The Probability Prediction Of Fault State Using Case-Based Reasoning With Random Forest
Semiconductor process is that have spent a lot of economic and time costs when an abnormality occurs. It is essential to manage the efficient process monitoring about the abnormality diagnosis of the equipment. Since the vacuum pump are used in a number of steps, it is possible to diagnose the overall condition of the process through the health monitoring via vacuum pump. In this paper, we performed an analysis using the case information to be generated in response to the aging of equipment in case based reasoning (CBM). In order to optimize the case which is defined as Initial state, Aging state, Fault state defined, we determine the two experimental conditions and analysis using a machine learning technique. In analysis, we used the Random Forest that there is strength in multiple and imbalance class. Based on the optimal case that selected as the experimental results, we can provide decision to assist with the diagnosis and prediction of equipment.
Keywords� Vacuum pump, Random Forest, case-based reasoning, Initial state, Aging state, Fault state.