Data Mining for Process and Quality Control in the Semiconductor Industry
Like in any other industry, manufacturing departments of semiconductor plants are evaluated by their ability to meet the delivery schedules. However, the final quantities and the flow times of individual semiconductor batches are affected by multiple uncertain factors, like material quality, process variability, equipment condition, and others. Thus, the tasks of predicting the batch quality (measured by yield) and its total flow time are an important part of the production planning activities. Beyond prediction, the plant management is interested in identifying the main causes of yield excursion and process delays.
In this paper, we are applying several methods of data mining and knowledge discovery to WIP (Work-in-Process) data, collected in a semiconductor plant. The information on each manufacturing batch includes its design parameters, process tracking data, line yield, etc. The data is prepared for data mining by converting a sequential dataset into a relational format. Classification models for predicting line yield and flow times are built from the pre-processed data by using the Info-Fuzzy Network (IFN) methodology. Fuzzy-based techniques of automated perception are used for post-processing the data mining results. We conclude the paper with a critical evaluation of the discovered knowledge and the methods used.
KeywordsData Mining Membership Function Mutual Information Association Rule Fuzzy Rule
Unable to display preview. Download preview PDF.
- Agrawal, R., Mehta, M., Shafer, J., and Srikant, R., “The Quest Data Mining System,” in Proceedings of KDD-96, pp. 244 – 249, 1996.Google Scholar
- Attneave, F., Applications of Information Theory to Psychology, Holt, Rinehart, and Winston, 1959.Google Scholar
- Breiman, L., Friedman, J.H., Olshen, R.A., & Stone, P.J., Classification and Regression Trees, Wadsworth, 1984.Google Scholar
- Fayyad, U. and Irani, K., “Multi-Interval Discretization of Continuous-Valued Attributes for Classification Learning,” in Proceedings of the 13th International Joint Conference on Artificial Intelligence, pp. 1022 – 1027, 1993.Google Scholar
- Korth, H.F. and Silberschatz, A., Database System Concepts, McGraw-Hill, Inc., 1991.Google Scholar
- Last, M. and Kandel, A., “Fuzzification and Reduction of Information-Theoretic Rule Sets,” in Data Mining and Computational Intelligence, pp. 63–93, Physica-Verlag, 2001.Google Scholar
- Mitchell, T.M., Machine Learning, McGraw-Hill, 1997.Google Scholar
- Pyle, D., Data Preparation for Data Mining, Morgan Kaufmann, 1999.Google Scholar
- Quinlan, J. R., C4. 5: Programs for Machine Learning, Morgan Kaufmann, 1993.Google Scholar
- Shenoi, S., “Multilevel Database Security Using Information Clouding,” in Proceedings of IEEE International Conference on Fuzzy Systems, pp. 483 – 488, 1993.Google Scholar
- Tobin, K. W., Karnowski, T.P., and Lakhani, F., “A Survey of Semiconductor Data Management Systems Technology,” in Proceedings of SPIE’s 25th Annual International Symposium on Microlithography, Santa Clara, CA, February 2000.Google Scholar
- Wang, L.-X., A Course in Fuzzy Systems and Control, Prentice-Hall, 1997.Google Scholar
- Zadeh, L. A., “A New Direction in System Analysis: From Computation with Measurements to Computation with Perceptions,” in New Directions in Rough Sets, Data Mining, and Granular-Soft Computing, pp. 10–11, Springer-Verlag, 1999.Google Scholar