MUMBAI, India, Aug. 31 -- Intellectual Property India has published a patent application (202641100738 A) filed by Vignan'S Nirula Institute Of Technology And Science For Women on August 20, 2026, for An Intelligent Machine Learning-Based System And Method For Predicting Agricultural Productivity Using Soil, Weather, And Crop Parameters.
Inventors include G. Rohini Phaneendra Kumari; Jagadish Kalava; Srikanth Kilaru; G. Kusuma Harinadh; Thalla Vasavi Surya Prabha; Bandla Himavarshini; Kilaru Vyshnavi; and Domathoti Nadiya.
The application for the patent was published on August 28, 2026, under issue no. 35/2026.
Abstract: This invention pertains to an intelligent machine learning based system and method for forecasting agricultural productivity based on a combination of various agricultural parameters like soil parameters, weather parameters, and other agricultural parameters. The invention is more specifically related to a predictive mechanism that is able to analyze various agricultural data from heterogeneous sources like soil testing laboratories, meteorological stations, agricultural research organizations, remote sensing satellites, IoT sensors, agricultural databases, and field measurements. The invention under consideration includes various computational modules like data acquisition from agriculture, intelligent data pre-processing, feature extraction, intelligent feature selection, creation of the machine learning models, productivity prediction, performance assessment, and decision support generation. Data pre-processing consists of missing value imputation, duplicate record elimination, noise filtering, outlier identification, normalization, feature scaling, and consistency testing of the data set. The intelligent feature selection methods help in selecting the most critical parameters related to soil, climate, and crops while eliminating any unnecessary data, thus helping to increase the prediction accuracy. The processed data set is used to train several machine learning models that can discover nonlinear relationships between agricultural parameters. The developed prediction model is tested against various performance measures in statistics, which include prediction accuracy, precision, recall, F1 measure, MAE, RMSE, MAPE, and coefficient of determination (R²) in order to ensure that the prediction model achieves reliable and robust performance under varying agricultural environments. The prediction results obtained from the prediction model are then provided to an intelligent decision support module that will aid the farmers, agricultural scientists, agricultural organizations, and decision makers in selecting crops, managing fertilizers, planning for irrigation and cultivation, allocating resources, and conducting risk assessment in agriculture. This invention provides a better approach in terms of prediction accuracy, computing speed, scalability, adaptability, and application than the traditional agricultural productivity prediction systems. Therefore, this invention contributes to precision agriculture, smart farming, agricultural resource efficiency, increased crop productivity, sustainability, and agricultural development.
Disclaimer: Curated by HT Syndication.
