MUMBAI, India, Aug. 12 -- Intellectual Property India has published a patent application (202611071304 A) filed by Deeksha Tripathi; Saroj Kr. Biswas; Sounak Majumdar; Nutan Singh; and Yogesh Rathia on June 08, 2026, for A Crop Yield Prediction System And Method For Smart Farming Across Multiple Administrative Regions.
Inventors include Deeksha Tripathi; Saroj Kr. Biswas; Sounak Majumdar; Nutan Singh; and Yogesh Rathia.
The application for the patent was published on August 07, 2026, under issue no. 32/2026.
Abstract: The present invention discloses a crop yield prediction system (100) and method for smart farming across multiple administrative regions. The system comprises a data interface (102) for receiving numerical climatic, fertilizer, and production attributes, and a cloud server (104) for storing the attributes as feature vectors. A pre- processing unit (106) detects missing values and generates completed feature vectors, while an anomaly processing unit (108) removes outlier feature vectors based on anomaly scores. A local–global feature weighting unit (110) computes global importance scores for attributes using cluster-based weighting, and a feature scaling unit (112) generates weighted feature vectors. A chronological training controller (114) optimizes system parameters. An ensemble classification unit (116) constructs multiple decision structures to predict crop labels. A crop recommender (118) outputs a recommended crop label for a target land parcel based on majority voting across ensemble predictions to enable accurate and robust crop yield prediction.
Disclaimer: Curated by HT Syndication.
