Devices that detect investment fraud based on behavioral sequences and contextual awareness

Abstract

A device for detecting investment fraud based on behavioral sequences and contextual awareness includes a data collection module, a behavior analysis module, a contextual awareness module, a risk assessment module, an indicator generation module, and a protection and warning module. The data collection module and the indicator generation module are connected to the behavior analysis module, the contextual awareness module, and the risk assessment module. The protection and warning module is also connected to the indicator generation module. After obtaining transaction data of at least two transactions performed by a bank customer within a predetermined time period through the data collection module, the behavior analysis module, the contextual awareness module, and the risk assessment module generate fraud identification results, anomaly detection results, and risk assessment results, respectively, based on the transaction data. The indicator generation module generates a fraud risk indicator based on the fraud identification results, anomaly detection results, and risk assessment results. The protection and warning module triggers a protection and warning mechanism when the fraud risk indicator exceeds a risk threshold. This technical means can achieve the technical effect of providing a proactive, real-time, and accurate investment fraud detection mechanism based on the bank customer's transaction behavior.

Claims

1. An apparatus for detecting investment fraud based on behavioral sequence and context awareness, the apparatus comprising at least: a data collection module for obtaining transaction data of at least two transactions when a bank customer performs at least two transactions within a predetermined time period; a behavioral analysis module connected to the data collection module for analyzing the transaction data to identify whether the at least two transactions conform to an investment fraud operation pattern and generating a fraud identification result; a context awareness module connected to the data collection module for generating an anomaly detection result based on the transaction data and using historical behavioral data of one of the bank customers; and a risk assessment module connected to the data collection module for performing a risk assessment on the transaction data to generate a risk assessment result. An indicator generation module, connected to the behavior analysis module, the context awareness module, and the risk assessment module, is used to generate a fraud risk indicator based on the fraud identification result, the anomaly detection result, and the risk assessment result; and a protection warning module, connected to the indicator generation module, is used to trigger a protection warning mechanism when the fraud risk indicator exceeds a risk threshold.

2. The device for detecting investment fraud based on behavioral sequence and context awareness as described in claim 1, wherein the behavioral analysis module analyzes the transaction operation data to generate a behavioral sequence feature, and uses an investment fraud behavior model based on the behavioral sequence feature to generate the fraud identification result.

3. The device for detecting investment fraud based on behavioral sequences and context awareness as described in claim 2, wherein the investment fraud behavior model is at least one rule or machine learning model of a known investment fraud pattern.

4. The device for detecting investment fraud based on behavioral sequence and context awareness as described in claim 1, wherein the context awareness module is further used to establish a historical behavioral parameter corresponding to the bank customer based on the historical behavioral data, and to use the historical behavioral parameter to generate the anomaly detection result based on the transaction operation data.

5. The device for detecting investment fraud based on behavioral sequences and context awareness as described in claim 4, wherein the historical behavioral parameters include at least time characteristics, amount characteristics, and target characteristics, and the context awareness module determines whether the transaction data deviates from the historical behavioral parameters to generate the anomaly detection result, or the historical behavioral parameters represent a typical historical sequence of most bank customers' behaviors, and the context awareness module generates the anomaly detection result based on the Livingston distance between the transaction data and the historical behavioral parameters.

6. The device for detecting investment fraud based on behavioral sequence and context awareness as described in claim 1, wherein the risk assessment module performs a risk assessment on one of the receiving accounts of the transaction object in the transaction operation data based on a pre-established account blacklist to generate a risk assessment result.

7. The device for detecting investment fraud based on behavioral sequence and context awareness as described in claim 6, wherein the risk assessment module is further used to analyze the distance and connection between the receiving account and each account in the account blacklist in the payment network and / or whether the receiving account has been transferred to a large number of accounts for the first time in the past period when the receiving account is not matched with the account blacklist in order to generate the risk assessment result.

8. The apparatus for detecting investment fraud based on behavioral sequence and context awareness as described in claim 1, wherein the indicator generation module scores or weights the fraud identification result, the anomaly detection result, and the risk assessment result to generate the fraud risk indicator, or provides the fraud identification result, the anomaly detection result, and the risk assessment result to an integrated learning model or a machine learning model to generate the fraud risk indicator.

9. The device for detecting investment fraud based on behavioral sequence and context awareness as described in claim 1, wherein the device further includes a threshold adjustment module connected to the indicator generation module, for adjusting the risk threshold according to the characteristics of the bank's customer group, the latest fraud patterns, or the model training results.

10. The device for detecting investment fraud based on behavioral sequence and context awareness as described in claim 1, wherein the protection warning module is further used to determine a risk level based on the risk threshold exceeded by the fraud risk indicator, and to trigger the corresponding protection warning mechanism based on the risk level.

Images