Integrating machine learning for rice yield prediction in Assam: a web application development approach

Student name: Ms Gayatri Bhagawati
Guide: Dr Ayushi Vijhani
Year of completion: 2024
Host Organisation: India meteorological Department
Supervisor (Host Organisation): Ms Suman Gurjar

---

Abstract:

Rice is an important staple crop in Assam, India, making yield forecast critical for good agricultural planning and food security. From 2005 to 2019, this study used machine learning to estimate rice output in 27 districts of Assam using meteorological data. The Multi-layer Perceptron (MLP) Regressor proved to be the most successful model, which was further refined using grid search for hyperparameter tuning.

A Flask-based web application was created that enables user interaction with the predictive model, enabling users to enter important features and get yield projections. The tool displayed strong predictive performance, allowing farmers and politicians to make educated decisions. The research presented here highlights machine learning's potential to boost agricultural productivity and sustainability.

Keywords: Rice yield prediction, Machine learning, Assam agriculture, Weather parameters, MLP Regressor.