To exactly and associatively infer the cause corresponding to an abnormality even to the input of symptom/sign data on which a noise is superimposed or in which a defect occurs by storing a cross-correlation matrix generated based on pair information as a association storing model and presuming the cause of the abnormality using the symptom data and the cross-correlation matrix as the association storing model.
A cross-correlation matrix storing means 2 stores a cross- correlation matrix which is generated base on pair information of the symptom data stored in a symptom/cause data storing means 1, and cause data as a cause of that symptom, as an association storing model. An abnormality diagnosis means 3 in which this association storing model is used presumes the cause of the abnormality as the status of monitoring object using the symptom data corresponding to time series data showing the status of the monitoring object and the cross-correlation matrix stored in the cross-correlation matrix storing means 2. Thus, even if the time series data is imperfect and the cause is unknown, the cause of the abnormality is presumed.
ISHIMARU KEIICHI
UDA HIROHISA
FUJI FACOM CORP