Robotics and Autonomous Systems 49 (2004) 137–152 Discovering useful and understandable patterns in manufacturing data Mark Last a,∗ , Abraham Kandel b a Department of Information Systems Engineering, Ben-Gurion University of the Negev, Beer-Sheva 84105, Israel b Department of Computer Science and Engineering, University of South Florida, 4202 E. Fowler Avenue, ENB 118, Tampa, FL 33620, USA Received 8 March 2002; received in revised form 5 August 2004 Abstract Accurate planning of produced quantities is a challenging task in semiconductor industry where the percentage of good parts (measured by yield) is affected by multiple factors. However, conventional data mining methods that are designed and tuned on “well-behaved” data tend to produce a large number of complex and hardly useful patterns when applied to manufacturing databases. This paper presents a novel, perception-based method, called Automated Perceptions Network (APN), for automated construction of compact and interpretable models from highly noisy data sets. We evaluate the method on yield data of two semiconductor products and describe possible directions for the future use of automated perceptions in data mining and knowledge discovery. © 2004 Elsevier B.V. All rights reserved. Keywords: Data mining; Knowledge discovery; Info-Fuzzy Network; Automated Perceptions Network; Yield management 1. Introduction The outgoing yield of manufactured batches is the leading efficiency criterion in semiconductor industry. The semiconductor yield is defined as a ratio between the actual and the maximum number of good parts (mi- croelectronic chips) in a finished batch. Since capital- ization expenses constitute a major part of manufac- ∗ Corresponding author. E-mail addresses: mlast@bgumail.bgu.ac.il (M. Last), kandel@csee.usf.edu (A. Kandel). turing costs in semiconductor industry, the direct cost of producing a single batch is almost fixed and rela- tively low. On the other hand, the income from a given batch is directly proportional to the number of good parts. Thus, there is a close relationship between the yield and the profitability of semiconductor companies and any information on the actual yield performance of these companies is considered highly confidential. The expected yield of every batch is an important decision parameter for the Production Control Depart- ment. An “optimistic” estimate of the outgoing yield may cause delays in the delivery of a customer order 0921-8890/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.robot.2004.09.002