Evaluation method, device, equipment, readable storage medium and program product

Abstract

The invention provides an evaluation method and device, equipment, a readable storage medium and a program product. The method comprises the steps of obtaining resource transfer related information of a resource transfer operation associated with a to-be-evaluated object within target time; determining resource transfer description data according to the resource transfer related information, the resource transfer description data is used for indicating one or more of a resource transfer time period, a type of a resource transfer object, a resource transfer type, position indication information of the resource transfer object, a resource transfer mode and a resource transfer frequency of a resource transfer operation of the to-be-evaluated object within a target time; determining semantic expression data of the resource transfer description data; and performing evaluation processing according to the semantic expression data to obtain an anomaly evaluation result. Through the method provided by the invention, the abnormal evaluation result can be determined according to the semantic expression data of the semantic dimension, the accuracy of the evaluation result is effectively improved, and the evaluation processing efficiency is also effectively improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to an evaluation method, an evaluation device, a computer equipment, a computer-readable storage medium, and a computer program product. Background Art

[0002] With the advancement of computer technology, online resource transfer technology has found an increasing number of applications. Online resource transfer technology enables resource transfers between a source and destination over the internet. This technology significantly improves resource transfer efficiency. However, in real-world applications, some objects exploit this technology to implement unusual resource transfer operations.

[0003] When determining abnormal resource transfer operations, professional technicians are typically required to analyze the relevant data of resource transfer operations associated with a specific object to obtain an assessment result for that object. This method of determining the assessment result relies heavily on the technical level of the professional technicians, which leads to low assessment efficiency and low accuracy of the assessment results. Summary of the Invention

[0004] The embodiments of the present application provide an evaluation method, apparatus, device, readable storage medium, and program product, which can determine abnormal evaluation results based on semantic expression data of semantic dimensions, effectively improving the accuracy of the evaluation results and the efficiency of the evaluation process.

[0005] In one aspect, an embodiment of the present application provides an evaluation method, the method comprising:

[0006] Obtain resource transfer related information of resource transfer operations associated with the object to be evaluated within the target time;

[0007] Determining resource transfer description data based on the resource transfer related information, the resource transfer description data being used to indicate characteristic information of the resource transfer operation of the object to be evaluated within the target time, the characteristic information including one or more of a time period of the resource transfer, a type of the resource transfer object, a resource transfer type, location indication information of the resource transfer object, a resource transfer method, and a number of resource transfers;

[0008] Determining semantic expression data of the resource transfer description data;

[0009] An evaluation process is performed based on the semantic expression data of the resource transfer description data to obtain an abnormality evaluation result, where the abnormality evaluation result is used to indicate whether there is an abnormality in the resource transfer operation of the object to be evaluated within the target time.

[0010] In one aspect, an embodiment of the present application provides an evaluation device, comprising:

[0011] An acquisition unit, configured to acquire resource transfer related information of a resource transfer operation associated with an object to be evaluated within a target time;

[0012] a determining unit, configured to determine resource transfer description data based on the resource transfer related information, the resource transfer description data being used to indicate characteristic information of the resource transfer operation of the object to be evaluated within the target time, the characteristic information including one or more of a time period of the resource transfer, a type of the resource transfer object, a resource transfer type, location indication information of the resource transfer object, a resource transfer method, and a number of resource transfers;

[0013] The determining unit is further configured to determine semantic expression data of the resource transfer description data;

[0014] A processing unit is used to perform evaluation processing based on the semantic expression data of the resource transfer description data to obtain an abnormality evaluation result, wherein the abnormality evaluation result is used to indicate whether there is an abnormality in the resource transfer operation of the object to be evaluated within the target time.

[0015] On the one hand, an embodiment of the present application provides a computer device, comprising: a processor, a communication interface and a memory, wherein the processor, the communication interface and the memory are interconnected, wherein the memory stores computer instructions, and the processor is used to call the computer instructions to implement the evaluation method provided in the embodiment of the present application.

[0016] Accordingly, an embodiment of the present application further provides a computer-readable storage medium, in which computer instructions are stored. When the computer-readable storage medium is executed on a computer device, the computer device implements the evaluation method provided in the embodiment of the present application.

[0017] Accordingly, an embodiment of the present application also provides a computer program product, which includes a computer program or computer instructions, and the computer program or computer instructions are stored in a computer-readable storage medium; the processor of a computer device reads the computer program or computer instructions from the computer-readable storage medium, and the processor executes the computer program or computer instructions, so that the computer device implements the evaluation method provided in the embodiment of the present application.

[0018] Through the evaluation method provided in the embodiment of the present application, resource transfer description data can be determined based on resource transfer related information, and the resource transfer description data includes characteristic information of the resource transfer operation of the object within the target time, that is, the resource transfer description data can describe the characteristics of multiple dimensions of the resource transfer operation to a certain extent; the semantic expression data of the resource transfer description data can be determined, and the semantic expression data can be a semantic feature vector, that is, the method provided in the present application can convert the description data into a semantic feature vector, which is convenient for subsequent analysis based on the feature vector; since the resource transfer description data can indicate the characteristic information of one or more resource transfer operations, the semantic expression data can include the semantic features of one or more characteristic information, and the abnormal evaluation result can be determined based on the semantic expression data, and the abnormal evaluation result can more accurately indicate whether there is an abnormality in the resource transfer operation of the object, that is, effectively improve the accuracy of the evaluation result; the method provided in the present application can also realize the automation of the evaluation process, effectively improving the efficiency of the evaluation process. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0020] Figure 1 This is a schematic diagram of the system architecture of an evaluation system provided in an embodiment of the present application;

[0021] Figure 2 This is a flow chart of an evaluation method provided in an embodiment of the present application;

[0022] Figure 3 is a schematic diagram of a method for determining semantic expression data provided in an embodiment of the present application;

[0023] Figure 4 is a schematic diagram of a semantic expression matrix provided in an embodiment of the present application;

[0024] Figure 5 is a schematic diagram of an evaluation method provided in an embodiment of the present application;

[0025] Figure 6 This is a flow chart of a model training method provided in an embodiment of the present application;

[0026] Figure 7 is a schematic diagram of a model training method provided in an embodiment of the present application;

[0027] Figure 8This is a structural block diagram of an evaluation device provided in an embodiment of the present application;

[0028] Figure 9 This is a structural block diagram of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0029] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0030] It should be noted that the terms "first" and "second" in the embodiments of this application are for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a technical feature designated as "first" or "second" may explicitly or implicitly include at least one such feature.

[0031] To facilitate understanding, we first provide a brief explanation of some nouns:

[0032] Resource transfer anomaly detection: Detects whether a resource transfer operation contains anomalies, and then performs corresponding intervention operations based on the detection results.

[0033] Text semantic conversion: converting text information into a vector that can express the semantics of the text, that is, using a numerical vector to represent the semantics of the text.

[0034] Embedding: It uses a certain mapping rule to extract as many features of the text as possible and display these features in the form of numbers.

[0035] Anomaly detection: is a technique for identifying items, events, or observations that do not match the expected pattern or other items in the dataset. It can also be seen as a classification problem in the case of severe data imbalance.

[0036] Object behavior data: This data records the behavior of an object performing a certain activity. Multiple pieces of object behavior data for the same object, recorded in chronological order within a certain time period, can form an object behavior list.

[0037] Online resource transfer technology enables resource transfers between different objects over the internet, significantly improving resource transfer efficiency. To ensure resource security, resource transfer operations associated with an object can be subject to resource transfer anomaly detection. During anomaly detection, professional technicians typically analyze and process data related to resource transfer operations associated with an object to assess the object's anomaly. However, this assessment method relies heavily on the professional knowledge and technical skills of the technicians, and different technicians may produce different assessment results, resulting in uncertainty. This assessment method is inefficient and generally lacks accuracy.

[0038] Based on this, an embodiment of the present application provides an evaluation method, which can obtain resource transfer related information of the resource transfer operation associated with the object to be evaluated within the target time; determine the resource transfer description data based on the resource transfer related information, the resource transfer description data is used to indicate the characteristic information of the resource transfer operation of the object to be evaluated within the target time, the characteristic information includes one or more of the time period of the resource transfer, the type of resource transfer object, the resource transfer type, the location indication information of the resource transfer object, the resource transfer method, and the number of resource transfers; determine the semantic expression data of the resource transfer description data; perform evaluation processing based on the semantic expression data of the resource transfer description data to obtain an abnormal evaluation result, and the abnormal evaluation result is used to indicate whether there is an abnormality in the resource transfer operation of the object to be evaluated within the target time. Through the method provided by the embodiment of the present application, the abnormal evaluation result can be determined based on the semantic expression data of the semantic dimension, which effectively improves the accuracy of the evaluation result and also effectively improves the efficiency of the evaluation processing.

[0039] The evaluation method provided in the embodiments of the present application can be applied to natural language processing technology in the field of artificial intelligence. Artificial Intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. Artificial intelligence technology is an interdisciplinary subject that covers a wide range of fields, including both hardware-level technology and software-level technology. Artificial intelligence software technology mainly includes computer vision technology, speech processing technology, natural language processing technology and machine learning / deep learning and other major directions. Natural language processing (Natural Language Processing, NLP) mainly studies various theories and methods that can enable effective communication between humans and computers using natural language. Natural language processing involves natural language, that is, the language used by people in daily life, and is closely related to linguistic research; it also involves computer science and mathematics. Natural language processing technology generally includes text processing, semantic understanding, machine translation, robot question answering, knowledge graph and other technologies.

[0040] The evaluation method provided by the embodiment of the present application can be applied to related scenarios of natural language processing technology. For example, a resource transfer application needs to monitor the resource transfer operations of the objects it uses to determine abnormal objects related to abnormal resource transfer operations. The relevant data of the resource transfer operations include numerical values ​​and natural language texts. The evaluation method provided by the embodiment of the present application can be used to obtain resource transfer related information of the resource transfer operations associated with the object to be evaluated, and the resource transfer description data can be determined based on the resource transfer related information. The resource transfer description data can indicate the characteristic information of the resource transfer operation of the object to be evaluated; the evaluation process can be performed based on the semantic expression data of the resource transfer description data to obtain an abnormal evaluation result indicating whether the resource transfer operation of the object to be evaluated is abnormal. Through the method provided by the embodiment of the present application, the evaluation result can be determined based on the semantic expression data of the description data, and abnormal evaluation is realized from the semantic dimension, which effectively improves the accuracy of the evaluation result.

[0041] The evaluation method provided in the embodiments of the present application can also be applied to the field of cloud computing. Cloud computing is a computing model that distributes computing tasks across a resource pool composed of a large number of computers, enabling various application systems to obtain computing power, storage space, and information services as needed. The cloud computing resource pool mainly includes: computing devices (virtualized machines, including operating systems), storage devices, and network devices. The evaluation method provided in the embodiments of the present application can be implemented based on cloud computing technology. Specifically, the device that implements the evaluation method provided in the embodiments of the present application can be a cloud computing device. For example, the cloud computing device can obtain resource transfer related information of resource transfer operations associated with the object to be evaluated within a target time from a cloud storage device; determine resource transfer description data based on the resource transfer related information, and the resource transfer description data can be used to indicate characteristic information of the resource transfer operations of the object to be evaluated within the target time; the cloud computing device can determine semantic expression data of the resource transfer description data, and perform evaluation processing based on the semantic expression data of the resource transfer description data to obtain an abnormal evaluation result. The cloud computing device can send the abnormal evaluation result to a processing device, and the processing device performs corresponding processing on the object to be evaluated based on the abnormal evaluation result. The method provided in the embodiments of the present application can effectively improve the accuracy of the evaluation results, thereby helping to improve the security of resources.

[0042] The following will introduce the architecture of the evaluation system provided in the embodiments of the present application with reference to the accompanying drawings.

[0043] See Figure 1 , which is a schematic diagram of the system architecture of an evaluation system provided by an embodiment of the present application, the evaluation system includes a terminal device 101 (such as Figure 1As shown, the terminal device 101 includes terminal device a and terminal device b), an evaluation device 102 and a database 103. The evaluation device 102 can exchange data with the terminal device 101 and the database 103.

[0044] The terminal device 101 can be used to implement resource transfer operations, such as Figure 1 As shown, terminal device a can perform resource transfer operations with terminal device b via the Internet. Terminal device 101 can be a handheld device (such as a smartphone, tablet computer), a computing device (such as a personal computer (PC), a vehicle terminal, an intelligent voice interaction device, a wearable device, or other intelligent device with resource transfer and communication functions, but is not limited thereto.

[0045] Evaluation device 102 can obtain resource transfer related information of a resource transfer operation associated with terminal device 101 and store the resource transfer related information in database 103; can perform evaluation processing based on the resource transfer related information to obtain an abnormality evaluation result corresponding to terminal device 101; and can also perform corresponding processing on terminal device 101 based on the abnormality evaluation result. Evaluation device 102 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, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0046] The database 103 is used to evaluate the relevant data of the device 102, such as: information related to resource transfer, abnormality assessment results, etc. The database 103 can be a local database in the device 102, or it can be a cloud database associated with the device 102 (i.e., a database deployed in the cloud). Specifically, it can be based on any deployment of a private cloud, a public cloud, a hybrid cloud, an edge cloud, etc., so that the cloud database focuses on different functions. For example, the database deployed in the private cloud has the user's personal equipment as the basic cloud hardware, and is more focused on serving a small number of users. The database deployed in the public cloud is based on a cloud platform provided by a third party, which allows the data stored in the database to be shared. Any user's data can be stored in the database, and any user can also use the data in the database.

[0047] The following will explain in detail Figure 1 The evaluation system shown works as follows:

[0048] The terminal device a in the terminal device 101 can implement a resource transfer operation with the terminal device b, and the evaluation device 102 can obtain the resource transfer related information to be stored of the resource transfer operation associated with the terminal device a from the terminal device a, or can also perform the following operations: Figure 1 As shown, the resource transfer related information to be stored is from terminal device b and the resource transfer operation associated with terminal device b. The evaluation device 102 can store the acquired resource transfer related information to be stored in the database 103.

[0049] When performing an evaluation on terminal device a, the evaluation device 102 can obtain resource transfer related information of the resource transfer operation associated with terminal device a within the target time (for example, within one hour, within one day) from the database 103; the evaluation device 102 can determine the resource transfer description data based on the resource transfer related information, and the resource transfer description data is used to indicate the characteristic information of the resource transfer operation of terminal device a within the target time, and the characteristic information may include one or more of the time period of the resource transfer, the type of resource transfer object, the resource transfer type, the location indication information of the resource transfer object, the resource transfer method, and the number of resource transfers; the evaluation device 102 can determine the semantic expression data of the resource transfer description data, and perform evaluation processing based on the semantic expression data of the resource transfer description data to obtain an abnormal evaluation result, and the abnormal evaluation result is used to indicate whether there is an abnormality in the resource transfer operation of terminal device a within the target time.

[0050] If the abnormality assessment result indicates that the resource transfer operation of terminal device a within the target time is abnormal, the assessment device 102 can suspend the resource transfer permission of terminal device a. The assessment method provided in the embodiment of the present application can determine the abnormality assessment result based on the semantic expression data of the resource transfer description data, and perform abnormality assessment from the semantic dimension of the description data, which is conducive to improving the accuracy of the assessment result and also helps to improve the security of resources.

[0051] It is understood that the schematic diagram of the evaluation system described in the embodiment of the present application is intended to more clearly illustrate the evaluation method of the embodiment of the present application and does not constitute a limitation on the evaluation method provided in the embodiment of the present application. For example, the evaluation method provided in the embodiment of the present application can be executed by the evaluation device 102 as well as by other devices that are different from the evaluation device 102 and can communicate with the terminal device 101 and the database 103. It is known to those skilled in the art that Figure 1 The number of terminal devices 101, evaluation devices 102, and databases 103 is merely illustrative. Any number of devices and nodes can be configured based on business implementation needs. Furthermore, as system architecture evolves and new business scenarios emerge, the evaluation methods provided in the embodiments of this application will also be applicable to similar technical problems.

[0052] It should be noted that the collection and processing of relevant data in this application (for example, information related to resource transfer) should be strictly in accordance with the requirements of relevant laws and regulations when applied in practice, and the informed consent or separate consent of the personal information subject should be obtained. Subsequent data use and processing should be carried out within the scope of authorization of laws and regulations and the personal information subject.

[0053] See Figure 2 , Figure 2 This is a flow chart of an evaluation method provided in an embodiment of the present application. The evaluation method can be implemented by the evaluation device 102 described above, or by other devices capable of implementing the evaluation method. The following description uses the evaluation method implemented by the evaluation device 102 as an example. The flow of the evaluation method provided in the embodiment of the present application includes, but is not limited to:

[0054] S201: Obtain resource transfer related information of a resource transfer operation associated with an object to be evaluated within a target time.

[0055] In an embodiment of the present application, the object to be evaluated is an object involved in a resource transfer operation. The object to be evaluated may hold a certain amount of resources, or may respond to an instruction to perform a resource transfer operation. When the evaluation device performs an evaluation process on the object to be evaluated, it may obtain resource transfer related information of the resource transfer operation associated with the object to be evaluated within a target time. The target time may be a time period determined according to application requirements, such as one hour, one day, etc. The resource transfer related information of the resource transfer operation associated with the object to be evaluated may include information such as the time when the resource transfer operation occurred, the resource transfer object, the resource transfer type, the location of the resource transfer object, and the resource transfer method. The time when the resource transfer operation occurs in the resource transfer related information matches the target time. For example, if the target time is "11:00 on January 1, 2024 - 13:00 on January 1, 2024", the resource transfer related information of the resource transfer operation associated with the object to be evaluated may be "at 12:00 on January 1, 2024, the first amount of resources of object A is transferred to object B through the first resource transfer method. Object B is a friend of object A, object A is located at place A, and object B is located at place A." Resource transfer related information may include data of multiple dimensions involved in a resource transfer operation. In some cases, the resource transfer related information may be the object behavior data of the object to be evaluated when performing the resource transfer operation. Through the method provided in the embodiment of the present application, resource transfer related information can be determined, which facilitates subsequent processing based on the resource transfer related information to obtain accurate abnormality assessment results.

[0056] It should be noted that the target time can be adaptively adjusted based on different actual application scenarios. When a terminal device performs a resource transfer operation, it can generate resource transfer-related information and send the resource transfer-related information to the evaluation device. The evaluation device can store the resource transfer-related information in an information database. The evaluation device can obtain resource transfer-related information that meets the requirements from the information database based on the target time and the object to be evaluated. The obtained resource transfer-related information can be one or more pieces. In some cases, the target time can also be part of the evaluation time. The evaluation time is the entire time, and the evaluation time can be divided to obtain multiple target times. For example, if the evaluation time is one day, the evaluation time can be divided to obtain target times 1-4, where target time 1 is "0:00-6:00", target time 2 is "6:00-12:00", target time 3 is "12:00-18:00", and target time 4 is "18:00-24:00". The evaluation device can obtain relevant information about resource transfers that meets the requirements according to each target time, and determine the semantic expression data of the resource transfer description data within the target time based on the acquired information; the evaluation device can evaluate and process the semantic expression data of the resource transfer description data within these four target times to obtain an abnormal evaluation result, which can indicate whether there is an abnormality in the resource transfer operation of the object to be evaluated within the evaluation time.

[0057] S202. Determine resource transfer description data based on the resource transfer related information, where the resource transfer description data is used to indicate characteristic information of the resource transfer operation of the object to be evaluated within the target time, and the characteristic information includes one or more of the time period of the resource transfer, the type of the resource transfer object, the resource transfer type, the location indication information of the resource transfer object, the resource transfer method, and the number of resource transfers.

[0058] In an embodiment of the present application, the evaluation device can determine resource transfer description data based on resource transfer related information. The resource transfer description data can be used to indicate characteristic information of the resource transfer operation of the object to be evaluated within the target time. The characteristic information may include one or more of the time period of resource transfer, the type of resource transfer object, the resource transfer type, the location indication information of the resource transfer object, the resource transfer method and the number of resource transfers. Among them, the time period of resource transfer is used to indicate the time period of the occurrence of the resource transfer operation; the type of resource transfer object can indicate the relationship between the resource transfer object and the object to be evaluated; the resource transfer type can be a transfer-out type, a transfer-in type, etc.; the location indication information of the resource transfer object can indicate the location relationship between the resource transfer object and the object to be evaluated; the resource transfer method can indicate the method of realizing resource transfer, such as "transfer" and the like; the number of resource transfers can indicate the frequency of resource transfer operations associated with the object to be evaluated within a certain period of time. For example: the resource transfer related information is "at 12:00 on January 1, 2024, the first amount of resources of object A will be transferred to object B through the first resource transfer method. Object B is a friend of object A, the location of object A is place A, and the location of object B is place A", then the resource transfer description data determined based on the resource transfer related information can be (6-12 o'clock (time period of resource transfer)_friend (type of resource transfer object)_transfer out (resource transfer type)_same city (location indication information of resource transfer object)_first resource transfer method (resource transfer method)_1 time (number of resource transfers)). It can be seen from the above example that the method provided in the present application can determine the resource transfer description data containing the characteristic information of the resource transfer operation based on the resource transfer related information containing natural language, and the resource transfer description data can be multiple characteristic semantics. The method provided in the embodiment of the present application can determine the resource transfer description data, which is conducive to the subsequent determination of accurate anomaly assessment results based on the semantic expression data of the resource transfer description data.

[0059] In one embodiment, determining resource transfer description data based on resource transfer-related information can be accomplished by determining a semantic mapping rule and processing the resource transfer-related information based on the semantic mapping rule to obtain the resource transfer description data. Semantic mapping rules can be constructed based on the feature dimensions required for anomaly assessment. For example, a semantic mapping rule can be constructed for "time period_friend / non-friend_transfer in / out_same city / non-same city_resource transfer method_number of resource transfers." Upon obtaining resource transfer-related information, the resource transfer-related information can be processed based on the semantic mapping rule to obtain the resource transfer description data.

[0060] In some cases, the process of processing the resource transfer related information based on the semantic mapping rule to obtain the resource transfer description data can be as follows: semantically dividing the resource transfer related information to obtain a plurality of shard information to be processed; for each shard information to be processed, semantically converting the shard information to be processed according to the semantic mapping rule to obtain a description text; and determining the resource transfer description data according to the description text corresponding to each shard information to be processed. For example: if the resource transfer related information is "at 12:00 on January 1, 2024, the first amount of resources of object A is transferred to object B (friend of object A) through the first resource transfer method, and the location of object A is place A, and the location of object B is place A", then the resource transfer related information is semantically divided to obtain a plurality of shard information to be processed, namely "at 12:00 on January 1, 2024", "through the first resource transfer method", "the first amount of resources", "transfer to", "object B", "object B (friend of object A)", "where object A is" and "where object A is" The location of object B is location A" and "the location of object B is location A"; taking the to-be-processed fragment information "January 1, 2024 12:00" as an example, assuming that the semantic mapping rule defines the "6-12 o'clock" time period, and the to-be-processed fragment information meets the semantic mapping rule, the to-be-processed information fragment "January 1, 2024 12:00" can be semantically converted according to the semantic mapping rule to obtain the "6-12 o'clock" description text; using the same method, the resource transfer description data can be determined as (6-12 o'clock_friend_transfer_same city_first resource transfer method_1 time). Through the method provided in the embodiment of the present application, the resource transfer description information can be determined based on the resource transfer related information, which is conducive to the subsequent use of the resource transfer description information for evaluation and processing to obtain accurate evaluation results.

[0061] It should be noted that different semantic mapping rules can be determined for different application scenarios. In order to further improve the processing efficiency of the evaluation method, a corpus can be constructed. The corpus is the most critical and basic part of the application related to natural semantic processing technology. It provides a method for constructing quantitative data. In the method provided in the embodiment of the present application, the corpus can include a general corpus and a black sample corpus of specially defined resource transfer operations; wherein, the general corpus can include semantics such as a certain time period, friends / non-friends, transfer in / out, whether in the same city, resource transfer method, number of transfers, etc. The black sample corpus is usually determined based on the relevant information of the abnormal resource transfer operation. The black sample corpus can be associated with the abnormal resource transfer operation to a certain extent, for example: the text corresponding to the number of transfers is "greater than 10 times", etc. In some cases, the black sample corpus can be determined according to certain rules (which can be rules summarized based on existing data). Through the method provided in the embodiment of the present application, multi-dimensional and extensive feature information can be obtained from the resource transfer related information, which increases the information dimension and is conducive to improving the accuracy of the evaluation results.

[0062] In one embodiment, the number of resource transfer operations associated with the object to be evaluated within the target time is N, where N is a positive integer. The implementation method for determining the resource transfer description data based on the resource transfer related information may be: performing statistical analysis on the resource transfer related information of the N resource transfer operations to obtain statistical analysis results, and determining the resource transfer description data based on the statistical analysis results; wherein the resource transfer related information includes one or more of the resource transfer time, resource transfer object, resource transfer type, location of the resource transfer object, and resource transfer method. Assuming that the object to be evaluated performs N resource transfer operations within the target time, where N is a positive integer, the evaluation device can obtain the resource transfer related information of the N resource transfer operations associated with the object to be evaluated within the target time. The resource transfer related information of the resource transfer operation may include one or more of the resource transfer time, resource transfer object, resource transfer type, location of the resource transfer object, and resource transfer method. The resource transfer related information includes the resource transfer time, which is used to indicate the time when the resource transfer operation occurs; the resource transfer object, which is used to indicate the operation object for resource transfer with the object to be evaluated; the resource transfer type, which is used to indicate the type of resource transfer operation performed with the object to be evaluated as the main body (transfer-in type or transfer-out type); the location of the resource transfer object, which is used to indicate the location of the operation object; and the resource transfer method, which is used to indicate the method of the resource transfer operation performed with the object to be evaluated as the main body.

[0063] The evaluation device can perform statistical analysis on the resource transfer related information of these N resource transfer operations to obtain statistical analysis results. Specifically, the evaluation device can analyze the resource transfer related information and merge multiple resource transfer related information with the same time period of resource transfer time, the same object type of resource transfer object, the same resource transfer type, the same location indication information corresponding to the location of the resource transfer object, and the same resource transfer method into one resource transfer related information, and determine the resource transfer related information obtained after the merger as the statistical analysis result.

[0064] For example, we now obtain the resource transfer related information of two resource transfer operations, among which the first resource transfer related information is "at 12:00 on January 1, 2024, the first amount of resources of object A is transferred to object B (a friend of object A) through the first resource transfer method, and the location of object A is place A, and the location of object B is place A", and the second resource transfer related information is "at 11:30 on January 1, 2024, the second amount of resources of object A is transferred to object C (a friend of object A) through the first resource transfer method, and the location of object A is place A, and the location of object C is place A". It can be seen that the first resource transfer related information is different from If the resource transfer time of the second resource transfer related information belongs to the same time period (all 6-12 o'clock), the object type of the resource transfer object is the same (all friends), the resource transfer type is the same (all transfer out), the location indication information corresponding to the location of the resource transfer object is the same (all in the same city), and the resource transfer method is the same (both are the first resource transfer method), then the first resource transfer information and the second resource transfer information can be merged to obtain the third resource transfer information. The third resource transfer information is "The resources of object A will be transferred to friends in the same city through the first resource transfer method from 11:30 to 12:00 on January 1, 2024, and the number of resource transfers is 2." The third resource transfer information is determined as statistical analysis information. After determining the statistical analysis information, the resource transfer description data can be determined based on the statistical analysis information.

[0065] If any of the N resource transfer related information does not satisfy the conditions that the resource transfer time belongs to the same time period, the resource transfer object has the same object type, the resource transfer type is the same, the location indication information corresponding to the location of the resource transfer object is the same, and the resource transfer method is the same, then no merging operation is performed and the resource transfer related information is directly determined as statistical analysis information. For example, if two resource transfer related information are determined, namely the fourth resource transfer related information and the fifth resource transfer related information, and if the fourth resource transfer related information and the fifth resource transfer related information do not satisfy the conditions that the resource transfer time belongs to the same time period, the resource transfer object has the same object type, the resource transfer type is the same, the location indication information corresponding to the location of the resource transfer object is the same, and the resource transfer method is the same, then the fourth resource transfer related information and the fifth resource transfer related information can be directly determined as the statistical analysis result, and first resource transfer description data can be determined based on the fourth resource transfer related information, and second resource transfer description data can be determined based on the fifth resource transfer related information. Subsequently, the semantic expression data of the first resource transfer description data and the semantic expression data of the second resource transfer description data can be used to jointly determine the evaluation result of the object to be evaluated. It should be noted that when multiple resource transfer description data are determined, the order of the resource transfer description data can be sorted from the beginning to the end according to the resource transfer time of the resource transfer operation corresponding to the resource transfer description data. Through the method provided in the embodiment of the present application, the resource transfer related information of multiple resource transfer operations can be processed to obtain resource transfer description data, which is conducive to comprehensively evaluating the characteristic information of multiple resource transfer operations and obtaining more accurate evaluation results.

[0066] In one embodiment, the number of resource transfer operations associated with the object to be evaluated within a target time is N, where N is a positive integer. The implementation method for determining resource transfer description data based on resource transfer related information may be: for any resource transfer related information among the N resource transfer related information, determining pending description data based on any resource transfer related information; performing statistical analysis and processing on the pending description data corresponding to the N pieces of resource transfer related information to obtain resource transfer description data; the pending description data includes characteristic information of the corresponding resource transfer operation, the characteristic information including one or more of the time period of the resource transfer, the type of the resource transfer object, the resource transfer type, the location indication information of the resource transfer object, and the resource transfer method. After the evaluation device obtains the resource transfer related information of the N resource transfer operations, it may process the N pieces of resource transfer related information separately to obtain pending description data corresponding to the N pieces of resource transfer related information, the description data including one or more of the time period of the resource transfer, the type of the resource transfer object, the resource transfer type, the location indication information of the resource transfer object, and the resource transfer method. The evaluation device can perform statistical analysis on the pending description data corresponding to N pieces of resource transfer related information. Specifically, the evaluation device can merge the pending description data with the same resource transfer time period, the same type of resource transfer object, the same resource transfer type, the same location indication information of the resource transfer object, and the same resource transfer method to obtain the merged description data, and determine the resource transfer description data based on the merged description data.

[0067] For example: obtaining the resource transfer related information of two resource transfer operations, processing the resource transfer related information of the two resource transfer operations respectively, and obtaining the first pending description data as (6-12 o'clock_friend_transfer out_same city_first resource transfer method), and the second pending description data as (6-12 o'clock_friend_transfer out_same city_first resource transfer method), then the first pending description data and the second pending description data can be merged to obtain the merged description data, which is (6-12 o'clock_friend_transfer out_same city_first resource transfer method_2 times), and the merged description data can be determined as the resource transfer description data.

[0068] For another example: the resource transfer related information of two resource transfer operations is obtained, and the resource transfer related information of the two resource transfer operations is processed respectively, and the third pending description data is obtained as (6-12 o'clock_friend_transfer out_same city_first resource transfer method), and the fourth pending description data is (6-12 o'clock_friend_transfer out_same city_second resource transfer method), then the third pending description data and the fourth pending description data can be directly determined as resource transfer description data. It should be noted that when multiple resource transfer description data are determined, the arrangement order of the resource transfer description data can be sorted from front to back according to the resource transfer time of the resource transfer operation corresponding to the resource transfer description data. The method provided by the embodiment of the present application can be conducive to the comprehensive evaluation of the characteristic information of multiple resource transfer operations to obtain more accurate evaluation results.

[0069] S203: Determine semantic expression data of the resource transfer description data.

[0070] In an embodiment of the present application, the evaluation device can determine the semantic expression data of the resource transfer description data, and the semantic expression data can be a semantic vector. For example: if the resource transfer description data is (6-12 o'clock_friend_transfer_same city_first resource transfer method_2 times), then the resource transfer description data can be [0.2, 0.3, 0.9, 0.1, 1, 0.8]. The semantic vector can express the characteristic information of the resource transfer operation included in the resource transfer description data from a mathematical dimension. Through the method provided in the embodiment of the present application, the semantic expression data can be determined, which facilitates the subsequent use of the semantic expression data to determine the evaluation results of the object to be evaluated, thereby improving the accuracy of the evaluation results.

[0071] In one embodiment, the resource transfer description data includes M resource transfer description texts, each corresponding to a type of characteristic information, where M is a positive integer. Determining the semantic expression data for the resource transfer description data can be accomplished by: determining a text feature vector for each resource transfer description text included in the resource transfer description data; and determining the semantic expression data for the resource transfer description data based on the text feature vectors of the resource transfer description texts included in the resource transfer description data. The resource transfer description data may include M resource transfer description texts (M is a positive integer), each corresponding to a type of characteristic information, and the characteristic information corresponding to one or more resource transfer description texts included in the resource transfer description data is different. For example, if the resource transfer description data is (6-12 pm_Friend_Transfer_Same City_First Resource Transfer Method_2 Times), the resource transfer description text included in the resource transfer description data is "6-12 pm," "Friend," "Transfer," "Same City," "First Resource Transfer Method," and "2 Times." The evaluation device can determine text feature vectors of each resource transfer description text included in the resource transfer description data; and based on the text feature vectors of the resource transfer description text included in the resource transfer description data, determine semantic expression data of the resource transfer description data. The method provided in the embodiments of the present application can determine accurate semantic expression data, which facilitates the subsequent determination of accurate evaluation results based on the semantic expression data.

[0072] In one embodiment, determining the text feature vectors of each resource transfer description text included in the resource transfer description data can be achieved by querying a database to determine whether a matching text feature vector that matches the target resource transfer description text exists in the database, the database including a mapping relationship between sample resource transfer description texts and text feature vectors, and the target resource transfer description text being any resource transfer description text included in the resource transfer description data; if a matching text feature vector that matches the target resource transfer description text exists in the database, the matching text feature vector is determined as the text feature vector of the target resource transfer description text; if a matching text feature vector that matches the target resource transfer description text does not exist in the database, text feature extraction is performed on the target resource transfer description text to obtain the text feature vector of the target resource transfer description text. After determining the resource transfer description data, the evaluation device can query the database to determine whether a matching text feature vector that matches the target resource transfer description text exists in the database, the database including a mapping relationship between sample resource transfer description texts and text feature vectors, for example, the database may include the sample resource transfer description text "Friends" and its text feature vector [0, 1, 0, 0, 0]. The target resource transfer description text is any resource transfer description text included in the resource transfer description data.

[0073] If a matching text feature vector that matches the target resource transfer description text exists in the database, the evaluation device can directly determine the matching text feature vector as the text feature vector of the target resource transfer description text. If a matching text feature vector that matches the target resource transfer description text does not exist in the database, the evaluation device can perform text feature extraction on the target resource transfer description text to obtain the text feature vector of the target resource transfer description text. The method provided in the embodiments of the present application can directly obtain text feature vectors from the database, which is beneficial for improving the processing efficiency of the evaluation method.

[0074] It should be noted that the process of extracting text features from the target resource transfer description text to obtain the target resource transfer description text's text feature vector can be a text semantic conversion process, and the description text can be converted into a text feature vector using an embedding method. The sample resource transfer description texts and their text feature vectors included in the database can also be determined using the embedding method.

[0075] In one embodiment, a method for determining text feature vectors of sample resource transfer description text in a database and determining text feature vectors of target resource transfer description text may include using the continuous bag of words (CBOW) model within the word2vec (text-to-vector) model to determine the text feature vectors. The word2vec model is a tool for converting text words into vector form; using this model, vector representations of individual text words can be obtained, laying the foundation for subsequent representations of descriptive data. The word2vec model is primarily divided into the continuous bag of words model and the continuous skip-gram (Skip-Gram) model. The CBOW model can infer target words from the original text, while the Skip-Gram model can infer the original sentence from the target word. CBOW is suitable for small databases, while the Skip-Gram model performs better in large corpus databases. The method provided in the embodiments of the present application may use the CBOW model or other models capable of performing text semantic conversion processing. The method provided in the embodiments of the present application can determine accurate text feature vectors, which helps improve the processing efficiency of the evaluation method and the accuracy of the evaluation results.

[0076] In one embodiment, determining the semantic expression data of the resource transfer description data based on the text feature vectors of the resource transfer description text included in the resource transfer description data can be achieved by: determining whether the resource transfer description text included in the resource transfer description data contains abnormal resource transfer description text, where the abnormal resource transfer description text is resource transfer description text in an abnormal text dataset, and the resource transfer description text in the abnormal text dataset corresponds to characteristic information of an abnormal resource transfer operation; if the resource transfer description text included in the resource transfer description data does not contain abnormal resource transfer description text, performing average weighted processing on the text feature vectors of the resource transfer description text included in the resource transfer description data; and determining the semantic expression data of the resource transfer description data based on the average weighted processing result. The evaluation device can determine whether one or more resource transfer description texts included in the resource transfer description data contain abnormal resource transfer description text, where the abnormal resource transfer description text, also known as a black sample semantic, is resource transfer description text in the abnormal text dataset, and the resource transfer description text in the abnormal text dataset corresponds to characteristic information of an abnormal resource transfer operation. The resource transfer description text in the abnormal text data set can be obtained by feature extraction of resource transfer related information of the abnormal resource transfer operation. The inclusion of abnormal resource transfer description text in the resource transfer description data can indicate that the resource transfer operation corresponding to the resource transfer description data is likely to be an abnormal resource transfer operation.

[0077] If the resource transfer description text included in the resource transfer description data does not contain abnormal resource transfer description text, the evaluation device can perform average weighted processing on the text feature vectors of the resource transfer description texts included in the resource transfer description data, that is, the weight data of each resource transfer description text included in the resource transfer description data is the same, and the evaluation data is weighted according to the text feature vector and weight data of each resource transfer description text to obtain the weighted text feature vector of each resource transfer description text; the evaluation device can determine the semantic expression data of the resource transfer description data based on the average weighted processing result, that is, determine the semantic expression data of the resource transfer description data based on the weighted text feature vector of each resource transfer description text.

[0078] For example: if it is determined that the resource transfer description text included in the resource transfer description data does not contain abnormal resource transfer description text, and the text feature vectors of the two resource transfer description texts included in the resource transfer description data are [1, 0, 0, 0, 0, 0] and [0, 1, 0, 0, 0, 0] respectively, and the weight data of the two resource transfer description texts are both 0.5, then it can be determined that the weighted text feature vectors of the two resource transfer description texts are [0.5, 0, 0, 0, 0] and [0, 0.5, 0, 0, 0, 0] respectively, thereby determining that the semantic expression data of the resource transfer description data is [0.5, 0.5, 0, 0, 0, 0]. By the method provided in the embodiment of the present application, the semantic expression data of the resource transfer description data can be reasonably determined, which is conducive to improving the accuracy of the evaluation results.

[0079] In one embodiment, the evaluation device may further perform the following steps: if the resource transfer description text included in the resource transfer description data contains abnormal resource transfer description text, performing non-average weighting processing on the text feature vectors of the resource transfer description text included in the resource transfer description data; and determining semantic expression data of the resource transfer description data based on the non-average weighting processing result; wherein the weight data of the abnormal resource transfer description text is greater than the weight data of the normal resource transfer description text, and the normal resource transfer description text is the resource transfer description text in the resource transfer description data other than the abnormal resource transfer description text. The evaluation device determines whether the resource transfer description text included in the resource transfer description data contains abnormal resource transfer description text. If the resource transfer description text included in the resource transfer description data contains abnormal resource transfer description text, the evaluation device may perform non-average weighting processing on the text feature vectors of the resource transfer description text included in the resource transfer description data, wherein the weight data of the abnormal resource transfer description text is greater than the weight data of the normal resource transfer description text, and the normal resource transfer description text is the resource transfer description text in the resource transfer description data other than the abnormal resource transfer description text.

[0080] For example: the resource transfer description data includes two resource transfer description texts, namely the first resource transfer description text and the second resource transfer description text. It is determined that the first resource transfer description text included in the resource transfer description data is an abnormal resource transfer description text, the text feature vector of the first resource transfer description text is [1, 0, 0, 0, 0, 0], and the text feature vector of the second resource transfer description text is [0, 1, 0, 0, 0, 0]. Then, the weight data of the abnormal resource transfer description text can be appropriately increased, or the weight data of the normal resource transfer description text can be appropriately reduced. Assuming that the weight data of the first resource transfer description text is determined to be 0.8 and the weight data of the second resource transfer description text is determined to be 0.2, the weighted text feature vectors of the two resource transfer description texts can be determined to be [0.8, 0, 0, 0, 0] and [0, 0.2, 0, 0, 0, 0] respectively, so that the semantic expression data of the resource transfer description data can be determined to be [0.8, 0.2, 0, 0, 0, 0]. Through the method provided in the embodiment of the present application, abnormal resource transfer description text can be introduced and corresponding differentiated processing can be performed, so that the semantic expression data of the resource transfer description data can be reasonably determined, which is conducive to improving the accuracy of the evaluation results.

[0081] See Figure 3 , this figure is a schematic diagram of a semantic expression data determination method provided by an embodiment of the present application. The evaluation device can obtain resource transfer related information of the resource transfer operation associated with the object to be evaluated within the target time, and can determine the resource transfer description information based on the resource transfer related information. The resource transfer description information can be used to indicate the characteristic information of the resource transfer operation of the object to be evaluated within the target time. The resource transfer description information may include one or more resource transfer description texts, and each resource transfer description text corresponds to a characteristic information. Assume that the resource transfer description texts included in the resource transfer description information are resource transfer description text A and resource transfer description text C. The evaluation device can determine the text feature vectors of each resource transfer description text included in the resource transfer description data. Specifically, the evaluation device can query the database (such as Figure 3 As shown, the database includes multiple sample resource transfer description texts and their text feature vectors, among which the text feature vector of resource transfer description text A is [1,0,0,0,0], and the text feature vector of resource transfer description text B is [0,1,0,0,0] 、 The text feature vector of the resource transfer description text C is [0,0,1,0,0] 、 The text feature vector of the resource transfer description text D is [0,0,0,1,0] 、The text feature vector of the resource transfer description text E is [0,0,0,0,1]), the text feature vector of the resource transfer description text A is determined to be [1,0,0,0,0], and the text feature vector of the resource transfer description text C is determined to be [0,0,1,0,0].

[0082] After determining the text feature vectors of each resource transfer description text in the resource transfer description data, the evaluation device can determine whether the resource transfer description texts included in the resource transfer description data contain abnormal resource transfer description texts. If there are no abnormal resource transfer description texts, the evaluation device can perform average weighted processing on the text feature vectors of each resource transfer description text, and determine the semantic expression data of the resource transfer description data based on the average weighted processing result, such as Figure 3 As shown, the text feature vector of the resource transfer description text A can be weighted (the weight data is 0.5) to obtain the weighted text feature vector of the resource transfer description text A [0.5, 0, 0, 0, 0]; the text feature vector of the resource transfer description text C can be weighted (the weight data is 0.5) to obtain the weighted text feature vector of the resource transfer description text C [0, 0, 0.5, 0, 0]; and the semantic expression data of the resource transfer description data is determined to be [0.5, 0, 0.5, 0, 0] based on the average weighted processing result.

[0083] If there is an abnormal resource transfer description text, the evaluation device can perform non-average weighted processing on the text feature vectors of each resource transfer description text, and determine the semantic expression data of the resource transfer description data based on the non-average weighted processing result, such as Figure 3 As shown, the text feature vector of the resource transfer description text A can be weighted (the weight data is 0.2) to obtain the weighted text feature vector of the resource transfer description text A [0.2, 0, 0, 0, 0]; the text feature vector of the resource transfer description text C can be weighted (the weight data is 0.8) to obtain the weighted text feature vector of the resource transfer description text C [0, 0, 0.8, 0, 0]; according to the average weighted processing result, the semantic expression data of the resource transfer description data is determined to be [0.2, 0, 0.8, 0, 0]. Through the method provided in the embodiment of the present application, the semantic expression data of the resource transfer description data can be accurately and reasonably determined, which is conducive to the subsequent determination of accurate evaluation results based on the semantic expression data.

[0084] In one embodiment, if resource transfer related information of multiple resource transfer operations associated with the object to be evaluated within the target time is obtained, in some cases, multiple resource transfer description data can be determined based on the multiple resource transfer related information obtained; the resource transfer times of the resource transfer operations corresponding to the multiple resource transfer description data can be arranged in sequence (if a resource transfer description data is determined based on multiple resource transfer related information, then the resource transfer time of the resource transfer operation corresponding to the resource transfer description data is the earliest time of the resource transfer information in the multiple resource transfer description information), and a resource transfer description data queue is obtained; for any resource transfer description data in the resource transfer description data queue, the method provided in the embodiment of the present application can be performed to obtain semantic expression data of the resource transfer description data; the semantic expression data of each resource transfer description data in the resource transfer description data queue can form a semantic expression matrix. Please refer to Figure 4 , which is a schematic diagram of a semantic expression matrix provided in an embodiment of the present application. Figure 4 It includes 4 resource transfer description data, namely resource transfer description data 1-4. The method provided in the embodiment of the present application can be used to determine the semantic expression data of each resource transfer description data (wherein, the semantic expression data of resource transfer description data 1 is [0.2, 0.1, 0, 0.4, 0.3], the semantic expression data of resource transfer description data 2 is [0.1, 0.1, 0.5, 0.2, 0.1], the semantic expression data of resource transfer description data 3 is [0.3, 0.3, 0.1, 0.1, 0.1], and the semantic expression data of resource transfer description data 4 is [0.1, 0.1, 0.1, 0.6, 0.1]). A semantic expression matrix can be determined based on the semantic expression data of each resource transfer description data. The semantic expression matrix can more comprehensively represent the characteristics of the resource transfer operations involved in the target time of the object to be evaluated. The evaluation device can use the semantic expression matrix to perform evaluation processing to obtain the evaluation result of the object to be evaluated. Through the method provided in the embodiment of the present application, the characteristics of the resource transfer operation can be used for evaluation processing to obtain a more accurate evaluation result.

[0085] S204 . Perform evaluation processing based on the semantic expression data of the resource transfer description data to obtain an abnormality evaluation result, where the abnormality evaluation result is used to indicate whether there is an abnormality in the resource transfer operation of the object to be evaluated within the target time.

[0086] In an embodiment of the present application, an evaluation device can perform evaluation processing based on the semantic expression data of the resource transfer description data to obtain an anomaly evaluation result, which indicates whether an anomaly exists in the resource transfer operation of the subject to be evaluated within the target time. Because the resource transfer description data may include one or more feature information of the resource transfer operation of the subject to be evaluated within the target time, the semantic expression data of the resource transfer description data may also include features of multiple dimensions. Compared to methods that only use numerical values ​​such as the number of resource transfers and the frequency of resource transfers for anomaly evaluation, the method provided in an embodiment of the present application can effectively increase the breadth of feature dimensions, thereby improving the accuracy of the anomaly evaluation results.

[0087] It should be noted that if there are multiple target times (each target time corresponds to a different time period), the evaluation device can use the method provided in the embodiment of the present application to determine the semantic expression data of the resource transfer description data within each target time, and perform evaluation processing based on the semantic expression data of the resource transfer description data within the multiple target times to obtain an anomaly evaluation result. The method provided in the embodiment of the present application can comprehensively determine the semantic expression data and determine an accurate anomaly evaluation result based on the semantic expression data.

[0088] In one embodiment, evaluating and processing the semantic expression data of resource transfer description data to obtain an abnormality evaluation result can be implemented by: inputting the semantic expression data of the resource transfer description data into an abnormality evaluation model for processing to obtain an abnormality evaluation result; wherein the abnormality evaluation model is obtained by adjusting the initial evaluation model based on difference data, and the difference data is determined based on the predicted evaluation result and the sample evaluation result; the predicted evaluation result is obtained by evaluating and processing the sample semantic expression data using the initial evaluation model, the sample semantic expression data is the semantic expression data of the sample resource transfer description data, the sample resource transfer description data is determined based on the sample resource transfer related information, and the sample resource transfer related information is the resource transfer related information of the sample resource transfer operation associated with the sample object within the reference time; the predicted evaluation result is a predicted evaluation result used to indicate whether the sample resource transfer operation has an abnormality, and the sample evaluation result is a marked evaluation result used to indicate whether the sample resource transfer operation has an abnormality. The evaluation device can input the semantic expression data of the resource transfer description data into the abnormality evaluation model for processing to obtain the abnormality evaluation result. The anomaly assessment model can be an Extreme Gradient Boosted (XGB) model using a Gradient Boost framework, a Deep Neural Networks (DNN) model using a deep learning framework, or other supervised classification models with better processing effects.

[0089] The XGB model includes multiple decision trees. During prediction, all decision trees in the model will score the predicted data, and then sum up all the scores to obtain a scoring result. The abnormal evaluation result of the object to be evaluated can be determined based on the scoring result. The DNN model is a multi-layer neural network, and the output features of the previous layer are used as the input of the next layer for feature learning. After layer-by-layer feature mapping, the features of the existing spatial samples can be mapped to another feature space, so that the existing input can have a better feature expression. The DNN model can represent actual complex nonlinear problems more carefully and efficiently, and can better implement the step of determining the abnormal evaluation result based on the semantic expression data in the embodiment of the present application.

[0090] The abnormality assessment model can be obtained by adjusting the initial assessment model of the difference data object, and the predicted assessment result can be obtained by evaluating and processing the sample semantic expression data using the initial assessment model. The sample semantic expression data is the semantic expression data of the sample resource transfer description data. The sample resource transfer description data can be determined based on the sample resource transfer related information. The sample resource transfer related information is the resource transfer related information of the sample resource transfer operation associated with the sample object within the reference time. The predicted assessment result is a predicted assessment result for indicating whether the sample resource transfer operation has an abnormality, and the sample assessment result is a marked assessment result for indicating whether the sample resource transfer operation has an abnormality. The device for performing model training can be the above-mentioned Figure 1 The evaluation device may also be other devices capable of model training. Through the method provided in the embodiment of the present application, the model can be used to determine the abnormal evaluation results of the object to be evaluated, which can effectively improve the processing efficiency of the evaluation method and the accuracy of the evaluation results.

[0091] In one embodiment, after determining the abnormality assessment result, the assessment device may further perform the following steps: if the abnormality assessment result indicates that the resource transfer operation of the subject to be assessed within the target time is abnormal, determining whether the resource transfer operation of the subject to be assessed within the target time is completed; if the resource transfer operation of the subject to be assessed within the target time is completed, suspending the resource transfer permission of the subject to be assessed; if the resource transfer operation of the subject to be assessed within the target time is not completed, terminating the uncompleted resource transfer operation. After determining the abnormality assessment result of the subject to be assessed, if the abnormality assessment result indicates that the resource transfer operation of the subject to be assessed within the target time is abnormal, the assessment device may determine whether the resource transfer operation of the subject to be assessed within the target time is completed; if the resource transfer operation of the subject to be assessed within the target time is completed, the assessment device may suspend the resource transfer permission of the subject to be assessed. For example, the assessment device may restrict or prohibit the resource transfer permission of the subject to be assessed to a certain extent, so that the subject to be assessed cannot perform certain resource transfer operations.

[0092] If the resource transfer operation of the object to be evaluated is not completed within the target time, the evaluation device can terminate the resource transfer operation that has not been completed. For example: the operation object can initiate a resource transfer operation to transfer the first amount of resources of the object to be evaluated to the target object. At this time, the first amount of resources of the object to be evaluated can be transferred to the intermediate device. After a certain time interval (for example: 1 hour, 1 day, etc.), the intermediate device transfers the first amount of resources to the target object, and the resource transfer operation is completed; if during the waiting time interval, the evaluation device determines that the abnormal evaluation result of the object to be evaluated indicates that there is an abnormality in the resource transfer operation of the object to be evaluated within the target time, the evaluation device can terminate the resource transfer operation that has not been completed, that is, the evaluation device can inform the intermediate device to stop transferring the first amount of resources to the target object. Through the method provided in the embodiment of the present application, the abnormal evaluation result can be used to perform corresponding processing, thereby effectively ensuring the security of resources.

[0093] See Figure 5, which is a schematic diagram of an evaluation method provided by an embodiment of the present application. The evaluation device can perform basic semantic preparation, that is, build a corpus, which can include a general corpus and a black sample corpus of specially defined resource transfer operations. The evaluation device can obtain resource transfer related information of the resource transfer operation of the object to be evaluated within the target time, and perform semantic matching based on the resource transfer related information and the corpus to obtain resource transfer description data. The resource transfer description data can be used to indicate the characteristic information of the resource transfer operation of the object to be evaluated within the target time. The characteristic information includes one or more of the time period of the resource transfer, the type of resource transfer object, the resource transfer type, the location indication information of the resource transfer object, the resource transfer method, and the number of resource transfers. The evaluation device can build a feature vector database, which can include a mapping relationship between sample resource transfer description texts and text feature vectors; the evaluation device can determine the text feature vector of each resource transfer description text based on the feature vector database and one or more resource transfer description texts included in the resource transfer description data, and determine the semantic expression data of the resource transfer description data. If multiple resource transfer description data are determined, the evaluation device can summarize the semantic expression data of each determined resource transfer description data to obtain a semantic expression matrix. The evaluation device can input the semantic expression matrix or semantic expression data into the anomaly evaluation model for processing to obtain an anomaly evaluation result of the object to be evaluated. The method provided by the embodiment of the present application can effectively improve the accuracy of the evaluation result, thereby ensuring the security of resources.

[0094] Through the evaluation method provided in the embodiment of the present application, resource transfer description data can be determined based on resource transfer related information. The resource transfer description data can describe the characteristics of the resource transfer operation to a certain extent, and is not limited to the number of resource transfers involved in the resource transfer operation; the semantic expression data of the resource transfer description data can be determined, and the semantic expression data can be a semantic feature vector, that is, the method provided in the present application can convert the description data into a semantic feature vector, which is convenient for subsequent analysis based on the feature vector; since the resource transfer description data can indicate the feature information of one or more resource transfer operations, the semantic expression data can include the semantic features of one or more feature information, and the abnormal evaluation result can be determined based on the semantic expression data. The abnormal evaluation result can more accurately indicate whether there is an abnormality in the resource transfer operation of the object, effectively improving the accuracy of the evaluation result; the abnormal evaluation result can be determined by using a model, effectively improving the processing efficiency of the evaluation method, and realizing the automation of the evaluation processing; corresponding processing can be performed based on the abnormal evaluation result, effectively ensuring the security of the resources and effectively reducing the loss of resources.

[0095] See Figure 6 , Figure 6This is a flow chart of a model training method provided in an embodiment of the present application. The model training method can be implemented by the above-mentioned evaluation device 102 or by other devices. Taking the evaluation device 102 as an example to implement the model training method, the process of the model training method provided in the embodiment of the present application includes but is not limited to:

[0096] S601. Obtain sample resource transfer related information of a sample resource transfer operation associated with a sample object within a reference time, and a sample evaluation result of the sample object, where the sample evaluation result is a marked evaluation result indicating whether the sample resource transfer operation is abnormal.

[0097] In an embodiment of the present application, the evaluation device can obtain sample resource transfer related information of a sample resource transfer operation associated with a sample object within a reference time, and the sample resource transfer related information is resource transfer related information of a sample resource transfer operation associated with a sample object within the reference time. The reference time can be determined according to the model training requirements, for example, the reference time can be one day or one week. The sample evaluation result of the sample object is a marked evaluation result indicating whether there is an abnormality in the sample resource transfer operation, for example, the sample evaluation result can be "normal" or "abnormal". Through the method provided in the embodiment of the present application, sample resource transfer related information and sample evaluation results can be obtained, which is conducive to the subsequent use of the acquired data for model training and improving the robustness of the model.

[0098] S602. Determine sample resource transfer description data based on the sample resource transfer related information, wherein the sample resource transfer description data is used to indicate sample characteristic information of the sample resource transfer operation of the sample object within the reference time, and the sample characteristic information includes one or more of the time period of resource transfer, the type of resource transfer object, the resource transfer type, the location indication information of the resource transfer object, the resource transfer method, and the number of resource transfers.

[0099] In an embodiment of the present application, the evaluation device can determine sample resource transfer description data based on sample resource conversion related information. The sample resource transfer description data can be used to indicate sample feature information of the sample resource transfer operation of the sample object within a reference time. The sample feature information may include one or more of the time period of the resource transfer, the type of resource transfer object, the resource transfer type, the location indication information of the resource transfer object, the resource transfer method, and the number of resource transfers. The method provided in the embodiment of the present application can determine the sample resource transfer description data, which is conducive to subsequently determining the prediction evaluation results based on the sample resource transfer description data, thereby realizing the model training process.

[0100] It should be noted that the evaluation device can obtain sample resource transfer related information of multiple sample resource transfer operations associated with the sample object within the reference time, and for multiple sample resource transfer related information; the evaluation device can perform statistical analysis and processing to obtain sample statistical analysis results, and determine one or more sample resource transfer description data based on the sample statistical analysis results.

[0101] In one embodiment, the implementation method for determining sample resource transfer description data based on sample resource transfer related information can be: processing the sample resource transfer related information based on semantic mapping rules to obtain sample resource transfer description data. The semantic mapping rules are determined based on application requirements, and the sample resource transfer related information can be mapped and processed based on the semantic mapping rules to obtain sample resource transfer description data. The method provided in the embodiment of the present application can determine the sample resource transfer description information, which is conducive to the subsequent use of the sample resource transfer description information for model training to obtain an anomaly assessment model with high prediction accuracy.

[0102] In one embodiment, the number of sample resource transfer operations associated with a sample object within a reference time is I, where I is a positive integer. The implementation method for determining the sample resource transfer description data based on the sample resource transfer related information may be: performing statistical analysis on the sample resource transfer related information of the I sample resource transfer operations to obtain sample statistical analysis results, and determining the sample resource transfer description data based on the sample statistical analysis results; wherein the sample resource transfer related information includes one or more of the resource transfer time, resource transfer object, resource transfer type, location of the resource transfer object, and resource transfer method. Assuming that the sample object performs I resource transfer operations within the target time, where I is a positive integer, the evaluation device may obtain the sample resource transfer related information of the I sample resource transfer operations associated with the sample object within the reference time. The sample resource transfer related information of the sample resource transfer operations may include one or more of the resource transfer time, resource transfer object, resource transfer type, location of the resource transfer object, and resource transfer method.

[0103] The evaluation device can perform statistical analysis on the sample resource transfer related information of these I sample resource transfer operations to obtain sample statistical analysis results. Specifically, the evaluation device can analyze the sample resource transfer related information and merge multiple sample resource transfer related information with the same time period of resource transfer time, the same object type of resource transfer object, the same resource transfer type, the same location indication information corresponding to the location of the resource transfer object, and the same resource transfer method into one sample resource transfer related information, and determine the sample resource transfer related information obtained after the merger as the sample statistical analysis result.

[0104] If the sample resource transfer related information of I sample resource transfer operations does not satisfy the resource transfer related information that the resource transfer time belongs to the same time period, the object type of the resource transfer object is the same, the resource transfer type is the same, the location indication information corresponding to the location of the resource transfer object is the same, and the resource transfer method is the same, then no merging operation is performed, and the sample resource transfer related information is directly determined as the sample statistical analysis information. It should be noted that when multiple sample resource transfer description data are determined, the arrangement order of the sample resource transfer description data can be sorted from front to back according to the resource transfer time of the sample resource transfer operation corresponding to the sample resource transfer description data. Through the method provided in the embodiment of the present application, the sample resource transfer related information of multiple sample resource transfer operations can be processed to obtain sample resource transfer description data, which is conducive to training with the integrated feature information of multiple sample resource transfer operations and is conducive to improving the accuracy of the evaluation results of the model.

[0105] In one embodiment, the number of sample resource transfer operations associated with a sample object within a reference time is I, where I is a positive integer. The implementation method for determining sample resource transfer description data based on the sample resource transfer related information may include: determining sample pending description data based on any sample resource transfer related information among the sample resource transfer related information of the I sample resource transfer operations; performing statistical analysis on the sample pending description data corresponding to the I sample resource transfer related information to obtain the sample resource transfer description data; the sample pending description data including sample feature information of the corresponding sample resource transfer operation, the sample feature information including one or more of the time period of the resource transfer, the type of the resource transfer object, the resource transfer type, the location indication information of the resource transfer object, and the resource transfer method. After obtaining the sample resource transfer related information of the I sample resource transfer operations, the evaluation device may process the I sample resource transfer related information separately to obtain the sample pending description data corresponding to the I sample resource transfer related information, the description data including one or more of the time period of the resource transfer, the type of the resource transfer object, the resource transfer type, the location indication information of the resource transfer object, and the resource transfer method. The evaluation device can perform statistical analysis on the sample pending description data corresponding to I pieces of sample resource transfer related information. Specifically, the evaluation device can merge the pending description data with the same resource transfer time period, the same type of resource transfer object, the same resource transfer type, the same location indication information of the resource transfer object, and the same resource transfer method to obtain the merged sample description data, and determine the sample resource transfer description data based on the merged sample description data. It should be noted that when multiple sample resource transfer description data are determined, the arrangement order of the sample resource transfer description data can be sorted from front to back according to the sample resource transfer time of the sample resource transfer operation corresponding to the sample resource transfer description data. The method provided by the embodiment of the present application can be beneficial for training with the integrated feature information of multiple sample resource transfer operations, which is conducive to more accurate evaluation results.

[0106] S603: Determine sample semantic expression data of the sample resource transfer description data.

[0107] In an embodiment of the present application, the sample resource transfer description data may include one or more intermediate resource transfer description texts, each of which corresponds to a type of sample feature information. The evaluation device can determine the text feature vectors of each intermediate resource transfer description text included in the sample resource transfer description data, and determine the sample semantic expression data of the sample resource transfer description data based on the text feature vectors of each intermediate resource transfer description text. The sample semantic expression data is also the semantic expression data of the sample resource transfer description data. The method provided in the embodiment of the present application can determine the sample semantic expression data, which facilitates the subsequent determination of the prediction evaluation results based on the sample semantic expression data.

[0108] It should be noted that if the evaluation device determines multiple sample resource transfer description data, the evaluation device can determine sample semantic expression data for each sample resource transfer description data, and determine a sample semantic expression matrix based on the sample semantic expression data for each sample resource transfer description data. This sample semantic expression matrix can be used to indicate the characteristics of the sample resource transfer operation of the sample object within the reference time, and a predicted evaluation result for the sample object can be determined based on this sample semantic expression matrix.

[0109] In one embodiment, the method for determining the text feature vectors of each intermediate resource transfer description text included in the sample resource transfer description data can be: querying a database to determine whether there is a matching text feature vector that matches the target intermediate resource transfer description text in the database, the database includes a mapping relationship between the sample resource transfer description text and the text feature vector, and the target intermediate resource transfer description text is any intermediate resource transfer description text included in the sample resource transfer description data; if there is a matching text feature vector that matches the target intermediate resource transfer description text in the database, then the matching text feature vector is determined as the text feature vector of the target intermediate resource transfer description text; if there is no matching text feature vector that matches the target intermediate resource transfer description text in the database, then text feature extraction processing is performed on the target intermediate resource transfer description text to obtain the text feature vector of the target intermediate resource transfer description text; and the target intermediate resource transfer description text and its text feature vector are added to the database.

[0110] After determining the sample resource transfer description data, the evaluation device may query a database to determine whether a matching text feature vector exists that matches the target intermediate resource transfer description text. The database includes mappings between multiple sample resource transfer description texts and text feature vectors. The target intermediate resource transfer description text is any intermediate resource transfer description text included in the sample resource transfer description data.

[0111] If a matching text feature vector that matches the target intermediate resource transfer description text exists in the database, the evaluation device can directly determine the matching text feature vector as the text feature vector of the target intermediate resource transfer description text. If a matching text feature vector that matches the target intermediate resource transfer description text does not exist in the database, the evaluation device can perform text feature extraction processing on the target resource transfer description text to obtain the text feature vector of the target resource transfer description text. The method provided in the embodiments of the present application can directly obtain text feature vectors from the database, which is conducive to improving the processing efficiency of the evaluation method.

[0112] It should be noted that the process of extracting text features from the target resource transfer description text to obtain the target resource transfer description text's text feature vector can be a text semantic conversion process, and the description text can be converted into a text feature vector using an embedding method. The sample resource transfer description texts and their text feature vectors included in the database can also be determined using the embedding method.

[0113] In one embodiment, the sample resource transfer description data includes a plurality of intermediate resource transfer description texts. Then, the implementation method of determining the sample semantic expression data of the sample resource transfer description data can be as follows: determining whether the intermediate resource transfer description texts included in the sample resource transfer description data contain abnormal resource transfer description texts, where the abnormal resource transfer description texts are resource transfer description texts in an abnormal text data set, and the resource transfer description texts in the abnormal text data set correspond to feature information of abnormal resource transfer operations; if the intermediate resource transfer description texts included in the sample resource transfer description data do not contain abnormal resource transfer description texts, then performing text feature vector analysis on the intermediate resource transfer description texts included in the sample resource transfer description data. Average weighted processing; determining sample semantic expression data of the sample resource transfer description data based on the average weighted processing result; if the intermediate resource transfer description text included in the sample resource transfer description data contains abnormal resource transfer description text, performing non-average weighted processing on the text feature vector of the intermediate resource transfer description text included in the sample resource transfer description data; determining sample semantic expression data of the sample resource transfer description data based on the non-average weighted processing result; wherein, the weight data of the abnormal resource transfer description text is greater than the weight data of the intermediate normal resource transfer description text, and the intermediate normal resource transfer description text is the intermediate resource transfer description text in the sample resource transfer description data excluding the abnormal resource transfer description text.

[0114] The evaluation device can determine whether there is an abnormal resource transfer description text among the multiple intermediate resource transfer description texts included in the sample resource transfer description data; if not, the evaluation device can perform average weighted processing on the text feature vectors of each intermediate resource transfer description text included in the sample resource transfer description data to obtain the weighted text feature vectors of each intermediate resource transfer description text, and determine the sample semantic expression data of the sample resource transfer description data based on the weighted text feature vectors of each intermediate resource transfer description text; if present, the evaluation device can perform non-average weighted processing on the text feature vectors of each intermediate resource transfer description text included in the sample resource transfer description data (wherein the weight data of the abnormal resource transfer description text is greater than the weight data of the intermediate resource transfer description text other than the abnormal resource transfer description text), obtain the weighted text feature vectors of each intermediate resource transfer description text, and determine the sample semantic expression data of the sample resource transfer description data based on the weighted text feature vectors of each intermediate resource transfer description text. Through the method provided in the embodiment of the present application, the sample semantic expression data of the sample resource transfer description data can be determined more reasonably, thereby helping to improve the prediction accuracy of the model.

[0115] S604: Input the sample semantic expression data into an initial evaluation model for evaluation processing to obtain a predicted evaluation result, where the predicted evaluation result is a predicted evaluation result indicating whether there is an abnormality in the sample resource transfer operation.

[0116] In an embodiment of the present application, the evaluation device can input the sample semantic expression data into the initial evaluation model for evaluation processing to obtain a predicted evaluation result, which is a predicted evaluation result for indicating whether there is an abnormality in the sample resource transfer operation. The initial evaluation model can be an extreme gradient boosting model (Extreme Gradient Boosted, XGB) model using gradient boost (Gradient Boost) as a framework, a deep neural network (Deep Neural Networks, DNN) model using a deep learning framework, or other supervised classification models with better processing effects. Through the method provided in the embodiment of the present application, a predicted evaluation result can be obtained, which is convenient for subsequent use of the predicted evaluation result for model training processing.

[0117] It should be noted that if there are multiple reference times (each corresponding to a different time period), the evaluation device can use the method provided in the embodiments of this application to determine the semantic expression data of the resource transfer description data within each reference time, and input the semantic expression data of the resource transfer description data within the multiple reference times into the initial evaluation model for evaluation processing to obtain a predicted evaluation result. The method provided in the embodiments of this application can effectively improve the accuracy of the evaluation results determined by the model.

[0118] S605: Determine difference data according to the prediction evaluation result and the sample evaluation result, and adjust the initial evaluation model using the difference data.

[0119] In an embodiment of the present application, the evaluation device can determine difference data based on the predicted evaluation results and sample evaluation results of the sample object, and the difference data can indicate the prediction accuracy of the initial evaluation model to a certain extent. The evaluation device can use the difference data to adjust the model parameters of the initial evaluation model. It should be noted that when the sample data is sufficient, the evaluation device can determine the predicted evaluation results and sample evaluation results of multiple sample objects, and determine the pending difference data based on the predicted evaluation results and sample evaluation results of each sample object; the evaluation device can determine the difference data based on the pending difference data of multiple sample objects, and use the difference data to train the model. Through the method provided in the embodiment of the present application, the initial evaluation model can be adjusted using difference data, which is beneficial to improving the prediction accuracy of the model. Since the sample semantic expression data can include features of multiple dimensions, the coverage and robustness of the model can also be effectively improved.

[0120] It should be noted that when using the relevant data of multiple sample objects to train the initial evaluation model, the multiple sample objects should include sample objects whose sample evaluation results indicate the presence of anomalies and sample objects whose sample evaluation results indicate the absence of anomalies. That is, the data used for model training should include good sample data and bad sample data. The training dataset corresponding to a sample object can be defined as {x, y}, where x represents the sample semantic expression data determined based on the sample resource transfer-related data associated with the sample object, and y is the sample evaluation result of the sample object (the presence of anomaly or the absence of anomaly), or the object label of the sample object (abnormal object or normal object). The initial evaluation model can be trained using the training datasets corresponding to multiple sample objects to obtain an anomaly evaluation model.

[0121] S606: Determine an abnormality assessment model based on the adjusted initial assessment model.

[0122] In an embodiment of the present application, the evaluation device can use the above-mentioned model training method to adjust the model parameters of the initial evaluation model multiple times. When the training requirements are met (for example, the number of model parameter adjustments reaches a predetermined number or the prediction accuracy of the adjusted initial evaluation model reaches a preset accuracy), the evaluation device can determine the adjusted initial evaluation model as an abnormal evaluation model. Through the method provided in the embodiment of the present application, an abnormal evaluation model can be determined, thereby effectively improving the processing efficiency of the evaluation method and improving the accuracy of the abnormal evaluation results of the object.

[0123] See Figure 7 , which is a schematic diagram of a model training method provided in an embodiment of the present application. Figure 7 As shown, the evaluation device can obtain sample resource transfer related information of the sample resource transfer operation associated with the sample object within the reference time, as well as the sample evaluation result of the sample object; the evaluation device can determine the sample resource transfer description data based on the sample resource transfer related information, and the sample resource transfer description data may include one or more sample resource transfer description texts, each sample resource transfer description text corresponding to a type of sample feature information. The evaluation device can determine the sample semantic expression data of the sample resource transfer description data, and input the sample semantic expression data of the sample resource transfer description data into the initial evaluation model for evaluation processing to obtain a predicted evaluation result. Difference data can be determined based on the sample evaluation result and the predicted evaluation result of the sample object, and the model parameters of the initial evaluation model can be adjusted using the difference data to obtain an adjusted initial evaluation model. An abnormal evaluation model can be determined based on the adjusted initial evaluation model. Through the method provided in the embodiment of the present application, a model training process can be implemented, and the prediction accuracy of the model can be effectively improved, which is conducive to improving the accuracy of the abnormal evaluation results.

[0124] Through the model training method provided in the embodiment of the present application, model training can be carried out using sample semantic expression data including features of multiple dimensions, effectively improving the coverage and robustness of the evaluation model; model training can be implemented using predicted evaluation results and sample evaluation results, effectively ensuring the prediction accuracy of the trained evaluation model, which is conducive to obtaining accurate abnormal evaluation results using the trained evaluation model, thereby ensuring the security of resources, terminating abnormal resource transfer operations in a timely manner, and reducing resource losses.

[0125] See Figure 8 , Figure 8 This is a block diagram of an evaluation device provided in an embodiment of the present application. The device includes:

[0126] An acquiring unit 801 is configured to acquire resource transfer related information of a resource transfer operation associated with an object to be evaluated within a target time;

[0127] a determining unit 802 configured to determine resource transfer description data based on the resource transfer related information, the resource transfer description data being used to indicate characteristic information of the resource transfer operation of the object to be evaluated within the target time, the characteristic information including one or more of a time period of the resource transfer, a type of the resource transfer object, a resource transfer type, location indication information of the resource transfer object, a resource transfer method, and a number of resource transfers;

[0128] The determining unit 802 is further configured to determine semantic expression data of the resource transfer description data;

[0129] The processing unit 803 is configured to perform evaluation processing based on the semantic expression data of the resource transfer description data to obtain an abnormality evaluation result, wherein the abnormality evaluation result is used to indicate whether there is an abnormality in the resource transfer operation of the object to be evaluated within the target time.

[0130] In one embodiment, when the processing unit 803 performs evaluation processing based on the semantic expression data of the resource transfer description data and obtains an abnormal evaluation result, it is specifically used to: input the semantic expression data of the resource transfer description data into an abnormal evaluation model for processing to obtain an abnormal evaluation result; wherein, the abnormal evaluation model is obtained by adjusting the initial evaluation model based on the difference data, and the difference data is determined based on the predicted evaluation result and the sample evaluation result; the predicted evaluation result is obtained by evaluating and processing the sample semantic expression data using the initial evaluation model, the sample semantic expression data is the semantic expression data of the sample resource transfer description data, the sample resource transfer description data is determined based on the sample resource transfer related information, and the sample resource transfer related information is the resource transfer related information of the sample resource transfer operation associated with the sample object within the reference time; the predicted evaluation result is a predicted evaluation result for indicating whether the sample resource transfer operation has an abnormality, and the sample evaluation result is a marked evaluation result for indicating whether the sample resource transfer operation has an abnormality.

[0131] In one embodiment, the resource transfer description data includes M resource transfer description texts, each of the resource transfer description texts corresponds to a type of feature information, M is a positive integer, and the determination unit 802, when determining the semantic expression data of the resource transfer description data, is specifically used to: determine the text feature vectors of each of the resource transfer description texts included in the resource transfer description data; and determine the semantic expression data of the resource transfer description data based on the text feature vectors of the resource transfer description texts included in the resource transfer description data.

[0132] In one embodiment, when determining the semantic expression data of the resource transfer description data based on the text feature vector of the resource transfer description text included in the resource transfer description data, the determination unit 802 is specifically used to: determine whether the resource transfer description text included in the resource transfer description data contains abnormal resource transfer description text, the abnormal resource transfer description text is the resource transfer description text in the abnormal text data set, and the resource transfer description text in the abnormal text data set corresponds to the characteristic information of the abnormal resource transfer operation; if the resource transfer description text included in the resource transfer description data does not contain the abnormal resource transfer description text, then perform average weighted processing on the text feature vector of the resource transfer description text included in the resource transfer description data; and determine the semantic expression data of the resource transfer description data based on the average weighted processing result.

[0133] In one embodiment, the determination unit 802 is further used to: if the resource transfer description text included in the resource transfer description data contains the abnormal resource transfer description text, perform non-average weighted processing on the text feature vector of the resource transfer description text included in the resource transfer description data; determine the semantic expression data of the resource transfer description data based on the non-average weighted processing result; wherein the weight data of the abnormal resource transfer description text is greater than the weight data of the normal resource transfer description text, and the normal resource transfer description text is the resource transfer description text in the resource transfer description data except the abnormal resource transfer description text.

[0134] In one embodiment, when determining the text feature vectors of each resource transfer description text included in the resource transfer description data, the determination unit 802 is specifically used to: query a database to determine whether there is a matching text feature vector that matches the target resource transfer description text in the database, the database including a mapping relationship between sample resource transfer description texts and text feature vectors, and the target resource transfer description text is any resource transfer description text included in the resource transfer description data; if there is a matching text feature vector that matches the target resource transfer description text in the database, the matching text feature vector is determined as the text feature vector of the target resource transfer description text; if there is no matching text feature vector that matches the target resource transfer description text in the database, text feature extraction processing is performed on the target resource transfer description text to obtain the text feature vector of the target resource transfer description text.

[0135] In one embodiment, the number of resource transfer operations associated with the object to be evaluated within the target time is N, where N is a positive integer. When the determination unit 802 determines the resource transfer description data based on the resource transfer related information, it is specifically used to: perform statistical analysis on the resource transfer related information of the N resource transfer operations to obtain statistical analysis results, and determine the resource transfer description data based on the statistical analysis results; wherein the resource transfer related information includes one or more of resource transfer time, resource transfer object, resource transfer type, location of resource transfer object, and resource transfer method.

[0136] In one embodiment, the processing unit 803 is further used to: if the abnormal evaluation result indicates that there is an abnormality in the resource transfer operation of the object to be evaluated within the target time, determine whether the resource transfer operation of the object to be evaluated within the target time is completed; if the resource transfer operation of the object to be evaluated within the target time is completed, suspend the resource transfer authority of the object to be evaluated; if the resource transfer operation of the object to be evaluated within the target time is not completed, terminate the resource transfer operation that has not been completed.

[0137] It is to be understood that the function of each functional unit of the evaluation device of the embodiment of the present application can be specifically implemented according to the evaluation method in the above-mentioned method embodiment, and its specific implementation process can refer to the relevant description in the above-mentioned evaluation method embodiment, and will not be repeated here. In the embodiment of the present application, the term "module" or "unit" refers to a computer program or a part of a computer program with a predetermined function, and works together with other related parts to achieve a predetermined goal, and can be implemented in whole or in part by using software, hardware (such as processing circuit or memory) or a combination thereof. Similarly, a processor (or multiple processors or memories) can be used to implement one or more modules or units. In addition, each module or unit can be a part of the overall module or unit that includes the module or unit function.

[0138] Through the evaluation device provided in the embodiment of the present application, resource transfer description data can be determined based on resource transfer related information. The resource transfer description data can describe the characteristics of the resource transfer operation to a certain extent, and is not limited to the number of resource transfers involved in the resource transfer operation; the semantic expression data of the resource transfer description data can be determined, and the semantic expression data can be a semantic feature vector, that is, the method provided in the present application can convert the description data into a semantic feature vector, which is convenient for subsequent analysis based on the feature vector; since the resource transfer description data can indicate the feature information of one or more resource transfer operations, the semantic expression data can include the semantic features of one or more feature information, and the abnormal evaluation result can be determined based on the semantic expression data. The abnormal evaluation result can more accurately indicate whether there is an abnormality in the resource transfer operation of the object, effectively improving the accuracy of the evaluation result; the abnormal evaluation result can be determined by using a model, effectively improving the processing efficiency of the evaluation method; corresponding processing can be performed based on the abnormal evaluation result, effectively ensuring the security of the resources and effectively reducing the loss of resources.

[0139] See Figure 9 , Figure 9 A structural block diagram of a computer device provided in an embodiment of the present application. Figure 9 The computer device shown can be the above Figure 1 The evaluation device 102 in the embodiment of the present application includes a computer device 901, a communication interface 902, and a memory 903. The processor 901, the communication interface 902, and the memory 903 may be connected via a bus or other means. The embodiment of the present application uses a bus connection as an example.

[0140] Among them, the processor 901 (or CPU (Central Processing Unit)) is the computing core and control core of the computer device. It can parse various instructions within the computer device and process various data of the computer device. For example, the CPU can be used to parse the power on and off instructions sent by the user to the computer device and control the computer device to perform power on and off operations; for another example, the CPU can transmit various interactive data between the internal structures of the computer device, etc. The communication interface 902 can optionally include a standard wired interface or a wireless interface (such as Wi-Fi, mobile communication interface, etc.), which is controlled by the processor 901 to send and receive data. The memory 903 (Memory) is a memory device in the computer device for storing programs and data. It can be understood that the memory 903 here can include both the built-in memory of the computer device and the extended memory supported by the computer device. The memory 903 provides storage space, which stores the operating system of the computer device, which may include but is not limited to: Android system, iOS system, Windows Phone system, etc., and this application is not limited to this.

[0141] In the embodiment of the present application, the processor 901 performs the following operations by running the executable program code in the memory 903:

[0142] Obtain resource transfer related information of resource transfer operations associated with the object to be evaluated within the target time;

[0143] Determining resource transfer description data based on the resource transfer related information, the resource transfer description data being used to indicate characteristic information of the resource transfer operation of the object to be evaluated within the target time, the characteristic information including one or more of a time period of the resource transfer, a type of the resource transfer object, a resource transfer type, location indication information of the resource transfer object, a resource transfer method, and a number of resource transfers;

[0144] Determining semantic expression data of the resource transfer description data;

[0145] An evaluation process is performed based on the semantic expression data of the resource transfer description data to obtain an abnormality evaluation result, where the abnormality evaluation result is used to indicate whether there is an abnormality in the resource transfer operation of the object to be evaluated within the target time.

[0146] In one embodiment, when the processor 901 performs evaluation processing based on the semantic expression data of the resource transfer description data and obtains an abnormal evaluation result, it is specifically used to: input the semantic expression data of the resource transfer description data into an abnormal evaluation model for processing to obtain an abnormal evaluation result; wherein, the abnormal evaluation model is obtained by adjusting the initial evaluation model based on the difference data, and the difference data is determined based on the predicted evaluation result and the sample evaluation result; the predicted evaluation result is obtained by evaluating and processing the sample semantic expression data using the initial evaluation model, the sample semantic expression data is the semantic expression data of the sample resource transfer description data, the sample resource transfer description data is determined based on the sample resource transfer related information, and the sample resource transfer related information is the resource transfer related information of the sample resource transfer operation associated with the sample object within the reference time; the predicted evaluation result is a predicted evaluation result for indicating whether the sample resource transfer operation has an abnormality, and the sample evaluation result is a marked evaluation result for indicating whether the sample resource transfer operation has an abnormality.

[0147] In one embodiment, the resource transfer description data includes M resource transfer description texts, each of the resource transfer description texts corresponds to a type of feature information, M is a positive integer, and when determining the semantic expression data of the resource transfer description data, the processor 901 is specifically used to: determine the text feature vectors of each of the resource transfer description texts included in the resource transfer description data; and determine the semantic expression data of the resource transfer description data based on the text feature vectors of the resource transfer description texts included in the resource transfer description data.

[0148] In one embodiment, when the processor 901 determines the semantic expression data of the resource transfer description data based on the text feature vector of the resource transfer description text included in the resource transfer description data, it is specifically used to: determine whether the resource transfer description text included in the resource transfer description data contains abnormal resource transfer description text, the abnormal resource transfer description text is the resource transfer description text in the abnormal text data set, and the resource transfer description text in the abnormal text data set corresponds to the characteristic information of the abnormal resource transfer operation; if the resource transfer description text included in the resource transfer description data does not contain the abnormal resource transfer description text, then perform average weighted processing on the text feature vector of the resource transfer description text included in the resource transfer description data; and determine the semantic expression data of the resource transfer description data based on the average weighted processing result.

[0149] In one embodiment, the processor 901 is further used to: if the resource transfer description text included in the resource transfer description data contains the abnormal resource transfer description text, perform non-average weighted processing on the text feature vector of the resource transfer description text included in the resource transfer description data; determine the semantic expression data of the resource transfer description data based on the non-average weighted processing result; wherein the weight data of the abnormal resource transfer description text is greater than the weight data of the normal resource transfer description text, and the normal resource transfer description text is the resource transfer description text in the resource transfer description data except the abnormal resource transfer description text.

[0150] In one embodiment, when determining the text feature vectors of each resource transfer description text included in the resource transfer description data, the processor 901 is specifically used to: query the database to determine whether there is a matching text feature vector that matches the target resource transfer description text in the database, the database includes a mapping relationship between sample resource transfer description texts and text feature vectors, and the target resource transfer description text is any resource transfer description text included in the resource transfer description data; if there is a matching text feature vector that matches the target resource transfer description text in the database, the matching text feature vector is determined as the text feature vector of the target resource transfer description text; if there is no matching text feature vector that matches the target resource transfer description text in the database, text feature extraction processing is performed on the target resource transfer description text to obtain the text feature vector of the target resource transfer description text.

[0151] In one embodiment, the number of resource transfer operations associated with the object to be evaluated within the target time is N, where N is a positive integer. When the processor 901 determines the resource transfer description data based on the resource transfer related information, it is specifically used to: perform statistical analysis on the resource transfer related information of the N resource transfer operations to obtain statistical analysis results, and determine the resource transfer description data based on the statistical analysis results; wherein the resource transfer related information includes one or more of resource transfer time, resource transfer object, resource transfer type, location of resource transfer object, and resource transfer method.

[0152] In one embodiment, the processor 901 is further used to: if the abnormal evaluation result indicates that there is an abnormality in the resource transfer operation of the object to be evaluated within the target time, determine whether the resource transfer operation of the object to be evaluated within the target time is completed; if the resource transfer operation of the object to be evaluated within the target time is completed, suspend the resource transfer authority of the object to be evaluated; if the resource transfer operation of the object to be evaluated within the target time is not completed, terminate the resource transfer operation that has not been completed.

[0153] In a specific implementation, the processor 901, communication interface 902, and memory 903 described in the embodiment of the present application can execute the implementation of the evaluation device 102 described in an evaluation method provided in an embodiment of the present application, and can also execute the implementation described in an evaluation device provided in an embodiment of the present application, which will not be repeated here.

[0154] Through the computer device provided by the embodiment of the present application, resource transfer description data can be determined based on resource transfer related information. The resource transfer description data can describe the characteristics of the resource transfer operation to a certain extent, and is not limited to the number of resource transfers involved in the resource transfer operation; the semantic expression data of the resource transfer description data can be determined, and the semantic expression data can be a semantic feature vector, that is, the method provided by the present application can convert the description data into a semantic feature vector, which is convenient for subsequent analysis based on the feature vector; since the resource transfer description data can indicate the feature information of one or more resource transfer operations, the semantic expression data can include the semantic features of one or more feature information, and the abnormal evaluation result can be determined based on the semantic expression data. The abnormal evaluation result can more accurately indicate whether there is an abnormality in the resource transfer operation of the object, effectively improving the accuracy of the evaluation result; the abnormal evaluation result can be determined using a model, effectively improving the processing efficiency of the evaluation method; corresponding processing can be performed based on the abnormal evaluation result, effectively ensuring the security of the resources and effectively reducing the loss of resources.

[0155] The present application also provides a computer-readable storage medium storing computer instructions that, when executed on a computer device, cause the computer device to execute the steps of each method embodiment of the present application to implement the evaluation method provided in the present application. The specific implementation method is described above and will not be repeated here.

[0156] The present application also provides a computer program product, which includes a computer program or computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer program or computer instructions from the computer-readable storage medium, and the processor executes the computer program or computer instructions, causing the computer device to perform the steps of each method embodiment of the present application to implement the evaluation method provided in the present application embodiment. The specific implementation method can be found in the above description and will not be repeated here.

[0157] It should be noted that for the aforementioned various method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that this application is not limited by the order of the actions described, because according to this application, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by this application.

[0158] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments. The technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a computer, a server or a network device, etc., specifically a processor in a computer device) to execute all or part of the steps of the above methods in each embodiment of the present application. Among them, the aforementioned storage medium may include: U disk, mobile hard disk, magnetic disk, optical disk, read-only memory (English: Read-Only Memory, abbreviated: ROM) or random access memory (English: Random Access Memory, abbreviated: RAM) and other media that can store program codes.

[0159] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the above embodiments, it should be understood that the technical solutions recorded in the above embodiments can still be modified, or some of the technical features therein can be replaced by equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An evaluation method, characterized in that: The method comprises: Obtain resource transfer related information of resource transfer operations associated with the object to be evaluated within the target time; Determining resource transfer description data based on the resource transfer related information, the resource transfer description data being used to indicate characteristic information of the resource transfer operation of the object to be evaluated within the target time, the characteristic information including one or more of a time period of the resource transfer, a type of the resource transfer object, a resource transfer type, location indication information of the resource transfer object, a resource transfer method, and a number of resource transfers; Determining semantic expression data of the resource transfer description data; An evaluation process is performed based on the semantic expression data of the resource transfer description data to obtain an abnormality evaluation result, where the abnormality evaluation result is used to indicate whether there is an abnormality in the resource transfer operation of the object to be evaluated within the target time.

2. The method according to claim 1, characterized in that The performing evaluation processing based on the semantic expression data of the resource transfer description data to obtain an abnormal evaluation result includes: Inputting the semantic expression data of the resource transfer description data into an anomaly assessment model for processing to obtain an anomaly assessment result; In which, the abnormal evaluation model is obtained by adjusting the initial evaluation model according to the difference data, and the difference data is determined according to the predicted evaluation result and the sample evaluation result; the predicted evaluation result is obtained by evaluating and processing the sample semantic expression data using the initial evaluation model, and the sample semantic expression data is the semantic expression data of the sample resource transfer description data, and the sample resource transfer description data is determined according to the sample resource transfer related information, and the sample resource transfer related information is the resource transfer related information of the sample resource transfer operation associated with the sample object within the reference time; the predicted evaluation result is a predicted evaluation result for indicating whether the sample resource transfer operation has an abnormality, and the sample evaluation result is a marked evaluation result for indicating whether the sample resource transfer operation has an abnormality.

3. The method according to claim 1 or 2, characterized in that The resource transfer description data includes M resource transfer description texts, each of which corresponds to one type of feature information, where M is a positive integer. The semantic expression data for determining the resource transfer description data includes: Determining a text feature vector of each resource transfer description text included in the resource transfer description data; The semantic expression data of the resource transfer description data is determined according to a text feature vector of the resource transfer description text included in the resource transfer description data.

4. The method according to claim 3, characterized in that The determining, based on the text feature vector of the resource transfer description text included in the resource transfer description data, the semantic expression data of the resource transfer description data comprises: determining whether the resource transfer description text included in the resource transfer description data contains abnormal resource transfer description text, wherein the abnormal resource transfer description text is a resource transfer description text in an abnormal text data set, and the resource transfer description text in the abnormal text data set corresponds to characteristic information of an abnormal resource transfer operation; If the resource transfer description text included in the resource transfer description data does not contain the abnormal resource transfer description text, performing average weighted processing on the text feature vectors of the resource transfer description text included in the resource transfer description data; The semantic expression data of the resource transfer description data is determined according to the average weighted processing result.

5. The method according to claim 4, characterized in that The method further comprises: If the resource transfer description text included in the resource transfer description data contains the abnormal resource transfer description text, performing non-average weighted processing on the text feature vector of the resource transfer description text included in the resource transfer description data; Determining semantic expression data of the resource transfer description data according to the non-average weighted processing result; The weight data of the abnormal resource transfer description text is greater than the weight data of the normal resource transfer description text, and the normal resource transfer description text is the resource transfer description text in the resource transfer description data except the abnormal resource transfer description text.

6. The method according to claim 3, characterized in that The determining of the text feature vectors of the resource transfer description texts included in the resource transfer description data includes: querying a database to determine whether there is a matching text feature vector in the database that matches the target resource transfer description text, wherein the database includes a mapping relationship between sample resource transfer description texts and text feature vectors, and the target resource transfer description text is any resource transfer description text included in the resource transfer description data; If a matching text feature vector matching the target resource transfer description text exists in the database, determining the matching text feature vector as the text feature vector of the target resource transfer description text; If there is no matching text feature vector matching the target resource transfer description text in the database, a text feature extraction process is performed on the target resource transfer description text to obtain a text feature vector of the target resource transfer description text.

7. The method according to claim 1 or 2, characterized in that The number of resource transfer operations associated with the object to be evaluated within the target time is N, where N is a positive integer. The resource transfer description data is determined based on the resource transfer related information, including: Performing statistical analysis on resource transfer related information of N resource transfer operations to obtain statistical analysis results, and determining resource transfer description data based on the statistical analysis results; The resource transfer related information includes one or more of resource transfer time, resource transfer object, resource transfer type, location of resource transfer object and resource transfer method.

8. The method according to claim 1 or 2, characterized in that The method further comprises: If the abnormality assessment result indicates that the resource transfer operation of the object to be assessed within the target time has an abnormality, determining whether the resource transfer operation of the object to be assessed within the target time has been completed; If the resource transfer operation of the object to be evaluated is completed within the target time, the resource transfer authority of the object to be evaluated is suspended; If the resource transfer operation of the object to be evaluated is not completed within the target time, the uncompleted resource transfer operation is terminated.

9. An evaluation device, characterized in that The device comprises: An acquisition unit, configured to acquire resource transfer related information of a resource transfer operation associated with an object to be evaluated within a target time; a determining unit, configured to determine resource transfer description data based on the resource transfer related information, the resource transfer description data being used to indicate characteristic information of the resource transfer operation of the object to be evaluated within the target time, the characteristic information including one or more of a time period of the resource transfer, a type of the resource transfer object, a resource transfer type, location indication information of the resource transfer object, a resource transfer method, and a number of resource transfers; The determining unit is further configured to determine semantic expression data of the resource transfer description data; A processing unit is used to perform evaluation processing based on the semantic expression data of the resource transfer description data to obtain an abnormality evaluation result, wherein the abnormality evaluation result is used to indicate whether there is an abnormality in the resource transfer operation of the object to be evaluated within the target time.

10. A computer device, characterized in that: include: A processor, a communication interface and a memory, wherein the processor, the communication interface and the memory are connected to each other, wherein the memory stores computer instructions, and the processor is used to call the computer instructions to implement the evaluation method according to any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, which, when executed on a computer device, enable the computer device to implement the evaluation method according to any one of claims 1 to 8.

12. A computer program product, characterized in that The computer program product comprises a computer program or computer instructions, and when the computer program or computer instructions are executed by a processor, the evaluation method according to any one of claims 1 to 8 is implemented.

Images