Development Of AI-ML Based Models for Predicting Prices of Agri-Horticultural Commodities Such as Pulses and Vegetables

Abstract

Agricultural economic markets are highly unstable and unpredictable. Commodities like rice, dal, wheat, onion, potato and tomato are most common commodities and highly used across India. This paper presents a strong system for forecasting price of these six commodities in state of Maharashtra as a whole. Factors that affect the price of these commodities like Weather, supply, transportation cost and government policies like Minimum Support Price (MSP) are collected, processed and trained for future predictions. Six Machine learning models are used for predictions (Extra Trees, Random Forest, Gradient Boosting, XGboost, LightGBM and stacking ensemble. Performance of each model is compared and model with best performance is used to forecast next 30 days forecasts.