Money laundering transaction identification method based on blockchain
By analyzing the blockchain transaction records, redefining the zero -finding address and establishing a user transaction network, solving the problem of difficulty finding the difficulty of finding zero address identification and incomplete user behavior characteristics in existing technology, and realizing more accurate and effective money laundering transaction behavior recognition identification Essence
Patent Information
- Application Number
- CN202210106228.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-01-28
- Publication Date
- 2025-01-24
- Estimated Expiration
- 2042-01-28
AI Technical Summary
In the prior art, when identifying money laundering transactions, it is difficult to accurately identify the change address, resulting in false positive and false negative errors in entity identification, and the address-based transaction graph cannot truly and comprehensively reflect the user's behavior characteristics, resulting in information loss.
By analyzing the transaction records on the chain, redefining the change address and adding it to the entity recognition, a user-based transaction network is established, user transaction characteristics are extracted, and money laundering transaction behavior is identified.
It significantly reduces false positive errors, reflects user transaction behavior more truly and comprehensively, fully extracts user characteristics, and improves the accuracy and effectiveness of money laundering transaction behavior identification.
Smart Images
Figure CN114445225B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of blockchain application, and in particular to a method for identifying money laundering transactions based on blockchain. Background Art
[0002] Due to the decentralization and anonymization mechanism, users participate in transactions mainly through addresses generated based on public keys, and transactions do not require any identity information. This trading method not only brings a large number of investors to cryptocurrencies, but also attracts many illegal activities, such as fraud, money laundering, extortion, and contraband trading. Among them, fraud and money laundering are the two most prominent forms of illegal Bitcoin transactions.
[0003] At present, the identification of money laundering transactions is mostly based on addresses. By establishing a transaction network with addresses as nodes, studying the transaction characteristics of addresses, and establishing an identification model for addresses for anti-money laundering identification. However, in real transactions, a user has multiple addresses. In a transaction related to money laundering, the user may also use multiple addresses to complete the corresponding operation. Therefore, the identification model based only on addresses cannot truly and comprehensively reflect the user's behavioral characteristics, and some information will be lost when analyzing user characteristics. If multiple addresses controlled by a user can be identified and a transaction graph based on the user is established for identification, the behavioral characteristics of money laundering transactions can be further revealed, and money laundering transactions can be identified more accurately.
[0004] In digital currency transactions, an entity is defined as an individual or organization that controls one or more addresses. Entity identification is to identify which addresses in the transaction network are controlled by the same person. The current methods for entity identification mainly include multi-input heuristics and change address heuristics. The multi-input heuristic is the most recognized and used method. This method assumes that a transaction can only be initiated by one user, no matter how many inputs are included. Therefore, the input addresses of a transaction should belong to the same entity. The change address heuristic is based on the principle of "change" in Bitcoin. It is believed that there is usually an address in the transaction output to return the unspent bitcoins of the sender. Therefore, this address and the input address of the corresponding transaction belong to the same entity. The problem is how to accurately identify the change address, because no matter which method is used, it will lead to false positives and false negatives. That is, the recognition result may mistakenly identify the payee address as the change address, and mistakenly identify the change address as the payee address. That is, there are the following problems in the prior art: the change address is difficult to identify in entity identification; the address-based transaction graph cannot truly and comprehensively reflect the user behavior characteristics; the address-based user feature analysis will lose some information.
[0005] The technical advantages of the present invention are as follows:
[0006] The present invention can significantly reduce false positives by parsing the on-chain transaction records and adding the redefined change address to entity recognition based on the on-chain transaction records. Secondly, establishing a user-based transaction network can reduce the scale of the transaction graph and simplify the problem on the one hand, and on the other hand, it can more realistically and comprehensively reflect the user's transaction behavior, fully extract user characteristics, and improve the accuracy and effectiveness of money laundering transaction behavior identification. Summary of the invention
[0007] The technical problem of the present invention is to provide a method for identifying money laundering transaction behavior based on blockchain, which can improve the accuracy of change address discrimination in entity identification and improve the comprehensiveness of user transaction behavior characteristics.
[0008] To achieve the above object, the technical solution adopted by the present invention is:
[0009] The method for identifying money laundering transaction behavior based on blockchain includes the following steps: S1 obtains and parses the target transaction records on the chain to obtain a transaction data set; wherein the transaction data set includes the payment information and receipt information of the target transaction; S2 identifies and marks the entities in the transaction data set; S3 establishes a user transaction graph corresponding to each entity based on the entity recognition result, extracts the transaction features in the user transaction graph, identifies money laundering transaction behavior, and outputs it.
[0010] Furthermore, the payment information includes: payment address, payment amount and payment time; the collection information includes: collection address, collection amount and collection time.
[0011] Further, S2 includes: S21 identifies the receiving address that is the change address in the transaction data set; if the change address is identified, execute S22; otherwise, execute S23; S22 marks the change address and the corresponding output address as the same entity; S23 marks the receiving address that is not identified as the change address as a separate entity.
[0012] Further, S21 includes: S211 determines whether the sending transaction record and the receiving transaction record of the payment address are both less than or equal to 1 time, if so, execute S212; if not, execute S23; S212 determines whether the payment address has a sending transaction record, if so, execute S213, if not, execute S23; S213 determines whether the time of sending the transaction in the payment address sending transaction record is later than the time of receiving the transaction at the payment address, and the amount of the sending transaction is less than the amount of the receiving transaction, if so, execute S214; if not, execute S23; S214 determines whether the payment address is used as input. Do the transaction input and its transaction input as output have an intersection? If so, execute S215; if not, execute S23; S215 determines whether the transaction output corresponding to the payment address has only one payment address that satisfies S211 to S214 at the same time. If so, execute S216; if not, execute S23; S216 determines whether the transaction corresponding to the payment address is the generation of Bitcoin. If not, execute S217; If so, execute S23; S217 determines whether the transaction output corresponding to the payment address is greater than 1. If so, identify the payment address as a change address and execute S22; if not, execute S23. BRIEF DESCRIPTION OF THE DRAWINGS
[0013] The present invention and its features, configurations and advantages will become more apparent by reading the detailed description of non-limiting embodiments with reference to the following drawings. The same reference numerals indicate the same parts throughout the drawings. The drawings are not drawn to scale, but rather to illustrate the subject matter of the present invention.
[0014] Figure 1 This is a flow chart of the method for identifying money laundering transaction behavior based on blockchain provided by the present invention. DETAILED DESCRIPTION
[0015] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, but they are not intended to limit the present invention.
[0016] The present invention first obtains public transaction records from the Bitcoin transaction system and parses them into address-based transaction data sets as test data, and collects labeled data sets that have been identified as money laundering behaviors from the outside as training data; then, based on entity recognition technology, the addresses are marked as corresponding users, a user data set is constructed and a user transaction graph is established, and a network feature analysis method is used to analyze and extract user transaction features from the transaction graph. Finally, a money laundering transaction behavior recognition model is established, the model effectiveness is evaluated and the model parameters are optimized, and finally predictions are made on the test set.
[0017] Predictive recognition methods such as Figure 1As shown, the following steps are included: S1 obtains and parses the target transaction record on the chain to obtain a transaction data set; wherein the transaction data set includes the payment information and receipt information of the target transaction; S2 identifies and marks the entities in the transaction data set; S3 obtains the user data set based on the entity recognition result, establishes the user transaction graph corresponding to each entity, extracts the transaction features in the user transaction graph, identifies the money laundering transaction behavior, and outputs it. The payment information includes: the payment address, the payment amount and the payment time; the receipt information includes: the receipt address, the receipt amount and the receipt time. S2 includes: S21 identifies the receipt address in the transaction data set as the change address; if the change address is identified, execute S22; otherwise, execute S23; S22 marks the change address and the corresponding output address as the same entity; S23 marks the receipt address that is not identified as the change address as an entity separately. S21 includes: S211 determines whether the sending transaction record and receiving transaction record of the payment address are both less than or equal to 1 time, if so, execute S212; if not, execute S23; S212 determines whether the payment address has a sending transaction record, if so, execute S213, if not, execute S23; S213 determines whether the time of sending transaction in the payment address sending transaction record is later than the time of receiving transaction at the payment address, and the amount of sending transaction is less than the amount of receiving transaction, if so, execute S214; if not, execute S23; S214 determines whether the payment address is used as input and its transaction Do the input and the transaction input as output have an intersection? If so, execute S215; if not, execute S23; S215 determines whether the transaction output corresponding to the payment address has only one payment address that satisfies S211 to S214 at the same time. If so, execute S216; if not, execute S23; S216 determines whether the transaction corresponding to the payment address is the generation of Bitcoin. If not, execute S217; If so, execute S23; S217 determines whether the transaction output corresponding to the payment address is greater than 1. If so, identify the payment address as a change address and execute S22; if not, execute S23.
[0018] Because mistakenly merging different entities will conceal the user's real transaction behavior, and splitting the same entity into several different entities for separate analysis will not cause serious information distortion (because the transaction behavior of each entity is real), so by adding the redefined change address to entity recognition, false positives can be significantly reduced. Secondly, establishing a user-based transaction network can reduce the scale of the transaction graph and simplify the problem on the one hand, and on the other hand, it can more truly and comprehensively reflect the user's transaction behavior, fully extract user characteristics, and identify money laundering transactions more accurately and effectively.
[0019] The preferred embodiments of the present invention are described above; it should be understood that the present invention is not limited to the above-mentioned specific embodiments, and the devices and structures not described in detail should be understood to be implemented in a common manner in the art; any technician familiar with the art can make many possible changes and modifications without departing from the technical solution of the present invention, or modify them into equivalent embodiments with equivalent changes, which does not affect the essential content of the present invention; therefore, any simple modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of protection of the technical solution of the present invention.
Claims
1. A method for identifying money laundering transactions based on blockchain, characterized in that: The following steps are involved: S1 obtains and parses the target transaction record on the chain to obtain a transaction data set; wherein the transaction data set includes payment information and receipt information of the target transaction; S2 identifies and labels entities in the transaction dataset; S3 establishes a user transaction graph corresponding to each entity based on the entity recognition result, extracts transaction features in the user transaction graph, identifies money laundering transaction behavior, and outputs it; The payment information includes: payment address, payment amount and payment time; the payment information includes: payment address, payment amount and payment time; The S2 includes: S21 identifies the payment address in the transaction data set as the change address; if the change address is identified, execute S22; otherwise, execute S23; S22 marks the change address and the corresponding output address as the same entity; S23 marks the receiving address that is not identified as a change address as an entity; The S21 includes: S211 determines whether the sending transaction record and the receiving transaction record of the payment address are both less than or equal to 1, if so, execute S212; if not, execute S23; S212 determines whether the payment address has a transaction record, if yes, execute S213, if no, execute S23; S213 determines whether the time of sending the transaction in the transaction record of the receiving address is later than the time of receiving the transaction, and the amount of the sending transaction is less than the amount of the receiving transaction. If so, execute S214; if not, execute S23; S214 determines whether the transaction input when the payment address is used as input and the transaction input when the payment address is used as output intersect, if so, execute S215; if not, execute S23; S215 determines whether the transaction output corresponding to the receiving address has only one receiving address that satisfies S211 to S214 at the same time. If so, execute S216; if not, execute S23; S216 determines whether the transaction corresponding to the payment address is the generation of Bitcoin. If not, execute S217; if yes, execute S23; S217 determines whether the transaction output corresponding to the payment address is greater than 1. If so, the payment address is identified as a change address and S22 is executed; if not, S23 is executed.