AI-Driven Market Demand Prediction: An Artificial Neural Network Model for Horticultural Commodities

The annual increase in population drives a corresponding rise in demand for goods and services, which in turn directly impacts agricultural needs at both regional and national levels. In Sikka Regency, the demand for horticultural commodities has grown alongside population expansion and increasing prosperity. However, the lack of proper planting distance management has disrupted efficient distribution, leading to logistical inefficiencies. As a result, it is crucial to develop an optimal market demand plan for horticultural commodities to prevent surplus accumulation. To address this challenge, this study proposes a predictive model for monthly market demand in 2024, focusing on 12 types of horticultural commodities in Sikka Regency. Food consumption time series data from 2019 to 2023 are used in the model. A back-propagation algorithm is utilized by an Artificial Neural Network (ANN) to make this prediction. With a Mean Absolute Error (MAE) of 0.166 and a Root Mean Square Error (RMSE) of 0.201, the model's performance evaluation showed an average training accuracy. The average testing accuracy, on the other hand, had an MAE of 0.303 and an RMSE of 0.338. These experimental results show that the suggested strategy might be a useful tool for horticulture commodities market participants and provide insightful information about market dynamics.