Research Article
An Efficient Approach for Food Demand Forecasting Using an Ensemble Technique and Statistical Analysis
@INPROCEEDINGS{10.1007/978-3-031-77075-3_6, author={Dudla Anil kumar and Bathula Thirupathi Rao and Bathini Rangaswamy and Kagitha Meghana}, title={An Efficient Approach for Food Demand Forecasting Using an Ensemble Technique and Statistical Analysis}, proceedings={Cognitive Computing and Cyber Physical Systems. 5th EAI International Conference, IC4S 2024, Bhimavaram, India, April 5--7, 2024, Proceedings, Part-I}, proceedings_a={IC4S}, year={2025}, month={2}, keywords={Food Demand Forecasting Machine Learning XGBoost CatBoost Customer Satisfaction Ensemble Technique Hyperparameter Tuning Data Visualization Gradient Boosting Restaurant Management and Optimization}, doi={10.1007/978-3-031-77075-3_6} }- Dudla Anil kumar
Bathula Thirupathi Rao
Bathini Rangaswamy
Kagitha Meghana
Year: 2025
An Efficient Approach for Food Demand Forecasting Using an Ensemble Technique and Statistical Analysis
IC4S
Springer
DOI: 10.1007/978-3-031-77075-3_6
Abstract
By utilizing machine learning, the current research offers a fresh approach to food demand forecasting in restaurants. It focuses on store-specific models that take into account a variety of variables, including location, weather, and events. Conventional methods frequently ignore store-specific subtleties in favor of relying exclusively on point-of-sale (POS) data. We provide a food demand forecasting model that utilizes machine learning methods, specifically XGBoost and CatBoost, to incorporate various data sources. The study goes into how the model was created, how real shop data was used to validate it, and how it was used to assess variables that affect consumer happiness, especially in the context of the fast-food trend. The study explores the variables influencing preferences for fast-food outlets and offers insights into consumer satisfaction and how it affects the fast-food restaurant industry as a whole. The benefits of the presented hypothesis, along with the algorithms and libraries used (Pandas, NumPy, Scikit-learn, and CatBoost) are described in depth. The study also emphasizes the significance of data visualization, ensemble techniques, hyperparameter adjustment, and other gradient-boosting libraries like XGBoost and CatBoost. Libraries for time series analysis are also thought to be useful for capturing data's temporal patterns. This all-encompassing method helps to create a strong and effective framework for food demand forecasting, which is crucial for restaurant management and optimization.
