MUMBAI, India, June 22 -- Intellectual Property India has published a patent application (202621048319 A) filed by Symbiosis International Deemed University on April 15, 2026, for A Stacking Ensemble Machine Learning System And Method For Agricultural Crop Price Prediction.
Inventors include Yash Khadgi; Rudrakshi Bhandekar; Vedant Nawghare; and Dr. Priya Dasarwar.
The application for the patent was published on June 12, 2026, under issue no. 24/2026.
Abstract: ABSTRACT A STACKING ENSEMBLE MACHINE LEARNING SYSTEM AND METHOD FOR AGRICULTURAL CROP PRICE PREDICTION The present invention discloses a stacking ensemble machine learning system (100) and method for agricultural crop price prediction. The system comprises a data collection module (110) for acquiring multi-source agricultural data including soil nutrients, weather conditions, and historical prices; a preprocessing module (120) implementing missing value imputation, categorical encoding, feature scaling, and outlier treatment; a feature selection module (130) employing correlation analysis, VIF assessment, and importance scoring; a model training module (140) comprising XGBoost (141), Ridge Regression (142), and Random Forest (143) base learners; a stacking ensemble module (150) with an XGBoost meta-learner (151) that aggregates base model predictions to generate optimized price forecasts; and a prediction output module (160) delivering results to end users. The system achieves R-squared accuracy of 91.36% and MAE of 0.18, significantly outperforming individual models, providing farmers and policymakers with reliable decision support for agricultural market planning. [
Disclaimer: Curated by HT Syndication.
