Prediction of Rice Production in Malaysia Using State Space Model

Abstract

Rice is not only a staple food in Malaysia but also the primary agricultural product, influencing the food security, employment rate, and economic stability. Hence, an accurate prediction of rice production can help in farm planning and policy construction. However, rice production is constantly evolving due to various factors, including climate change, uncertain pricing, and limited technological advancements. This paper proposes a state-space model to predict rice production, capturing the underlying dynamics of rice production data. A historical rice production dataset from 1980 to 2019 was analysed, and descriptive statistics were calculated to understand the trend in rice production. The state space model is formulated after standardising the data. The model parameters are estimated through an optimisation process to reduce the prediction error. The accuracy of the proposed model is evaluated using a mean square error (MSE) metric. The results showed that the proposed model can accurately fit the historical rice production data, with the smallest MSE value compared to the simple moving average and the exponential smoothing method. Therefore, the state space model demonstrates its effectiveness in predicting rice production, supporting data-driven decision-making in agricultural planning and policy formation.