Analytics in Manufacturing: A Case Study from Aluminum Recycling

• Activity Number: 21
• Type: Topic Contributed
• Date/Time: Sunday, August 1, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): J. Michael Hardin*+ and Michael D. Conerly
• Companies: The University of Alabama and The University of Alabama
• Keywords: logistic regression ; data warehousing ; model recentency ; real-time Instrumentation

Abstract

Statistical methods have played an important role in manufacturing over the past 20 years. In particular, the use of quality control methodology has enabled companies to decrease defective rates, and improve their products and customer satisfaction. Recently, however, other analytical methods have been introduced to address business problems. Often referred to as Business Analytics, these methods include techniques from statistical learning theory and machine learning. Through the use of extensive digital sensors and the ability to capture their measurements in data warehouses, the potential exists to build predictive models to further exist manufactures in achieving greater efficiency in their processes. In this paper we present the results of a project involving the development of a predictive model to assist in decision support related to the recycling of aluminum in coils.