Sri Lanka is a country that entirely depends on rice as a primary food source and the production of rice mainly relies on the changes of weather. Due to its rising demand, prediction of paddy yield in the future years is necessary and important. This research aims to investigate the trend of the yield of rice in Sri Lanka, by analysing the regional climatic data. In addition, this study focuses on identifying weather elements which affect the paddy growth in Sri Lanka. This study was conducted in the Jaffna, Vavuniya, Mannar, Batticaloa and Trincomalee districts, based on the collected yield data during the Yala and Maha seasons. As the initial step, the weather and paddy yield data of the past 11 years were collected from the Department of Meteorology, Sri Lanka and the Department of Census and Statistics respectively. Then the collected regional weather raw data were pre-processed, and null values were filled with the corresponding mean values. In addition, a feature scaling technique was used to normalize the weather and yield data. The prediction was done using linear regression and the performance of the model was measured  using Root Mean Squared Error (RMSE). In the evaluation of the proposed study, 74 data samples were used to train the prediction model and 24 samples were used for testing. The testing results showed that the proposed study closely resembles the yield prediction, with the least (0.86) RMSE. In addition, the results showed the correlation between paddy growth and weather elements: rainfall, relative humidity, and temperature, which are +0.46, +0.35, and -0.25, respectively. On this basis, rainfall positively influences the paddy yield while temperature influences it negatively. This prediction study could be beneficial for farmers, as it predicts the future demand of rice - based on the weather changes and identifies the correlation between paddy growth and weather elements in the local regions of Sri Lanka.

Keywords:  yield prediction; linear regression; data mining; Dry Zone agriculture.