PROBLEM TO BE SOLVED: To enable appropriate selection of representative data to be used for abnormality detection among a large number of data that is the result of status monitoring of a target device.;SOLUTION: From an analysis result data group 21 obtained by traveling a railway vehicle equipped with a diagnosis target device, analysis result data close to the center point O of a data distribution area is selected as a primitive data group 31. With respect to each analysis result data of an initial analysis result data group 21, an abnormality determination is performed on this primitive data group 31. A representative data group 331 is selected by repeating an addition of the analysis result data having the abnormality and the abnormality degree satisfying the minimum condition to a primitive data group 31. In second and subsequent runs, by performing the abnormality determination on an existing representative data group 330, a part of an analysis result data group 20 obtained by a relevant run is selected and added/updated to the existing representative data group 330.;SELECTED DRAWING: Figure 5;COPYRIGHT: (C)2019,JPO&INPIT
展开▼