Deep learning for credit controls

Patent metadata

Field	Value
Publication number	US-12079705-B2
Application number	US-202318112582-A
Country	US
Kind code	B2
Filing date	Feb 22, 2023
Priority date	Mar 23, 2017
Publication date	Sep 3, 2024
Grant date	Sep 3, 2024

Abstract

Systems and methods are provided to identify abnormal transaction activity by a participant that is inconsistent with current conditions. Historical participant and external data are identified. A recurrent neural network identifies patterns in the historical participant and external data. A new transaction by the participant is received. The new transaction is compared using the patterns to the historical participant and external data. An abnormality score is generated. An alert is generated if the abnormality score exceeds a threshold.

First claim

What is claimed is:

1. A computer implemented method comprising:
identifying, by a processor coupled with a data transaction processing system, using a structured neural network comprising a layered plurality of interconnected processing nodes, one or more patterns in historic participant transaction data for a participant in the data transaction processing system and historic external market factor data including data indicative of characteristics of a financial derivative product traded on an exchange for a time period that corresponds to the historic participant transaction data that occurs during the time period, the one or more patterns indicative of a historical normal activity by the participant in relation to the historic external market factor data, wherein at least a subset of the interconnections of the layered plurality of interconnected processing nodes are dynamically weighted;
receiving, by the processor, from the participant, data indicative of a new transaction;
calculating, by the processor, current external market factor data;
comparing, by the processor, the data indicative of the new transaction and the current external market factor data with the one or more patterns;
generating, by the processor, an abnormality score for the new transaction based on the comparison; and
generating, by the processor, an alert when the abnormality score exceeds a first threshold.

2. The computer implemented method of claim 1 , wherein identifying the one or more patterns comprises:
encoding, by the processor, using the structured neural network, the historic participant transaction data and the historic external market factor data using a plurality of first layers of the layered plurality of interconnected processing nodes;
decoding, by the processor, using the structured neural network, the encoded data using a plurality of second layers of the layered plurality of interconnected processing nodes;
comparing, by the processor, using the structured neural network, the decoded data with the historic participant transaction data and the historic external market factor data; and
identifying, by the processor, using the structured neural network, the one or more patterns in the layered plurality of interconnected processing nodes when the decoded data is within a predefined distance of the historic participant transaction data and the historic external market factor data.

3. The computer implemented method of claim 2 , wherein the plurality of first layers comprise a decreasing number of nodes in each layer of the plurality of first layers, and the plurality of second layers comprise an increasing number of nodes in each layer of the plurality of second layers.

4. The computer implemented method of claim 2 , wherein only outputs of a smallest layer of the plurality of first layers is connected to a largest layer of the plurality of second layers.

5. The computer implemented method of claim 1 , wherein the layered plurality of interconnected processing nodes comprises a plurality of long short term memory nodes.

6. The computer implemented method of claim 1 , wherein the interconnected processing nodes comprise long short term memory units.

7. The computer implemented method of claim 1 , further comprising:
updating, by the processor, the historic participant transaction data and the historic external market factor data with the data indicative of the new transaction and the current external market factor data.

8. The computer implemented method of claim 1 , further comprising:
prohibiting, by the processor, the new transaction from being processed when the abnormality score exceeds a second threshold.

9. The computer implemented method of claim 1 , wherein the historic participant transaction data comprises data relating to a single or related set of products.

10. A computer implemented method comprising:
calculating, by a risk processor, a plurality of risk profiles for a participant in a data transaction processing system for a plurality of time periods, the plurality of risk profiles based on a plurality of historical participant parameters;
calculating, by the risk processor, a plurality of external risk profiles for the plurality of time periods, the plurality of external risk profiles based on a plurality of historical external market parameters including data indicative of characteristics of a financial derivative product traded on an exchange that correspond to the plurality of historical participant parameters that occurs during the plurality of time periods;
identifying, by the risk processor using a structured neural network comprising a layered plurality of interconnected processing nodes, a plurality of patterns between the plurality of risk profiles and the plurality of external risk profiles, wherein at least a subset of the interconnections of the layered plurality of interconnected processing nodes are dynamically weighted;
receiving, by the risk processor, data for a new transaction from the participant;
calculating, by the risk processor, a current external factor risk profile as a function of current external market parameters;
generating, by the risk processor, a current risk profile for the participant comprising at least data for the new transaction;
comparing, by the risk processor, the current risk profile, the current external factor risk profile, the plurality of risk profiles, and the plurality of external risk profiles using the plurality of patterns;
calculating, by the risk processor, an abnormality score based on the comparison; and
generating, by the risk processor, an alert when the abnormality score exceeds a threshold.

11. The computer implemented method of claim 10 , wherein calculating the plurality of risk profiles comprises calculating using a value at risk factor that indicates a probability of occurrence of loss by a predefined factor by a predefined amount of time.

12. The computer implemented method of claim 10 , wherein identifying the plurality of patterns comprises:
encoding, by the risk processor using the structured neural network, the plurality of risk profiles and plurality of external risk profiles using a plurality of first layers of the layered plurality of interconnected processing nodes;
decoding, by the risk processor using the structured neural network, the encoded data using a plurality of second layers of the layered plurality of interconnected processing nodes;
comparing, by the risk processor using the structured neural network, the decoded data with the plurality of risk profiles and plurality of external risk profiles; and
identifying, by the risk processor using the structured neural network, the plurality of patterns in the layered plurality of interconnected processing nodes when the decoded data is within a predefined distance of the plurality of risk profiles and the plurality of external risk profiles.

13. The computer implemented method of claim 12 , wherein the plurality of first layers comprise a decreasing number of nodes in each layer of the plurality of first layers, and the plurality of second layers comprise an increasing number of nodes in each layer of the plurality of second layers.

14. The computer implemented method of claim 12 , wherein the layered plurality of interconnected processing nodes comprises a plurality of long short term memory nodes.

15. The computer implemented method of claim 10 , further comprising:
updating, by the risk processor using the structured neural network, the plurality of patterns using the data indicative of the new transaction and the current external market parameters.

Assignees

• Chicago Mercantile Exchange Inc

Inventors

• Studnitzer Ari
• Geddes David
• Singh Inderdeep

Classifications

• G06N3/0442

  characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU] · CPC title

• G06N3/0455

  Auto-encoder networks; Encoder-decoder networks · CPC title

• G06Q40/04 Primary

  Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange · CPC title

• G06N3/0895

  Weakly supervised learning, e.g. semi-supervised or self-supervised learning · CPC title

• G06N3/09

  Supervised learning · CPC title