Bond related transaction identification method and device

Abstract

The embodiment of the present application provides a bond-related transaction identification method and device, which can be used in the field of financial technology. The method includes: searching for a connection path between a node where a first user is located and a node where a second user is located in a preset event map, and obtaining at least one basic space vector corresponding to the connection path; if the basic space vector corresponding to the connection path does not contain a target bond identifier, then obtaining the multidimensional transaction feature attributes corresponding to each basic space vector; if the existence of an association attribute between the first user and the second user is identified based on the multidimensional transaction feature attributes corresponding to each basic space vector, then determining the connection path corresponding to the association attribute as the association transaction path between the first user and the second user. The present application can effectively improve the accuracy and effectiveness of bond-related transaction identification, thereby effectively reducing the risk of illegal related transactions, and improving the accuracy and reliability of risk warning, avoidance or processing of illegal related transactions.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, in particular to the fields of big data, artificial intelligence and financial technology, and specifically to a method and device for identifying bond-related transactions. Background Art

[0002] In the fields of big data, artificial intelligence, and financial technology such as bank asset management, bond investment transactions have a relatively stable return rate and can effectively support the net value or annualized rate of return of wealth management products. Therefore, the scale of bond investment by wealth management products in the field of bank asset management is relatively large, which is specifically manifested in the number of bond transactions and the amount of bond transactions are higher than other types of investment targets (such as funds, deposits, trust products, etc.). It is precisely because of the above characteristics of bond investment transactions that illegal transactions that use related transactions as a means to achieve interest transfer are prone to occur in bond investment transactions. Among them, the more hidden forms of association in illegal transactions include: concealment through equivalent bond replacement, or concealment through cross-diversity bond trading methods, etc.

[0003] At present, the existing method of identifying bond-related transactions is usually to directly obtain the transaction data of the two parties regarding a certain bond. However, due to the single angle of association search, this method of identifying related transactions cannot be applied to transactions through equivalent bond replacement and illegal association scenarios through cross-diversity bond transactions. Therefore, it will affect the accuracy and comprehensiveness of bond-related transaction identification, and thus cannot guarantee the compliance of bond transactions. Summary of the invention

[0004] In response to the problems in the prior art, the present application provides a bond-related transaction identification method and device, which can effectively improve the accuracy and effectiveness of bond-related transaction identification, thereby effectively reducing the risk of illegal related transactions and improving the accuracy and reliability of risk warning, avoidance or processing of illegal related transactions.

[0005] In order to solve the above technical problems, this application provides the following technical solutions:

[0006] In a first aspect, the present application provides a method for identifying bond-related transactions, comprising:

[0007] Searching for a connection path between a node where a first user is located and a node where a second user is located in a preset event graph, and obtaining at least one basic space vector corresponding to the connection path, wherein the basic space vector is used to store a bond identifier and two transaction parties of a single bond transaction; wherein the first user is one of the two transaction parties in the obtained at least one basic space vector, and the second user is one of the two transaction parties in the obtained at least one basic space vector;

[0008] If the target bond identifier is not included in the basic space vector corresponding to the connection path, then the multi-dimensional transaction feature attribute corresponding to each of the basic space vectors is obtained;

[0009] According to the multi-dimensional transaction feature attributes corresponding to each of the basic space vectors, it is identified whether there is an associated attribute between the first user and the second user. If so, the connection path corresponding to the associated attribute is determined as the associated transaction path between the first user and the second user.

[0010] Furthermore, the event graph is a multi-dimensional vectorized event graph;

[0011] Wherein, each node in the multidimensional vectorized event graph corresponds to each user identifier one by one; the edge in the multidimensional vectorized event graph is used to represent the basic space vector between the nodes at both ends thereof;

[0012] The basic space vector includes a first vector unit and a second vector unit storing the same bond identifier;

[0013] The first vector unit is further used to store the selling party of the two transaction parties, and the second vector unit is further used to store the buying party of the two transaction parties;

[0014] The first vector unit and the second vector unit respectively correspond to pre-stored multi-dimensional transaction feature attributes.

[0015] Furthermore, the edges in the multidimensional vectorized event graph are also used to represent the multidimensional space vectors between the nodes at both ends thereof;

[0016] There is a one-to-one correspondence between each of the multidimensional space vectors and each of the basic space vectors;

[0017] The multidimensional space vector includes a third vector unit and a fourth vector unit storing the same bond identifier;

[0018] The third vector unit is also used to store the seller and the multi-dimensional transaction feature attributes;

[0019] The fourth vector unit is also used to store the buyer and the multi-dimensional transaction feature attributes;

[0020] Correspondingly, if the basic space vector corresponding to the connection path does not contain the target bond identifier, the multi-dimensional transaction feature attributes corresponding to each of the basic space vectors are obtained, including:

[0021] It is searched whether the target bond identifier is included in at least one basic space vector corresponding to the connection path. If not, at least one multidimensional space vector corresponding to the connection path is extracted from the multidimensional vectorized event map.

[0022] Further, the step of identifying whether there is an associated attribute between the first user and the second user according to the multi-dimensional transaction feature attributes corresponding to each of the basic space vectors, and if so, determining the connection path corresponding to the associated attribute as the associated transaction path between the first user and the second user, includes:

[0023] Based on a preset support vector machine, a dimensionality reduction process is performed on the multidimensional transaction feature attributes extracted from each of the multidimensional space vectors to obtain a dimensionality reduction attribute corresponding to the multidimensional transaction feature attribute;

[0024] Obtaining similarities between the same dimensionality reduction attributes in each of the multidimensional space vectors;

[0025] If there is a dimension reduction attribute whose similarity is greater than the similarity threshold, it is determined that there is an associated attribute between the first user and the second user, and the connection path corresponding to the associated attribute is determined as the associated transaction path between the first user and the second user;

[0026] Output the transaction data corresponding to each of the multidimensional space vectors on the associated transaction path.

[0027] Furthermore, it also includes:

[0028] If the basic space vector corresponding to the connection path contains the target bond identifier, the connection path is determined as the associated transaction path between the first user and the second user, and the associated bond is the bond corresponding to the target bond identifier;

[0029] The transaction data corresponding to each of the basic space vectors on the associated transaction path and the target bond identifier are output.

[0030] Furthermore, before searching the connection path between the node where the first user is located and the node where the second user is located in the preset event graph and obtaining at least one basic space vector corresponding to the connection path, the method further includes:

[0031] Standardizing the transaction data of each bond transaction to form standard meta-events and transaction characteristic attribute data corresponding to each bond transaction, wherein the standard meta-event includes the bond identifier of a single bond transaction, the seller and the buyer of the transaction;

[0032] Based on a preset event graph architecture, basic space vectors corresponding to each of the standard meta-events are generated to form a corresponding event graph;

[0033] According to the transaction characteristic attribute data corresponding to each of the bond transactions, the multi-dimensional transaction characteristic attributes corresponding to each of the basic space vectors are generated.

[0034] Furthermore, before the transaction data of each bond transaction is standardized to form the standard meta-event and transaction characteristic attribute data corresponding to each bond transaction, the method further includes:

[0035] The transaction data of each bond transaction is collected based on a preset stream processing platform, and the transaction data of each bond transaction is formatted to form transaction data that meets the preset format requirements.

[0036] In a second aspect, the present application provides a bond-related transaction identification device, comprising:

[0037] A vector search module, used to search for a connection path between a node where a first user is located and a node where a second user is located in a preset event graph, and obtain at least one basic space vector corresponding to the connection path, wherein the basic space vector is used to store the bond identifier and the two parties of a single bond transaction; wherein the first user is one of the two parties of the transaction in at least one of the obtained basic space vectors, and the second user is one of the two parties of the transaction in at least one of the obtained basic space vectors;

[0038] An attribute acquisition module, for acquiring the multi-dimensional transaction feature attributes corresponding to each of the basic space vectors if the basic space vector corresponding to the connection path does not contain the target bond identifier;

[0039] The association identification module is used to identify whether there is an association attribute between the first user and the second user based on the multi-dimensional transaction feature attributes corresponding to each of the basic space vectors. If so, the connection path corresponding to the association attribute is determined as the association transaction path between the first user and the second user.

[0040] In a third aspect, the present application provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the method for identifying bond-related transactions is implemented when the processor executes the program.

[0041] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which implements the bond-related transaction identification method when executed by a processor.

[0042] It can be seen from the above technical solution that the present application provides a bond-related transaction identification method and device, which searches for a connection path between a node where a first user is located and a node where a second user is located in a preset event graph, and obtains at least one basic space vector corresponding to the connection path, and the basic space vector is used to store the bond identification and transaction parties of a single bond transaction; wherein the first user is one of the transaction parties in at least one of the obtained basic space vectors, and the second user is one of the transaction parties in at least one of the obtained basic space vectors; if the basic space vector corresponding to the connection path does not contain the target bond identification, then the multidimensional transaction feature attributes corresponding to each of the basic space vectors are obtained; the first user and the second user are identified according to the multidimensional transaction feature attributes corresponding to each of the basic space vectors. Whether there is an associated attribute between the two users, if so, the connection path corresponding to the associated attribute is determined as the associated transaction path between the first user and the second user. When the direct relationship between the two parties to the bond transaction cannot be found, multi-dimensional attribute association identification can be performed to obtain the indirect association relationship between the two parties to the bond transaction, and the illegal bond related transactions that are concealed by the application of transaction objects and transaction methods can be effectively and accurately identified. In other words, the accuracy and effectiveness of bond related transaction identification can be effectively improved, and then the risk of illegal related transactions can be effectively reduced, the accuracy and reliability of risk warning, avoidance or processing of illegal related transactions can be improved, the operational reliability of the bond trading platform and the compliance of bond transactions can be improved, the legitimate rights and interests of bond users and issuing institutions can be guaranteed, and the user experience of bond users and issuing institutions can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0044] Figure 1 This is a first flow chart of the bond-related transaction identification method in the embodiment of the present application.

[0045] Figure 2 This is a first example flow chart of a connection path in the bond-related transaction identification method in the embodiment of the present application.

[0046] Figure 3 This is a second example flow chart of the connection path in the bond-related transaction identification method in the embodiment of the present application.

[0047] Figure 4It is a specific flow chart of step 100 in the bond-related transaction identification method in the embodiment of the present application.

[0048] Figure 5 It is a specific flow chart of step 200 in the bond-related transaction identification method in the embodiment of the present application.

[0049] Figure 6 It is a specific flow chart of step 300 in the bond-related transaction identification method in the embodiment of the present application.

[0050] Figure 7 This is a second flow chart of the bond-related transaction identification method in the embodiment of the present application.

[0051] Figure 8 This is a third flow chart of the bond-related transaction identification method in the embodiment of the present application.

[0052] Fig. 9 This is a fourth flow chart of the bond-related transaction identification method in the embodiment of the present application.

[0053] Fig.10 It is a structural diagram of a bond-related transaction identification device in an embodiment of the present application.

[0054] Fig.11 It is a structural diagram of the bond-related transaction identification system in the application example of this application.

[0055] Fig.12 It is a schematic diagram of the internal structure of the data collection component in the bond-related transaction identification system in the application example of this application.

[0056] Fig.13 It is a schematic diagram of the internal structure of the multi-dimensional vectorized event graph processing component of the transaction event in the bond-related transaction identification system in the application example of this application.

[0057] Fig.14 It is a schematic diagram of the internal structure of the related transaction analysis component in the bond related transaction identification system in the application example of this application.

[0058] Fig.15 It is a schematic diagram of the structure of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0060] It should be noted that the bond-related transaction identification method and device disclosed in the present application can be used in the fields of big data, artificial intelligence and financial technology, and can also be used in any field other than big data, artificial intelligence and finance. The application field of the bond-related transaction identification method and device disclosed in the present application is not limited.

[0061] In bond investment transactions, since related-party transactions can be concealed based on the characteristics of bond investment transactions, it is difficult to directly analyze related-party transactions. Specifically,

[0062] 1. Concealment through the subject matter of the transaction

[0063] The subject of investment in bond investment transactions is bonds. The actual annualized yields (including discounts and premiums) of bonds issued by the same bond issuer or by issuers with the same credit rating are equivalent to a certain extent. Therefore, when related transactions occur, they can be concealed by transactions involving equivalent bond replacement. For example: a financial investment manager sells bond B1 from financial product A1 to financial product A2, and financial product A2 then sells the equivalent bond B2 to financial product A3. Among them, bonds B1 and B2 are equivalent to a certain extent, and the concealment of related transactions between financial product A1 and financial product A3 is achieved through equivalent transactions.

[0064] 2. Concealment through transactions

[0065] There are various ways of bond investment transactions, including bond trading, repurchase (pledge), repurchase (buyout), bond lending, pledge and other trading methods. Bond trading across trading methods provides a favorable place for concealing related transactions. For example, financial product A1 buys bond B1 from the primary market through bond trading, and after buying bond B1, trades bond B1 to financial product A2 through positive repurchase transactions. After that, financial product A3 borrows bond B1 from financial product A2 through bond lending or other possible trading methods, which finally realizes the concealment of related transactions between financial product A1 and financial product A3.

[0066] With regard to the above two concealed methods of related-party transactions, in the existing transaction system, conventional analysis methods cannot effectively identify related-party transactions, resulting in the inability to expose business risks.

[0067] Based on this, in view of the problem that the existing bond-related transaction identification methods cannot guarantee the accuracy and comprehensiveness of bond-related transaction identification, and thus cannot guarantee the compliance of bond transactions, the present application provides a bond-related transaction identification method, a bond-related transaction identification device, an electronic device and a computer-readable storage medium, by searching for a connection path between a node where a first user is located and a node where a second user is located in a preset event graph, and obtaining at least one basic space vector corresponding to the connection path, the basic space vector being used to store the bond identification and the two transaction parties of a single bond transaction; wherein the first user is one of the two transaction parties in at least one of the obtained basic space vectors, and the second user is one of the two transaction parties in at least one of the obtained basic space vectors; if the basic space vector corresponding to the connection path does not contain the target bond identification, then the multi-dimensional transaction feature attributes corresponding to each of the basic space vectors are obtained. property; according to the multi-dimensional transaction characteristic attributes corresponding to each of the basic space vectors, it is identified whether there is an associated attribute between the first user and the second user. If so, the connection path corresponding to the associated attribute is determined as the associated transaction path between the first user and the second user. When the direct relationship between the two parties to the bond transaction cannot be found, multi-dimensional attribute association identification can be performed to obtain the indirect association relationship between the two parties to the bond transaction, and illegal bond-related transactions that are concealed by the application of transaction objects and transaction methods can be effectively and accurately identified. In other words, the accuracy and effectiveness of bond-related transaction identification can be effectively improved, and then the risk of illegal related transactions can be effectively reduced, the accuracy and reliability of risk warning, avoidance or processing of illegal related transactions can be improved, the operational reliability of the bond trading platform and the compliance of bond transactions can be improved, the legitimate rights and interests of bond users and issuing institutions, etc. can be guaranteed, and the user experience of bond users and issuing institutions, etc. can be improved.

[0068] The details are described in detail through the following embodiments and application examples.

[0069] In order to solve the problem that the existing bond-related transaction identification method cannot guarantee the accuracy and comprehensiveness of bond-related transaction identification, and thus cannot guarantee the compliance of bond transactions, the present application provides an embodiment of a bond-related transaction identification method, see Figure 1 The bond-related transaction identification method specifically includes the following contents:

[0070] Step 100: Search for a connection path between a node where a first user is located and a node where a second user is located in a preset event graph, and obtain at least one basic space vector corresponding to the connection path, wherein the basic space vector is used to store the bond identifier and the two parties to a transaction of a single bond transaction; wherein the first user is one of the two parties to the transaction in at least one of the basic space vectors obtained, and the second user is one of the two parties to the transaction in at least one of the basic space vectors obtained.

[0071] In one or more embodiments of the present application, the event graph is composed of nodes and edges connecting the nodes, and each node is used to represent a user, and the edge is used to represent the correspondence between the users at both ends, that is, it can contain basic space vectors.

[0072] In step 100, firstly, a bond-related transaction identification instruction is received, and the user identifiers of the two users to be identified and the bond identifier of the target bond are read from the bond-related transaction identification instruction; the two users can be confirmed as the first user and the second user to be processed currently. It can be understood that the first user and the second user are only expressions to distinguish between non-identical users, and the expression does not limit the transaction role, that is: the first user can be a bond seller or a bond buyer, and similarly, the second user can also be a bond seller or a bond buyer. Based on this, compared with the existing bond-related transaction identification method that can only search for related transactions between the two parties to the transaction, the present application can search for related transaction data about a certain bond between any two users of transaction roles, and can further explore bond-related transactions to improve the accuracy of bond-related transaction identification.

[0073] In one example, if the user A, user B and the bond with the bond code 123 to be identified are obtained according to the currently received bond-related transaction identification instruction, the node where user A is located and the node where user B is located are found in the event graph, and then all the connection paths between the nodes where user A and B are located are found. In one example, see Figure 2 , there are three connection paths between the nodes where users A and B are located, the first one is "AB", the second one is "BEA", and the third one is "ACDB". Then the basic space vectors represented by the edges corresponding to these three connection paths are obtained, which are:

[0074] (1) The basic space vector X4 of the first connection path “AB” is: (A, 232)->(B, 232); that is, user A is the seller, B is the buyer, and the bond code of the transaction is 232;

[0075] (2) There are two basis space vectors for the second connection path “BEA”, namely:

[0076] The basic space vector X6 of BE is: (B, 123)>(E, 123); that is, user B is the seller, E is the buyer, and the bond code of the transaction is 123;

[0077] The basic space vector X5 of EA is: (E, 561)->(A, 561); that is, user E is the seller, A is the buyer, and the bond code of the transaction is 561;

[0078] (3) The third connection path “ACDB” has three basic space vectors, namely:

[0079] The basic space vector X1 of AC is: (A, 567)->(C, 567); that is, user A is the seller, C is the buyer, and the bond code of the transaction is 567;

[0080] The basic space vector X2 of CD is: (C, 123)->(D, 123); that is, user C is the seller, D is the buyer, and the bond code of the transaction is 123;

[0081] The basic space vector X3 of DB is: (D, 345)->(B, 345); that is, user D is the seller, B is the buyer, and the bond code of the transaction is 345.

[0082] Step 200: If the basic space vector corresponding to the connection path does not contain the target bond identifier, then the multi-dimensional transaction feature attributes corresponding to each of the basic space vectors are obtained.

[0083] It can be understood that the multi-dimensional transaction characteristic attributes include: transaction date, bond credit rating, bond issuer, actual bond yield, institution to which the selling product belongs, institution to which the buying product belongs, industry to which the bond belongs and bond classification, etc. At least two of the transaction characteristic attributes are included, and preferably all of them are included to further improve the comprehensiveness and accuracy of bond-related transaction identification.

[0084] With the aforementioned Figure 2Taking a specific example, for the first connection path "AB", since the basic space vector therein does not include the bond with the bond code 123, it can be determined that the first connection path "AB" is a connection path without a direct correlation relationship; and for the second connection path "BEA" and the third connection path "ACDB", since the basic space vectors therein both include the bond with the bond code 123, it is shown that these two connection paths are paths with a direct correlation relationship, which is suitable for the identification scenario of concealed related transactions by replacing transactions with equivalent bonds. Therefore, the complete transaction information corresponding to the two connection paths can be directly output so that risk control personnel can quickly and reliably perform risk control processing for corresponding illegal related transactions.

[0085] As for the first connection path "AB", since the second connection path "BEA" and the third connection path "ACDB" have determined that there is a transaction relationship between A and B regarding the bond with bond code 123, in order to further improve identification efficiency, no processing can be performed on the first connection path "AB".

[0086] In another example, see Figure 3 , there is only one connection path "AFGB" between the nodes where users A and B are located, and there are three basic space vectors of this connection path "AFGB", which are:

[0087] The basic space vector X7 of AF is: (A, 567)->(F, 567); that is, user A is the seller, F is the buyer, and the bond code of the transaction is 567;

[0088] The basic space vector X8 of FG is: (F, 234)>(G, 234); that is, user F is the seller, G is the buyer, and the bond code of the transaction is 234;

[0089] GB's basic space vector X9 is: (G, 345)->(B, 345); that is, user G is the seller, B is the buyer, and the bond code of the transaction is 345.

[0090] It can be seen that for the connection path "AFGB", since the basic space vectors therein do not contain bonds with bond code 123, it means that this connection path is a path in which there is no direct association between users A and B, and therefore meets the execution conditions of step 200: if the basic space vector corresponding to the connection path does not contain the target bond identifier.

[0091] Therefore, for the connection path "AFGB", it can be known that the multi-dimensional transaction feature attributes corresponding to each of the basic space vectors AF, FG and GB are obtained, and then step 300 is executed to perform indirect relationship identification, so as to be applicable to the violation association scenarios related to attributes such as cross-diversity bond trading methods.

[0092] Step 300: Identify whether there is an associated attribute between the first user and the second user based on the multidimensional transaction feature attributes corresponding to each of the basic space vectors. If so, determine the connection path corresponding to the associated attribute as the associated transaction path between the first user and the second user.

[0093] In an implementation example of step 300, referring to Table 1, the multi-dimensional transaction characteristic attributes include transaction date, bond credit rating, bond issuer, bond actual yield, seller product institution, buyer product institution, bond industry, and bond classification. The identifiers of each characteristic attribute in Table 1 are only identifiers, not specific content. For example, the specific content actually corresponding to transaction date 1-1 may be "November 03, 2020 13:00".

[0094] Table 1

[0095]

[0096] Based on the above Table 1, the association attribute judgment can be performed according to the similarity rules corresponding to each type of transaction feature attribute. For example, the similarity rule corresponding to the transaction date can be "time difference is less than 10 minutes". If more than half of the transaction feature attributes between X7-X9 among the eight transaction feature attributes are judged to be similar through the association attribute judgment, then the connection path "AFGB" is determined as the association transaction path between user A and user B. In other words, by applying this method, although the direct association relationship between the target bond (bond code 123) is not found in the connection path "AFGB", according to the similarity judgment of the feature attributes, it can still be determined that there is an indirect association relationship between user A and user B, and then the risk avoidance or control of illegal bond-related transactions can be carried out based on the indirect association relationship.

[0097] In order to further improve the execution efficiency of step 300, the multi-dimensional transaction feature attributes may be subjected to dimensionality reduction processing, which will be described in detail in subsequent embodiments.

[0098] From the above description, it can be seen that the bond-related transaction identification method provided in the embodiment of the present application can perform multi-dimensional attribute association identification when the direct relationship between the two parties to the bond transaction cannot be found, so as to obtain the indirect association relationship between the two parties to the bond transaction, and can effectively and accurately identify the hidden illegal bond-related transactions using the transaction subject matter and transaction method. In other words, it can effectively improve the accuracy and effectiveness of bond-related transaction identification, and thus effectively reduce the risk of illegal related transactions, improve the accuracy and reliability of risk warning, avoidance or processing of illegal related transactions, improve the operational reliability of the bond trading platform and the compliance of bond transactions, ensure the legitimate rights and interests of bond users and issuers, and improve the user experience of bond users and issuers.

[0099] In order to improve the efficiency of obtaining the multi-dimensional transaction feature attributes corresponding to each of the basic space vectors, in one embodiment of the bond-related transaction identification method provided in the present application, the bond-related transaction identification method further specifically includes the following contents:

[0100] The event graph is a multidimensional vectorized event graph; wherein each node in the multidimensional vectorized event graph corresponds to each user identifier one by one; and the edge in the multidimensional vectorized event graph is used to represent the basic space vector between the nodes at both ends thereof;

[0101] The basic space vector includes a first vector unit and a second vector unit storing the same bond identifier;

[0102] The first vector unit is also used to store the selling party of the transaction, which may be: (buying party, bond code);

[0103] The second vector unit is also used to store the buyer of the transaction parties, which may be specifically: (buyer, bond code).

[0104] The first vector unit and the second vector unit respectively correspond to pre-stored multi-dimensional transaction feature attributes.

[0105] Based on the above multi-dimensional vectorized event graph, see Figure 4 Step 100 of the bond-related transaction identification method provided in this application specifically includes the following contents:

[0106] Step 110: receiving a bond-related transaction identification instruction, and reading the user identifiers of two users to be identified and the bond identifier of the target bond from the bond-related transaction identification instruction; the two users can be confirmed as the first user and the second user to be processed currently.

[0107] Step 120: Determine the node where the first user is located and the node where the second user is located in the multi-dimensional vectorized event graph according to the user identifier of the first user and the user identifier of the second user.

[0108] Step 130: Determine all connection paths between the node where the first user is located and the node where the second user is located from the multi-dimensional vectorized event graph, and extract at least one basic space vector corresponding to each of the connection paths.

[0109] Specifically, firstly, through the direct association between vectors, that is, according to the basic attributes (seller, bond code) -> (buyer, bond code), we can identify the vectors that can be associated with each other. We can obtain all the vectors where the end point of one vector is the starting point of another vector, that is, identify the basic associated transactions.

[0110] From the above description, it can be seen that the bond-related transaction identification method provided in the embodiment of the present application, by setting the basic space vector to include a first vector unit and a second vector unit, and the correspondence between the first vector unit and the second vector unit and the pre-stored multi-dimensional transaction feature attributes, can effectively improve the efficiency, reliability and accuracy of obtaining the multi-dimensional transaction feature attributes corresponding to each of the basic space vectors.

[0111] In order to improve the efficiency of obtaining the multi-dimensional transaction feature attributes corresponding to each of the basic space vectors, in one embodiment of the bond-related transaction identification method provided in the present application, the bond-related transaction identification method further specifically includes the following contents:

[0112] The edge line in the multidimensional vectorized event graph is also used to represent the multidimensional space vector between the nodes at both ends thereof; each of the multidimensional space vectors corresponds to each of the basic space vectors one by one; the multidimensional space vector includes a third vector unit and a fourth vector unit storing the same bond identifier;

[0113] The third vector unit is also used to store the seller and the multi-dimensional transaction feature attributes, such as: (seller, transaction date, bond code, bond issuer, bond yield, bond credit rating, ...);

[0114] The fourth vector unit is also used to store the buyer and the multi-dimensional transaction feature attributes, such as: (buyer, transaction date, bond code, bond issuer, bond yield, bond credit rating, ...).

[0115] Based on multi-dimensional vectorized event graph, see Figure 5 , step 200 in the bond-related transaction identification method provided in this application specifically includes the following contents:

[0116] Step 210: searching whether the target bond identifier is included in at least one basic space vector corresponding to the connection path, if not, executing step 220;

[0117] Step 220: Extract at least one multidimensional space vector corresponding to the connection path from the multidimensional vectorized event graph.

[0118] Specifically, after the multi-dimensional vectorized event graph is established, specific transaction events are represented by multi-dimensional vectors, for example: (sold product, transaction date, bond code, bond issuer, bond yield, bond credit rating, ...) -> (purchased product, transaction date, bond code, bond issuer, bond yield, bond credit rating, ...). For vectors that cannot directly use basic attributes for related transaction identification, support vector machines (SVM) are used to classify the extended attributes (bond issuer, bond credit rating, bond industry, bond yield, ...) one by one.

[0119] From the above description, it can be seen that the bond-related transaction identification method provided in the embodiment of the present application can effectively improve the acquisition efficiency, reliability and accuracy of the multidimensional transaction feature attributes corresponding to each of the basic space vectors by setting a multidimensional space vector for storing multidimensional transaction feature attributes in the graph.

[0120] In order to improve the efficiency and accuracy of obtaining bond-related transaction identification results, in one embodiment of the bond-related transaction identification method provided in this application, see Figure 6 Step 300 in the bond-related transaction identification method specifically includes the following contents:

[0121] Step 310: Based on a preset support vector machine, dimensionality reduction processing is performed on the multidimensional transaction feature attributes extracted from each of the multidimensional space vectors to obtain the reduced-dimensional attributes corresponding to the multidimensional transaction feature attributes.

[0122] Step 320: Obtain the similarity between the same dimensionality reduction attributes in each of the multidimensional space vectors;

[0123] Step 330: If there is a dimension reduction attribute with a similarity greater than a similarity threshold, it is determined that there is an associated attribute between the first user and the second user, and the connection path corresponding to the associated attribute is determined as the associated transaction path between the first user and the second user.

[0124] Step 340: Output the transaction data corresponding to each of the multidimensional space vectors on the associated transaction path.

[0125] Specifically, the extended attributes that belong to the same category after classification by the support vector machine (SVM) can be normalized to achieve attribute dimension reduction, and then re-queried based on the normalized result.

[0126] For example, referring to the aforementioned Table 1, the eight transaction feature attributes can be reduced to two transaction feature attributes based on a preset dimensionality reduction rule (indicating which transaction feature attributes can be normalized and the normalization method), and then the associated attributes are judged according to the similarity rules corresponding to the two reduced attributes. If the two reduced attributes are both judged to be similar, the connection path "AFGB" is determined as the associated transaction path between user A and user B.

[0127] From the above description, it can be seen that the bond-related transaction identification method provided in the embodiment of the present application can effectively improve the efficiency and accuracy of obtaining bond-related transaction identification results by performing dimensionality reduction processing on the multi-dimensional transaction feature attributes extracted from each of the multi-dimensional space vectors based on a preset support vector machine, and performing similarity recognition on the reduced dimensionality attributes.

[0128] In order to improve the acquisition efficiency and comprehensive applicability of bond-related transaction identification results, in one embodiment of the bond-related transaction identification method provided in this application, see Figure 7 The bond-related transaction identification method further specifically includes the following contents after step 100:

[0129] Step 400: If the basic space vector corresponding to the connection path includes a target bond identifier, the connection path is determined as an associated transaction path between the first user and the second user, and the associated bond is the bond corresponding to the target bond identifier.

[0130] Step 500: Output the transaction data corresponding to each of the basic space vectors on the associated transaction path and the target bond identifier.

[0131] From the above description, it can be seen that the bond-related transaction identification method provided in the embodiment of the present application, after identifying the target bond identifier in the basic space vector corresponding to the connection path, directly determines the connection path as the related transaction path between the first user and the second user, which can effectively improve the efficiency of obtaining the bond-related transaction identification results and the comprehensiveness of applicability.

[0132] In order to improve the efficiency and accuracy of event graph application, in one embodiment of the bond-related transaction identification method provided in this application, see Figure 8 The bond-related transaction identification method may further specifically include the following contents before step 100:

[0133] Step 020: Standardize the transaction data of each bond transaction to form a standard meta-event and transaction characteristic attribute data corresponding to each bond transaction, wherein the standard meta-event includes the bond identification of a single bond transaction, the seller and the buyer among the two parties of the transaction.

[0134] Step 030: Based on the preset event graph architecture, basic space vectors corresponding to each of the standard meta-events are generated respectively to form a corresponding event graph.

[0135] Step 040: Generate multi-dimensional transaction characteristic attributes corresponding to each of the basic space vectors according to the transaction characteristic attribute data corresponding to each of the bond transactions.

[0136] From the above description, it can be seen that the bond-related transaction identification method provided in the embodiment of the present application can effectively improve the efficiency and reliability of event graph construction, and then effectively improve the efficiency and accuracy of event graph application, so as to effectively improve the efficiency and accuracy of bond-related transaction identification.

[0137] In order to improve the efficiency and convenience of constructing an event graph using the transaction data of each bond transaction, in one embodiment of the bond-related transaction identification method provided in the present application, see Fig. 9 The bond-related transaction identification method further includes the following contents before step 020:

[0138] Step 010: Based on a preset stream processing platform, the transaction data of each bond transaction is collected, and the format of the transaction data of each bond transaction is converted to form transaction data that meets the preset format requirements.

[0139] From the above description, it can be seen that the bond-related transaction identification method provided in the embodiment of the present application can effectively improve the efficiency and reliability of collecting transaction data of each bond transaction, and can effectively improve the efficiency and convenience of constructing an event graph using the transaction data of each bond transaction.

[0140] From the software level, in order to solve the problem that the existing bond-related transaction identification method cannot guarantee the accuracy and comprehensiveness of bond-related transaction identification, and thus cannot guarantee the compliance of bond transactions, the present application provides an embodiment of a bond-related transaction identification device for executing all or part of the contents of the bond-related transaction identification method, see Fig.10 The bond-related transaction identification device specifically includes the following contents:

[0141] The vector search module 10 is used to search for a connection path between a node where a first user is located and a node where a second user is located in a preset event graph, and obtain at least one basic space vector corresponding to the connection path, wherein the basic space vector is used to store the bond identifier and the two parties to a transaction of a single bond transaction; wherein the first user is one of the two parties to the transaction in at least one of the basic space vectors obtained, and the second user is one of the two parties to the transaction in at least one of the basic space vectors obtained.

[0142] In one or more embodiments of the present application, the event graph is composed of nodes and edges connecting the nodes, and each node is used to represent a user, and the edge is used to represent the correspondence between the users at both ends, that is, it can contain basic space vectors.

[0143] In the vector search module 10, a bond-related transaction identification instruction is first received, and the user identifiers of the two users to be identified and the bond identifier of the target bond are read from the bond-related transaction identification instruction; the two users can be confirmed as the first user and the second user to be processed. It can be understood that the first user and the second user are only expressions to distinguish between non-identical users, and the expression does not limit the transaction role, that is: the first user can be a bond seller or a bond buyer, and similarly, the second user can also be a bond seller or a bond buyer. Based on this, compared with the existing bond-related transaction identification method that can only search for related transactions between the two parties to the transaction, the present application can search for related transaction data about a certain bond between any two users of transaction roles, and can further explore bond-related transactions to improve the accuracy of bond-related transaction identification.

[0144] The attribute acquisition module 20 is used to acquire the multi-dimensional transaction feature attributes corresponding to each of the basic space vectors if the basic space vector corresponding to the connection path does not contain the target bond identifier.

[0145] It can be understood that the multi-dimensional transaction characteristic attributes include: transaction date, bond credit rating, bond issuer, actual bond yield, institution to which the selling product belongs, institution to which the buying product belongs, industry to which the bond belongs and bond classification, etc. At least two of the transaction characteristic attributes are included, and preferably all of them are included to further improve the comprehensiveness and accuracy of bond-related transaction identification.

[0146] The association identification module 30 is used to identify whether there is an association attribute between the first user and the second user based on the multi-dimensional transaction feature attributes corresponding to each of the basic space vectors. If so, the connection path corresponding to the association attribute is determined as the association transaction path between the first user and the second user.

[0147] The embodiment of the bond-related transaction identification device provided in the present application can be specifically used to execute the processing flow of the embodiment of the bond-related transaction identification method in the above-mentioned embodiment. Its functions are not repeated here, and reference can be made to the detailed description of the above-mentioned method embodiment.

[0148] From the above description, it can be seen that the bond-related transaction identification device provided in the embodiment of the present application can perform multi-dimensional attribute association identification when the direct relationship between the two parties to the bond transaction cannot be found, so as to obtain the indirect association relationship between the two parties to the bond transaction, and can effectively and accurately identify the hidden illegal bond-related transactions using transaction objects and transaction methods. In other words, it can effectively improve the accuracy and effectiveness of bond-related transaction identification, and thus effectively reduce the risk of illegal related transactions, improve the accuracy and reliability of risk warning, avoidance or processing of illegal related transactions, improve the operational reliability of the bond trading platform and the compliance of bond transactions, ensure the legitimate rights and interests of bond users and issuing institutions, and improve the user experience of bond users and issuing institutions.

[0149] In order to improve the efficiency of obtaining the multi-dimensional transaction feature attributes corresponding to each of the basic space vectors, in one embodiment of the bond-related transaction identification device provided in the present application, the bond-related transaction identification device further specifically includes the following contents:

[0150] The event graph is a multidimensional vectorized event graph; wherein each node in the multidimensional vectorized event graph corresponds to each user identifier one by one; and the edge in the multidimensional vectorized event graph is used to represent the basic space vector between the nodes at both ends thereof;

[0151] The basic space vector includes a first vector unit and a second vector unit storing the same bond identifier;

[0152] The first vector unit is also used to store the selling party of the transaction, which may be: (buying party, bond code);

[0153] The second vector unit is also used to store the buyer of the transaction parties, which may be specifically: (buyer, bond code).

[0154] The first vector unit and the second vector unit respectively correspond to pre-stored multi-dimensional transaction feature attributes.

[0155] Based on the above multi-dimensional vectorized event graph, the vector search module 10 in the bond-related transaction identification device provided by the present application is specifically used to perform the following contents:

[0156] Step 110: receiving a bond-related transaction identification instruction, and reading the user identifiers of two users to be identified and the bond identifier of the target bond from the bond-related transaction identification instruction; the two users can be confirmed as the first user and the second user to be processed currently.

[0157] Step 120: Determine the node where the first user is located and the node where the second user is located in the multi-dimensional vectorized event graph according to the user identifier of the first user and the user identifier of the second user.

[0158] Step 130: Determine all connection paths between the node where the first user is located and the node where the second user is located from the multi-dimensional vectorized event graph, and extract at least one basic space vector corresponding to each of the connection paths.

[0159] Specifically, firstly, through the direct association between vectors, that is, according to the basic attributes (seller, bond code) -> (buyer, bond code), we can identify the vectors that can be associated with each other. We can obtain all the vectors where the end point of one vector is the starting point of another vector, that is, identify the basic associated transactions.

[0160] From the above description, it can be seen that the bond-related transaction identification device provided in the embodiment of the present application, by setting the basic space vector to include a first vector unit and a second vector unit, and the correspondence between the first vector unit and the second vector unit and the pre-stored multi-dimensional transaction feature attributes, can effectively improve the efficiency, reliability and accuracy of obtaining the multi-dimensional transaction feature attributes corresponding to each of the basic space vectors.

[0161] In order to improve the efficiency of obtaining the multi-dimensional transaction feature attributes corresponding to each of the basic space vectors, in one embodiment of the bond-related transaction identification device provided in the present application, the bond-related transaction identification device further specifically includes the following contents:

[0162] The edge line in the multidimensional vectorized event graph is also used to represent the multidimensional space vector between the nodes at both ends thereof; each of the multidimensional space vectors corresponds to each of the basic space vectors one by one; the multidimensional space vector includes a third vector unit and a fourth vector unit storing the same bond identifier;

[0163] The third vector unit is also used to store the seller and the multi-dimensional transaction feature attributes, such as: (seller, transaction date, bond code, bond issuer, bond yield, bond credit rating, ...);

[0164] The fourth vector unit is also used to store the buyer and the multi-dimensional transaction feature attributes, such as: (buyer, transaction date, bond code, bond issuer, bond yield, bond credit rating, ...).

[0165] Based on the multi-dimensional vectorized event graph, the attribute acquisition module 20 in the bond-related transaction identification device provided by the present application is specifically used to perform the following contents:

[0166] Step 210: searching whether the target bond identifier is included in at least one basic space vector corresponding to the connection path, if not, executing step 220;

[0167] Step 220: Extract at least one multidimensional space vector corresponding to the connection path from the multidimensional vectorized event graph.

[0168] Specifically, after the multi-dimensional vectorized event graph is established, specific transaction events are represented by multi-dimensional vectors, for example: (sold product, transaction date, bond code, bond issuer, bond yield, bond credit rating, ...) -> (purchased product, transaction date, bond code, bond issuer, bond yield, bond credit rating, ...). For vectors that cannot directly use basic attributes for related transaction identification, support vector machines (SVM) are used to classify the extended attributes (bond issuer, bond credit rating, bond industry, bond yield, ...) one by one.

[0169] From the above description, it can be seen that the bond-related transaction identification device provided in the embodiment of the present application can effectively improve the efficiency, reliability and accuracy of obtaining the multidimensional transaction characteristic attributes corresponding to each of the basic space vectors by setting a multidimensional space vector for storing multidimensional transaction characteristic attributes in the graph.

[0170] In order to improve the efficiency and accuracy of obtaining the bond-related transaction identification results, in one embodiment of the bond-related transaction identification device provided in the present application, the association identification module 30 in the bond-related transaction identification device is specifically used to perform the following contents:

[0171] Step 310: Based on a preset support vector machine, dimensionality reduction processing is performed on the multidimensional transaction feature attributes extracted from each of the multidimensional space vectors to obtain the reduced-dimensional attributes corresponding to the multidimensional transaction feature attributes.

[0172] Step 320: Obtain the similarity between the same dimensionality reduction attributes in each of the multidimensional space vectors;

[0173] Step 330: If there is a dimension reduction attribute with a similarity greater than a similarity threshold, it is determined that there is an associated attribute between the first user and the second user, and the connection path corresponding to the associated attribute is determined as the associated transaction path between the first user and the second user.

[0174] Step 330: Output the transaction data corresponding to each of the multidimensional space vectors on the associated transaction path.

[0175] Specifically, the extended attributes that belong to the same category after classification by the support vector machine (SVM) can be normalized to achieve attribute dimension reduction, and then re-queried based on the normalized result.

[0176] For example, referring to the aforementioned Table 1, the eight transaction feature attributes can be reduced to two transaction feature attributes based on a preset dimensionality reduction rule (indicating which transaction feature attributes can be normalized and the normalization method), and then the associated attributes are judged according to the similarity rules corresponding to the two reduced attributes. If the two reduced attributes are both judged to be similar, the connection path "AFGB" is determined as the associated transaction path between user A and user B.

[0177] From the above description, it can be seen that the bond-related transaction identification device provided in the embodiment of the present application can effectively improve the efficiency and accuracy of obtaining bond-related transaction identification results by performing dimensionality reduction processing on the multi-dimensional transaction feature attributes extracted from each of the multi-dimensional space vectors based on a preset support vector machine, and performing similarity recognition on the reduced dimensionality attributes.

[0178] In order to improve the acquisition efficiency and comprehensive applicability of bond-related transaction identification results, in one embodiment of the bond-related transaction identification device provided in the present application, the steps in the bond-related transaction identification device are also specifically used to perform the following contents:

[0179] Step 400: If the basic space vector corresponding to the connection path includes a target bond identifier, the connection path is determined as an associated transaction path between the first user and the second user, and the associated bond is the bond corresponding to the target bond identifier.

[0180] Step 500: Output the transaction data corresponding to each of the basic space vectors on the associated transaction path and the target bond identifier.

[0181] From the above description, it can be seen that the bond-related transaction identification device provided in the embodiment of the present application, after identifying the target bond identifier in the basic space vector corresponding to the connection path, directly determines the connection path as the related transaction path between the first user and the second user, which can effectively improve the efficiency of obtaining the bond-related transaction identification results and the comprehensiveness of applicability.

[0182] In order to improve the efficiency and accuracy of event graph application, in one embodiment of the bond-related transaction identification device provided in the present application, the bond-related transaction identification device is also used to perform the following contents:

[0183] Step 020: Standardize the transaction data of each bond transaction to form a standard meta-event and transaction characteristic attribute data corresponding to each bond transaction, wherein the standard meta-event includes the bond identification of a single bond transaction, the seller and the buyer among the two parties of the transaction.

[0184] Step 030: Based on the preset event graph architecture, basic space vectors corresponding to each of the standard meta-events are generated respectively to form a corresponding event graph.

[0185] Step 040: Generate multi-dimensional transaction characteristic attributes corresponding to each of the basic space vectors according to the transaction characteristic attribute data corresponding to each of the bond transactions.

[0186] From the above description, it can be seen that the bond-related transaction identification device provided in the embodiment of the present application can effectively improve the efficiency and reliability of event graph construction, and then effectively improve the efficiency and accuracy of event graph application, so as to effectively improve the efficiency and accuracy of bond-related transaction identification.

[0187] In order to improve the efficiency and convenience of constructing an event graph using the transaction data of each bond transaction, in one embodiment of the bond-related transaction identification device provided in the present application, the bond-related transaction identification device is also used to perform the following contents:

[0188] Step 010: Based on a preset stream processing platform, the transaction data of each bond transaction is collected, and the format of the transaction data of each bond transaction is converted to form transaction data that meets the preset format requirements.

[0189] From the above description, it can be seen that the bond-related transaction identification device provided in the embodiment of the present application can effectively improve the efficiency and reliability of collecting transaction data of each bond transaction, and can effectively improve the efficiency and convenience of constructing an event graph using the transaction data of each bond transaction.

[0190] In order to further illustrate the present scheme, the present application also provides a specific application example of a bond related transaction identification method implemented by a bond related transaction identification system, which involves the analysis of related transactions in bond investment transactions in the field of bank asset management. The bond related transaction identification system uses the transaction subject information and transaction order information in bond transactions as characteristic attributes to establish an event graph of bond transactions. In view of the defect that the existing event graph cannot accurately express multiple characteristic attributes, the idea of ​​multi-dimensional vectors is introduced and combined with the event graph to improve the accuracy of the bond transaction event graph and further realize the effective analysis of related transactions.

[0191] In view of the phenomenon that the related transactions of bond investment transactions in the field of asset management and financial management cannot be effectively discovered due to the two types of hidden transaction methods in the existing technology, it is planned to introduce the multi-dimensional attribute characteristics of bond transactions through the vectorized event graph mechanism, dig out the event correlation relationship based on multi-dimensional feature attributes, and improve the effectiveness of related transaction analysis. It has the following characteristics:

[0192] 1. Collect bond transactions of different trading methods through event graphs to form standard trading events. For example, convert complex transactions such as pledge, repurchase, and lending into simple buy and sell standard trading events, reduce the complexity of event analysis, and establish a basic standardized bond trading event graph.

[0193] 2. On the basis of the transaction event map, the characteristics of the transaction target bonds (such as: bond issuer, issuer rating, bond classification, bond credit rating, bond yield) and the characteristics of bond transactions (such as: bond seller, bond buyer, transaction date) are comprehensively introduced, and a multidimensional space vector is established based on the characteristic attributes of bond transactions, and the transaction information is expressed in detail using the multidimensional space vector method.

[0194] 3. When analyzing related transactions, support vector machines are used to normalize similar bond transaction feature attributes, and the original multi-dimensional space vectors are reduced in dimension through normalization to improve the data analysis performance. After the dimension is reduced, the related transaction analysis is performed between the reduced-dimensional space vectors through the characteristics between the vectors, that is, the transaction vectors with continuous endings are extracted as suspicious related transactions.

[0195] The bond related transaction identification system includes three parts: transaction event data collection component, transaction event multi-dimensional vectorized event graph processing component and related transaction analysis component. The system of the application example of this application supports independent deployment. Fig.11 When the system is running, the data collection component first collects or receives the transaction data pushed by the transaction system, and then the transaction event multi-dimensional vectorized event graph processing component processes the transaction data into standard transaction events, and establishes a vectorized event graph based on transaction events and transaction attributes. Finally, when related transaction analysis is required, the related transaction query analysis component performs related transaction analysis based on the vectorized event graph. The following is a detailed description of each internal component:

[0196] 1. Data collection component

[0197] The data collection component collects transaction data based on the stream processing platform (Kafka). The trading system uploads the transaction data or transaction log in the specified data format by connecting to the process platform client. After the data is uploaded, the stream platform (Kafka) consumer deployed by this system obtains the bond transaction data uploaded by the trading system through a scheduled task. Fig.12 The internal structure of the data acquisition component in the system of the application example of the present application is schematically shown.

[0198] 2. Transaction event multi-dimensional vectorized event graph processing component

[0199] The multidimensional vector component of trading events specifically includes two parts: standardization of bond trading events and establishment of multidimensional vector event graph. Fig.13 The internal structure of the multi-dimensional vectorized event graph processing component of the transaction event in the application example system of the present application is schematically shown.

[0200] 1) Standardization of bond trading events: Mainly used to process different types of bond trading (such as repurchase, bond lending and pledge) into standard bond trading transactions, simplifying event processing. For example, financial product A1 and counterparty financial product A2 initiate a reverse repurchase transaction of bond B1, which is actually equivalent to financial product A2 selling bond B1 to financial product A1. After processing, a standard bond trading event in the format of [seller, bond, buyer] is formed.

[0201] 2) Establishment of multi-dimensional vectorized event graph: Extract transaction events through event graph, extract standard transaction events into meta-events that meet the event graph, and expand meta-events into multi-dimensional space vectors according to the transaction attributes of bonds during the underlying data processing. For example, the basic meta-event [seller, bond, buyer] is expanded to the basic space vector (seller, bond code) -> (buyer, bond code). On the basis of the basic space vector, the characteristic attributes of bond transactions are continuously introduced to perform multi-dimensional expansion of the vector. For example, the bond transaction date is introduced, and the space vector becomes (seller, bond code, transaction date) -> (buyer, bond code, transaction date). The bond transaction characteristic attributes introduced in this system mainly include: bond credit rating, bond issuer, bond actual yield, seller product institution, buyer product institution, bond industry, bond classification, etc.

[0202] 3. Related transaction analysis component

[0203] After establishing a multi-dimensional vectorized event graph through the transaction event multi-dimensional vectorized event graph processing component, the associated transaction analysis component is used to analyze bond transaction associated events. Fig.14 The internal structure of the associated transaction analysis component in the application example system of this application is schematically shown as follows:

[0204] 1) Association vector query

[0205] After the multi-dimensional vectorized event map is established, specific transaction events are represented by multi-dimensional vectors, for example: (sold product, transaction date, bond code, bond issuer, bond yield, bond credit rating, ...) -> (purchased product, transaction date, bond code, bond issuer, bond yield, bond credit rating, ...). First, through the direct association between vectors, that is, according to the basic attributes (seller, bond code) -> (buyer, bond code) can be mutually associated vector identification. Get all vectors where the end point of vector 1 is the starting point of vector 2, that is, identify the basic associated transactions.

[0206] 2) Feature attribute classification SVM

[0207] For vectors that cannot directly use basic attributes to identify related transactions, support vector machines (SVM) are used to classify extended attributes (bond issuer, bond credit rating, bond industry, bond yield, etc.) one by one.

[0208] 3) Directed dimensionality reduction of multi-dimensional transaction event vectors

[0209] For the extended attributes that belong to the same category after classification by the support vector machine SVM, they are normalized and the normalized result is brought back into step 1) for query, that is, the query is performed again through step 1), and the extended attributes that belong to the same category after classification by the SVM are considered to be the same value for query analysis.

[0210] In summary, the bond-related transaction identification system and method provided by the application example of this application flexibly combines the software and hardware resource availability of the service provider with the offline and registration of the service, and mainly achieves the following effects:

[0211] 1. The characteristic attributes of bond trading are introduced into the event graph mechanism through multi-dimensional vectors, so that the event graph can accurately express bond trading events.

[0212] 2. When analyzing related transactions, the characteristic attributes of bond transactions are normalized through support vector machines to reduce the complexity of related transaction analysis. At the same time, similar bonds are associated through feature normalization, which can effectively solve the problem of related transactions not being discovered by replacing bonds.

[0213] From the hardware level, in order to solve the problem that the existing bond-related transaction identification method cannot guarantee the accuracy and comprehensiveness of bond-related transaction identification, and thus cannot guarantee the compliance of bond transactions, the present application provides an embodiment of an electronic device for implementing all or part of the contents of the bond-related transaction identification method, and the electronic device specifically includes the following contents:

[0214] Fig.15FIG. 9 is a schematic block diagram of the system structure of the electronic device 9600 according to an embodiment of the present application. Fig.15 As shown, the electronic device 9600 may include a central processor 9100 and a memory 9140; the memory 9140 is coupled to the central processor 9100. It is worth noting that Fig.15 is exemplary; other types of structures may also be used to supplement or replace this structure to implement telecommunication functions or other functions.

[0215] In one embodiment, the bond-related transaction identification function may be integrated into a central processing unit, wherein the central processing unit may be configured to perform the following control:

[0216] Step 100: Search for a connection path between a node where a first user is located and a node where a second user is located in a preset event graph, and obtain at least one basic space vector corresponding to the connection path, wherein the basic space vector is used to store the bond identifier and the two parties to a transaction of a single bond transaction; wherein the first user is one of the two parties to the transaction in at least one of the basic space vectors obtained, and the second user is one of the two parties to the transaction in at least one of the basic space vectors obtained.

[0217] In one or more embodiments of the present application, the event graph is composed of nodes and edges connecting the nodes, and each node is used to represent a user, and the edge is used to represent the correspondence between the users at both ends, that is, it can contain basic space vectors.

[0218] In step 100, firstly, a bond-related transaction identification instruction is received, and the user identifiers of the two users to be identified and the bond identifier of the target bond are read from the bond-related transaction identification instruction; the two users can be confirmed as the first user and the second user to be processed currently. It can be understood that the first user and the second user are only expressions to distinguish between non-identical users, and the expression does not limit the transaction role, that is: the first user can be a bond seller or a bond buyer, and similarly, the second user can also be a bond seller or a bond buyer. Based on this, compared with the existing bond-related transaction identification method that can only search for related transactions between the two parties to the transaction, the present application can search for related transaction data about a certain bond between any two users of transaction roles, and can further explore bond-related transactions to improve the accuracy of bond-related transaction identification.

[0219] Step 200: If the basic space vector corresponding to the connection path does not contain the target bond identifier, then the multi-dimensional transaction feature attributes corresponding to each of the basic space vectors are obtained.

[0220] It can be understood that the multi-dimensional transaction characteristic attributes include: transaction date, bond credit rating, bond issuer, actual bond yield, institution to which the selling product belongs, institution to which the buying product belongs, industry to which the bond belongs and bond classification, etc. At least two of the transaction characteristic attributes are included, and preferably all of them are included to further improve the comprehensiveness and accuracy of bond-related transaction identification.

[0221] Step 300: Identify whether there is an associated attribute between the first user and the second user based on the multidimensional transaction feature attributes corresponding to each of the basic space vectors. If so, determine the connection path corresponding to the associated attribute as the associated transaction path between the first user and the second user.

[0222] From the above description, it can be seen that the electronic device provided in the embodiment of the present application can perform multi-dimensional attribute association identification when the direct relationship between the two parties to the bond transaction cannot be found, so as to obtain the indirect association relationship between the two parties to the bond transaction, and can effectively and accurately identify the hidden illegal bond-related transactions using the transaction subject matter and transaction method. In other words, it can effectively improve the accuracy and effectiveness of bond-related transaction identification, and thus effectively reduce the risk of illegal related transactions, improve the accuracy and reliability of risk warning, avoidance or processing of illegal related transactions, improve the operational reliability of the bond trading platform and the compliance of bond transactions, ensure the legitimate rights and interests of bond users and issuers, and improve the user experience of bond users and issuers.

[0223] In another embodiment, the bond-related transaction identification device can be configured separately from the central processor 9100. For example, the bond-related transaction identification device can be configured as a chip connected to the central processor 9100, and the bond-related transaction identification function can be implemented under the control of the central processor.

[0224] like Fig.15 As shown, the electronic device 9600 may also include: a communication module 9110, an input unit 9120, an audio processor 9130, a display 9160, and a power supply 9170. It is worth noting that the electronic device 9600 does not necessarily have to include Fig.15 In addition, the electronic device 9600 may also include Fig.15 For components not shown, reference may be made to the prior art.

[0225] like Fig.15 As shown, the central processing unit 9100 is sometimes also referred to as a controller or an operation control, and may include a microprocessor or other processor device and / or logic device. The central processing unit 9100 receives input and controls the operation of various components of the electronic device 9600.

[0226] The memory 9140 may be, for example, one or more of a cache, a flash memory, a hard drive, a removable medium, a volatile memory, a non-volatile memory or other suitable devices. The above-mentioned information related to the failure may be stored, and a program for executing the relevant information may also be stored. The CPU 9100 may execute the program stored in the memory 9140 to implement information storage or processing, etc.

[0227] The input unit 9120 provides input to the central processing unit 9100. The input unit 9120 is, for example, a key or a touch input device. The power supply 9170 is used to provide power to the electronic device 9600. The display 9160 is used to display display objects such as images and texts. The display may be, for example, an LCD display, but is not limited thereto.

[0228] The memory 9140 may be a solid-state memory, such as a read-only memory (ROM), a random access memory (RAM), a SIM card, etc. It may also be a memory that saves information even when the power is off, can be selectively erased, and is provided with more data, examples of which are sometimes referred to as EPROMs, etc. The memory 9140 may also be some other type of device. The memory 9140 includes a buffer memory 9141 (sometimes referred to as a buffer). The memory 9140 may include an application / function storage unit 9142, which is used to store application programs and function programs or processes for executing the operation of the electronic device 9600 through the central processor 9100.

[0229] The memory 9140 may also include a data storage unit 9143 for storing data, such as contacts, digital data, pictures, sounds, and / or any other data used by the electronic device. The driver storage unit 9144 of the memory 9140 may include various drivers for communication functions of the electronic device and / or for executing other functions of the electronic device (such as messaging applications, address book applications, etc.).

[0230] The communication module 9110 is a transmitter / receiver 9110 that transmits and receives signals via an antenna 9111. The communication module (transmitter / receiver) 9110 is coupled to the central processor 9100 to provide input signals and receive output signals, which may be the same as the case of a conventional mobile communication terminal.

[0231] Based on different communication technologies, multiple communication modules 9110 may be provided in the same electronic device, such as a cellular network module, a Bluetooth module and / or a wireless LAN module, etc. The communication module (transmitter / receiver) 9110 is also coupled to a speaker 9131 and a microphone 9132 via an audio processor 9130 to provide an audio output via the speaker 9131 and receive an audio input from the microphone 9132, thereby realizing a common telecommunication function. The audio processor 9130 may include any suitable buffer, decoder, amplifier, etc. In addition, the audio processor 9130 is also coupled to the central processor 9100, so that recording can be performed on the local machine through the microphone 9132, and the sound stored on the local machine can be played through the speaker 9131.

[0232] The embodiments of the present application also provide a computer-readable storage medium capable of implementing all the steps of the bond-related transaction identification method in the above embodiments. The computer-readable storage medium stores a computer program. When the computer program is executed by a processor, all the steps of the bond-related transaction identification method in the above embodiments are implemented. For example, when the processor executes the computer program, the following steps are implemented:

[0233] Step 100: Search for a connection path between a node where a first user is located and a node where a second user is located in a preset event graph, and obtain at least one basic space vector corresponding to the connection path, wherein the basic space vector is used to store the bond identifier and the two parties to a transaction of a single bond transaction; wherein the first user is one of the two parties to the transaction in at least one of the basic space vectors obtained, and the second user is one of the two parties to the transaction in at least one of the basic space vectors obtained.

[0234] In one or more embodiments of the present application, the event graph is composed of nodes and edges connecting the nodes, and each node is used to represent a user, and the edge is used to represent the correspondence between the users at both ends, that is, it can contain basic space vectors.

[0235] In step 100, firstly, a bond-related transaction identification instruction is received, and the user identifiers of the two users to be identified and the bond identifier of the target bond are read from the bond-related transaction identification instruction; the two users can be confirmed as the first user and the second user to be processed currently. It can be understood that the first user and the second user are only expressions to distinguish between non-identical users, and the expression does not limit the transaction role, that is: the first user can be a bond seller or a bond buyer, and similarly, the second user can also be a bond seller or a bond buyer. Based on this, compared with the existing bond-related transaction identification method that can only search for related transactions between the two parties to the transaction, the present application can search for related transaction data about a certain bond between any two users of transaction roles, and can further explore bond-related transactions to improve the accuracy of bond-related transaction identification.

[0236] Step 200: If the basic space vector corresponding to the connection path does not contain the target bond identifier, then the multi-dimensional transaction feature attributes corresponding to each of the basic space vectors are obtained.

[0237] It can be understood that the multi-dimensional transaction characteristic attributes include: transaction date, bond credit rating, bond issuer, actual bond yield, institution to which the selling product belongs, institution to which the buying product belongs, industry to which the bond belongs and bond classification, etc. At least two of the transaction characteristic attributes are included, and preferably all of them are included to further improve the comprehensiveness and accuracy of bond-related transaction identification.

[0238] Step 300: Identify whether there is an associated attribute between the first user and the second user based on the multidimensional transaction feature attributes corresponding to each of the basic space vectors. If so, determine the connection path corresponding to the associated attribute as the associated transaction path between the first user and the second user.

[0239] From the above description, it can be seen that the computer-readable storage medium provided in the embodiment of the present application can perform multi-dimensional attribute association identification when the direct relationship between the two parties to the bond transaction cannot be found, so as to obtain the indirect association relationship between the two parties to the bond transaction, and can effectively and accurately identify the hidden illegal bond-related transactions using the transaction subject matter and transaction method. In other words, it can effectively improve the accuracy and effectiveness of bond-related transaction identification, and thus effectively reduce the risk of illegal related transactions, improve the accuracy and reliability of risk warning, avoidance or processing of illegal related transactions, improve the operational reliability of the bond trading platform and the compliance of bond transactions, ensure the legitimate rights and interests of bond users and issuers, and improve the user experience of bond users and issuers.

[0240] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, devices, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0241] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (apparatus), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0242] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0243] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0244] The present invention uses specific embodiments to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A method for identifying bond-related transactions, characterized in that: include: Searching for all connection paths between the node where the first user is located and the node where the second user is located in a preset event graph, and obtaining a basic space vector corresponding to each of the connection paths, wherein the basic space vector is used to store the bond identifier and the two parties of a single bond transaction; wherein the first user is one of the two parties of the transaction in the obtained basic space vector, and the second user is one of the two parties of the transaction in the obtained basic space vector; If the target bond identifier is not included in the basic space vectors corresponding to the connection paths, then the multi-dimensional transaction feature attributes corresponding to the basic space vectors are obtained; According to the multi-dimensional transaction feature attributes corresponding to each of the basic space vectors, it is identified whether there is an associated attribute between the first user and the second user. If so, the connection path corresponding to the associated attribute is determined as the associated transaction path between the first user and the second user.

2. The bond-related transaction identification method according to claim 1, characterized in that: The event graph is a multi-dimensional vectorized event graph; Wherein, each node in the multidimensional vectorized event graph corresponds to each user identifier one by one; the edge in the multidimensional vectorized event graph is used to represent the basic space vector between the nodes at both ends thereof; The basic space vector includes a first vector unit and a second vector unit storing the same bond identifier; The first vector unit is further used to store the selling party of the two transaction parties, and the second vector unit is further used to store the buying party of the two transaction parties; The first vector unit and the second vector unit respectively correspond to pre-stored multi-dimensional transaction feature attributes.

3. The bond-related transaction identification method according to claim 2, characterized in that: The edge in the multi-dimensional vectorized event graph is also used to represent the multi-dimensional space vector between the nodes at both ends thereof; There is a one-to-one correspondence between each of the multidimensional space vectors and each of the basic space vectors; The multidimensional space vector includes a third vector unit and a fourth vector unit storing the same bond identifier; The third vector unit is also used to store the seller and the multi-dimensional transaction feature attributes; The fourth vector unit is also used to store the buyer and the multi-dimensional transaction feature attributes; Correspondingly, if the basic space vectors corresponding to the connection paths do not contain the target bond identifier, the multi-dimensional transaction feature attributes corresponding to the basic space vectors are obtained, including: It is searched whether the target bond identifier is included in the basic space vector corresponding to each of the connection paths. If not, each multidimensional space vector corresponding to the connection path is extracted from the multidimensional vectorized event map.

4. The bond-related transaction identification method according to claim 3, characterized in that: The step of identifying whether there is an associated attribute between the first user and the second user according to the multi-dimensional transaction feature attributes corresponding to each of the basic space vectors, and if so, determining the connection path corresponding to the associated attribute as the associated transaction path between the first user and the second user, includes: Based on a preset support vector machine, a dimensionality reduction process is performed on the multidimensional transaction feature attributes extracted from each of the multidimensional space vectors to obtain a dimensionality reduction attribute corresponding to the multidimensional transaction feature attribute; Obtaining similarities between the same dimensionality reduction attributes in each of the multidimensional space vectors; If there is a dimension reduction attribute whose similarity is greater than the similarity threshold, it is determined that there is an associated attribute between the first user and the second user, and the connection path corresponding to the associated attribute is determined as the associated transaction path between the first user and the second user; Output the transaction data corresponding to each of the multidimensional space vectors on the associated transaction path.

5. The bond-related transaction identification method according to claim 1, characterized in that: Also includes: If the basic space vector corresponding to the connection path contains the target bond identifier, the connection path is determined as the associated transaction path between the first user and the second user, and the associated bond is the bond corresponding to the target bond identifier; The transaction data corresponding to each of the basic space vectors on the associated transaction path and the target bond identifier are output.

6. The bond-related transaction identification method according to claim 1, characterized in that: Before searching for all connection paths between the node where the first user is located and the node where the second user is located in the preset event graph and obtaining the basic space vector corresponding to each of the connection paths, the method further includes: Standardizing the transaction data of each bond transaction to form standard meta-events and transaction characteristic attribute data corresponding to each bond transaction, wherein the standard meta-event includes the bond identifier of a single bond transaction, the seller and the buyer of the transaction; Based on a preset event graph architecture, basic space vectors corresponding to each of the standard meta-events are generated to form a corresponding event graph; According to the transaction characteristic attribute data corresponding to each of the bond transactions, the multi-dimensional transaction characteristic attributes corresponding to each of the basic space vectors are generated.

7. The bond-related transaction identification method according to claim 6, characterized in that: Before the transaction data of each bond transaction is standardized to form the standard meta-event and transaction characteristic attribute data corresponding to each bond transaction, the method further includes: The transaction data of each bond transaction is collected based on a preset stream processing platform, and the transaction data of each bond transaction is formatted to form transaction data that meets the preset format requirements.

8. A bond-related transaction identification device, characterized in that: include: A vector search module, used to search for all connection paths between the node where the first user is located and the node where the second user is located in a preset event graph, and obtain the basic space vector corresponding to each of the connection paths, wherein the basic space vector is used to store the bond identifier and the two parties of a single bond transaction; wherein the first user is one of the two parties of the transaction in the obtained basic space vector, and the second user is one of the two parties of the transaction in the obtained basic space vector; An attribute acquisition module, for acquiring the multi-dimensional transaction feature attribute corresponding to each of the basic space vectors if none of the basic space vectors corresponding to the connection paths contains the target bond identifier; The association identification module is used to identify whether there is an association attribute between the first user and the second user based on the multi-dimensional transaction feature attributes corresponding to each of the basic space vectors. If so, the connection path corresponding to the association attribute is determined as the association transaction path between the first user and the second user.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the bond-related transaction identification method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the bond-related transaction identification method according to any one of claims 1 to 7 is implemented.

Images