Red onion price factors correlation identification and price prediction using multiple machine learning models for Jaffna district Sri Lanka

Red onion is a commercial vegetable crop that was widely farmed in Sri Lanka. Red onion prices are fluctuating, and it is difficult to identify price patterns in Sri Lanka. In the Jaffna district, red onion cultivation farmers have been facing losses for the last ten years due to traditional price forecasting techniques. There are several factors influencing the price of red onions in the market. In the research on fertilizer prices, the big onion price, labour cost, rainfall, and United States dollar exchange were considered important to the red onion price for data pre-processing and modelling. The research aims to identify the correlation between factors and predict the price using different machine-learning techniques. The auto-regression (AR), long short-term memory (LSTM), and random forest regression supervised machine learning models were used to identify high-accuracy, good-fit red onion prices. The performance of the models was evaluated based on mean absolute percentage error (MAPE). The auto-regression model performed well compared to other models. The model's accuracy will help predict the price of red onions for months and years.