System analysis method, system analysis apparatus, and program

A system analysis apparatus, comprising:

• at least a processor;
• and
• a memory in circuit communication with the processor,
• wherein the processor is configured to execute program instructions stored in the memory to implement:
• a time-series information acquisition part that acquires time-series information of sensor values outputted by a plurality of sensors arranged in a system, wherein the sensor values indicate an individual one of the sensor values outputted by the respective sensors indicates an abnormality and/or whether an individual relationship between sensor values outputted by different sensors indicates an abnormality;
• a change point estimation part that estimates a change point group of change points in the time-series information, wherein the each of change points indicates a time point at which state of the system has changed;
• a relevance level estimation part that estimates an abnormality level for each change points included in the change point group and a similarity level between abnormality level vectors of two arbitrary time points included in the change point group, wherein the abnormality level vector is composed of abnormality levels;
• a clustering part that generates groups of change point groups by classifying the change point group into a plurality of groups based on the time-series information and the similarity level between abnormality level vectors;
• and
• an output part that generates and outputs output information, which is information relating to an abnormality per group of the change point groups.

A non-transient computer-readable storage medium stored a program, causing a computer to execute:

• (a) acquiring time-series information of sensor values outputted by a plurality of sensors arranged in a system, wherein the sensor values indicate an individual one of the sensor values outputted by the respective sensors indicates an abnormality and/or whether an individual relationship between sensor values outputted by different sensors indicates an abnormality;
• (b) estimating a change point group of change points in the time-series information, wherein the each of change points indicates a time point at which state of the system has changed;
• (c) estimating an abnormality level for each change points included in the change point group and a similarity level between abnormality level vectors of two arbitrary time points included in the change point group, wherein the abnormality level vector is composed of abnormality levels;
• (d) generating groups of change point groups by classifying the change point group into a plurality of groups based on the time-series information and the abnormality level;
• and
• (e) generating and outputting output information, which is information relating to an abnormality per group of the change point groups.