ...