A data processing method, apparatus, device, and readable storage medium

Abstract

The present application discloses a data processing method, apparatus, device and readable storage medium. The method includes: obtaining asset transaction data and the type of transaction behavior to which the asset transaction data belongs; determining the transaction characteristics corresponding to the asset transaction data according to the asset transaction data and the type of transaction behavior; obtaining a feature evaluation matrix corresponding to the type of transaction behavior, and sorting the feature text according to the feature evaluation matrix to obtain a target transaction message corresponding to the asset transaction data. By adopting the present application, the generation efficiency and quality of the message can be improved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and particularly to a data processing method, apparatus, device, and readable storage medium. Background Art

[0002] In an anti-money laundering system, the generation of anti-money laundering messages is an essential part. Currently, for the generation of anti-money laundering messages, manual writing is mainly adopted, which takes a large amount of time, has extremely low timeliness, and is slow, making it unfavorable for quickly completing the reporting of anti-money laundering operations. Summary of the Invention

[0003] Embodiments of this application provide a data processing method, apparatus, device, and readable storage medium, which can improve the generation efficiency and quality of messages.

[0004] On the one hand, embodiments of this application provide a data processing method, including:

[0005] Obtain asset transaction data and the type of transaction behavior to which the asset transaction data belongs;

[0006] Determine the transaction characteristics corresponding to the asset transaction data according to the asset transaction data and the type of transaction behavior;

[0007] Obtain a feature evaluation matrix corresponding to the type of transaction behavior, sort the feature texts according to the feature evaluation matrix, and obtain a target transaction message corresponding to the asset transaction data; the feature text is the text matching the transaction characteristics; the target transaction message is used to identify the transaction legality of the asset transaction data.

[0008] On the one hand, embodiments of this application provide a data processing apparatus, including:

[0009] A data acquisition module, configured to obtain asset transaction data and the type of transaction behavior to which the asset transaction data belongs;

[0010] A feature determination module, configured to determine the transaction characteristics corresponding to the asset transaction data according to the asset transaction data and the type of transaction behavior;

[0011] A message generation module, configured to obtain a feature evaluation matrix corresponding to the type of transaction behavior, sort the feature texts according to the feature evaluation matrix, and obtain a target transaction message corresponding to the asset transaction data; the feature text is the text matching the transaction characteristics; the target transaction message is used to identify the transaction legality of the asset transaction data.

[0012] Among them, the feature determination module includes:

[0013] A set acquisition unit, configured to obtain a set of behavior type characteristics corresponding to the type of transaction behavior;

[0014] A data matching unit for matching asset transaction data with a set of behavior type features;

[0015] A feature determination unit for, if there is a behavior type feature in the set of behavior type features that matches the asset transaction data, using the behavior type feature that matches the asset transaction data as the transaction feature corresponding to the asset transaction data.

[0016] Among them, the feature evaluation matrix is the matrix included in the feature evaluation model; the number of transaction features is N, and N is a positive integer;

[0017] The message generation module includes:

[0018] A matrix acquisition unit for obtaining the feature evaluation matrix corresponding to the transaction behavior type through the feature evaluation model; the feature evaluation matrix includes feature evaluation values corresponding to one or more behavior type features respectively; the feature evaluation value is used to characterize the matching degree between the behavior type feature and the transaction behavior type.

[0019] An evaluation value determination unit for obtaining the feature evaluation values corresponding to N transaction features respectively in the feature evaluation matrix as N target feature evaluation values;

[0020] A sequence generation unit for generating a feature evaluation value sequence including N target feature evaluation values according to the numerical sizes corresponding to the N target feature evaluation values respectively;

[0021] A text sorting unit for sorting the feature texts corresponding to N transaction features respectively according to the feature evaluation value sequence;

[0022] A message generation unit for generating a target transaction message corresponding to the asset transaction data according to the sorted N feature texts.

[0023] Among them, the message generation unit includes:

[0024] A rule acquisition subunit for obtaining a message generation rule; the message generation rule is used to specify the maximum number of words included in the target transaction message;

[0025] A text determination subunit for determining target feature texts among the sorted N feature texts according to the maximum number of words;

[0026] A text splicing subunit for splicing the target feature texts in the sorting order among the sorted N feature texts to obtain a target transaction message corresponding to the asset transaction data; the number of words included in the target transaction message is less than or equal to the maximum number of words.

[0027] Among them, the text determination subunit is further used to select P feature texts among the sorted N feature texts; P is a positive integer less than N;

[0028] The text determination subunit is further configured to obtain the text word count of P feature texts and match the text word count with the maximum word count;

[0029] The text determination subunit is further configured to, if the text word count is less than the maximum word count, determine candidate feature texts from the remaining feature texts according to the sorting order among the sorted N feature texts; the candidate feature texts are the feature texts other than the P feature texts among the N feature texts;

[0030] The text determination subunit is further configured to determine target feature texts according to the candidate feature texts and the P feature texts;

[0031] The text determination subunit is further configured to, if the text word count is greater than or equal to the maximum word count, determine the P feature texts as the target feature texts.

[0032] Wherein, the text determination subunit is further configured to determine the word count difference between the maximum word count and the text word count;

[0033] The text determination subunit is further configured to obtain the candidate text word count corresponding to the candidate feature texts and match the candidate text word count with the word count difference;

[0034] The text determination subunit is further configured to, if the candidate text word count is less than or equal to the word count difference, determine the candidate feature texts and the P feature texts as the target feature texts;

[0035] The text determination subunit is further configured to, if the candidate text word count is greater than the word count difference, determine the P feature texts as the target feature texts.

[0036] Wherein, the device further includes:

[0037] The modified feature acquisition module is configured to obtain the modified transaction features corresponding to the asset transaction data in response to a selection operation for the behavior type features;

[0038] The message update module is configured to sort the modified feature texts according to the feature evaluation matrix and update the target transaction message corresponding to the asset transaction data to a modified transaction message according to the sorting result; the modified feature texts are the texts matching the modified transaction features;

[0039] The parameter adjustment module is configured to adjust the parameters in the feature evaluation matrix in the feature evaluation model according to the modified transaction message to obtain an adjusted feature evaluation matrix; the adjusted feature evaluation matrix is used to sort the feature texts corresponding to the next asset transaction data.

[0040] Wherein, the parameter adjustment module includes:

[0041] An assignment unit, configured to set a transaction feature to a valid value, set a first remaining feature to an invalid value, and generate a transaction feature matrix including the valid value corresponding to the transaction feature and the invalid value corresponding to the first remaining feature; the first remaining feature is a feature other than the transaction feature in the behavior type features.

[0042] The assignment unit is further configured to set a corrected transaction feature to a valid value, set a second remaining feature to an invalid value, and generate a corrected feature matrix including the valid value corresponding to the corrected transaction feature and the invalid value corresponding to the second remaining feature; the second remaining feature is a feature other than the corrected transaction feature in the behavior type features.

[0043] A matrix adjustment unit, configured to adjust parameters of a feature evaluation matrix according to the corrected feature matrix and the transaction feature matrix, to obtain an adjusted feature evaluation matrix.

[0044] Wherein, the matrix adjustment unit includes:

[0045] An operator sub-unit, configured to perform a subtraction operation on the corrected feature matrix and the transaction feature matrix, to obtain an adjusted feature matrix corresponding to the asset transaction data.

[0046] A matrix determination sub-unit, configured to perform an addition operation on the adjusted feature matrix and the feature evaluation matrix, and use the matrix obtained by the addition operation as the adjusted feature evaluation matrix.

[0047] Wherein, the apparatus further includes:

[0048] A suspicious text acquisition module, configured to acquire suspicious feature text in a target transaction message; the suspicious feature text is text that matches a suspicious transaction feature; the transaction features include suspicious transaction features.

[0049] A text matching module, configured to match the suspicious feature text with legal feature text.

[0050] A data determination module, configured to determine that the asset transaction data is legal asset transaction data if the suspicious feature text meets the legal conditions indicated by the legal feature text.

[0051] The data determination module is further configured to determine that the asset transaction data is illegal asset transaction data if the suspicious feature text does not meet the legal conditions indicated by the legal feature text.

[0052] An embodiment of the present application provides a computer device on the one hand, including: a processor and a memory;

[0053] The memory stores a computer program, and when the computer program is executed by the processor, the processor is caused to execute the method in the embodiment of the present application.

[0054] One aspect of the embodiments of the present application provides a computer-readable storage medium storing a computer program, which when loaded and executed by a processor, can execute the method in the embodiments of the present application.

[0055] One aspect of the present application provides a computer program product or a computer program, which includes computer instructions stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the method provided in one aspect of the embodiments of the present application.

[0056] In the embodiments of the present application, after obtaining the asset transaction data, the transaction characteristics corresponding to the asset transaction data can be determined according to the type of transaction behavior to which the asset transaction data belongs; subsequently, the feature texts matching the transaction characteristics can be obtained, and these feature texts can be sorted according to the feature evaluation matrix corresponding to the type of transaction behavior, so that the target transaction message corresponding to the asset transaction data can be generated. It can be seen that after obtaining the asset transaction data and the type of transaction behavior in the present application, the transaction characteristics corresponding to the asset transaction data can be obtained. Subsequently, according to the feature evaluation matrix, the feature texts corresponding to the transaction characteristics can be sorted, so that the feature texts corresponding to the transaction characteristics that more conform to the type of transaction behavior are arranged more forward. When generating the target transaction message, the feature texts arranged forward can be preferentially selected, so that the target transaction message can more conform to the characteristics of the type of transaction behavior. That is to say, the present application can realize automatic message generation, improve the message generation efficiency, and at the same time improve the quality of the generated message. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained according to these drawings.

[0058] Figure 1 is a network architecture diagram provided by the embodiments of the present application;

[0059] Figure 2 is a scenario schematic diagram provided by the embodiments of the present application;

[0060] Figure 3 is a scenario schematic diagram for correcting a message provided by the embodiments of the present application;

[0061] Figure 4It is a flowchart of a data processing method provided by an embodiment of the present application;

[0062] Figure 5 It is a schematic flowchart of a process for adjusting a feature evaluation matrix provided by an embodiment of the present application;

[0063] Figure 6 It is a system framework diagram provided by an embodiment of the present application;

[0064] Figure 7 It is a logic flowchart of generating a target transaction message provided by an embodiment of the present application;

[0065] Figure 8 It is a schematic structural diagram of a data processing device provided by an embodiment of the present application;

[0066] Figure 9 It is a schematic structural diagram of a computer device provided by an embodiment of the present application. Detailed implementation manners

[0067] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0068] Artificial Intelligence (AI) is to use a digital computer or a machine controlled by a digital computer to simulate, extend, and expand human intelligence, a theory, method, technology, and application system that can perceive the environment, acquire knowledge, and use knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technology in computer science. It attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a way similar to human intelligence. Artificial intelligence also studies the design principles and implementation methods of various intelligent machines, enabling the machines to have the functions of perception, reasoning, and decision-making.

[0069] Artificial intelligence technology is an interdisciplinary subject, involving a wide range of fields, including both hardware-level technologies and software-level technologies. Artificial intelligence basic technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation / interaction systems, and mechatronics. Artificial intelligence software technologies mainly include several major directions such as computer vision technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0070] With the research and progress of artificial intelligence technology, artificial intelligence technology has been studied and applied in multiple fields, such as common smart homes, smart wearable devices, virtual assistants, smart speakers, smart marketing, driverless, autonomous driving, drones, robots, smart healthcare, smart customer service, etc. It is believed that with the development of technology, artificial intelligence technology will be applied in more fields and play an increasingly important role.

[0071] The solution provided by the embodiments of this application belongs to Machine Learning (ML) under the field of artificial intelligence.

[0072] Machine Learning (ML) is an interdisciplinary subject involving multiple fields such as probability theory, statistics, approximation theory, convex analysis, and algorithm complexity theory. It specifically studies how computers simulate or implement human learning behaviors to acquire new knowledge or skills and reorganize the existing knowledge structure to continuously improve their own performance. Machine learning is the core of artificial intelligence and the fundamental way to make computers intelligent, and its applications cover all fields of artificial intelligence. Machine learning and deep learning usually include technologies such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and rote learning.

[0073] Please refer to Figure 1 , Figure 1 which is a network architecture diagram provided by the embodiments of this application. As Figure 1 shown, the network architecture may include a business server 1000 and a user terminal cluster. The user terminal cluster may include one or more user terminals, and the number of user terminals will not be limited here. As Figure 1 shown, multiple user terminals may include user terminal 100a, user terminal 100b, user terminal 100c, …, user terminal 100n; as Figure 1 shown, user terminal 100a, user terminal 100b, user terminal 100c, …, user terminal 100n may be respectively connected to the business server 1000 through a network connection, so that each user terminal can perform data interaction with the business server 1000 through this network connection.

[0074] It can be understood that each user terminal as Figure 1 shown may be installed with a target application. When the target application runs on each user terminal, it can be respectively connected to Figure 1Data interaction is carried out between the business servers 1000 shown, so that the business servers 1000 can receive business data from each user terminal. Among them, the target application can include applications with functions of displaying data information such as text, images, audio, and video. For example, the application can be a text editing application (such as a message editing application), which can be used for users to input data and obtain a document corresponding to the data. For example, for a message editing application, a user can input asset transaction data in the message editing application, so as to obtain a message corresponding to the asset transaction data. It should be understood that the user can modify the message through the message editing application to obtain the final corrected message. It should be understood that the business servers 1000 in this application can obtain business data according to these applications. For example, the business data can be the asset transaction data input by the user and the transaction behavior type corresponding to the asset transaction data.

[0075] Subsequently, for the obtained asset transaction data and transaction behavior type, the business server 1000 can obtain a set of behavior type characteristics corresponding to the transaction behavior type, and match the asset transaction data with the set of behavior type characteristics, so as to select, from the set of behavior type characteristics, the behavior type characteristics that match the asset transaction data, and use the behavior type characteristics that match the asset transaction data as the transaction characteristics of the asset transaction data. Further, the business server 1000 can obtain the characteristic text corresponding to the transaction characteristics, and obtain the characteristic evaluation matrix corresponding to the transaction behavior type, where the characteristic evaluation matrix includes the characteristic evaluation values of the characteristic texts corresponding to each behavior type characteristic in the set of behavior type characteristics (a characteristic evaluation value can represent the matching degree between a behavior type characteristic and the transaction behavior type, and the larger the characteristic evaluation value, the higher the matching degree), that is, the characteristic evaluation matrix includes the characteristic evaluation value of the characteristic text corresponding to the transaction characteristics. Then, through the characteristic evaluation matrix, the characteristic text corresponding to the transaction characteristics can be sorted. Further, the business server 1000 can splice the characteristic texts in order according to the arrangement order of the characteristic texts, so as to generate a target transaction message composed of one or more characteristic texts. Subsequently, the business server 1000 can return the target transaction message to the user terminal.

[0076] An embodiment of this application can select a user terminal as the target user terminal from multiple user terminals. The user terminal can include: intelligent terminals with data processing functions (such as text data display function, video data playback function, music data playback function) such as smart phones, tablet computers, laptop computers, desktop computers, smart TVs, smart speakers, desktop computers, smart watches, in-vehicle devices, etc., but is not limited thereto. For example, an embodiment of this application can Figure 1The user terminal 100a shown is used as the target user terminal. The target application can be integrated into the target user terminal. At this time, the target user terminal can perform data interaction with the service server 1000 through the target application.

[0077] For example, when a user uses the target application (such as a message editing application) in the user terminal, after the user enters asset transaction data in the message editing application, the user can select the type of transaction behavior to which the asset transaction data belongs. For example, the type of transaction behavior selected by the user is the type of "crimes of disrupting the financial management order"; subsequently, the user terminal can send the asset transaction data and the type of transaction behavior to which the asset transaction data belongs, that is, the type of "crimes of disrupting the financial management order" to the service server 1000 together; subsequently, the service server 1000 can obtain the set of behavior type characteristics corresponding to the type of "crimes of disrupting the financial management order". The set of behavior type characteristics can include one or more behavior type characteristics. For example, the set of behavior type characteristics can include behavior type characteristic a, behavior type characteristic b, and behavior type characteristic c; further, the service server 1000 can obtain the characteristic evaluation matrix corresponding to the type of "crimes of disrupting the financial management order". The characteristic evaluation matrix can include one or more characteristic evaluation values. A characteristic evaluation value can be used to represent the matching degree between a behavior type characteristic and the type of "crimes of disrupting the financial management order". The higher the characteristic evaluation value, the higher the matching degree. For example, the characteristic evaluation matrix can be as shown in the characteristic evaluation matrix A1:

[0078]

[0079] Among them, as shown in the characteristic evaluation matrix A1, the characteristic evaluation matrix can include the characteristic evaluation value 100, the characteristic evaluation value 96, and the characteristic evaluation value 98. Among them, the characteristic evaluation value 100 can be used to represent the matching degree between the behavior type characteristic a and the type of "crimes of disrupting the financial management order", the characteristic evaluation value 96 can be used to represent the matching degree between the behavior type characteristic b and the type of "crimes of disrupting the financial management order", and the characteristic evaluation value 98 can be used to represent the matching degree between the behavior type characteristic c and the type of "crimes of disrupting the financial management order". It should be understood that since 100>98>96, the behavior type characteristic a has the highest matching degree with the type of "crimes of disrupting the financial management order".

[0080] Further, the business server 1000 can match the asset transaction data with the set of behavior type characteristics corresponding to the type of "crimes of disrupting the order of financial management" (i.e., {behavior type characteristic a, behavior type characteristic b, behavior type characteristic c}). According to the matching result, if it is determined that behavior type characteristic a and behavior type characteristic b in the set of behavior type characteristics match the asset transaction data, then behavior type characteristic a and behavior type characteristic b can be determined as the transaction characteristics corresponding to the asset transaction data. Subsequently, the business server 1000 can obtain, in the feature evaluation matrix A1, the feature evaluation values corresponding to the transaction characteristics (i.e., behavior type characteristic a and behavior type characteristic b) of the asset transaction data (i.e., feature evaluation value 100 and feature evaluation value 96). Based on the feature evaluation value 100 and feature evaluation value 96, the feature text 1 corresponding to behavior type characteristic a and the feature text 2 corresponding to behavior type characteristic b can be sorted. For example, the feature text 1 and feature text 2 can be sorted in descending order of the feature evaluation value, and the sorted feature text sequence is {feature text 1, feature text 2}. Subsequently, the business server 1000 can splice the feature texts in order according to the arrangement order of the feature texts, and the spliced text obtained is "feature text 1 + feature text 2", and this spliced text can be used as the target transaction message corresponding to the asset transaction data. For example, if feature text 1 is "a 20-year-old young person" and feature text 2 is "transfer 700,000 yuan", then the spliced text can be "A 20-year-old young person transfers 700,000 yuan", and this spliced text "A 20-year-old young person transfers 700,000 yuan" can be used as the target transaction message corresponding to the asset transaction data. The business server 1000 can return the target transaction message "A 20-year-old young person transfers 700,000 yuan" to the user terminal, and the user terminal can display the target transaction message "A 20-year-old young person transfers 700,000 yuan", and the user can view the target transaction message "A 20-year-old young person transfers 700,000 yuan" through the user terminal.

[0081] Optionally, it can be understood that when splicing the feature texts in the arrangement order of the feature texts, if there is a situation where the feature evaluation values corresponding to the feature texts are the same, then the splicing can be performed according to the specified order corresponding to the feature texts, where the specified order corresponding to the feature texts can be an artificially specified order. For example, in the feature evaluation matrix A1, the feature evaluation value corresponding to behavior type characteristic a is 100, and the feature evaluation value corresponding to behavior type characteristic b is also 100. Since the feature evaluation values of behavior type characteristic a and behavior type characteristic b are the same, when splicing feature text 1 and feature text 2, the specified order corresponding to the feature texts (feature text 1 is arranged before feature text 2) can be obtained, and feature text 1 and feature text 2 can be spliced into "feature text 1 + feature text 2" according to the specified order corresponding to the feature texts.

[0082] Optionally, it can be understood that when splicing the feature texts in the arranged order of the feature texts to generate the target transaction message of the asset transaction data, the maximum limit word count corresponding to the target transaction message can be obtained first, and the feature texts constituting the target transaction message can be filtered according to the maximum limit word count. For example, if the maximum limit word count is 10, the feature text 1 "a young person of twenty years old" can be used as a component of the target transaction. The text word count of the feature text 1 is 7, and the remaining word count is 3. However, the text word count of the feature text 2 "transfer 700,000 yuan" is 6, so the feature text 2 can be deleted and the feature text 2 may not be used as a component of the target transaction message. That is to say, the target transaction message is composed of the feature text 1, and the target transaction message is "a young person of twenty years old".

[0083] It should be understood that after obtaining the asset transaction data and its corresponding transaction behavior type, the present application can obtain the set of behavior type features corresponding to the transaction behavior type, and match the asset transaction data with the set of behavior type features, so as to determine the transaction features corresponding to the asset transaction data. Through the feature evaluation matrix corresponding to the transaction behavior type, the feature texts corresponding to the transaction features can be sorted, so as to generate a target transaction message that conforms to the transaction features, which can improve the quality of the target transaction message. It can be understood that this process does not require manual participation, can realize automatic message generation, can improve the message generation efficiency, and at the same time, can also improve the quality of the generated message.

[0084] Optionally, it can be understood that the network architecture may include multiple business servers. A user terminal can be connected to one business server. Each business server can obtain the business data in the user terminal connected to it (such as the asset transaction data input by the user and its corresponding transaction behavior type), and process the business data according to the function of the target application (for example, if the target application is a message editing application, the business server can generate the target transaction message corresponding to the business data according to the business data).

[0085] Optionally, it can be understood that the user terminal can also obtain the business data (such as the asset transaction data input by the user and its corresponding transaction behavior type), and process the business data according to the function of the target application (for example, if the target application is a message editing application, the business server can generate the target transaction message corresponding to the business data according to the business data).

[0086] It can be understood that the method provided by the embodiments of the present application can be executed by a computer device, which includes but is not limited to a user terminal or a service server. Among them, the service server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms.

[0087] Among them, the user terminal and the service server can be directly or indirectly connected through wired or wireless communication methods, and the present application does not limit this.

[0088] For ease of understanding, please also refer to Figure 2 , Figure 2 which is a schematic diagram of a scenario provided by the embodiments of the present application. Among them, as Figure 2 shown, the user terminal A can be any user terminal in the user terminal cluster corresponding to the above Figure 1 corresponding embodiment. For example, the user terminal can be the user terminal 100a; as Figure 2 shown, the service server can be the service server 1000 in the above Figure 1 corresponding embodiment.

[0089] As Figure 2 shown, the user a (which can refer to the audit user who audits and reports asset transaction data) can input the asset transaction data to be audited and reported in the target application of the user terminal A (such as Figure 2 the message generation system shown), and select the transaction behavior type to which the asset transaction data belongs from multiple transaction behavior types (including transaction behavior type 1, transaction behavior type 2, and transaction behavior type 3); as Figure 2As shown, the type of transaction behavior selected by user a is transaction behavior type 2. After user a clicks the confirmation control, user terminal A can respond to this trigger operation of user a and send a message generation request to the business server (simultaneously sending the asset transaction data and the transaction behavior type 2 to which the asset transaction data belongs); further, after receiving the message generation request, the business server can obtain the set of behavior type characteristics corresponding to the transaction behavior type 2; further, the business server can match the set of behavior type characteristics with the asset transaction data. Among them, the set of behavior type characteristics includes multiple behavior type characteristics (including behavior type characteristic 1, behavior type characteristic 2, behavior type characteristic 3, behavior type characteristic 4, behavior type characteristic 5, and behavior type characteristic 6). Through the matching, among these multiple behavior type characteristics, the behavior type characteristic that matches the asset transaction data can be determined, and the behavior type characteristic that matches the asset transaction data is used as the transaction characteristic of the asset transaction data. For example, the behavior type characteristic 1 is "abnormal transaction object", and if all the transaction objects included in the asset transaction data are foreigners, it can be determined that the behavior type characteristic 1 "abnormal transaction object" is the behavior type characteristic that matches the asset transaction data, and the behavior type characteristic 1 "abnormal transaction object" is determined as the transaction characteristic of the asset transaction data.

[0090] Further, after determining the transaction characteristics of the asset transaction data (including behavior type characteristic 1, behavior type characteristic 2, and behavior type characteristic 3) through matching, the characteristic evaluation matrix corresponding to the transaction behavior type 2 can be obtained. The characteristic evaluation matrix can be as shown in characteristic evaluation matrix B1:

[0091]

[0092] Among them, the characteristic evaluation matrix B1 can include the characteristic evaluation values corresponding to behavior type characteristic 1, behavior type characteristic 2, behavior type characteristic 3, behavior type characteristic 4, behavior type characteristic 5, and behavior type characteristic 6 respectively. The characteristic evaluation value corresponding to behavior type characteristic 1 is 100, the characteristic evaluation value corresponding to behavior type characteristic 2 is 92, the characteristic evaluation value corresponding to behavior type characteristic 3 is 97, the characteristic evaluation value corresponding to behavior type characteristic 4 is 96, the characteristic evaluation value corresponding to behavior type characteristic 5 is 99, and the characteristic evaluation value corresponding to behavior type characteristic 6 is 100. Through the characteristic evaluation matrix B1, the characteristic texts corresponding to the transaction characteristics of the asset transaction data (including behavior type characteristic 1, behavior type characteristic 2, and behavior type characteristic 3) can be sorted. Among them, the characteristic text can refer to the text pre-configured for the behavior type characteristic. A characteristic text has an association relationship with a behavior type characteristic, that is, a behavior type characteristic can be configured to be associated with a characteristic text.

[0093] For example, the feature text associated with behavior type feature 1 is feature text 1 "All transaction counterparts are foreigners"; the feature text associated with behavior type feature 2 is feature text 2 "Frequent transactions and large amounts"; the feature text associated with behavior type feature 3 is feature text 3 "The transaction entity deliberately avoids attention and disperses transactions for handling at multiple outlets"; since the feature evaluation value 100 of behavior type feature 1 is greater than the feature evaluation value 97 of behavior type feature 2, and the feature evaluation value 100 of behavior type feature 1 is greater than the feature evaluation value 97 of behavior type feature 3, the feature text 1 "All transaction counterparts are foreigners" corresponding to the behavior type feature 1 can be arranged in the first place; and because the feature evaluation values of behavior type feature 2 and behavior type feature 3 are the same, the feature text 2 can be arranged before the feature text 3 according to the text specified order of feature text 2 and feature text 3, and the arrangement order of the feature texts can be obtained as {feature text 1, feature text 2, feature text 3}.

[0094] Further, the service server can obtain the message generation rule and determine that the maximum number of words for message generation specified in the message generation rule is 100. Since the word count of feature text 1 "All transaction counterparts are foreigners" is 9, the word count of feature text 2 "Frequent transactions and large amounts" is 9, and the word count of feature text 3 "The transaction entity deliberately avoids attention and disperses transactions for handling at multiple outlets" is 22, the total word count of feature text 1 + feature text 2 + feature text 3 is 9 + 9 + 22 = 40. The total word count 40 is much less than the maximum word count 100, so the feature text 1, feature text 2, and feature text 3 can be concatenated in order, and the concatenated text obtained is "All transaction counterparts are foreigners, Frequent transactions and large amounts, The transaction entity deliberately avoids attention and disperses transactions for handling at multiple outlets"; the service server can use this concatenated text "All transaction counterparts are foreigners, Frequent transactions and large amounts, The transaction entity deliberately avoids attention and disperses transactions for handling at multiple outlets" as the target transaction message of the asset transaction data and return the target transaction message to the user terminal A. As Figure 2 shown, the user terminal A can display the target transaction message on the display interface. At the same time, as Figure 2As shown, the user terminal A can also display the behavior type characteristics corresponding to the behavior type 2 (including behavior type characteristic 1, behavior type characteristic 2, behavior type characteristic 3, behavior type characteristic 4, behavior type characteristic 5, and behavior type characteristic 6) below the target transaction message in the display interface. It should be understood that since the target transaction message is composed of the characteristic texts corresponding to behavior type characteristic 1, behavior type characteristic 2, and behavior type characteristic 3, the marked behavior type characteristic 1, behavior type characteristic 2, and behavior type characteristic 3 can be displayed. Thus, when user a views the target transaction message through the user terminal A, it can be determined that the target transaction message is composed of the characteristic texts corresponding to behavior type characteristic 1, behavior type characteristic 2, and behavior type characteristic 3.

[0095] Optionally, it can be understood that user a can modify the target transaction message to obtain a modified transaction message, and the user terminal A can send the modified transaction message to the service server. The service server can adjust the characteristic evaluation values in the characteristic evaluation matrix B1 according to the modified transaction message, so that the characteristic evaluation values in the adjusted characteristic evaluation matrix can better meet the manual requirements. Thus, the transaction message obtained through the adjusted characteristic evaluation matrix can better meet the needs of the user.

[0096] For ease of understanding, please also refer to Figure 3 , Figure 3 which is a schematic diagram of a scenario for modifying a message provided by an embodiment of the present application. As Figure 3 shown, user a can select from behavior type characteristic 1, behavior type characteristic 2, behavior type characteristic 3, behavior type characteristic 4, behavior type characteristic 5, and behavior type characteristic 6 to determine the behavior type characteristic that matches the asset transaction data as the transaction characteristic corresponding to the asset transaction data. As Figure 3As shown, the behavior type features selected by user a are behavior type feature 1, behavior type feature 3, and behavior type feature 5. It can be seen that among the behavior type features selected by user a (including behavior type feature 1, behavior type feature 3, and behavior type feature 5) and the behavior type features determined by the business server (including behavior type feature 1, behavior type feature 2, and behavior type feature 3), there are the same behavior type features 1 and behavior type 3. Then, the feature text 2 "Frequent transactions, large amounts" corresponding to behavior type feature 2 in the target transaction message generated by the business server can be deleted. At the same time, the feature text 1 "All transaction objects are foreigners" corresponding to behavior type feature 1 and the feature text 3 "The transaction subject deliberately avoids attention and disperses transactions among multiple outlets" corresponding to behavior type feature 3 are retained. It should be understood that the behavior type features selected by user a also include behavior type feature 5. In the feature evaluation matrix B1, the evaluation value of behavior type feature 5 is 98. Since the evaluation value 98 is less than 100 but greater than 97, the feature text 5 "The addresses left in the account opening materials of multiple people are the same" corresponding to behavior type feature 5 can be placed after feature text 1 and before feature text 3. Then, the arrangement order corresponding to feature text 1, feature text 3, and feature text 5 is {feature text 1, feature text 5, feature text 3}; it should be understood that since the text word count of this feature text 5 "The addresses left in the account opening materials of multiple people are the same" is 12, the total text word count of feature text 1, feature text 3, and feature text 5 is 9 + 22 + 12 = 43. Since the total text word count 43 is less than the maximum word count 100, the concatenated text "All transaction objects are foreigners, the addresses left in the account opening materials of multiple people are the same, the transaction subject deliberately avoids attention and disperses transactions among multiple outlets" composed of feature text 1 + feature text 5 + feature text 3 can be used as the corrected transaction message.

[0097] Furthermore, the user terminal A can return the corrected transaction message to the business server, and the business server can adjust the feature evaluation matrix B1 according to the behavior type features (behavior type feature 1, behavior type feature 3, and behavior type feature 5 selected by user a) corresponding to the feature text in the corrected transaction message to obtain an adjusted feature evaluation matrix, that is, the feature evaluation matrix B2.

[0098] Among them, it can be understood that the evaluation values and the content included in the feature text in the above feature evaluation matrix B1 and feature evaluation matrix B2 are all illustrative examples for easy understanding and have no reference significance. For the specific implementation method of adjusting the feature evaluation matrix to obtain the adjusted feature evaluation matrix, reference can be made to the description in the subsequent Figure 4 corresponding embodiments.

[0099] It should be understood that by obtaining the asset transaction data and the type of transaction behavior to which the asset transaction data belongs, the transaction characteristics of the asset transaction data can be determined according to the set of behavior type characteristics corresponding to the type of transaction behavior. And according to the characteristic evaluation matrix corresponding to the type of transaction behavior, the characteristic texts corresponding to the transaction characteristics can be sorted, so that the characteristic texts corresponding to the transaction characteristics that are more in line with the type of transaction behavior are arranged more forward. When generating the target transaction message, the characteristic texts arranged forward can be preferentially selected, so that the target transaction message can be more in line with the characteristics of the type of transaction behavior. The whole process does not require manual participation, can realize automatic generation of messages, improve the generation efficiency of messages, and at the same time can also improve the quality of messages. At the same time, it should be understood that for the automatically generated target transaction message, it can be manually corrected to obtain a corrected transaction message, and the characteristic evaluation matrix can be adjusted through the corrected transaction message, so that the parameters (evaluation values) in the characteristic evaluation matrix are more in line with the manual requirements, thereby ensuring the quality of the message.

[0100] Further, please refer to Figure 4 , Figure 4 which is a flowchart of a data processing method provided by an embodiment of the present application. Among them, this method can be executed by a user terminal (for example, any user terminal in the user terminal cluster in the corresponding embodiment above Figure 1 ), or can be executed by a business server (for example, the business server in the corresponding embodiment above Figure 1 ), or can be jointly executed by the business server and the user terminal. Hereinafter, taking this method being executed by the user terminal in the area as an example for description, where this data processing method can at least include the following steps S101-step S103:

[0101] Step S101, obtain asset transaction data and the type of transaction behavior to which the asset transaction data belongs.

[0102] In the present application, the asset transaction data may refer to the transaction data of the user for the asset. The asset transaction data may include the basic information of the user, the basic information of the transaction object, the transaction amount, the transaction institution, the transaction location, etc.; and the type of transaction behavior may refer to the type of illegal fund transaction to which the asset transaction belongs. The type of illegal fund transaction may include the type of illegal drug transaction, the type of illegal organization transaction, the type of illegal transportation transaction, the type of unknown asset source, and the type of disrupting the asset management order, etc., and no further examples will be given here.

[0103] In this application, the auditing user who audits asset transaction data can upload the asset transaction data to be audited and reported through a target application (such as a message editing application) in the user terminal, and select the transaction behavior type to which the asset transaction data belongs. The user terminal can obtain the asset transaction data and the transaction behavior type to which it belongs.

[0104] Step S102: Determine the transaction characteristics corresponding to the asset transaction data according to the asset transaction data and the transaction behavior type.

[0105] In this application, after obtaining the asset transaction data and the transaction behavior type to which it belongs, the behavior type feature set corresponding to the transaction behavior type can be obtained. According to the behavior type feature set, the transaction characteristics corresponding to the asset transaction data can be determined. The specific method for determining the transaction characteristics corresponding to the asset transaction data can be as follows: the behavior type feature set corresponding to the transaction behavior type can be obtained; subsequently, the asset transaction data can be matched with the behavior type feature set; if there is a behavior type feature in the behavior type feature set that matches the asset transaction data, the behavior type feature that matches the asset transaction data can be used as the transaction characteristic corresponding to the asset transaction data. It should be understood that the behavior type feature set may include one or more behavior type features, and the one or more behavior type features can be defined manually. For each transaction behavior type, the behavior type features corresponding to the transaction behavior type can be defined manually; by matching the asset transaction data with these behavior type features, it can be determined whether there is a behavior type feature in these behavior type features that matches the asset transaction data. If so, the behavior type feature that matches the asset transaction data can be determined as the transaction characteristic corresponding to the asset transaction data. For example, the behavior type feature set includes the behavior type feature "age greater than 70", and the asset transaction data contains the text "the user's age is 75 years old". Then, through matching, it can be determined that the text "the user's age is 75 years old" in the asset transaction data conforms to the behavior type feature "age greater than 70", and the behavior type feature "age greater than 70" can be used as the transaction characteristic of the asset transaction data.

[0106] Step S103: Obtain the feature evaluation matrix corresponding to the transaction behavior type, sort the feature text according to the feature evaluation matrix to obtain the target transaction message corresponding to the asset transaction data; the feature text is the text that matches the transaction characteristics; the target transaction message is used to identify the transaction legality of the asset transaction data.

[0107] In this application, the feature evaluation matrix may refer to the matrix included in the feature evaluation model. Through this feature evaluation matrix, the feature evaluation matrix can be obtained. Among them, the feature evaluation matrix may include feature evaluation values corresponding to one or more behavior type features respectively. A feature evaluation value can be used to represent the matching degree between a behavior type feature and the transaction behavior type. It should be understood that the higher the feature evaluation value, the more the behavior type feature conforms to the transaction behavior type. Hereinafter, taking the number of transaction features as N (N is a positive integer) as an example, the specific method for determining the target transaction message corresponding to the asset transaction data according to the feature evaluation matrix will be described. The specific method may be that after obtaining the feature evaluation matrix through the feature evaluation model, the feature evaluation values corresponding to N transaction features can be obtained in the feature evaluation matrix as N target feature evaluation values. Subsequently, according to the numerical sizes corresponding to the N target feature evaluation values respectively, a feature evaluation value sequence including the N target feature evaluation values can be generated. According to the feature evaluation value sequence, the feature texts corresponding to the N transaction features can be sorted, and the target transaction message corresponding to the asset transaction data can be generated according to the sorted N feature texts.

[0108] It should be understood that the feature text may refer to the text pre-configured for the behavior type feature. Each behavior type feature may be configured with an associated feature text, and through the feature evaluation values in the feature evaluation matrix, these feature texts can be sorted. For example, the feature text associated with the behavior type feature "age greater than 65" is "elderly person", and the feature evaluation value corresponding to the behavior type feature "age greater than 65" is 100; the feature text associated with the behavior type feature "the names of the transaction objects are not Chinese names" is "all transaction objects are foreigners", and the feature evaluation value corresponding to the behavior type feature "the names of the transaction objects are not Chinese names" is 98. Then, because the feature evaluation value 100 is greater than the feature evaluation value 98, the feature text "elderly person" can be arranged before the feature text "all transaction objects are foreigners".

[0109] It should be understood that after obtaining the sorted N feature texts, these sorted N feature texts can be concatenated in order to obtain a concatenated text, which can be used as the target transaction message for the asset transaction data. For example, the N feature texts include feature text 1 "All trading objects are foreigners", feature text 2 "Frequent trading and large amounts", and feature text 3 "The trading entity deliberately avoids attention". After sorting according to the feature evaluation values, the arrangement order is {feature text 2, feature text 3, feature text 1}. Then, after concatenating feature text 1, feature text 2, and feature text 3 in the arrangement order of {feature text 2, feature text 3, feature text 1}, the obtained concatenated text is "Frequent trading and large amounts, the trading entity deliberately avoids attention, all trading objects are foreigners", and this concatenated text can be used as the target transaction message.

[0110] Among them, it should be understood that when concatenating the sorted feature texts in order, the maximum word count (i.e., the maximum number of characters limit) in the message generation rule can be obtained, and the sorted feature texts can be filtered according to this maximum word count, and the target transaction message can be generated from the filtered feature texts. The specific method can be to obtain the message generation rule; among them, the message generation rule is used to specify the maximum word count included in the target transaction message; P feature texts can be selected from the sorted N feature texts; P is a positive integer less than N; subsequently, the text word count of the P feature texts can be obtained and matched with the maximum word count; if the text word count of the P feature texts is less than the maximum word count, then according to the sorting order among the sorted N feature texts, candidate feature texts can be determined from the remaining feature texts; among them, the candidate feature texts are the feature texts among the N feature texts except the P feature texts; according to the candidate feature texts and the P feature texts, the target feature texts can be determined; and if the text word count is greater than or equal to the maximum word count, then the P feature texts can be determined as the target feature texts. Further, according to the sorting order among the sorted N feature texts, the target feature texts can be concatenated to obtain the target transaction message corresponding to the asset transaction data; the message word count included in the target transaction message is less than or equal to the maximum word count.

[0111] Among them, the specific method for determining the target feature texts according to the candidate feature texts and the P feature texts can be to determine the word count difference between the maximum word count and the text word count; subsequently, the candidate text word count corresponding to the candidate feature texts can be obtained and matched with the word count difference; if the candidate text word count is less than or equal to the word count difference, then the candidate feature texts and the P feature texts can be determined as the target feature texts; and if the candidate text word count is greater than the word count difference, then the P feature texts can be determined as the target feature texts.

[0112] It should be understood that after sorting the feature texts corresponding to the transaction features according to the feature evaluation matrix, the first P feature texts can be selected for splicing in the order of arrangement of the feature texts. At this time, if the text word count of the first P feature texts has reached the maximum word count, the (P + 1)-th feature text and the subsequent feature texts can be no longer considered, and the spliced text composed of the first P feature texts can be directly used as the target transaction message of the asset transaction data; if the text word count of the first P feature texts is less than the maximum word count, the word count difference between the maximum word count and the text word count of the first P feature texts can be obtained. If the candidate text word count of the (P + 1)-th feature text is less than or equal to the word count difference, it can be determined that the total text word count of the (P + 1)-th feature text meets the limit condition of the maximum word count, and the (P + 1)-th feature text can also be used as a component of the target transaction message; similarly, if the total text word count of the first (P + 1) feature texts is less than the maximum word count, the word count difference between the total text word count of the first (P + 1) feature texts and the maximum word count can be continuously obtained. If the candidate text word count of the (P + 2)-th feature text is less than or equal to the word count difference, the (P + 2)-th feature text can also be used as a component of the target transaction message. That is to say, through the maximum word count, the feature texts can be filtered. If the text word count of the first P feature texts has reached the limit condition of the maximum word count, the (P + 1)-th feature text and the subsequent feature texts are filtered.

[0113] It should be understood that the above sorting of the feature texts through the feature evaluation matrix is in a descending order. For the sorting of the feature texts, an ascending order can also be used. After sorting in ascending order, the feature texts can be selected from the back to the front in the order of arrangement to form the target transaction message. That is to say, it can be understood that whether in the ascending order or the descending order, the feature texts with higher feature evaluation values are preferentially selected for splicing. If there are feature texts with the same feature evaluation value, they can be sorted according to the original text specified order of the feature texts. For example, the feature evaluation values of feature text 1 and feature text 2 are both 100, but the original text specified order of feature text 1 and feature text 2 is {feature text 2, feature text 1}, then feature text 2 can be placed before feature text 1 according to this order {feature text 2, feature text 1}.

[0114] It should be understood that for the specific method of confirming the transaction features and sorting the feature texts corresponding to the transaction features through the feature evaluation matrix, it can be as shown in formula (1):

[0115] Y = X S ·X m Formula (1)

[0116] Among them, X in formula (1) Smay refer to the feature evaluation matrix, X m may refer to the transaction feature matrix generated by transaction features, and Y may refer to the target feature evaluation matrix for evaluating the transaction features of asset transaction data.

[0117] It should be understood that the feature evaluation matrix includes the feature evaluation values of one or more behavior type features of the transaction behavior type. Hereinafter, taking the behavior type features including behavior type feature 1, behavior type feature 2, and behavior type feature 3 as an example, formula (1) will be described. The feature evaluation matrices for behavior type feature 1, behavior type feature 2, and behavior type feature 3 can be as shown in feature evaluation matrix C1:

[0118]

[0119] Among them, 98 in feature evaluation matrix C1 can be the feature evaluation value of behavior type feature 1, 100 can be the feature evaluation value of behavior type feature 2, and 97 can be the feature evaluation value of behavior type feature 3.

[0120] Further, after matching the asset transaction data with the behavior type feature 1, behavior type feature 2, and behavior type feature 3, it can be determined that the transaction features of the asset transaction data are behavior type feature 1 and behavior type feature 3. Then, behavior type feature 1 and behavior type feature 3 can be set as valid values. For example, behavior type feature 1 and behavior type feature 3 are set to 1; and since behavior type feature 2 is not the transaction feature of the asset transaction data, behavior type feature 2 can be set as an invalid value. For example, behavior type feature 2 can be set to 0, so as to obtain a transaction feature matrix including the valid values (for example, 1) corresponding to behavior type feature 1 and behavior type feature 3, and the invalid value (for example, 0) corresponding to behavior type feature 2. This transaction feature matrix can be as shown in transaction feature matrix C2:

[0121]

[0122] It should be understood that the above feature evaluation matrix C1 and this transaction feature matrix can be multiplied for processing, and thus the target feature evaluation matrix C3 corresponding to the asset transaction data can be obtained. The target feature evaluation matrix C3 can be as follows:

[0123]

[0124] It should be understood that through the feature evaluation values 98 and 97 in the target feature evaluation matrix C3, the feature texts corresponding to the transaction features (i.e., behavior type feature 1 and behavior type feature 3) of the asset transaction data can be sorted.

[0125] It can be understood that after generating the target transaction message, the target transaction message can be used to determine the legality of the asset transaction data. The specific method for determining the legality of the asset transaction data through the target transaction message can be to obtain the suspicious feature text in the target transaction message; wherein, the suspicious feature text is the text that matches the suspicious transaction feature; the transaction feature can include the suspicious transaction feature; subsequently, the suspicious feature text can be matched with the legal feature text; if the suspicious feature text meets the legal conditions indicated by the legal feature text, it can be determined that the asset transaction data is legal asset transaction data; and if the suspicious feature text does not meet the legal conditions indicated by the legal feature text, it can be determined that the asset transaction data is illegal asset transaction data. For example, the transaction feature includes the suspicious transaction feature "the transaction amount is greater than 300,000 yuan", and the feature text of the suspicious transaction feature "the transaction amount is huge, exceeding 300,000 yuan" can be used as the suspicious feature text; and because the legal feature text is "the transaction amount is not greater than 250,000 yuan", it can be determined that the suspicious feature text "the transaction amount is huge, exceeding 300,000 yuan" does not meet the legal conditions indicated by the legal feature text "the transaction amount is not greater than 250,000 yuan", and it can be determined that the asset transaction data corresponding to the target transaction message is illegal asset transaction data.

[0126] It can be understood that after the user terminal presents the target transaction message corresponding to the asset transaction data in the terminal display interface, the user can adjust the target transaction message to obtain a corrected transaction message, and the corrected transaction message can be used to adjust the feature evaluation matrix so that the feature evaluation values in the feature evaluation matrix are more in line with the user's requirements for the behavior type features and feature texts. Among them, the specific method for the user to adjust the target transaction message can be that the user can directly perform text addition, text deletion, or text modification on the text in the target transaction message; the user can also modify the feature text in the target transaction message by selecting the corrected transaction feature corresponding to the asset transaction data in the behavior type features. For example, for the target transaction message composed of the above-mentioned behavior type feature 1 and behavior type feature 3, in addition to behavior type feature 1 and behavior type feature 3, the corrected transaction feature selected by the user also includes behavior type feature 2, then the feature text corresponding to behavior type feature 2 will also be added to the target transaction message accordingly. It should be noted that if the user adds behavior type feature 2, but at this time the text word count of the target transaction message has reached the maximum word count for message generation, the user can add the feature text of behavior type feature 2 to the message by deleting the feature texts of other behavior type features (for example, deleting behavior type feature 1).

[0127] Among them, for the specific implementation method of adjusting the feature evaluation matrix through the corrected transaction message, reference can be made to the description of the corresponding embodiment in the subsequent Figure 5 description of the corresponding embodiment.

[0128] In this application, by obtaining asset transaction data and the type of transaction behavior to which the asset transaction data belongs, the transaction characteristics of the asset transaction data can be determined according to the set of behavior type characteristics corresponding to the type of transaction behavior. And according to the characteristic evaluation matrix corresponding to the type of transaction behavior, the characteristic text corresponding to the transaction characteristics can be sorted, so that the characteristic text corresponding to the transaction characteristics that more conform to the type of transaction behavior is arranged more forward. When generating the target transaction message, the characteristic text arranged forward can be preferentially selected, so that the target transaction message can better conform to the characteristics of the type of transaction behavior. It can be seen that the whole process of generating the message does not require manual participation, and the automatic generation of the message can be realized, improving the generation efficiency of the message. At the same time, it should be understood that for the automatically generated target transaction message, it can be manually corrected to obtain a corrected transaction message, and through this corrected transaction message, the characteristic evaluation matrix can be adjusted to make the parameters (evaluation values) in the characteristic evaluation matrix more conform to the manual requirements for the behavior type characteristics and the characteristic text, thereby ensuring the quality of the generated message.

[0129] Further, please refer to Figure 5 , Figure 5 which is a schematic flowchart of a process for adjusting a characteristic evaluation matrix provided by an embodiment of this application. As Figure 5 shown, this process may include:

[0130] Step S201, in response to a selection operation for behavior type characteristics, obtain the corrected transaction characteristics corresponding to the asset transaction data.

[0131] In this application, the behavior type characteristics may be one or more. The one or more behavior type characteristics may be characteristics defined manually for the type of transaction behavior, and the type of transaction behavior may refer to the type of transaction behavior to which the asset transaction data belongs.

[0132] In this application, the user terminal may display the target transaction message for the asset transaction data on the terminal display interface, and the auditing user may correct the target transaction message. For example, when the user terminal displays the target transaction message, it may display all the behavior type characteristics of the type of transaction behavior (and mark the behavior type characteristics that are the transaction characteristics of the asset transaction data), and the auditing user may select these all behavior type characteristics to obtain the corrected transaction characteristics of the asset transaction data selected by the auditing user, and the user terminal may obtain the corrected transaction characteristics selected by the auditing user.

[0133] Step S202, sort the corrected characteristic text according to the characteristic evaluation matrix, and update the target transaction message corresponding to the asset transaction data to a corrected transaction message according to the sorting result; the corrected characteristic text is the text that matches the corrected transaction characteristics.

[0134] In this application, for the specific implementation of sorting the corrected feature text according to the feature evaluation matrix, reference can be made to the above Figure 3 description of sorting the feature text according to the feature evaluation matrix in the corresponding embodiment, which will not be elaborated here.

[0135] Further, the target transaction message can be updated according to the sorting result to obtain a corrected transaction message. For example, taking the behavior type features including behavior type feature 1, behavior type feature 2, and behavior type feature 3, and the feature evaluation matrix being the above feature evaluation matrix C1 as an example, the target transaction message consists of the feature text "abnormal transaction object" corresponding to behavior type feature 1 and the feature text "abnormal transaction location" corresponding to behavior type feature 3, that is, the target transaction message is "abnormal transaction object, abnormal transaction location"; and the corrected transaction features selected by the auditing user are behavior type feature 1, behavior type feature 2, and behavior type feature 3 (that is, compared with the transaction features, the corrected transaction features selected by the auditing user also include behavior type feature 2), then through the feature evaluation matrix C1, it can be determined that the feature text "all transaction objects are foreigners" of this behavior type feature 2 should be ranked first (in descending order), so the corrected transaction message can be "all transaction objects are foreigners, abnormal transaction object, abnormal transaction location".

[0136] Step S203: Adjust the parameters of the feature evaluation matrix in the feature evaluation model according to the corrected transaction message to obtain an adjusted feature evaluation matrix; the adjusted feature evaluation matrix is used to sort the feature text corresponding to the next asset transaction data.

[0137] In this application, the specific method for adjusting the parameters of the feature evaluation matrix in the feature evaluation model according to the corrected transaction message can be to set the transaction feature as a valid value and set the first remaining feature as an invalid value, so as to generate a transaction feature matrix including the valid value corresponding to the transaction feature and the invalid value corresponding to the first remaining feature; where the first remaining feature can be the feature other than the transaction feature in the behavior type features; similarly, the corrected transaction feature can be set as a valid value and the second remaining feature can be set as an invalid value, so as to generate a corrected feature matrix including the valid value corresponding to the corrected transaction feature and the invalid value corresponding to the second remaining feature; where the second remaining feature can be the feature other than the corrected transaction feature in the behavior type features; then, according to the corrected feature matrix and the transaction feature matrix, the parameters of the feature evaluation matrix are adjusted to obtain an adjusted feature evaluation matrix.

[0138] Among them, the specific method for adjusting the parameters of the feature evaluation matrix according to the corrected feature matrix and the transaction feature matrix can be to perform a subtraction operation on the corrected feature matrix and the transaction feature matrix to obtain the adjusted feature matrix corresponding to the asset transaction data; subsequently, an addition operation can be performed on the adjusted feature matrix and the feature evaluation matrix, and the matrix obtained from the addition operation is used as the adjusted feature evaluation matrix.

[0139] The specific method for adjusting the feature evaluation matrix can be as shown in formula (2):

[0140]

[0141] Among them, X s can be used to represent the feature evaluation matrix, P i can be used to represent the adjusted feature evaluation matrix corresponding to the asset transaction data i, can be used to represent the value obtained by summing the adjusted feature evaluation matrices corresponding to n asset transaction data; X sn can be used to represent the adjusted feature evaluation matrix.

[0142] It should be understood that within a period of time (for example, one hour, 30 minutes, 20 hours, 3 days, etc.), the asset transaction data of a certain transaction behavior type may include n, then the adjusted evaluation matrix corresponding to each asset transaction data can be determined, and the adjusted evaluation matrices of these n asset transaction data are summed, so that the total adjusted evaluation matrix can be obtained; further, the total adjusted evaluation matrix can be added to the feature evaluation matrix X s to obtain the adjusted feature evaluation matrix X sn .

[0143] For example, taking n in formula (2) as 1, the behavior type features include behavior type feature 1, behavior type feature 2, and behavior type feature 3, the transaction feature is behavior type feature 1 and behavior type feature 3, the feature evaluation matrix is the above feature evaluation matrix C1, and the transaction feature matrix is the above transaction feature matrix C2 as an example. If the corrected transaction feature is behavior type feature 1, behavior type feature 2, and behavior type feature 3, then the behavior type feature 1, behavior type feature 2, and behavior type feature 3 can be set as valid values. For example, the behavior type feature 1, behavior type feature 2, and behavior type feature 3 are all set to 1; thus, a corrected feature matrix containing the valid values (for example, 1) corresponding to behavior type feature 1, behavior type feature 2, and behavior type feature 3 can be obtained, and the corrected feature matrix can be as shown in the corrected feature matrix C4:

[0144]

[0145] Further, the transaction feature matrix C2 and the corrected feature matrix C4 can be subjected to a subtraction operation to obtain an adjusted feature matrix C5 corresponding to the asset transaction data. The adjusted feature matrix C5 can be shown as follows:

[0146]

[0147] Further, the feature evaluation matrix C1 and the adjusted feature evaluation matrix C5 can be added to obtain an adjusted feature evaluation matrix C6. The adjusted feature evaluation matrix C6 can be shown as follows:

[0148]

[0149] It can be seen that in the adjusted feature evaluation matrix C6, the feature evaluation value of the behavior type feature 2 selected by the auditing user is adjusted and increased. Therefore, when generating a message, the feature text corresponding to the behavior type feature with a high feature evaluation value can be preferentially selected. That is, the probability of selecting the feature text of the behavior type feature 2 will be higher, and the target transaction message generated through the feature evaluation matrix will be more in line with the manual requirements for the behavior type feature and the corresponding feature text, that is, the quality of the generated target transaction message will be higher.

[0150] In this application, by responding to the user's selection operation of the behavior type feature, the corrected transaction feature selected by the user can be obtained. According to the corrected transaction feature, a corrected transaction message can be generated, and the feature evaluation matrix can be adjusted through the corrected transaction message, so that the feature evaluation matrix can learn the features selected manually and make it more in line with the manual requirements, thereby ensuring the quality of the automatically generated target transaction message.

[0151] Further, please refer to Figure 6 , Figure 6 which is a system framework diagram provided by an embodiment of this application. As shown in Figure 6 , the system may include a message output module, a feature evaluation model, and a message generation application. The functions of the message output module, the feature evaluation model, and the message generation application will be described below:

[0152] Message generation application. The message generation application can be used for the user to input asset transaction data and can also be used for the user to select the transaction behavior type to which the asset transaction data belongs.

[0153] Message Output Module. The Message Output Module can be used to receive the asset transaction data and the type of transaction behavior of the message generation application, obtain the set of behavior type features corresponding to the type of transaction behavior, and determine the transaction features of the asset transaction data according to the set of behavior type features. At the same time, the Message Output Module can obtain the feature evaluation matrix from the Feature Evaluation Model. In the Message Output Module, the feature text corresponding to the transaction features can be sorted through the feature evaluation matrix, so as to generate the target transaction message of the asset transaction data. The Message Output Module can feedback the target transaction message to the message generation application and receive the corrected transaction message (i.e., the corrected transaction message after manual correction) by the user through the message generation application for the target transaction message.

[0154] Feature Evaluation Model. The Feature Evaluation Model can be used to provide the feature evaluation matrix of any type of transaction behavior to the Message Output Module. At the same time, the Feature Evaluation Model can receive the corrected transaction message feedback by the Message Output Module and adjust the feature evaluation matrix according to the corrected transaction message.

[0155] It should be understood that the above Message Output Module, message generation application, and Feature Evaluation Model can all be integrated into the user terminal or the business server. For the specific implementation methods of the functions of the Message Output Module, message generation application, and Feature Evaluation Model (for example, the specific implementation method of generating the target transaction message of the asset transaction data, or the specific implementation method of correcting the target transaction message to obtain the corrected transaction message and adjusting the feature evaluation matrix through the corrected transaction message), reference can be made to the description in the corresponding embodiments above, which will not be elaborated here. Figure 4 - Figure 5 The description in the corresponding embodiments above will not be elaborated here.

[0156] Further, please refer to Figure 7 , Figure 7 which is a logic flow chart for generating a target transaction message provided by an embodiment of the present application. As Figure 7 shown, the logic flow may include the following steps:

[0157] Step S301, obtain the feature evaluation matrix.

[0158] In the present application, after obtaining the asset transaction data and the type of transaction behavior to which the asset transaction data belongs, the feature evaluation matrix corresponding to the type of transaction behavior can be obtained.

[0159] Step S302, sort the feature text according to the feature evaluation matrix.

[0160] In the present application, the feature text of the transaction features corresponding to the asset transaction data can be sorted according to the magnitude of the feature evaluation values in the feature evaluation matrix (it can be sorted in descending order or in ascending order).

[0161] Step S303: Sort the feature texts with the same feature evaluation values in the original specified order.

[0162] In this application, if there are feature texts with the same feature evaluation values, they can be sorted and concatenated in the original specified order (which can be a manually specified order) of the feature texts.

[0163] Step S304: Determine whether the number of words is greater than or equal to the maximum word count.

[0164] In this application, after sorting the feature texts, the first P feature texts can be selected in order and concatenated in order to generate a target transaction message. Here, the number of words refers to the text word count of the first P feature texts. If the text word count of the first P feature texts is less than the maximum word count, step S305 can be entered; if the text word count of the first P feature texts is greater than or equal to the maximum word count, step S309 can be entered, and the feature text obtained by concatenating the first P feature texts is used as the target transaction message.

[0165] Step S305: Determine whether all the feature texts have been added.

[0166] In this application, if the text word count of the above-mentioned first P feature texts is less than the maximum word count, it can be determined whether all the feature texts of the transaction features have been added at this time. If there are P transaction features, there are also P feature texts in total. Then, all the feature texts have been added at this time, and the concatenated text composed of the first P feature texts can be determined as the target transaction message (that is, step S309 can be entered). If the feature text has not been added yet, step S306 can be entered.

[0167] Step S306: Add the feature texts corresponding to the transaction features in order.

[0168] In this application, if the feature text has not been added yet, the (P + 1)-th feature text can be obtained and added after the P-th feature text.

[0169] Step S307: Determine whether the number of words is greater than the maximum word count.

[0170] In this application, the number of words (i.e., the text word count) of the first (P + 1) feature texts can be obtained. If the number of words of the first (P + 1) feature texts is greater than the maximum word count, step S308 can be entered; if the number of words of the first (P + 1) feature texts is less than or equal to the maximum word count, step S305 can be entered.

[0171] Step S308: Remove the last added feature text.

[0172] In this application, if the number of words in the first P + 1 feature texts is greater than the maximum word count, it can be indicated that after adding the (P + 1)-th feature text, the text word count no longer meets the limit condition of the maximum word count. Then, the (P + 1)-th feature text can be deleted.

[0173] Step S309: Generate a target transaction message.

[0174] Among them, for the specific implementation manners of steps S301 - S309, reference can be made to the descriptions of steps S101 - S103 for generating a target transaction message in the corresponding embodiments above. Details will not be elaborated here. Figure 4 The descriptions of steps S101 - S103 for generating a target transaction message in the corresponding embodiments above. Details will not be elaborated here.

[0175] In this application, by obtaining asset transaction data and the type of transaction behavior to which the asset transaction data belongs, the transaction characteristics of the asset transaction data can be determined according to the set of behavior type characteristics corresponding to the type of transaction behavior. And according to the characteristic evaluation matrix corresponding to the type of transaction behavior, the feature texts corresponding to the transaction characteristics can be sorted, so that the feature texts corresponding to the transaction characteristics that more conform to the type of transaction behavior are arranged more forward. When generating a target transaction message, the feature texts arranged forward can be preferentially selected, so that the target transaction message can better conform to the characteristics of the type of transaction behavior. It can be seen that the entire process of generating the message does not require manual participation, and automated message generation can be realized, improving the message generation efficiency.

[0176] Further, please refer to Figure 8 , Figure 8 which is a schematic structural diagram of a data processing device provided by an embodiment of this application. The data processing device can be a computer program (including program code) running in a computer device. For example, the data processing device is an application software. The data processing device can be used to execute Figure 4 or Figure 5 the method shown. As Figure 8 shown, the data processing device 1 can include: a data acquisition module 11, a feature determination module 12, and a message generation module 13.

[0177] The data acquisition module 11 is used to obtain asset transaction data and the type of transaction behavior to which the asset transaction data belongs;

[0178] The feature determination module 12 is used to determine the transaction characteristics corresponding to the asset transaction data according to the asset transaction data and the type of transaction behavior;

[0179] The message generation module 13 is configured to obtain a feature evaluation matrix corresponding to the transaction behavior type, sort the feature text according to the feature evaluation matrix, and obtain a target transaction message corresponding to the asset transaction data; the feature text is text that matches the transaction feature; the target transaction message is used to identify the transaction legality of the asset transaction data.

[0180] Among them, for the specific implementation manners of the data acquisition module 11, the feature determination module 12, and the message generation module 13, reference can be made to the descriptions of steps S101 - S103 in the corresponding embodiments above, which will not be elaborated here. Figure 4 Please refer to

[0181] Please refer to Figure 8 , the feature determination module 12 may include: a set acquisition unit 121, a data matching unit 122, and a feature determination unit 123.

[0182] The set acquisition unit 121 is configured to obtain a behavior type feature set corresponding to the transaction behavior type;

[0183] The data matching unit 122 is configured to match the asset transaction data with the behavior type feature set;

[0184] The feature determination unit 123 is configured to, if there is a behavior type feature in the behavior type feature set that matches the asset transaction data, use the behavior type feature that matches the asset transaction data as the transaction feature corresponding to the asset transaction data.

[0185] Among them, for the specific implementation manners of the set acquisition unit 121, the data matching unit 122, and the feature determination unit 123, reference can be made to the description in step S102 of the corresponding embodiments above, which will not be elaborated here. Figure 4 Please refer to

[0186] Among them, the feature evaluation matrix is the matrix included in the feature evaluation model; the number of transaction features is N, and N is a positive integer;

[0187] Please refer to Figure 8 , the message generation module 13 may include: a matrix acquisition unit 131, an evaluation value determination unit 132, a sequence generation unit 133, and a message generation unit 134.

[0188] The matrix acquisition unit 131 is configured to obtain a feature evaluation matrix corresponding to the transaction behavior type through the feature evaluation model; the feature evaluation matrix includes feature evaluation values corresponding to one or more behavior type features respectively; the feature evaluation value is used to characterize the matching degree between the behavior type feature and the transaction behavior type;

[0189] The evaluation value determination unit 132 is configured to obtain the feature evaluation values corresponding to N transaction features in the feature evaluation matrix as N target feature evaluation values;

[0190] The sequence generation unit 133 is configured to generate a feature evaluation value sequence including the N target feature evaluation values according to the numerical magnitudes corresponding to the N target feature evaluation values;

[0191] The text sorting unit 134 is configured to sort the feature texts corresponding to the N transaction features according to the feature evaluation value sequence;

[0192] The message generation unit 135 is configured to generate a target transaction message corresponding to the asset transaction data according to the sorted N feature texts.

[0193] Among them, for the specific implementation manners of the matrix acquisition unit 131, the evaluation value determination unit 132, the sequence generation unit 133, the text sorting unit 134, and the message generation unit 135, reference may be made to the description in step S103 of the corresponding embodiment above, which will not be elaborated here. Figure 4 For details, please refer to

[0194] Please refer to Figure 8 , the message generation unit 135 may include: a rule acquisition subunit 1351, a text determination subunit 1352, and a text splicing subunit 1353.

[0195] The rule acquisition subunit 1351 is configured to acquire a message generation rule; the message generation rule is used to specify the maximum number of words included in the target transaction message;

[0196] The text determination subunit 1352 is configured to determine target feature texts from the sorted N feature texts according to the maximum number of words;

[0197] The text splicing subunit 1353 is configured to splice the target feature texts in the sorting order of the sorted N feature texts to obtain a target transaction message corresponding to the asset transaction data; the number of words in the target transaction message is less than or equal to the maximum number of words.

[0198] Among them, for the specific implementation manners of the rule acquisition subunit 1351, the text determination subunit 1352, and the text splicing subunit 1353, reference may be made to the description in step S103 of the corresponding embodiment above, which will not be elaborated here. Figure 4 For details, please refer to

[0199] Among them, the text determination subunit is further configured to select P feature texts from the sorted N feature texts; P is a positive integer less than N;

[0200] The text determination subunit 1352 is further configured to obtain the text word count of P feature texts and match the text word count with the maximum word count;

[0201] The text determination subunit 1352 is further configured to, if the text word count is less than the maximum word count, determine candidate feature texts from the remaining feature texts according to the sorting order among the sorted N feature texts; the candidate feature texts are the feature texts other than the P feature texts among the N feature texts;

[0202] The text determination subunit 1352 is further configured to determine target feature texts according to the candidate feature texts and the P feature texts;

[0203] The text determination subunit 1352 is further configured to, if the text word count is greater than or equal to the maximum word count, determine the P feature texts as the target feature texts.

[0204] Among them, the text determination subunit 1352 is further configured to determine the word count difference between the maximum word count and the text word count;

[0205] The text determination subunit 1352 is further configured to obtain the candidate text word count corresponding to the candidate feature texts and match the candidate text word count with the word count difference;

[0206] The text determination subunit 1352 is further configured to, if the candidate text word count is less than or equal to the word count difference, determine the candidate feature texts and the P feature texts as the target feature texts;

[0207] The text determination subunit 1352 is further configured to, if the candidate text word count is greater than the word count difference, determine the P feature texts as the target feature texts.

[0208] Please refer to Figure 8 that the data processing device 1 may further include: a correction feature acquisition module 14, a message update module 15, and a parameter adjustment module 16.

[0209] The correction feature acquisition module 14 is configured to obtain the corrected transaction features corresponding to the asset transaction data in response to a selection operation for the behavior type features;

[0210] The message update module 15 is configured to sort the corrected feature texts according to the feature evaluation matrix, and update the target transaction message corresponding to the asset transaction data to a corrected transaction message according to the sorting result; the corrected feature texts are the texts matching the corrected transaction features;

[0211] The parameter adjustment module 16 is configured to adjust the parameters of the feature evaluation matrix in the feature evaluation model according to the corrected transaction message to obtain an adjusted feature evaluation matrix; the adjusted feature evaluation matrix is used to sort the feature texts corresponding to the next asset transaction data.

[0212] Among them, for the specific implementation manners of the correction feature acquisition module 14, the message update module 15, and the parameter adjustment module 16, reference may be made to the descriptions of steps S201 - S203 in the corresponding embodiments above, and details will not be elaborated here. Figure 5

[0213] Please refer to Figure 8 , the parameter adjustment module 16 may include: an assignment unit 161 and a matrix adjustment unit 162.

[0214] The assignment unit 161 is configured to set the transaction feature to a valid value, set the first remaining feature to an invalid value, and generate a transaction feature matrix including the valid value corresponding to the transaction feature and the invalid value corresponding to the first remaining feature; the first remaining feature is the feature other than the transaction feature in the behavior type feature;

[0215] The assignment unit 161 is further configured to set the corrected transaction feature to a valid value, set the second remaining feature to an invalid value, and generate a corrected feature matrix including the valid value corresponding to the corrected transaction feature and the invalid value corresponding to the second remaining feature; the second remaining feature is the feature other than the corrected transaction feature in the behavior type feature;

[0216] The matrix adjustment unit 162 is configured to perform parameter adjustment on the feature evaluation matrix according to the corrected feature matrix and the transaction feature matrix to obtain an adjusted feature evaluation matrix.

[0217] Among them, for the specific implementation manners of the assignment unit 161 and the matrix adjustment unit 162, reference may be made to the description of step S203 in the corresponding embodiments above, and details will not be elaborated here. Figure 5

[0218] Please refer to Figure 8 , the matrix adjustment unit 162 may include: an operator unit 1621 and a matrix determination unit 1622.

[0219] The operator unit 1621 is configured to perform a subtraction operation on the corrected feature matrix and the transaction feature matrix to obtain an adjusted feature matrix corresponding to the asset transaction data;

[0220] The matrix determination unit 1622 is configured to perform an addition operation on the adjusted feature matrix and the feature evaluation matrix, and use the matrix obtained from the addition operation as the adjusted feature evaluation matrix.

[0221] Among them, for the specific implementation manners of the operator unit 1621 and the matrix determination unit 1622, reference may be made to the description of step S203 in the corresponding embodiments above, and details will not be elaborated here. Figure 5

[0222] Please refer toFigure 8 , the data processing device 1 may further include: a suspicious text acquisition module 17, a text matching module 18, and a data determination module 19.

[0223] The suspicious text acquisition module 17 is configured to acquire suspicious feature text in a target transaction message; the suspicious feature text is text that matches suspicious transaction features; the transaction features include suspicious transaction features;

[0224] The text matching module 18 is configured to match the suspicious feature text with legal feature text;

[0225] The data determination module 19 is configured to determine that the asset transaction data is legal asset transaction data if the suspicious feature text meets the legal conditions indicated by the legal feature text;

[0226] The data determination module 19 is further configured to determine that the asset transaction data is illegal asset transaction data if the suspicious feature text does not meet the legal conditions indicated by the legal feature text.

[0227] Among them, for the specific implementation manners of the suspicious text acquisition module 17, the text matching module 18, and the data determination module 19, reference may be made to the description in step S103 in the corresponding embodiment above, which will not be elaborated here. Figure 4 The description in step S103 in the corresponding embodiment above, which will not be elaborated here.

[0228] In this application, by acquiring asset transaction data and the transaction behavior type to which the asset transaction data belongs, the transaction features of the asset transaction data can be determined according to the set of behavior type features corresponding to the transaction behavior type, and according to the feature evaluation matrix corresponding to the transaction behavior type, the feature text corresponding to the transaction features can be sorted, so that the feature text corresponding to the transaction features that more conform to the transaction behavior type is arranged more forward. When generating a target transaction message, the feature text arranged forward can be preferentially selected, so that the target transaction message can better conform to the features of the transaction behavior type. It can be seen that the entire process of generating the message does not require manual participation, and automatic message generation can be realized, improving the message generation efficiency. At the same time, it should be understood that for the automatically generated target transaction message, manual correction can be performed to obtain a corrected transaction message, and through the corrected transaction message, the feature evaluation matrix can be adjusted to make the parameters (evaluation values) in the feature evaluation matrix more conform to the needs of manual for behavior type features and feature text, thereby ensuring the quality of the generated message.

[0229] Further, please refer to Figure 9 , Figure 9 is a schematic structural diagram of a computer device provided by an embodiment of the present application. As Figure 9 shown, the above Figure 8The device 1 in the corresponding embodiment can be applied to the above computer device 1000, which may include: a processor 1001, a network interface 1004, and a memory 1005. In addition, the computer device 1000 further includes: a user interface 1003 and at least one communication bus 1002. Among them, the communication bus 1002 is used to realize the connection and communication between these components. Among them, the user interface 1003 may include a display screen (Display) and a keyboard (Keyboard). Optionally, the user interface 1003 may further include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface). The memory 1005 may be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. Optionally, the memory 1005 may further be at least one storage device located far from the aforementioned processor 1001. As Figure 9 shown, the memory 1005, as a computer-readable storage medium, may include an operating system, a network communication module, a user interface module, and a device control application program.

[0230] In Figure 9 the computer device 1000 shown, the network interface 1004 can provide network communication functions; while the user interface 1003 is mainly used to provide an input interface for users; and the processor 1001 can be used to call the device control application program stored in the memory 1005 to achieve:

[0231] Obtain asset transaction data and the type of transaction behavior to which the asset transaction data belongs;

[0232] Determine the transaction characteristics corresponding to the asset transaction data according to the asset transaction data and the type of transaction behavior;

[0233] Obtain the feature evaluation matrix corresponding to the type of transaction behavior, sort the feature text according to the feature evaluation matrix, and obtain the target transaction message corresponding to the asset transaction data; the feature text is the text that matches the transaction characteristics; the target transaction message is used to identify the transaction legality of the asset transaction data.

[0234] In one embodiment, when the processor 1001 executes to determine the transaction characteristics corresponding to the asset transaction data according to the asset transaction data and the type of transaction behavior, the following steps are specifically executed:

[0235] Obtain the set of behavior type characteristics corresponding to the type of transaction behavior;

[0236] Match the asset transaction data with the set of behavior type characteristics;

[0237] If there is a behavior type feature in the behavior type feature set that matches the asset transaction data, then use the behavior type feature that matches the asset transaction data as the transaction feature corresponding to the asset transaction data.

[0238] Among them, the feature evaluation matrix is the matrix included in the feature evaluation model; the number of transaction features is N, and N is a positive integer.

[0239] In one embodiment, when the processor 1001 executes to obtain the feature evaluation matrix corresponding to the transaction behavior type, sort the feature text according to the feature evaluation matrix, and obtain the target transaction message corresponding to the asset transaction data, the following steps are specifically executed:

[0240] Through the feature evaluation model, obtain the feature evaluation matrix corresponding to the transaction behavior type; the feature evaluation matrix includes feature evaluation values corresponding to one or more behavior type features respectively; the feature evaluation value is used to characterize the matching degree between the behavior type feature and the transaction behavior type.

[0241] In the feature evaluation matrix, obtain the feature evaluation values corresponding to the N transaction features respectively as N target feature evaluation values.

[0242] According to the numerical sizes corresponding to the N target feature evaluation values respectively, generate a feature evaluation value sequence including the N target feature evaluation values.

[0243] According to the feature evaluation value sequence, sort the feature texts corresponding to the N transaction features respectively, and generate the target transaction message corresponding to the asset transaction data according to the sorted N feature texts.

[0244] In one embodiment, when the processor 1001 executes to sort the feature texts corresponding to the N transaction features respectively according to the feature evaluation value sequence, and generate the target transaction message corresponding to the asset transaction data according to the sorted N feature texts, the following steps are specifically executed:

[0245] Obtain the message generation rule; the message generation rule is used to specify the maximum number of words included in the target transaction message.

[0246] According to the maximum number of words, determine the target feature text among the sorted N feature texts.

[0247] According to the sorting order among the sorted N feature texts, splice the target feature texts to obtain the target transaction message corresponding to the asset transaction data; the number of words included in the target transaction message is less than or equal to the maximum number of words.

[0248] In one embodiment, when the processor 1001 executes to determine the target feature text among the sorted N feature texts according to the maximum number of words, the following steps are specifically executed:

[0249] Select P feature texts from the sorted N feature texts; P is a positive integer less than N;

[0250] Obtain the text word count of the P feature texts, and match the text word count with the maximum word count;

[0251] If the text word count is less than the maximum word count, determine candidate feature texts from the remaining feature texts according to the sorting order among the sorted N feature texts; the candidate feature texts are the feature texts among the N feature texts except the P feature texts;

[0252] Determine the target feature texts according to the candidate feature texts and the P feature texts;

[0253] If the text word count is greater than or equal to the maximum word count, determine the P feature texts as the target feature texts.

[0254] In one embodiment, when the processor 1001 executes to determine the target feature texts according to the candidate feature texts and the P feature texts, it specifically performs the following steps:

[0255] Determine the word count difference between the maximum word count and the text word count;

[0256] Obtain the candidate text word count corresponding to the candidate feature texts, and match the candidate text word count with the word count difference;

[0257] If the candidate text word count is less than or equal to the word count difference, determine the candidate feature texts and the P feature texts as the target feature texts;

[0258] If the candidate text word count is greater than the word count difference, determine the P feature texts as the target feature texts.

[0259] In one embodiment, the processor 1001 also specifically performs the following steps:

[0260] Respond to the selection operation for the behavior type feature, and obtain the corrected transaction feature corresponding to the asset transaction data;

[0261] Sort the corrected feature texts according to the feature evaluation matrix, and update the target transaction message corresponding to the asset transaction data to the corrected transaction message according to the sorting result; the corrected feature texts are the texts matching the corrected transaction feature;

[0262] Adjust the parameters of the feature evaluation matrix in the feature evaluation model according to the corrected transaction message to obtain the adjusted feature evaluation matrix; the adjusted feature evaluation matrix is used to sort the feature texts corresponding to the next asset transaction data.

[0263] In one embodiment, when the processor adjusts the parameters of the feature evaluation matrix in the feature evaluation model according to the corrected transaction message to obtain the adjusted feature evaluation matrix, the following steps are specifically executed:

[0264] Set the transaction feature to a valid value, set the first remaining feature to an invalid value, and generate a transaction feature matrix including the valid value corresponding to the transaction feature and the invalid value corresponding to the first remaining feature; the first remaining feature is the feature other than the transaction feature in the behavior type feature;

[0265] Set the corrected transaction feature to a valid value, set the second remaining feature to an invalid value, and generate a corrected feature matrix including the valid value corresponding to the corrected transaction feature and the invalid value corresponding to the second remaining feature; the second remaining feature is the feature other than the corrected transaction feature in the behavior type feature;

[0266] According to the corrected feature matrix and the transaction feature matrix, adjust the parameters of the feature evaluation matrix to obtain the adjusted feature evaluation matrix.

[0267] In one embodiment, when the processor 1001 adjusts the parameters of the feature evaluation matrix according to the corrected feature matrix and the transaction feature matrix to obtain the adjusted feature evaluation matrix, the following steps are specifically executed:

[0268] Perform a subtraction operation on the corrected feature matrix and the transaction feature matrix to obtain an adjusted feature matrix corresponding to the asset transaction data;

[0269] Perform an addition operation on the adjusted feature matrix and the feature evaluation matrix, and use the matrix obtained from the addition operation as the adjusted feature evaluation matrix.

[0270] In one embodiment, the processor 1001 also specifically executes the following steps:

[0271] Obtain the suspicious feature text in the target transaction message; the suspicious feature text is the text that matches the suspicious transaction feature; the transaction feature includes the suspicious transaction feature;

[0272] Match the suspicious feature text with the legal feature text;

[0273] If the suspicious feature text meets the legal conditions indicated by the legal feature text, determine that the asset transaction data is legal asset transaction data;

[0274] If the suspicious feature text does not meet the legal conditions indicated by the legal feature text, determine that the asset transaction data is illegal asset transaction data.

[0275] It should be understood that the computer device 1000 described in the embodiments of the present application can execute the foregoing Figure 4 or Figure 5For the description of the data processing method in the corresponding embodiment, the foregoing can also be executed. Figure 8 For the description of the data processing apparatus 1 in the corresponding embodiment, it will not be repeated here. In addition, the description of the beneficial effects of adopting the same method will not be repeated either.

[0276] In addition, it should be noted here that: The embodiment of the present application also provides a computer-readable storage medium, and the computer-readable storage medium stores a computer program executed by the aforementioned computer device 1000 for data processing. The computer program includes program instructions. When the aforementioned processor executes the program instructions, it can execute the foregoing Figure 4 or Figure 5 For the description of the data processing method in the corresponding embodiment, therefore, it will not be repeated here. In addition, the description of the beneficial effects of adopting the same method will not be repeated either. For the technical details not disclosed in the embodiment of the computer-readable storage medium involved in the present application, please refer to the description of the method embodiment of the present application.

[0277] The aforementioned computer-readable storage medium may be the data processing apparatus provided in any of the foregoing embodiments or an internal storage unit of the aforementioned computer device, such as the hard disk or memory of the computer device. The computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk equipped on the computer device, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. Further, the computer-readable storage medium may also include both the internal storage unit and the external storage device of the computer device. The computer-readable storage medium is used to store the computer program and other programs and data required by the computer device. The computer-readable storage medium may also be used to temporarily store the data that has been output or will be output.

[0278] In one aspect of the present application, there is provided a computer program product or a computer program. The computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the method provided in one aspect of the embodiment of the present application.

[0279] In the description, claims, and drawings of the embodiments of the present application, the terms "first", "second", etc. are used to distinguish different objects, rather than to describe a specific order. In addition, the term "comprising" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, device, product, or equipment that includes a series of steps or units is not limited to the listed steps or modules, but may optionally further include steps or modules not listed, or may optionally further include other step units inherent to these processes, methods, devices, products, or equipment.

[0280] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of each example have been generally described according to their functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. A professional technician can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0281] The methods and related devices provided by the embodiments of the present application are described with reference to the method flowcharts and / or structural schematic diagrams provided by the embodiments of the present application. Specifically, each process and / or block of the method flowchart and / or structural schematic diagram, as well as the combination of the 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 the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the functions specified in Figure 1 one process or multiple processes and / or structural schematic Figure 1 one block or multiple blocks. These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured product including an instruction device that implements the functions specified in Figure 1 one process or multiple processes and / or structural schematic Figure 1 one block or multiple blocks. These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one process or multiple processes and / or structural schematic one block or multiple blocks.

[0282] The above-disclosed are only the preferred embodiments of the present application. Of course, the scope of rights of the present application cannot be limited thereby. Therefore, equivalent changes made according to the claims of the present application still fall within the scope covered by the present application.

Claims

1. A data processing method, characterized in that, Including: Obtain asset transaction data and the type of transaction behavior to which the asset transaction data belongs; Obtain the set of behavior type features corresponding to the type of transaction behavior; Match the asset transaction data with the set of behavior type features; if there is a behavior type feature in the set of behavior type features that matches the asset transaction data, then use the behavior type feature that matches the asset transaction data as the transaction feature corresponding to the asset transaction data; Obtain the feature evaluation matrix corresponding to the type of transaction behavior, sort the feature text according to the feature evaluation matrix, and obtain the target transaction message corresponding to the asset transaction data; the feature text is the text that matches the transaction feature; the target transaction message is used to identify the transaction legality of the asset transaction data; the feature evaluation matrix includes feature evaluation values corresponding to one or more behavior type features respectively; one feature evaluation value is used to characterize the matching degree of one behavior type feature and the type of transaction behavior; the sorting process of the feature text depends on the respective feature evaluation values included in the feature evaluation matrix; In response to a selection operation for the behavior type feature, obtain the corrected transaction feature corresponding to the asset transaction data; Sort the corrected feature text according to the feature evaluation matrix, and update the target transaction message corresponding to the asset transaction data to a corrected transaction message according to the sorting result; the corrected feature text is the text that matches the corrected transaction feature; Adjust the parameters of the feature evaluation matrix in the feature evaluation model according to the corrected transaction message, and obtain an adjusted feature evaluation matrix; the adjusted feature evaluation matrix is used to sort the feature text corresponding to the next asset transaction data.

2. The method according to claim 1, wherein The feature evaluation matrix is the matrix included in the feature evaluation model; the number of transaction features is N, and N is a positive integer; The obtaining the feature evaluation matrix corresponding to the type of transaction behavior, sorting the feature text according to the feature evaluation matrix, and obtaining the target transaction message corresponding to the asset transaction data includes: Through the feature evaluation model, obtain the feature evaluation matrix corresponding to the type of transaction behavior; In the feature evaluation matrix, obtain the feature evaluation values corresponding to the N transaction features respectively as N target feature evaluation values; Generate a feature evaluation value sequence including the N target feature evaluation values according to the numerical magnitudes corresponding to the N target feature evaluation values respectively; Sort the feature texts corresponding to the N transaction features according to the feature evaluation value sequence, and generate the target transaction message corresponding to the asset transaction data according to the sorted N feature texts.

3. The method according to claim 2, characterized in that, The generating the target transaction message corresponding to the asset transaction data according to the sorted N feature texts includes: Obtain a message generation rule; the message generation rule is used to specify the maximum number of words included in the target transaction message; Determine the target feature text among the sorted N feature texts according to the maximum number of words; Concatenate the target feature texts according to the sorting order among the N sorted feature texts to obtain the target transaction message corresponding to the asset transaction data; the number of message words included in the target transaction message is less than or equal to the maximum number of words.

4. The method according to claim 3, wherein The determining the target feature texts from the N sorted feature texts according to the maximum number of words includes: Select P feature texts from the N sorted feature texts; P is a positive integer less than N; Obtain the text word amounts of the P feature texts, and match the text word amounts with the maximum number of words; If the text word amount is less than the maximum number of words, determine candidate feature texts from the remaining feature texts according to the sorting order among the N sorted feature texts; the candidate feature texts are the feature texts among the N feature texts other than the P feature texts; Determine the target feature texts according to the candidate feature texts and the P feature texts; If the text word amount is greater than or equal to the maximum number of words, determine the P feature texts as the target feature texts.

5. The method according to claim 4, wherein The determining the target feature texts according to the candidate feature texts and the P feature texts includes: Determine the difference in the number of words between the maximum number of words and the text word amount; Obtain the candidate text word amount corresponding to the candidate feature texts, and match the candidate text word amount with the difference in the number of words; If the candidate text word amount is less than or equal to the difference in the number of words, determine the candidate feature texts and the P feature texts as the target feature texts; If the candidate text word amount is greater than the difference in the number of words, determine the P feature texts as the target feature texts.

6. The method according to claim 1, wherein The adjusting the parameter of the feature evaluation matrix in the feature evaluation model according to the corrected transaction message to obtain the adjusted feature evaluation matrix includes: Set the transaction feature to a valid value and set the first remaining feature to an invalid value to generate a transaction feature matrix including the valid value corresponding to the transaction feature and the invalid value corresponding to the first remaining feature; the first remaining feature is the feature other than the transaction feature in the behavior type feature; Set the corrected transaction feature to the valid value and set the second remaining feature to the invalid value to generate a corrected feature matrix including the valid value corresponding to the corrected transaction feature and the invalid value corresponding to the second remaining feature; the second remaining feature is the feature other than the corrected transaction feature in the behavior type feature; Adjust the parameter of the feature evaluation matrix according to the corrected feature matrix and the transaction feature matrix to obtain the adjusted feature evaluation matrix.

7. The method according to claim 6, characterized in that, The adjusting the parameter of the feature evaluation matrix according to the corrected feature matrix and the transaction feature matrix to obtain the adjusted feature evaluation matrix includes: Perform a subtraction operation on the corrected feature matrix and the transaction feature matrix to obtain the adjusted feature matrix corresponding to the asset transaction data; Perform an addition operation on the adjusted feature matrix and the feature evaluation matrix, and use the matrix obtained from the addition operation as the adjusted feature evaluation matrix.

8. The method according to claim 1, characterized in that The method further includes: Obtain the suspicious feature text in the target transaction message; the suspicious feature text is the text that matches the suspicious transaction feature; the transaction feature includes the suspicious transaction feature; Match the suspicious feature text with the legal feature text; If the suspicious feature text meets the legal conditions indicated by the legal feature text, determine that the asset transaction data is legal asset transaction data; If the suspicious feature text does not meet the legal conditions indicated by the legal feature text, determine that the asset transaction data is illegal asset transaction data.

9. A data processing device, characterized in that, It includes: A data acquisition module for acquiring asset transaction data and the transaction behavior type to which the asset transaction data belongs; A feature determination module for obtaining the set of behavior type features corresponding to the transaction behavior type; matching the asset transaction data with the set of behavior type features; if there is a behavior type feature in the set of behavior type features that matches the asset transaction data, use the behavior type feature that matches the asset transaction data as the transaction feature corresponding to the asset transaction data; A message generation module for obtaining the feature evaluation matrix corresponding to the transaction behavior type, sorting the feature text according to the feature evaluation matrix to obtain the target transaction message corresponding to the asset transaction data; the feature text is the text that matches the transaction feature; the target transaction message is used to identify the transaction legality of the asset transaction data; the feature evaluation matrix includes feature evaluation values respectively corresponding to one or more behavior type features; one feature evaluation value is used to represent the matching degree between the behavior type feature and one transaction behavior type; the sorting process of the feature text depends on the respective feature evaluation values included in the feature evaluation matrix; A modified feature acquisition module for obtaining the modified transaction feature corresponding to the asset transaction data in response to a selection operation on the behavior type feature; A message update module for sorting the modified feature text according to the feature evaluation matrix, and updating the target transaction message corresponding to the asset transaction data to a modified transaction message according to the sorting result; the modified feature text is the text that matches the modified transaction feature; A parameter adjustment module for adjusting the parameter of the feature evaluation matrix in the feature evaluation model according to the modified transaction message to obtain an adjusted feature evaluation matrix; the adjusted feature evaluation matrix is used to sort the feature text corresponding to the next asset transaction data.

10. A computer device, characterized in that, It includes: A processor, a memory, and a network interface; The processor is connected to the memory and the network interface, wherein the network interface is used to provide network communication functions, the memory is used to store program codes, and the processor is used to call the program codes to execute the method according to any one of claims 1-8.

11. A computer-readable storage medium, characterized in that, A computer program is stored in the computer-readable storage medium, and the computer program is adapted to be loaded and executed by a processor to perform the method according to any one of claims 1-8.

Images