Hotspot account identification method, device and electronic device

Abstract

The present application discloses a hotspot account identification method, device and electronic device, which can be applied to the field of artificial intelligence or finance. Obtain a transaction map within a pre-constructed target time period, split the target time period to obtain sub-time periods, obtain the attribute information of the entity, and based on the transaction map, count the parameter information of the entity within the sub-time period, obtain the preset maximum concurrent processing number, analyze the attribute information of the entity, the parameter information and the preset maximum concurrent processing number to filter out hotspot entities from the entity, and use the hotspot entity as a hotspot account. In the present invention, hotspot accounts can be filtered out through the above steps, and then corresponding hotspot account processing strategies can be set for hotspot accounts to avoid the problem of excessive processing pressure on the bank system caused by frequent operations of hotspot accounts, which affects the stability of bank services.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and more particularly, to a hotspot account identification method, device, and electronic device. Background Art

[0002] Hot accounts refer to accounts with high frequency of deductions and deposits. In other words, the data of hot accounts is hot data and will be updated frequently, which may cause the bank system to encounter database processing bottlenecks and a sudden increase in transaction pressure.

[0003] If hotspot accounts can be identified, corresponding hotspot account processing policies can be set for these accounts, thereby avoiding the problem of excessive processing pressure on the banking system caused by frequent operations on hotspot accounts, which affects the stability of banking services. Summary of the Invention

[0004] In view of this, the present invention provides a hotspot account identification method, device and electronic device to solve the problem of the urgent need for hotspot account identification.

[0005] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0006] A hotspot account identification method, comprising:

[0007] Obtaining a pre-built transaction graph within a target time period; the transaction graph includes transaction relationships between entities;

[0008] Splitting the target time period to obtain sub-time periods;

[0009] Obtaining attribute information of the entity;

[0010] Based on the transaction graph, collecting parameter information of the entity in the sub-time period; the parameter information includes transaction inflow information and transaction outflow information; the transaction inflow information is calculated based on the number of transactions with an inflow relationship; the transaction outflow information is the number of transactions with an outflow relationship;

[0011] Get the preset maximum number of concurrent processing;

[0012] The attribute information of the entity, the parameter information and the preset maximum concurrent processing quantity are analyzed to filter out hot entities from the entities, and use the hot entities as hot accounts.

[0013] Optionally, obtain a pre-built transaction graph for the target time period, including:

[0014] Obtaining a maximum transaction processing time from historical transaction records, and determining a target time period based on the maximum transaction processing time;

[0015] Collecting transaction information of entities that have conducted transactions within the target time period;

[0016] Determining transaction direction information in the transaction information;

[0017] When the transaction direction information is transfer-in, establishing a transaction relationship from the transfer-in entity to the entity;

[0018] When the transaction direction information is transfer-out, a transaction relationship is established from the entity to the transfer-out entity, where the transaction relationship includes the transaction direction and / or the number of transactions.

[0019] Optionally, based on the transaction graph, collecting statistics on parameter information of the entity in the sub-time period includes:

[0020] The number of first transactions in which the transaction relationship of the entity is a transfer-out relationship in the transaction graph within the sub-time period is used as the transaction transfer-out information;

[0021] Get the maximum and minimum number of transactions;

[0022] Determine a target transfer-out entity corresponding to the entity; the target transfer-out entity is the transaction transferor of the entity;

[0023] Obtaining the number of second transactions in the transaction graph in which the transaction relationship of the target transfer-out entity is a transfer-in relationship within the sub-time period;

[0024] Obtaining, in the transaction graph within the sub-time period, the number of third transactions in which the transaction relationship between the target transfer-out entity and the entity is a transfer-in relationship;

[0025] Based on the maximum number of transactions, the minimum number of transactions, the second number of transactions, and the third number of transactions, the transaction transfer-in information of the entity is determined.

[0026] Optionally, analyzing the attribute information of the entity, the parameter information, and the preset maximum concurrent processing quantity to filter out hotspot entities from the entities includes:

[0027] The product of the preset maximum concurrent processing quantity and the preset ratio is used as the set threshold;

[0028] Calculating a target number of transactions for the entity based on the entity's inbound and outbound transaction information;

[0029] Filter out entities whose target number is greater than the set threshold and use them as target entities;

[0030] Calling a preset hot account identification model to process the attribute information of the target entity, the parameter information, and the preset maximum concurrent processing quantity to obtain a hot account identification result of the target entity;

[0031] The target entity whose hotspot account identification result is a preset result is used as the hotspot entity.

[0032] Optionally, the generation process of the preset hotspot account identification model includes:

[0033] Obtain attribute samples and parameter samples of entity samples;

[0034] Obtaining an account tag for the entity sample;

[0035] Get the maximum concurrent processing volume sample;

[0036] The preset hotspot account recognition model is trained using the maximum concurrent processing volume sample, the attribute sample of the entity sample, the parameter sample and the account label until the preset training stop condition is met.

[0037] A hotspot account identification device, comprising:

[0038] A graph acquisition module is used to acquire a pre-built transaction graph within a target time period; the transaction graph includes transaction relationships between entities;

[0039] A time splitting module is used to split the target time period into sub-time periods;

[0040] An attribute acquisition module, used to acquire attribute information of the entity;

[0041] an information statistics module, configured to collect parameter information of the entity within the sub-time period based on the transaction graph; the parameter information includes transaction inflow information and transaction outflow information; the transaction inflow information is calculated based on the number of transactions with an inflow relationship; the transaction outflow information is the number of transactions with an outflow relationship;

[0042] The quantity acquisition module is used to obtain the preset maximum concurrent processing quantity;

[0043] The entity screening module is used to analyze the attribute information of the entity, the parameter information and the preset maximum concurrent processing quantity to screen out hot entities from the entities and use the hot entities as hot accounts.

[0044] Optionally, the atlas acquisition module includes:

[0045] A time determination submodule is used to obtain the maximum transaction processing time in the historical transaction records and determine the target time period based on the maximum transaction processing time;

[0046] An information statistics submodule, used to collect transaction information of entities that have conducted transactions within the target time period;

[0047] A first information determination submodule, configured to determine transaction direction information in the transaction information;

[0048] The relationship construction submodule is used to construct a transaction relationship from the receiving entity to the entity when the transaction direction information is transfer-in, and to construct a transaction relationship from the entity to the transferring-out entity when the transaction direction information is transfer-out. The transaction relationship includes the transaction direction and / or the number of transactions.

[0049] Optionally, the information statistics module includes:

[0050] A second information determination submodule is configured to use, within the sub-time period, the number of first transactions in which the transaction relationship of the entity is a transfer-out relationship in the transaction graph as the transaction transfer-out information;

[0051] The first transaction number acquisition submodule is used to obtain the maximum and minimum transaction numbers;

[0052] A first entity determination submodule is configured to determine a target transfer-out entity corresponding to the entity; the target transfer-out entity is the transaction transferor of the entity;

[0053] A second transaction number acquisition submodule is configured to acquire, within the sub-time period, the number of second transactions in which the transaction relationship of the target transfer-out entity is a transfer-in relationship in the transaction graph;

[0054] A third transaction number acquisition submodule, configured to acquire, within the sub-time period, the number of third transactions in the transaction graph in which the transaction relationship between the target transfer-out entity and the entity is a transfer-in relationship;

[0055] The third information determination submodule is configured to determine the entity's transaction transfer information based on the maximum number of transactions, the minimum number of transactions, the second number of transactions, and the third number of transactions.

[0056] Optionally, the entity screening module includes:

[0057] A threshold calculation submodule is used to set the threshold by multiplying the preset maximum number of concurrent processing by the preset ratio;

[0058] A number calculation submodule, configured to calculate a target number of transactions for the entity based on the entity's transaction transfer-in information and transaction transfer-out information;

[0059] An entity screening submodule, configured to screen out entities whose target number is greater than a set threshold and use them as target entities;

[0060] An account identification submodule, configured to call a preset hotspot account identification model, process the attribute information of the target entity, the parameter information, and the preset maximum concurrent processing quantity, and obtain a hotspot account identification result of the target entity;

[0061] The second entity determination submodule is configured to use the target entity whose hotspot account identification result is a preset result as a hotspot entity.

[0062] An electronic device comprising: a memory and a processor;

[0063] Wherein, the memory is used to store programs;

[0064] The processor calls the program and is used to execute the above-mentioned hotspot account identification method.

[0065] Compared with the prior art, the present invention has the following beneficial effects:

[0066] The present invention provides a method, device and electronic device for identifying hot accounts, which obtains a transaction map within a pre-constructed target time period, splits the target time period to obtain sub-time periods, obtains attribute information of the entity, and based on the transaction map, counts parameter information of the entity within the sub-time period, obtains a preset maximum number of concurrent processing, analyzes the attribute information of the entity, the parameter information and the preset maximum number of concurrent processing, to filter out hot entities from the entity, and use the hot entities as hot accounts. In the present invention, hot accounts can be filtered out through the above steps, and then corresponding hot account processing strategies can be set for hot accounts to avoid the problem of excessive processing pressure on the bank system caused by frequent operations of hot accounts, which affects the stability of bank services. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0068] Figure 1 A method flow chart of a hotspot account identification method provided by an embodiment of the present invention;

[0069] Figure 2 A schematic diagram of the structure of the entity graph provided by an embodiment of the present invention;

[0070] Figure 3 A flowchart of another hotspot account identification method provided by an embodiment of the present invention;

[0071] Figure 4 A method flow chart of another hotspot account identification method provided by an embodiment of the present invention;

[0072] Figure 5 A flowchart of another hotspot account identification method provided by an embodiment of the present invention;

[0073] Figure 6 A flowchart of a fifth hotspot account identification method provided by an embodiment of the present invention;

[0074] Figure 7 A schematic structural diagram of a hotspot account identification device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0075] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0076] Hot accounts refer to accounts with high frequency of deductions and deposits. In other words, the data of hot accounts is hot data and will be updated frequently, which may cause the bank system to encounter database processing bottlenecks and a sudden increase in transaction pressure.

[0077] If hotspot accounts can be identified, corresponding hotspot account processing policies can be set for these accounts, thereby avoiding the problem of excessive processing pressure on the banking system caused by frequent operations on hotspot accounts, which affects the stability of banking services.

[0078] When identifying hot accounts, you can rely on manual experience, but this method has a large delay.

[0079] To this end, the present invention provides a hotspot account identification method, device and electronic device, which obtains a transaction map within a pre-constructed target time period, splits the target time period to obtain sub-time periods, obtains the attribute information of the entity, and based on the transaction map, counts the parameter information of the entity within the sub-time period, obtains the preset maximum concurrent processing number, analyzes the attribute information of the entity, the parameter information and the preset maximum concurrent processing number to filter out hotspot entities from the entity, and uses the hotspot entity as a hotspot account. In the present invention, hotspot accounts can be filtered out through the above steps, and then corresponding hotspot account processing strategies can be set for hotspot accounts to avoid the problem of excessive processing pressure on the bank system caused by frequent operations of hotspot accounts, which affects the stability of bank services.

[0080] It should be noted that the hot account identification method, device, and electronic device provided by the present invention can be used in the fields of artificial intelligence or finance. The above is only an example and does not limit the application field of the hot account identification method, device, and electronic device provided by the present invention.

[0081] Based on the above content, the present invention provides a hotspot account identification method, referring to Figure 1 , which may include:

[0082] S11. Obtain a pre-built transaction graph within a target time period.

[0083] In this embodiment, the transaction graph includes transaction relationships between entities. Entities can be users who can generate transactions, such as user A, user B, etc.

[0084] Transaction relationships between entities can be either incoming or outgoing, such as when conducting transactions involving two entities, such as transfers or remittances. For example, if user A transfers money to user B, the transaction relationship for user A is an outgoing one, while for user B, the transaction relationship is an incoming one.

[0085] In practical applications, refer to Figure 3 , step S11 may include:

[0086] S21. Obtain a maximum transaction processing time in historical transaction records, and determine a target time period based on the maximum transaction processing time.

[0087] In this embodiment, each transaction has a corresponding transaction processing time. In this embodiment, the maximum transaction processing time T2 recorded in the historical transaction record is obtained, such as 8 minutes.

[0088] When identifying hot accounts, the statistical data is generally the transaction fluctuations in the period before the account became a hot account.

[0089] You can use a fixed time interval T1, such as 5 minutes. If the current time is 12:00, divide it into five time intervals: [11:35-11:40], [11:40-11:45], [11:45-11:50], [11:50-11:55], and [11:55-12:00]. Analyze the trading fluctuations in these time intervals to identify hot accounts.

[0090] If T1 is greater than the maximum transaction processing time T2, then 11:35-12:00 in this embodiment is the target time period.

[0091] When T1 is less than the maximum transaction processing time T2, T1 = T2 can be used to select a time interval. The time period consisting of the minimum and maximum values of the time interval is the target time period.

[0092] S22. Collecting transaction information of entities that have conducted transactions within the target time period.

[0093] Specifically, the transaction information within the target time period is counted. The transaction information may include the transaction object (such as entity A and entity B), transaction direction information (such as entity A transfers money to entity B, the transaction direction is entity A→entity B), transaction amount, etc.

[0094] S23: Determine the transaction direction information in the transaction information.

[0095] The transaction direction information is as shown above: entity A → entity B.

[0096] S24: When the transaction direction information is transfer-in, establish a transaction relationship from the transfer-in entity to the entity.

[0097] The transaction relationship includes transaction direction and / or transaction number.

[0098] S25: When the transaction direction information is transfer-out, establish a transaction relationship from the entity to the transfer-out entity.

[0099] The transaction relationship includes transaction direction and / or transaction number.

[0100] Specifically, based on the entity, if the transaction direction is inbound, an edge is constructed, which is a directed edge and has a direction from inbound entity to entity. If the transaction direction is outbound, an edge is constructed, which has a direction from entity to outbound entity.

[0101] If an edge exists, and if there is subsequent transaction information, the number of transactions on that edge is increased by 1. This completes the construction of entities and edges in the transaction graph.

[0102] For example, if entity C transfers money to entity A twice, and entity A transfers money to entity B once, then we can construct an edge between entity C and entity A with 2 transactions on the edge, and construct an edge between entity A and entity B with 1 transaction on the edge. Figure 2 .

[0103] It should be noted that for the same transaction, if entity C transfers money to entity A twice, for entity C, the transaction direction information is transfer-out, and for entity A, the transaction direction information is transfer-in.

[0104] S12: Split the target time period into sub-time periods.

[0105] Specifically, the target time period is split into sub-time periods according to the time interval T1. For example, if the current time is 12:00, the target time period is divided into five time periods: [11:35-11:40], [11:40-11:45], [11:45-11:50], [11:50-11:55], and [11:55-12:00]. Each time period is a sub-time period.

[0106] S13. Obtain attribute information of the entity.

[0107] In this embodiment, each entity in the aforementioned transaction graph has corresponding attribute information, such as account type, account balance, customer age, occupation, marital status, education level, etc. The account type and account balance can be stored in the account information table, while the customer age, occupation, marital status, and education level can be stored in the customer information to which the account belongs.

[0108] S14. Based on the transaction graph, collect parameter information of the entity in the sub-time period.

[0109] The parameter information includes transaction transfer-in information and transaction transfer-out information.

[0110] The DNum of the transaction transfer-out information is the number of transactions in the transfer-out relationship. For example, if entity A transfers funds to entities C, D, and F three, two, and four times, respectively, the DNum of entity A's transaction transfer-out information is 3 + 2 + 4 = 9.

[0111] That is, the transaction transfer information DNum directly calculates the total number of transactions on the edge with the entity as the starting point in the transaction graph.

[0112] The transaction transfer-in information CNum is calculated based on the number of transactions whose transaction relationship is a transfer-in relationship. The number of transactions whose transaction relationship is a transfer-in relationship is the total number of transactions on the edge with the entity as the end point.

[0113] In practical applications, refer to Figure 4 , step S14 may include:

[0114] S41. The number of first transactions in which the transaction relationship of the entity is a transfer-out relationship in the transaction graph within the sub-time period is used as the transaction transfer-out information.

[0115] In this embodiment, the transaction transfer-out information DNum directly calculates the total number of transactions of the edge starting from the entity in the transaction graph, that is, the first transaction number.

[0116] S42. Obtain the maximum number of transactions and the minimum number of transactions.

[0117] In this embodiment, the maximum and minimum transaction counts are the maximum and minimum transaction counts for edges in the transaction graph. For example, if there are five edges, and the transaction counts on each edge are 1, 2, 3, 4, and 5, respectively, the maximum and minimum transaction counts are 5 and 1, respectively.

[0118] S43: Determine the target transfer-out entity corresponding to the entity.

[0119] The target transfer-out entity is the entity that transfers out of the transaction. In this embodiment, for an entity, once a transaction with that entity is determined as the destination, the other party to that transaction is the target transfer-out entity. For example, if entity A transfers funds to entity B, and entity B is used as the reference, then entity A is the target transfer-out entity for entity B.

[0120] S44. Obtain the number of second transactions in the transaction graph within the sub-time period, in which the transaction relationship of the target transfer-out entity is a transfer-in relationship.

[0121] Specifically, transactions whose destination is the target transfer-out entity are selected from the transaction graph. The sum of the number of transactions along the edges of these transactions is the second number of transactions. For example, with entity B as the reference, entity A → entity B, then entity A is the target transfer-out entity. The number of edges from entity A → entity B is n(AB), and the total number of transactions with entity A as the destination is n(A). n(A) is the second number of transactions.

[0122] In this embodiment, the second transaction count for the target transfer-out entity is determined because a transaction relationship exists between entity B and entity A, and the transaction direction is from entity A to entity B. Therefore, entity A's transaction status will also affect entity B's transaction status. For example, if one entity encounters a performance bottleneck, the processing time for its associated accounts will also increase. Therefore, when screening hot accounts, the transaction status of the associated accounts must also be considered.

[0123] S45. Obtain the number of third transactions in the transaction graph within the sub-time period, in which the transaction relationship between the target transfer-out entity and the entity is a transfer-in relationship.

[0124] Specifically, taking entity B as the benchmark, the number of transactions in which entity B is the transferee, that is, the transaction relationship is the transfer-in relationship, is counted.

[0125] For example, entity A→entity B, then the number of edges from entity A→entity B is n(AB), which is the number of third transactions.

[0126] S46. Determine the transaction transfer information of the entity based on the maximum transaction number, the minimum transaction number, the second transaction number, and the third transaction number.

[0127] Specifically, taking entity A→entity B and entity G→entity B as examples, the transaction transfer information CNum of entity B is:

[0128]

[0129] Among them, a1 and a2 are weight coefficients, which are the weighted sum of the impact of entity A and entity G on the number of transactions of entity B.

[0130] It should be noted that if there are other entities transferred to entity B, the weighted summation can be continued, and the sum of all weight coefficients is 1.

[0131] S15. Obtain the preset maximum number of concurrent processing.

[0132] In this embodiment, the preset maximum concurrent processing quantity is the maximum quantity that the bank can process concurrently.

[0133] S16: Analyze the attribute information of the entity, the parameter information and the preset maximum concurrent processing quantity to filter out hot entities from the entities, and use the hot entities as hot accounts.

[0134] In this embodiment, in order to improve the accuracy of hotspot account identification, a model is used to identify hotspot accounts.

[0135] In practical applications, refer to Figure 5 , step S16 may include:

[0136] S51: The product of the preset maximum concurrent processing quantity and the preset ratio is used as a set threshold.

[0137] S52: Calculate the target number of the entity based on the transaction transfer-in information and transaction transfer-out information of the entity.

[0138] S53: Filter out entities whose target number is greater than the set threshold and use them as target entities.

[0139] In this embodiment, in order to reduce the amount of calculation when identifying hot accounts, a preset ratio (such as 20%) is set according to the bank's preset maximum concurrent processing number, and then the product of the preset maximum concurrent processing number and the preset ratio is calculated and used as the set threshold.

[0140] Calculate the entity's transaction inflow information CNum + transaction outflow information DNum to obtain the target number. If the target number is not greater than the set threshold, the entity will not be predicted as a hot account.

[0141] If the target number is greater than the set threshold, the entity is determined as the target entity.

[0142] S54: Calling a preset hotspot account identification model, processing the attribute information of the target entity, the parameter information and the preset maximum concurrent processing quantity, and obtaining a hotspot account identification result of the target entity.

[0143] In this embodiment, the preset hotspot account identification model is trained. Figure 6 , the specific generation process includes:

[0144] S61. Obtain attribute samples and parameter samples of the entity sample.

[0145] Specifically, the attribute sample refers to the above-mentioned attribute information, and the parameter sample refers to the above-mentioned parameter information.

[0146] S62: Obtain the account tag of the entity sample.

[0147] In this embodiment, the entity samples can be manually labeled with account tags, and the account tags are divided into 0 and 1. The account tag of a hot account is 1, and the account tag of a non-hot account is 0.

[0148] S63. Obtain a maximum concurrent processing volume sample.

[0149] In this embodiment, the maximum concurrent processing amount sample may be the aforementioned preset maximum concurrent processing amount.

[0150] S64. Use the maximum concurrent processing volume sample, the attribute sample of the entity sample, the parameter sample and the account label to train the preset hot account identification model until the preset training stop condition is met.

[0151] Specifically, the preset hot account identification model M1 can be a lightGBM tree model. Using the aforementioned maximum concurrent processing volume sample, attribute samples of the entity sample, parameter samples, and account labels as the training set, the lightGBM tree model is iteratively optimized. When the loss function is less than a preset threshold, training is stopped to obtain the preset hot account identification model M1.

[0152] After training the preset hot account identification model M1, the target entity's attribute information, parameter information, and the preset maximum number of concurrent transactions are input into the preset hot account identification model M1. The preset hot account identification model M1 analyzes transaction fluctuations in different sub-time periods to obtain a hot account identification result for the target entity. The hot account identification result can be 0 or 1, where 1 represents a hot account and 0 represents a non-hot account.

[0153] S55: The target entity whose hotspot account identification result is a preset result is used as a hotspot entity.

[0154] Among them, the preset result is 1, that is, the target entity with the hot account identification result of 1 is filtered out and used as the hot entity, and the hot entity is the hot account.

[0155] In this embodiment, a transaction graph within a pre-constructed target time period is obtained, the target time period is split to obtain sub-time periods, the attribute information of the entity is obtained, and based on the transaction graph, the parameter information of the entity within the sub-time period is counted to obtain the preset maximum concurrent processing number. The attribute information of the entity, the parameter information and the preset maximum concurrent processing number are analyzed to filter out hot entities from the entity, and use the hot entities as hot accounts. In the present invention, hot accounts can be filtered out through the above steps, and then corresponding hot account processing strategies can be set for hot accounts to avoid the problem of excessive processing pressure on the bank system caused by frequent operations of hot accounts, which affects the stability of bank services.

[0156] In addition, the present invention can identify hot accounts in advance, reduce the subjectivity and delay of manual settings, avoid performance bottlenecks in the banking system, reduce system pressure and improve stability, thereby improving customer user experience.

[0157] Optionally, based on the above-mentioned embodiment of the hotspot account identification method, another embodiment of the present invention provides a hotspot account identification device, referring to Figure 7 , which may include:

[0158] A graph acquisition module 11 is configured to acquire a pre-built transaction graph within a target time period; the transaction graph includes transaction relationships between entities;

[0159] A time splitting module 12 is used to split the target time period into sub-time periods;

[0160] An attribute acquisition module 13 is used to acquire attribute information of the entity;

[0161] An information statistics module 14 is configured to collect parameter information of the entity within the sub-time period based on the transaction graph; the parameter information includes transaction inflow information and transaction outflow information; the transaction inflow information is calculated based on the number of transactions with an inflow relationship; the transaction outflow information is the number of transactions with an outflow relationship;

[0162] The number acquisition module 15 is used to obtain the preset maximum number of concurrent processing;

[0163] The entity screening module 16 is used to analyze the attribute information of the entity, the parameter information and the preset maximum concurrent processing quantity to screen out hot entities from the entities and use the hot entities as hot accounts.

[0164] Furthermore, the atlas acquisition module includes:

[0165] A time determination submodule is used to obtain the maximum transaction processing time in the historical transaction records and determine the target time period based on the maximum transaction processing time;

[0166] An information statistics submodule, used to collect transaction information of entities that have conducted transactions within the target time period;

[0167] A first information determination submodule, configured to determine transaction direction information in the transaction information;

[0168] The relationship construction submodule is used to construct a transaction relationship from the receiving entity to the entity when the transaction direction information is transfer-in, and to construct a transaction relationship from the entity to the transferring-out entity when the transaction direction information is transfer-out. The transaction relationship includes the transaction direction and / or the number of transactions.

[0169] Furthermore, the information statistics module includes:

[0170] A second information determination submodule is configured to use, within the sub-time period, the number of first transactions in which the transaction relationship of the entity is a transfer-out relationship in the transaction graph as the transaction transfer-out information;

[0171] The first transaction number acquisition submodule is used to obtain the maximum and minimum transaction numbers;

[0172] A first entity determination submodule is configured to determine a target transfer-out entity corresponding to the entity; the target transfer-out entity is the transaction transferor of the entity;

[0173] A second transaction number acquisition submodule is configured to acquire, within the sub-time period, the number of second transactions in which the transaction relationship of the target transfer-out entity is a transfer-in relationship in the transaction graph;

[0174] A third transaction number acquisition submodule, configured to acquire, within the sub-time period, the number of third transactions in the transaction graph in which the transaction relationship between the target transfer-out entity and the entity is a transfer-in relationship;

[0175] The third information determination submodule is configured to determine the entity's transaction transfer information based on the maximum number of transactions, the minimum number of transactions, the second number of transactions, and the third number of transactions.

[0176] Furthermore, the entity screening module includes:

[0177] A threshold calculation submodule is used to set the threshold by multiplying the preset maximum number of concurrent processing by the preset ratio;

[0178] A number calculation submodule, configured to calculate a target number of transactions for the entity based on the entity's transaction transfer-in information and transaction transfer-out information;

[0179] An entity screening submodule, configured to screen out entities whose target number is greater than a set threshold and use them as target entities;

[0180] An account identification submodule, configured to call a preset hotspot account identification model, process the attribute information of the target entity, the parameter information, and the preset maximum concurrent processing quantity, and obtain a hotspot account identification result of the target entity;

[0181] The second entity determination submodule is configured to use the target entity whose hotspot account identification result is a preset result as a hotspot entity.

[0182] Furthermore, a model generation module is included, which is specifically used to:

[0183] Obtain attribute samples and parameter samples of entity samples, obtain account labels of the entity samples, obtain maximum concurrent processing volume samples, and use the maximum concurrent processing volume samples, attribute samples, parameter samples and account labels of the entity samples to train the preset hotspot account recognition model until the preset training stop conditions are met.

[0184] In this embodiment, a transaction graph within a pre-constructed target time period is obtained, the target time period is split to obtain sub-time periods, the attribute information of the entity is obtained, and based on the transaction graph, the parameter information of the entity within the sub-time period is counted to obtain the preset maximum concurrent processing number. The attribute information of the entity, the parameter information and the preset maximum concurrent processing number are analyzed to filter out hot entities from the entity, and use the hot entities as hot accounts. In the present invention, hot accounts can be filtered out through the above steps, and then corresponding hot account processing strategies can be set for hot accounts to avoid the problem of excessive processing pressure on the bank system caused by frequent operations of hot accounts, which affects the stability of bank services.

[0185] In addition, the present invention can identify hot accounts in advance, reduce the subjectivity and delay of manual settings, avoid performance bottlenecks in the banking system, reduce system pressure and improve stability, thereby improving customer user experience.

[0186] It should be noted that, for the working process of each module and sub-module in this embodiment, please refer to the corresponding description in the above embodiment, which will not be repeated here.

[0187] Optionally, based on the above-mentioned embodiments of the hotspot account identification method and apparatus, another embodiment of the present invention provides an electronic device, including: a memory and a processor;

[0188] Wherein, the memory is used to store programs;

[0189] The processor calls the program and is used to execute the above-mentioned hotspot account identification method.

[0190] In this embodiment, a transaction graph within a pre-constructed target time period is obtained, the target time period is split to obtain sub-time periods, the attribute information of the entity is obtained, and based on the transaction graph, the parameter information of the entity within the sub-time period is counted to obtain the preset maximum concurrent processing number. The attribute information of the entity, the parameter information and the preset maximum concurrent processing number are analyzed to filter out hot entities from the entity, and use the hot entities as hot accounts. In the present invention, hot accounts can be filtered out through the above steps, and then corresponding hot account processing strategies can be set for hot accounts to avoid the problem of excessive processing pressure on the bank system caused by frequent operations of hot accounts, which affects the stability of bank services.

[0191] In addition, the present invention can identify hot accounts in advance, reduce the subjectivity and delay of manual settings, avoid performance bottlenecks in the banking system, reduce system pressure and improve stability, thereby improving customer user experience.

[0192] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A hotspot account identification method, characterized in that: include: Obtaining a pre-built transaction graph within a target time period; the transaction graph includes transaction relationships between entities; Splitting the target time period to obtain sub-time periods; Obtaining attribute information of the entity; Based on the transaction graph, collecting parameter information of the entity in the sub-time period; the parameter information includes transaction inflow information and transaction outflow information; the transaction inflow information is calculated based on the number of transactions with an inflow relationship; the transaction outflow information is the number of transactions with an outflow relationship; Get the preset maximum number of concurrent processing; Analyzing the attribute information of the entity, the parameter information, and the preset maximum concurrent processing quantity to filter out hot entities from the entities, and using the hot entities as hot accounts; The step of collecting statistics on parameter information of the entity in the sub-time period based on the transaction graph includes: The number of first transactions in which the transaction relationship of the entity is a transfer-out relationship in the transaction graph within the sub-time period is used as the transaction transfer-out information; Get the maximum and minimum number of transactions; Determine a target transfer-out entity corresponding to the entity; the target transfer-out entity is the transaction transferor of the entity; Obtaining the number of second transactions in the transaction graph in which the transaction relationship of the target transfer-out entity is a transfer-in relationship within the sub-time period; Obtaining, in the transaction graph within the sub-time period, the number of third transactions in which the transaction relationship between the target transfer-out entity and the entity is a transfer-in relationship; Based on the maximum number of transactions, the minimum number of transactions, the second number of transactions, and the third number of transactions, the transaction transfer-in information of the entity is determined.

2. The hotspot account identification method according to claim 1, characterized in that: Get a pre-built transaction graph for the target time period, including: Obtaining a maximum transaction processing time from historical transaction records, and determining a target time period based on the maximum transaction processing time; Collecting transaction information of entities that have conducted transactions within the target time period; Determining transaction direction information in the transaction information; When the transaction direction information is transfer-in, establishing a transaction relationship from the transfer-in entity to the entity; When the transaction direction information is transfer-out, a transaction relationship is established from the entity to the transfer-out entity, where the transaction relationship includes the transaction direction and / or the number of transactions.

3. The hotspot account identification method according to claim 1, characterized in that: Analyzing the attribute information of the entity, the parameter information, and the preset maximum concurrent processing quantity to filter out hot entities from the entity, including: The product of the preset maximum concurrent processing quantity and the preset ratio is used as the set threshold; Calculating a target number of transactions for the entity based on the entity's inbound and outbound transaction information; Filter out entities whose target number is greater than the set threshold and use them as target entities; Calling a preset hot account identification model to process the attribute information of the target entity, the parameter information, and the preset maximum concurrent processing quantity to obtain a hot account identification result of the target entity; The target entity whose hotspot account identification result is a preset result is used as the hotspot entity.

4. The hotspot account identification method according to claim 3, characterized in that: The generation process of the preset hotspot account identification model includes: Obtain attribute samples and parameter samples of entity samples; Obtaining an account tag for the entity sample; Get the maximum concurrent processing volume sample; The preset hotspot account recognition model is trained using the maximum concurrent processing volume sample, the attribute sample of the entity sample, the parameter sample and the account label until the preset training stop condition is met.

5. A hotspot account identification device, characterized in that: include: A graph acquisition module is used to acquire a pre-built transaction graph within a target time period; the transaction graph includes transaction relationships between entities; A time splitting module is used to split the target time period into sub-time periods; An attribute acquisition module, used to acquire attribute information of the entity; an information statistics module, configured to collect parameter information of the entity within the sub-time period based on the transaction graph; the parameter information includes transaction inflow information and transaction outflow information; the transaction inflow information is calculated based on the number of transactions with an inflow relationship; the transaction outflow information is the number of transactions with an outflow relationship; The quantity acquisition module is used to obtain the preset maximum concurrent processing quantity; An entity screening module, configured to analyze the attribute information of the entity, the parameter information, and the preset maximum concurrent processing quantity, so as to screen out hot entities from the entity and use the hot entities as hot accounts; Wherein, the information statistics module includes: A second information determination submodule is configured to use, within the sub-time period, the number of first transactions in which the transaction relationship of the entity is a transfer-out relationship in the transaction graph as the transaction transfer-out information; The first transaction number acquisition submodule is used to obtain the maximum and minimum transaction numbers; A first entity determination submodule is configured to determine a target transfer-out entity corresponding to the entity; the target transfer-out entity is the transaction transferor of the entity; A second transaction number acquisition submodule is configured to acquire, within the sub-time period, the number of second transactions in which the transaction relationship of the target transfer-out entity is a transfer-in relationship in the transaction graph; A third transaction number acquisition submodule, configured to acquire, within the sub-time period, the number of third transactions in the transaction graph in which the transaction relationship between the target transfer-out entity and the entity is a transfer-in relationship; The third information determination submodule is configured to determine the entity's transaction transfer information based on the maximum number of transactions, the minimum number of transactions, the second number of transactions, and the third number of transactions.

6. The hotspot account identification device according to claim 5, characterized in that: The atlas acquisition module includes: A time determination submodule is used to obtain the maximum transaction processing time in the historical transaction records and determine the target time period based on the maximum transaction processing time; An information statistics submodule, used to collect transaction information of entities that have conducted transactions within the target time period; A first information determination submodule, configured to determine transaction direction information in the transaction information; The relationship construction submodule is used to construct a transaction relationship from the receiving entity to the entity when the transaction direction information is transfer-in, and to construct a transaction relationship from the entity to the transferring-out entity when the transaction direction information is transfer-out. The transaction relationship includes the transaction direction and / or the number of transactions.

7. The hotspot account identification device according to claim 5, characterized in that: The entity screening module includes: A threshold calculation submodule is used to set the threshold by multiplying the preset maximum number of concurrent processing by the preset ratio; A number calculation submodule, configured to calculate a target number of transactions for the entity based on the entity's transaction transfer-in information and transaction transfer-out information; An entity screening submodule, configured to screen out entities whose target number is greater than a set threshold and use them as target entities; An account identification submodule, configured to call a preset hotspot account identification model, process the attribute information of the target entity, the parameter information, and the preset maximum concurrent processing quantity, and obtain a hotspot account identification result of the target entity; The second entity determination submodule is configured to use the target entity whose hotspot account identification result is a preset result as a hotspot entity.

8. An electronic device, characterized in that: include: memory and processor; Wherein, the memory is used to store programs; The processor calls the program and is used to execute the hotspot account identification method as described in any one of claims 1-4.

Images