Data Mining Forecasting Methods for Predicting the Availability of Sustainable Rice Commodities in Bali Province

This research aims to prepare the capacity of local governments to provide sufficient strategic food production, especially rice, which is the commodity most consumed by the people of Gianyar Regency every year. This research is descriptive research with a quantitative approach. The data used in the analysis is secondary data in the form of rice production data (tons), rice production (tons), harvested area (ha), public consumption of rice commodities (kg) and population (people) from 2018 to 2022. obtained from the Central Statistics Agency (BPS) and the Bali Province Agriculture and Food Security Service. The results of this research show that the forecasting method using the Backpropagation algorithm can be used to provide an approach to forecasting rice harvest area, rice production, rice production, community consumption levels, and population in 2023 and 2024 in Gianyar Regency. The Backpropagation architectural design uses 12 input layers, 10 hidden layers, and 1 output with Mean Absolute Percentage Error (MAPE) results of 0.22 and Mean Absolute Deviation (MAD) of 339 with a very good forecasting category. The average ratio of normative per capita consumption to food availability, especially rice commodities in Gianyar Regency in 2023 will experience a high deficit condition with a Cnorm Ratio value (z) = 1.52 and in 2024 will experience a high deficit condition with a Cnorm Ratio value (z) = 1.57. The forecasting method using the Backpropagation Algorithm can still be developed by focusing on each district by calculating the overall food commodities in that district.