CARAT Classification: “Unknown”
- Unidentified patterns that do not meet any of the defined criteria for radio, non-radio, or measurement equipment.
- Can also arise from a new behavior that represents a new pattern in the known categories. Using AI/ML, CARAT also identifies such patterns and integrates them into its analyses after review.
- It may also happen that the traces do not contain enough information to accurately identify the problem.
AI/ML-Methods in CARAT
To further enhance the analysis of these transactions, an in depth analysis based on different approaches and ML models is defined. Combining the power of Unsupervised ML Models to uncover hidden connections and cluster the similar data together, along with the Root-Cause Analysis results of the Supervised ML Models, we are able to identify new emerging Patterns or update old ones, thus keeping our Pattern database up to date.
