Sequence Analysis Technique to Detect Performance Patterns in Complex Manufacturing Process

Many modern manufacturing plants deal with large scale multidimensional data that are daily collected from hundreds of operational units in a production line. Wafer fabrication in the semiconductor industry is probably one of the most complex manufacturing processes. Maintaining high yields through the statistical process control as a sole monitoring method is obviously inefficient in such highly complicated operations. We thus present in this paper a sequence analysis method, which is one of the advanced data mining techniques, to identify and extract patterns from manufacturing data that can reveal and differentiate low performance processes from the normal ones. We also provide the program coding, based on the R language, for easy reexamination and repetition to process engineers who are interested in applying our proposed sequence analysis method to their complex data at hand.