Authors: Lior Rokach Roni Romano Oded Maimon
Publish Date: 2008/01/23
Volume: 19, Issue: 3, Pages: 313-325
Abstract
Data mining techniques can be used for discovering interesting patterns in complicated manufacturing processes These patterns are used to improve manufacturing quality Classical representations of quality data mining problems usually refer to the operations settings and not to their sequence This paper examines the effect of the operation sequence on the quality of the product using data mining techniques For this purpose a novel decision tree framework for extracting sequence patterns is developed The proposed method is capable to mine sequence patterns of any length with operations that are not necessarily immediate precedents The core induction algorithmic framework consists of four main steps In the first step all manufacturing sequences are represented as string of tokens In the second step a large set of regular expressionbased patterns are induced by employing a sequence patterns In the third step we use feature selection methods to filter out the initial set and leave only the most useful patterns In the last stage we transform the quality problem into a classification problem and employ a decision tree induction algorithm A comparative study performed on benchmark databases illustrates the capabilities of the proposed framework
Keywords: