B. Flores Barrio
Time series of non-negative integers (counts) arise in many areas like finance, epidemiology, transport or retail. Having models that adequately represent the time series and provide accurate forecasts is essential in those domains. We propose Bayesian state-space models that are flexible enough to adequately forecast high and low count series and exploit cross-series relationships with a multivariate approach. This is illustrated with a large scale sales forecasting problem faced by a major retail company, integrated within its inventory management planning methodology. The company has hundreds of shops in several countries, each one with thousands of references.
Palabras clave: Count time series, Sales forecasting, Bayesian analysis, Dynamic generalized linear models, Inventory management, Retail.
Programado
Sesión Invitada Math-In. Industrial Applications I
8 de junio de 2022 12:40
Aula Magna
