Extracting the Rules of KPIs for Equipment Management Based on Rough Set Theory
This paper practically collects manufacturing supplier dataset in Taiwan. The dataset includes production records, and there are 18 attributes such as production scheduling, scheduled downtime, process, etc. For comparison, decision tree, naive bayesian, and multi-layer perceiving are utilized to compare with the proposed procedure in classification accuracy. The results show that the correct rate of rough set theory is not only superior to decision trees, naive bayesian, and multi-layer perceiving (MLP), and the proposed procedure can be easy to understand and produce fewer rules. In managerial implication, the results can generate predictive model, and classifiable rule which can help manufacturing to find out key related factors in equipment throughput and use the rules generated as a capacity planning assessment.
J. W. Wang and C. H. Cheng, "Extracting the Rules of KPIs for Equipment Management Based on Rough Set Theory", Advanced Materials Research, Vols. 314-316, pp. 2358-2361, 2011