# US 2022/0036200 A1 · Rules and machine learning to provide regulatory complied fraud detection systems

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2022/0036200 A1
- Document ID: US20220036200A1
- Status: Published application
- Kind: A1
- Application number: 16/940,944
- Filed: 2020-07-28
- Published: 2022-02-03
- Earliest claimed: 2020-07-28
- Assignees: International Business Machines Corporation
- Applicants: International Business Machines Corporation
- Inventors: Yi-Hui Ma; Willie Robert Patten, JR.
- Web page: https://patentgrep.com/patent/US20220036200A1

## CPC

- G06N 5/01: Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
- G06N 20/20: Ensemble learning
- G06N 5/025: Extracting rules from data
- G06N 5/04: Inference or reasoning models
- G06N 20/00: Machine learning
- G06Q 20/4016: {involving fraud or risk level assessment in transaction processing}
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR

## Abstract

An approach is provided that receives data sets pertaining to entities. The received data sets are analyzed using a trained artificial intelligence (AI) system, with the analysis results in a fraud probability level. In response to the probability level indicating a high probability of fraud, a rule-based decision tree is applied to the received set of data. A behavior pattern is identified based on a result of the rule-based decision tree and this behavior pattern is used to identify a possible fraudulent event pertaining one of the entities.

## Claims

### Claim 1

A method implemented by an information handling system that includes a processor and a memory accessible by the processor, the method comprising:
receiving one or more sets of data pertaining to a plurality of entities;
analyzing the received sets of data using a trained artificial intelligence (AI) system, wherein the analysis results in a fraud probability level;
applying a rule-based decision tree to the received set of data based on the resulting fraud probability level;
identifying a pattern based on a result of the rule-based decision tree; and
detecting a possible fraudulent event pertaining to a behavior pattern corresponding to a selected one of the plurality of entities.

### Claim 2 (depends on 1)

The method of claim 1 wherein the fraud probability level corresponds to a confidence score in response to the AI system's analysis of a plurality of behavior patterns included in the sets of data.

### Claim 3 (depends on 2)

The method of claim 2 further comprising:
applying the rule-based decision tree in response to the confidence score reaching a predetermined threshold.

### Claim 4 (depends on 1)

The method of claim 1 further comprising
forming the behavior pattern from a plurality of logic branches included in the sets of data that reach a result.

### Claim 5 (depends on 4)

The method of claim 4 further comprising
training the AI system using the formed behavior pattern.

### Claim 6 (depends on 1)

The method of claim 1 further comprising
retaining AI resulting data that includes an answer corresponding to each of the plurality of entities, a confidence value corresponding to each of the answers, and one or more sets of supporting passages found in the sets of data by the AI system; and
retaining decision-tree resulting data that includes a plurality of branches taken by the rule-based decision tree along with a data-based reasoning corresponding to one or more of the plurality of branches; and
associating the retained decision-tree resulting data with the AI resulting data that corresponds to the same entity.

### Claim 7 (depends on 6)

The method of claim 6 further comprising:
forming one or more behavior patterns from the plurality of branches included in the decision-tree resulting data, wherein at least one of the behavior branches is selected from the group consisting of (1) an unusual number of counter party occurrences indicting fraud, (2) an association with a high risk network indicating fraud, (3) a high but consistent transaction volume indicating non-fraud, and (4) a normal seasonal behavior indicating non-fraud.

### Claim 8

An information handling system comprising:
one or more processors;
a memory coupled to at least one of the processors; and
a set of computer program instructions stored in the memory and executed by at least one of the processors in order to perform actions comprising:
receiving one or more sets of data pertaining to a plurality of entities;
analyzing the received sets of data using a trained artificial intelligence (AI) system, wherein the analysis results in a fraud probability level;
applying a rule-based decision tree to the received set of data based on the resulting fraud probability level;
identifying a pattern based on a result of the rule-based decision tree; and
detecting a possible fraudulent event pertaining to a behavior pattern corresponding to a selected one of the plurality of entities.

### Claim 9 (depends on 8)

The information handling system of claim 8 wherein the fraud probability level corresponds to a confidence score in response to the AI system's analysis of a plurality of behavior patterns included in the sets of data.

### Claim 10 (depends on 9)

The information handling system of claim 9 wherein the actions further comprise:
applying the rule-based decision tree in response to the confidence score reaching a predetermined threshold.

### Claim 11 (depends on 8)

The information handling system of claim 8 wherein the actions further comprise
forming the behavior pattern from a plurality of logic branches included in the sets of data that reach a result.

### Claim 12 (depends on 11)

The information handling system of claim 11 wherein the actions further comprise
training the AI system using the formed behavior pattern.

### Claim 13 (depends on 8)

The information handling system of claim 8 wherein the actions further comprise
retaining AI resulting data that includes an answer corresponding to each of the plurality of entities, a confidence value corresponding to each of the answers, and one or more sets of supporting passages found in the sets of data by the AI system; and
retaining decision-tree resulting data that includes a plurality of branches taken by the rule-based decision tree along with a data-based reasoning corresponding to one or more of the plurality of branches; and
associating the retained decision-tree resulting data with the AI resulting data that corresponds to the same entity.

### Claim 14 (depends on 13)

The information handling system of claim 13 wherein the actions further comprise:
forming one or more behavior patterns from the plurality of branches included in the decision-tree resulting data, wherein at least one of the behavior branches is selected from the group consisting of (1) an unusual number of counter party occurrences indicting fraud, (2) an association with a high risk network indicating fraud, (3) a high but consistent transaction volume indicating non-fraud, and (4) a normal seasonal behavior indicating non-fraud.

### Claim 15

A computer program product stored in a computer readable storage medium, comprising computer program code that, when executed by an information handling system, performs actions comprising:
receiving one or more sets of data pertaining to a plurality of entities;
analyzing the received sets of data using a trained artificial intelligence (AI) system, wherein the analysis results in a fraud probability level;
applying a rule-based decision tree to the received set of data based on the resulting fraud probability level;
identifying a pattern based on a result of the rule-based decision tree; and
detecting a possible fraudulent event pertaining to a behavior pattern corresponding to a selected one of the plurality of entities.

### Claim 16 (depends on 15)

The computer program product of claim 15 wherein the fraud probability level corresponds to a confidence score in response to the AI system's analysis of a plurality of behavior patterns included in the sets of data.

### Claim 17 (depends on 16)

The computer program product of claim 16 wherein the actions further comprise:
applying the rule-based decision tree in response to the confidence score reaching a predetermined threshold.

### Claim 18 (depends on 15)

The computer program product of claim 15 wherein the actions further comprise
forming the behavior pattern from a plurality of logic branches included in the sets of data that reach a result.

### Claim 19 (depends on 18)

The computer program product of claim 18 wherein the actions further comprise
training the AI system using the formed behavior pattern.

### Claim 20 (depends on 15)

The computer program product of claim 15 wherein the actions further comprise
retaining AI resulting data that includes an answer corresponding to each of the plurality of entities, a confidence value corresponding to each of the answers, and one or more sets of supporting passages found in the sets of data by the AI system; and
retaining decision-tree resulting data that includes a plurality of branches taken by the rule-based decision tree along with a data-based reasoning corresponding to one or more of the plurality of branches; and
associating the retained decision-tree resulting data with the AI resulting data that corresponds to the same entity.
