Interaction of Forecast and Response Time Series in Collaborative Inventory Management Setting

• Activity Number: 129
• Type: Contributed
• Date/Time: Monday, July 30, 2012 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistical Learning and Data Mining
• Author(s): Burcu Aydin*+ and J. Steve Marron
• Companies: Hewlett-Packard Labs and The University of North Carolina at Chapel Hill
• Keywords: Object Oriented Data Analysis ; Forecasting ; Collaborative Inventory Management ; Time Series Data ; Complex Data ; Business Applications

Abstract

Object Oriented Data Analysis (OODA) is a novel approach in statistics to explore interactions within complex structured data. OODA aims to construct the methodologies required to analyze the rich data structures that become increasingly available to researchers due to technological developments achieved in the last decades.

One of the application areas is investigating implications of forecast sharing in a Collaborative Inventory Management (CIM) setting. CIM is a widely used practice in industry, where significant amounts of procurement spend is governed by this process in many companies.

Some of the well-known forecast drivers, together with correlations generated by the interactive nature of shared forecast data and the inherent white noise create a rich and complex data set. In this work, the OODA ideas are used to analyze the interactions of forecast values with the response series in a supply chain setting with information sharing. The ultimate aim is to develop tools for determining causes of forecast errors, and their relationship with the collaborative nature of the process.