Anomaly detection based on cluster transitions
US9787704B2 · kind B2 · utility
Assignee
Inventors
Key dates
| Filing date | Mar 6, 2015 |
| Grant date | Oct 10, 2017 |
| Priority date | — |
| Expiry date | Mar 6, 2035 |
Classification
- Technology area (CPC H)Electricity
- CPC primaryH04L67/1095
- WIPO fieldDigital communication
- WIPO sectorElectrical engineering
Abstract
Systems and methods may include receiving first data of components, which may represent performance characteristics of the components at a first time. The systems and methods may include performing a first cluster analysis of the first data to identify clusters of the components with similar characteristics. The systems and methods may include receiving second data of the components, which may represent performance characteristics of the components at a second time. The systems and methods may include performing a second cluster analysis of the second data to identify clusters of the components with similar characteristics. The systems and methods may include determining whether a component transitioned from a cluster identified in the first cluster analysis to a different cluster identified in the second cluster analysis. The systems and methods may include determining that an anomaly occurred in response to determining that the component transitioned from the cluster to the different cluster.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.