An efficient machine learning model for multi-crop price prediction usingRandom forest

Abstract

This paper introduces the development and design of an application in Streamlit for multi-crop market price forecasting based on machine learning (ML). We specifically adopt the applied statistical model of Random Forest and a descriptive data set related to agriculture in India. Our app supports farmers and other market stakeholders with accurate and real-time price forecasts for different crops across multiple regional markets. Farmers and stakeholders can provide crop, location, and other parameters to customize and personalize their market forecasting through an easy-to-use, interactive user interface. The application also offers analysis of crop profitability for regional markets and helps the farmer optimize crops based on optimal yield and margin for maximum returns on investment. The user interface uses rich bar charts and trend graphs for more meaningful interpretation and action planning. The application is a lightweight and robust system that supports raw use on personal computers or local servers in remote and isolated areas of the world. Our integrated model of prediction, user interaction, multi-crop integration, and visualization, offers a unique and valuable service and tool at absolutely no cost to Indian agriculture and the agriculture sector. It offers a free resource for sustainability and enhanced engagement between farmers and local markets