摘要 |
<P>PROBLEM TO BE SOLVED: To perform prediction using prediction models which are small in number than process treatment apparatuses and chambers. Ž<P>SOLUTION: With respect to process data gathered by a process data gathering unit 25, a model selector 25 calculates a feature quantity through principal component analysis, classifies it into one of one or more groups classified on a principal component plane, and selects a prediction model corresponding to the group as the classification destination. Then a prediction unit 27 inputs process data for prediction to the prediction model to obtain an estimate of measurement data. The groups on the principal component plane show differences in operation state among process treatment apparatuses. Therefore, only a prediction model corresponding to an operation state is prepared to handle any of processing states of the process treatment apparatuses, and internal states of the process treatment apparatuses or characteristic values of manufactured products is predicted using prediction models which are less in number than the process treatment apparatuses and chambers. Ž<P>COPYRIGHT: (C)2010,JPO&INPIT Ž
|