CPC G06Q 40/12 (2013.12) [G06F 18/2431 (2023.01); G06N 5/04 (2013.01)] | 16 Claims |
1. A method for anomaly detection, the method comprising:
receiving financial data;
applying a set of dynamic or interdependent algorithms to the received financial data to obtain a set of outcomes; and
using a machine learning (ML) classifier to classify the outcomes as: (i) algorithm compliant, wherein classification as algorithm compliant indicates that a particular outcome resulting from application of one or more of the set of dynamic or interdependent algorithms is within norms of an organization or controller associated with the financial data, (ii) potentially algorithm non-compliant, wherein classification as potentially algorithm non-compliant indicates that a particular outcome is likely outside the norms of the organization or controller associated with the financial data, or (iii) algorithm non-compliant, wherein classification as algorithm non-compliant indicates that a particular outcome is outside the norms of the organization or controller associated with the financial data, wherein:
the classifier is trained to perform such classification by simultaneously analyzing a plurality of outcomes resulting from application of a plurality of dynamic or interdependent algorithms from the set to financial data, and
classification of the outcomes as potentially algorithm non-compliant or algorithm non-compliant indicates a latent anomaly in the received financial data.
|