MUMBAI, India, July 24 -- Intellectual Property India has published a patent application (202641085153 A) filed by Jntuh University College Of Engineering on July 11, 2026, for Walmart Sales Forecasting Using Machine Learning And Time-Series Models.
Inventors include Dr. K. Santhi Sree; Ujjelli Lokesh Reddy; Kathi Mounika; Pedapalli Sobhitha; Mellacheruvu Venkata Gaayatri; C Shiva Rama Krishna; M. Srimani; Gurri Suvidha Reddy; Erri Suvarna; Hanumandla Naga Shiva; Chirra Karthik; V. Nithin; and Kayam Sravani.
The application for the patent was published on July 17, 2026, under issue no. 29/2026.
Abstract: The present invention discloses a hybrid retail sales forecasting system integrating machine learning and time-series modelling for accurate weekly sales prediction. The system comprises a Data Preprocessing Module for preparing historical sales data, a Feature Engineering Module for generating temporal predictive features, a Prophet Forecasting Module for modelling trend, seasonality, holiday effects, and changepoints, a Residual Computation Module for calculating forecasting errors, an XGBoost Residual Prediction Module for learning nonlinear residual patterns, a Hybrid Forecast Generation Module for combining Prophet forecasts with predicted residual values, an Evaluation Module for assessing forecasting performance using Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and Mean Absolute Percentage Error (MAPE), and a Forecast Visualization Module for presenting forecast results. The integrated framework combines statistical time-series decomposition with machine learning-based residual correction to improve forecasting accuracy, reduce prediction error, and generate scalable, interpretable, and reliable weekly sales forecasts suitable for inventory planning, supply chain management, and retail business decision support.
Disclaimer: Curated by HT Syndication.
