Application of the Generalized State Space Model for Predicting and Analyzing Production Quantity in a Retail Bakery

Abstract

A production quantity prediction system, independent of subjective judgment, was developed using a generalized state space model to reduce food waste in retail bakeries. The model was constructed through formulation, time series cross-validation, and parameter estimation. The results suggest that the model successfully captured the patterns underlying production quantity decisions. Furthermore, it was confirmed that both the day of the week and total precipitation significantly influenced production quantity. Future work will focus on refining the dataset to develop a more reliable and generalized model. Key words : SDGs, food waste reduction, time series analysis, state space model, feature selection