MUMBAI, India, June 30 -- Intellectual Property India has published a patent application (202641056388 A) filed by Soumyadip Chattopadhyay; Anurag Sinha; and Dr. Iniyan S on May 04, 2026, for Machine Learning Framework For Yield Prediction Using Soil Health And Climate Data.
Inventors include Soumyadip Chattopadhyay; Anurag Sinha; and Dr. Iniyan S.
The application for the patent was published on June 26, 2026, under issue no. 26/2026.
Abstract: Under climate changing condition, sustainable agriculture and intelligent decision making is very necessary. Change in weather and soil distribution in the earth had made the work even tougher specially for the statistical procedures which are being used now for forecasting the yield of crops. In our model we had integrated climate and soil health for predicting the yield of the crops. For the project we have taken two crops i.e coni and soyabean and collected the climate and soil health dataset required for these crops. Our datasets contain of weekly weather data, characteristics of crops and soil at many levels. We have removed the non-informative features and zero-valued predominant features by using data preprocessing to make our model more robust. We have used many regression models like linear regression, decision tree, k-nearest neighbor, support vector regression, random forest, gradient boosting regression methods, ensemble regression and XGBoost. Out of all of these models XGBoost regression model was having the highest performance metrics, so we fine-tuned our model with that. We have built a platform where farmers can give input of the soil and climate health and they can get the predicted output of the yield of the crops.
Disclaimer: Curated by HT Syndication.
