Data processing method and device, computer equipment, medium and product
By obtaining and processing real-time and historical data feature information of transaction objects, and using the multi-task model of the task processing cluster for scoring integration, the problem of insufficient timeliness and accuracy of classification results in transaction security management is solved, and the effect of transaction security management is improved .
Patent Information
- Application Number
- CN202310542250.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-05-12
- Publication Date
- 2024-11-12
AI Technical Summary
In the transaction security management of existing technology, the timeliness and accuracy of classification results are poor, resulting in poor transaction security management results.
When the data processing conditions are triggered, the feature information of real-time and historical data is obtained from the database based on the object identification information, and the feature processing is performed. Multiple task models of the task processing cluster are called to process the target feature information, and the standard scoring results are integrated. To determine the transaction operation management of the object.
It realizes timely and accurate classification results of transaction objects, improves the effectiveness of transaction security management, and ensures the accurate real-time rating and accuracy of objects.
Smart Images

Figure CN118941286A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of Internet technology, and in particular to a data processing method, a data processing device, a computer equipment, a computer-readable storage medium, and a computer program product. Background Art
[0002] With the explosive development of the big data era, a series of security issues have emerged. Among them, a typical security issue is the transaction security of individuals, merchants and other objects.
[0003] At present, most of the transaction security issues for the above objects are solved by analyzing the historical transaction information of the objects to determine the classification results of the objects, and then managing the transactions securely according to the classification results of the objects. However, this analysis and processing method of transaction data generally has the problem of poor timeliness and accuracy of classification results, which in turn makes the transaction security management effect poor. Therefore, how to determine the classification results of objects in a timely and accurate manner has become a technical problem that needs to be solved urgently. Summary of the invention
[0004] The embodiments of the present application provide a data processing method, apparatus, computer equipment, medium, and product, which can timely and accurately determine the classification results of objects, thereby improving the security management effect of transactions.
[0005] On the one hand, an embodiment of the present application discloses a data processing method, the method comprising:
[0006] When a data processing condition is triggered, basic feature information and derived feature information of the first object are obtained from a database based on identification information of the first object, wherein the basic feature information is obtained by performing feature extraction processing on real-time data of the first object, and the derived feature information is obtained by performing feature derivation processing on historical data of the first object;
[0007] Performing feature processing on the basic feature information and the derived feature information to obtain target feature information of the first object;
[0008] Calling N task processing models included in the task processing cluster to process the target feature information to obtain N standard scoring results, where N is an integer greater than 1;
[0009] The N standard scoring results are integrated to obtain a target scoring result of the first object, and the target scoring result is used to manage the transaction operation of the first object.
[0010] In one aspect, an embodiment of the present application discloses a data processing device, the device comprising:
[0011] an acquisition unit, configured to acquire basic feature information and derived feature information of the first object from a database based on identification information of the first object when a data processing condition is triggered, wherein the basic feature information is obtained by performing feature extraction processing on real-time data of the first object, and the derived feature information is obtained by performing feature derivation processing on historical data of the first object;
[0012] A processing unit, configured to perform feature processing on the basic feature information and the derived feature information to obtain target feature information of the first object;
[0013] The processing unit is further used to call N task processing models included in the task processing cluster to process the target feature information to obtain N standard scoring results, where N is an integer greater than 1;
[0014] The processing unit is further used to integrate the N standard scoring results to obtain a target scoring result of the first object, and the target scoring result is used to manage the transaction operation of the first object.
[0015] On the one hand, an embodiment of the present application discloses a computer device, which includes a processor suitable for implementing one or more computer programs; and a computer storage medium, wherein the computer storage medium stores one or more computer programs, and the one or more computer programs are suitable for being loaded by the processor and executed with the above-mentioned data processing method.
[0016] In one aspect, the present application discloses a computer-readable storage medium storing one or more computer programs, wherein the one or more computer programs are suitable for being loaded by a processor and executing the above-mentioned data processing method.
[0017] The present application discloses a computer program product, which includes a computer program stored in a computer-readable storage medium. A processor of a computer device reads the computer program from the computer-readable storage medium, and the processor executes the computer program, so that the computer device executes the above-mentioned data processing method.
[0018] In the embodiment of the present application, when the data processing condition is triggered, the basic feature information and derived feature information of the first object are obtained from the database based on the identification information of the first object. The basic feature information is obtained by performing feature extraction processing on the real-time data of the first object, and the derived feature information is obtained by performing feature derivation processing on the historical data of the first object; since the basic feature information and the derived feature information are real-time data and historical data respectively, the embodiment of the present application can achieve timeliness. The basic feature information and the derived feature information are feature processed to obtain the target feature information of the first object. Processing is performed immediately after acquisition, rather than offline processing, which further reflects the timeliness of the embodiment of the present application. The N task processing models included in the task processing cluster are called to process the target feature information to obtain N standard scoring results, where N is an integer greater than 1. The N standard scoring results are integrated and processed to obtain the target scoring result of the first object, and the target scoring result is used to manage the transaction operation of the first object. In this process, since the target feature information is processed by the task processing cluster, each task processing model has a different division of labor, so the first object can be analyzed from multiple dimensions, thereby improving the accuracy of the classification result of the object. As can be seen from the above, the embodiments of the present application can ensure the timeliness of object classification results by combining real-time data and historical data, and can achieve quasi-real-time rating of objects; at the same time, the target feature information is processed in multiple dimensions by the task processing cluster, thereby ensuring the accuracy of the object classification results. 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 drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0020] Figure 1 It is a schematic diagram of a network architecture of a data processing system disclosed in an embodiment of the present application;
[0021] Figure 2 It is a flowchart of a data processing method disclosed in an embodiment of the present application;
[0022] Figure 3 It is a flowchart of another data processing method disclosed in an embodiment of the present application;
[0023] Figure 4 is a schematic diagram of a task processing cluster disclosed in an embodiment of the present application;
[0024] Figure 5It is a flowchart of another data processing method disclosed in an embodiment of the present application;
[0025] Figure 6 is a structural schematic diagram of a data processing device disclosed in an embodiment of the present application;
[0026] Figure 7 It is a structural schematic diagram of a computer device disclosed in an embodiment of the present application. DETAILED DESCRIPTION
[0027] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0028] In order to more clearly understand the technical solution provided by the embodiments of the present application, some key terms involved in the embodiments of the present application are first introduced here:
[0029] 1. Artificial Intelligence
[0030] Artificial Intelligence (AI) refers to the theory, method, technology and application system of simulating, extending and expanding human intelligence, perceiving the environment, acquiring knowledge and using knowledge to obtain the best results using digital computers or machines controlled by digital computers. AI technology is a comprehensive subject, and the fields involved are relatively wide. Among them, basic artificial intelligence technologies generally include technologies such as sensors, special artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. Artificial intelligence software technology mainly includes computer vision technology, speech processing technology, natural language processing technology, and machine learning / deep learning. The data processing method provided in the embodiment of the present application mainly relates to machine learning (ML) technology in AI technology, which mainly uses machine learning methods to build task processing clusters.
[0031] Machine learning is a multi-disciplinary subject that involves probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory and other disciplines. It specializes in studying how computers simulate or implement human learning behavior to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Machine learning is the core of artificial intelligence and the fundamental way to make computers intelligent. Its applications are spread across all areas of artificial intelligence. Machine learning and deep learning usually include artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and self-learning.
[0032] 2. Cloud Technology
[0033] Cloud technology refers to a hosting technology that unifies hardware, software, network and other resources in a wide area network or local area network to achieve data computing, storage, processing and sharing. Cloud technology is a general term for network technology, information technology, integration technology, management platform technology, application technology, etc. based on the cloud computing business model application. It can form a resource pool, which is used on demand and is flexible and convenient. Cloud computing technology will become an important support. Cloud storage is a new concept extended and developed from the concept of cloud computing. Distributed cloud storage system (hereinafter referred to as storage system) refers to a storage system that uses cluster applications, grid technology and distributed storage file systems to bring together a large number of different types of storage devices (storage devices are also called storage nodes) in the network through application software or application interfaces to work together and provide data storage and business access functions to the outside world. In short, a database can be regarded as an electronic filing cabinet - a place to store electronic files. Users can add, query, update, delete and other operations on the data in the files. The so-called "database" is a collection of data that is stored together in a certain way, can be shared with multiple users, has as little redundancy as possible, and is independent of the application program.
[0034] 3. Blockchain
[0035] Blockchain is a new application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of the information (anti-counterfeiting) and generate the next block.
[0036] The embodiment of the present application provides a data processing scheme, which includes: ① when the data processing condition is triggered, the basic feature information and derived feature information of the first object are obtained from the database based on the identification information of the first object, the basic feature information is obtained by performing feature extraction processing on the real-time data of the first object, and the derived feature information is obtained by performing feature derivation processing on the historical data of the first object; since the basic feature information and the derived feature information are real-time data and historical data respectively, the embodiment of the present application can achieve timeliness. ② The basic feature information and the derived feature information are feature processed to obtain the target feature information of the first object. Processing is performed immediately after acquisition, rather than offline processing, which further reflects the timeliness of the embodiment of the present application. ③ The N task processing models included in the task processing cluster are called to process the target feature information to obtain N standard scoring results, where N is an integer greater than 1. ④ The N standard scoring results are integrated and processed to obtain the target scoring result of the first object, and the target scoring result is used to manage the transaction operation of the first object. In these two steps, since the target feature information is processed by the task processing cluster, each task processing model has different division of labor, so the first object can be analyzed from multiple dimensions, thereby improving the accuracy of the classification result of the object. As can be seen from the above, the embodiment of the present application can ensure the timeliness of the object classification results by combining real-time data and historical data. At the same time, by processing the target feature information in multiple dimensions through the task processing cluster, the accuracy of the object classification results is ensured, thereby achieving quasi-real-time rating of the object.
[0037] In one possible implementation, the data processing method provided in the embodiment of the present application can be implemented based on artificial intelligence (AI) technology. Specifically, it relates to machine learning (ML) technology in AI technology, for example, a task processing cluster is constructed using a machine learning method. In one possible implementation, the data processing method provided in the present application can be implemented based on cloud technology and / or blockchain technology. Specifically, it may involve one or more of cloud storage (Cloud storage), cloud database (Cloud Database), and big data (Big data) in cloud technology. For example, the data required to execute the data processing method (such as basic feature information and derived feature information of the first object) is obtained from the cloud database.
[0038] In a possible implementation, the data processing method provided by the present application can also be implemented based on blockchain technology. For example, the data required to execute the data processing method can be stored in the form of blocks on the blockchain; the data generated by executing the data processing method (such as the target feature information of the first object, the target scoring result of the first object, etc.) can be stored in the form of blocks on the blockchain; in addition, the data processing device that executes the data processing method can be a node device in the blockchain network.
[0039] The data processing system proposed in the embodiment of the present application will be introduced in detail below.
[0040] See also Figure 1, is a schematic diagram of the architecture of a data processing system disclosed in an embodiment of the present application, and the data processing system may include a first server 101, a second server 102, and a client terminal 103. The first server 101 is mainly responsible for the entry management of objects, such as providing account registration services to multiple objects. After registering an account, the object can conduct transactions such as transfer, payment, and refund with other objects. The first server 101 can record the entry data of the object and send the object entry data within a period of time to the second server 102, so that the second server can obtain the identification information of the first object from the object entry data. The second server 102 is mainly used to obtain the relevant feature information of the first object (including basic feature information with high real-time performance and derived feature information derived from features) from the database in a timely manner according to the identification information of the first object, and analyze and process the basic feature information and derived feature information of the first object to obtain the target scoring result of the first object. According to the target scoring result, the classification result of the first object in the transaction security dimension can be determined, and further, the transaction authority of the first object can be managed according to the classification result, including the closure, downgrade, and reopening of the transaction authority. The client terminal 103 acts directly on the object, including the object logging into the object's account information through the client terminal 103 and conducting transactions through the client terminal 103. In one embodiment, the client terminal 103 includes, but is not limited to: smart phones, tablet computers, smart wearable devices, smart voice interaction devices, smart home appliances, personal computers, vehicle terminals, smart cameras, virtual reality devices (such as AR (Augmented Reality, augmented reality) devices) and other devices, and this application does not limit this. The first server 101 and the second server 102 can be independent physical servers, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN (Content Delivery Network, content distribution network), and big data and artificial intelligence platforms, and this embodiment of the application does not limit this. The first server 101, the second server 102 and the client terminal 103 can establish a direct communication connection through wired communication, or can establish an indirect communication connection through wireless communication, and this embodiment of the application does not limit this.
[0041] In the embodiment of the present application, in combination with the data processing system, the data processing method may include: when the first server 101 sends the object entry data of the first object to the second server 102, determining that the data processing condition is triggered, the first server 101 obtains the identification information of the first object from the object entry data, and then obtains the basic feature information and derived feature information of the first object from the database based on the identification information of the first object, the basic feature information is obtained by feature extraction processing of the real-time data of the first object, and the derived feature information is obtained by feature derivation processing of the historical data of the first object. Then, the basic feature information and the derived feature information are processed to obtain the target feature information of the first object; the N task processing models included in the task processing cluster are called to process the target feature information to obtain N standard scoring results, where N is an integer greater than 1. Finally, the N standard scoring results are integrated to obtain the target scoring result of the first object, and the target scoring result is used to manage the transaction operation of the first object. Optionally, the target scoring result may be sent to the first server 101 first, and the first server 101 then manages the transaction operation corresponding to the client terminal 103. Through the data processing method provided in the embodiment of the present application, the classification result of the object (such as the first object) can be determined timely and accurately.
[0042] In one implementation, the above data processing system can be combined with an actual application scenario. Assume that in an actual application scenario, there is an object A that is first settled in the application platform (the application platform can support transactions, such as payment applications) through the client terminal 103, and the object A enters the corresponding object settlement data, including but not limited to the identification information of the object A. The corresponding first server 101 in the application platform (which can be understood as the backend system of the application platform) sends the object settlement data of the object A to the second server 102, and the second server 102 obtains the identification information of the object A through the object settlement data, and obtains the basic feature information and derived feature information of the object A from different databases based on the identification information of the object A; then the basic feature information and the derived feature information are processed to obtain the target feature information of the object A; then the N task processing models included in the task processing cluster are called to process the target feature information to obtain N standard scoring results; finally, the N standard scoring results are integrated to obtain the target scoring result of the object A, and the target scoring result is used to manage the transaction operation of the object A. Furthermore, after object A is registered in the application platform, object A can be scored again during the subsequent use of object A. For example, when a preset interval (such as every few seconds or minutes) is reached, the data processing condition is determined to be triggered, and then the above-mentioned scoring process is executed, so that object A can be scored in seconds, improving the timeliness of the scoring; and in this way, it can be applied to the scoring of any object, so it is also more versatile.
[0043] In a possible implementation, the second server 102 can be understood as a real-time classification device (also understood as a real-time scoring device), which can complete the classification processing of any object in quasi-real time, for example, obtain a target scoring result, and classify the object into a specified category according to the target scoring result. More specifically, any object is scored in real time to better and more timely manage the transaction operations of each object.
[0044] It should be noted that in the present application, the relevant data involved in the data processing process, such as the identification information, basic feature information, and derived feature information of the first object. When the above embodiments of the present application are applied to specific products or technologies, user permission or consent must be obtained, and the relevant data collection, use, and processing processes must comply with relevant national and regional laws, regulations, and standards, and comply with the principles of legality, legitimacy, and necessity, and do not involve the acquisition of data types prohibited or restricted by laws and regulations. In some optional embodiments, the relevant data involved in the embodiments of the present application are obtained after the object has been authorized separately. In addition, when obtaining the separate authorization of the object, the purpose of the relevant data involved is indicated to the object.
[0045] The data processing method proposed in the embodiment of the present application is introduced in detail below.
[0046] See also Figure 2 , is a flow chart of a data processing method disclosed in an embodiment of the present application. The data processing method can be executed by a computer device, and the computer device can be Figure 1 Corresponding to the second server 102 shown, the data processing method may include but is not limited to the following steps S201-S204:
[0047] S201: When a data processing condition is triggered, basic feature information and derived feature information of the first object are obtained from a database based on identification information of the first object, wherein the basic feature information is obtained by performing feature extraction processing on real-time data of the first object, and the derived feature information is obtained by performing feature derivation processing on historical data of the first object.
[0048] The data processing trigger conditions in the present embodiment include but are not limited to the following two:
[0049] The first is to determine that the data processing condition is triggered when the object entry data sent by the first server is received. The first server may send the object entry data when a settled object is detected, or it may send it at a time interval, for example, once every second or every two seconds. In this way, the computer device (that is, the second server) can grasp the object entry data of the client terminal in a timely manner. Furthermore, each time the object entry data is received, the identification information of the object can be obtained from the object entry data. In the present application, the first object is used as an example for explanation. Optionally, as time goes by, more new objects are settled, so the identification information of the second object, the third object, and so on can also be obtained.
[0050] The second type is to determine the triggering of the data processing condition when the preset interval time is reached. This situation belongs to the data processing condition triggered by the computer device side itself. It can be understood that after scoring the newly settled object once, the embodiment of the present application can score it multiple times, or for some retained objects, the embodiment of the present application can also score them multiple times. The data triggering condition at this time can be set by the developer, for example, triggered by a timer, such as reaching a preset interval time, triggering a data processing, and the preset interval time can be flexibly set. The preset interval time can be one second, two seconds, or one minute, two minutes, and so on. If the preset interval time is one second, the object score is performed once every one second.
[0051] The first object may refer to an individual or a merchant. The identification information of the first object includes but is not limited to a text identifier, a digital identifier, etc. For example, when the first object is an individual, the text identifier may refer to a user nickname; when the first object is a merchant, the text identifier may refer to a merchant name, etc. The digital identifier may refer to the virtual account information of the first object. The first object may register an account, and after registering an account, may conduct transactions such as transfers, payments, and refunds with other objects.
[0052] When the data processing condition is triggered, the basic feature information of the first object is obtained from the first database through the real-time data query interface based on the identification information of the first object. The basic feature information is obtained by performing feature extraction processing on the real-time data of the first object and is stored in the first database. The basic feature information belongs to real-time feature information. The basic feature information may include the account information and transaction information of the first object at the current moment. Based on the real-time feature information, the timeliness of the acquired data can be guaranteed. At the same time, the derived feature information of the first object is obtained from the second database based on the identification information of the first object. The derived feature information is obtained by performing feature derivation processing on the historical data of the first object. The derived feature information can also be understood as an extended feature information, which is obtained by performing derivation processing on some historical data using a preset derivation method. The above-mentioned basic feature information and derived feature information can be understood as being stored in two databases respectively, or in one database. When it is necessary to score the object, it can be directly obtained from the database.
[0053] In a possible implementation, the derived feature information is mainly determined based on the historical data of the first object, and the historical data may include the account information and transaction information of the first object. The transaction information may be the data generated when the first pair of objects transfer, pay, and refund with other objects. The generation process of the derived feature information for the first object may include: obtaining the historical data of the first object, the historical data includes the account information and transaction information of the first object; then using a preset feature derivation method to perform feature derivation processing on the account information and transaction information of the first object to obtain the derived feature information of the first object, and storing the derived feature information of the first object in the second database. Among them, the preset feature derivation method includes one or more of the data analysis model derivation method, the association derivation method, and the clustering derivation method. The above is a schematic explanation of the generation process of the derived feature information of the first object. The derived feature information of any object can be generated according to the above method.
[0054] The following takes the first object as an example to explain the preset feature derivation method:
[0055] 1. Data analysis model derivation method. In the embodiments of the present application, the data analysis model derivation method may refer to the RFM analysis model derivation method. The RFM analysis model is an important tool and means to measure customer value and customer ability to create benefits. Among them, R represents the time interval between the last consumption (Recency), F represents the consumption frequency (Frequency), and M represents the consumption amount (Monetary): the total amount of consumption within a period of time ("a period of time" is usually determined according to different application scenarios).
[0056] In the present application, when the RFM analysis model derivation method is used for feature derivation, the transaction information of the first object is determined according to the account information of the first object, as shown in Table 1, including the transaction business category and the transaction time category. Then the corresponding derivative algorithm is set to combine the three in a many-to-many form, that is, the transaction business category and the transaction time category are processed by the derivative algorithm to obtain the derived feature information. As shown in Table 1, the transaction business categories include commercial payments, corporate account splitting, refunds, and corporate payments to change, etc., the transaction time categories include the past 30 days, the past 60 days, the past 90 days, and the past 180 days, etc., and the derivative algorithms include average (AVG), maximum (MAX), minimum (MIN), standard deviation (STD), etc. In the actual derivative processing process, any one or more derivative algorithms can be used to process the transaction business category and the transaction time category to obtain the derived feature information.
[0057] Table 1
[0058] Transaction business category Trading time category Derivative Algorithm Enterprise Account Sharing Last 30 days AVG Commercial Payments Nearly 60 days MAX Refund Nearly 90 days MIN Business Payment to Change Nearly 180 days STD
[0059] 2. Association derivation method: The account information of other objects associated with the first object is queried based on the account information of the first object, and the transaction information of other objects associated with the first object is queried based on the transaction information of the first object, and then the account information and transaction information of other objects associated with the first object are integrated to obtain the derived feature information of the first object.
[0060] 3. Clustering derivation method. According to the account information of the first object, the account information of other objects associated with it is queried, and according to the transaction information of the first object, the transaction information of other objects associated with it is queried, and then the transaction information of other objects and the transaction of the first object are clustered, and the derived feature information of the first object is determined according to the clustering result.
[0061] In addition to the above-mentioned preset derivation feature methods, optionally, simple polynomial feature derivation methods, cross feature derivation methods, exponential feature derivation methods, etc. can be used for feature derivation processing. Some common derivation methods such as genetic algorithm, ant colony optimization algorithm, particle swarm algorithm, hill climbing method, etc. can also be used for feature derivation processing.
[0062] S202: Perform feature processing on the basic feature information and the derived feature information to obtain target feature information of the first object.
[0063] The processing here may include feature splicing processing and feature conversion processing. After obtaining the basic feature information and derived feature information of the first object, the basic feature information and the derived feature information may be first subjected to feature splicing processing, and then the spliced feature information may be subjected to feature conversion processing to obtain the target feature information of the first object. Alternatively, after obtaining the basic feature information and the derived feature information of the first object, the basic feature information and the derived feature information may be first subjected to feature conversion processing respectively, and then the basic feature information after feature conversion and the derived feature information after feature conversion may be spliced to obtain the target feature information of the first object. In the present application, any one of the above two methods may be used to determine the target feature information.
[0064] Furthermore, the feature splicing processing method includes horizontal splicing, vertical splicing and cross splicing, etc., and the feature conversion processing includes feature format conversion, that is, converting data in one format into data in another format. In the present application, the basic feature information and the derived feature information can be converted into a unified format so that each task processing model in the task processing cluster can be recognized.
[0065] S203: calling N task processing models included in the task processing cluster to process the target feature information to obtain N standard scoring results, where N is an integer greater than 1.
[0066] The concept of cluster indicates that there are at least two task processing models, so N of the N task processing models included in the task processing cluster is an integer greater than 1. In the present application, the task processing models included in the task processing cluster are of different types, including but not limited to annotation registration processing models, non-standard information processing models, and comprehensive processing models.
[0067] Since a single task processing model is not comprehensive and accurate enough in its analysis capability when classifying objects, this application adopts multiple task processing models to classify objects. Multiple task processing models embody a task processing cluster concept, indicating that the cluster includes at least 2 or more task processing models. By processing the relevant data of the object through such a task processing cluster, the relevant data of the object can be comprehensively analyzed, thereby improving the accuracy of object classification.
[0068] In a possible implementation, the sub-feature information corresponding to each task processing model can be determined from the target feature information according to the model type of the N task processing models included in the task processing cluster. That is, the tasks implemented by different task processing models are different, and correspondingly, the feature information required by different types of task processing models is also different. Then, each of the N task processing models is called to score the corresponding sub-feature information to obtain N first scoring results, that is, one task processing model outputs a first scoring result. After each task processing model outputs the first scoring result, in order to ensure that the results of multiple task processing models can be compared at the same scale, the present application sets a standard score calculation rule, which can make the first scoring result of each task processing model a score at the same scale. The specific implementation process includes: ① According to the set M standard score segments, the N first scoring results are divided into intervals to obtain M reference score segments, where M is an integer greater than or equal to 1; the number of reference score segments is consistent with the number of standard score segments, so that the corresponding relationship can be found. ② For any one of the N first scoring results, taking the scoring result to be processed as an example, determine the first score segment in the M reference score segments where the scoring result to be processed is located, for example, the scoring result to be processed processes the i-th score segment in the M reference score segments. ③ Determine the corresponding second score segment from the set M standard score segments according to the first score segment in which the scoring result to be processed is located. For example, if the scoring result to be processed is in the i-th score segment in the M reference score segments, similarly, find the i-th score segment in the M standard score segments, so as to find the corresponding score segments of the two. ④ Use the standard score calculation rule to calculate the lower limit value of the first score segment, the difference between the upper limit value and the lower limit value of the first score segment, the lower limit value of the second score segment, and the scoring result to be processed, and obtain the standard scoring result of the scoring result to be processed. For any one of the N first scoring results, this method can be used for score conversion to obtain N standard scoring results. Among them, the standard score calculation rule for any first scoring result is as follows:
[0069] Standard scoring result = lower limit of the second score segment + (200 / difference between upper and lower limits of the first score segment) * (first scoring result - lower limit of the first score segment)
[0070] For example, the set M standard score segments are divided into 5 segments, namely 0-200, 200-400, 400-600, 600-800, and 800-1000. According to these 5 standard score segments, the N first scoring results are divided into score segments, and 5 corresponding reference score segments are obtained, namely 0-150, 150-250, 250-300, 300-400, and 400-500. At this time, there is a first scoring result 275, which needs to be converted into a score. First, the first score segment found is 250-300, which belongs to the third score segment. Then, the third score segment is found from the set M standard score segments, that is, the second score segment is 400-600, and the above formula is used for calculation, that is, the standard scoring result corresponding to 275 = 400 + (200 / 50) * (275-250) = 500.
[0071] In this way, each first scoring result can be converted to obtain a score under a unified scale, which is convenient for comparison between subsequent task processing models. For example, a target model is determined from the task model cluster, and the target feature information of the first object is processed using the target model, so that the classification result of the object can be quickly determined. Alternatively, based on the accumulation of time, the task processing models that are prone to high scores are marked, and the objects are first classified using these marks, which can also improve the efficiency of determining the object classification results. Furthermore, since each task processing model in the task processing cluster can be changed and added in a timely manner, it has the properties of easy expansion and easy change, which can greatly improve the model scalability and model effectiveness.
[0072] S204: Integrate the N standard scoring results to obtain a target scoring result of the first object, where the target scoring result is used to manage the transaction operation of the first object.
[0073] In one possible implementation, after obtaining N standard scoring results, a voting mechanism can be used to integrate the N standard scoring results to obtain a target scoring result for the first object. In the present application, the voting mechanism can be divided into several categories. The first category is to select the largest standard scoring result among the N standard scoring results as the target scoring result for the first object. The second category is to calculate the mean of the N standard scoring results and use the mean as the target scoring result for the first object; when calculating the mean, it can also include removing the highest score and the lowest score and then calculating the mean. The third category is to determine the number of N standard scoring results that are greater than a set threshold. If the number reaches a preset ratio, the first object is assigned a value, and the assignment is used as the target scoring result for the first object. The above is just an exemplary description of several ways to determine the target scoring result for the first object. Other ways that can achieve similar effects can also be used, which will not be listed here one by one.
[0074] Furthermore, after determining the target scoring result of the first object, the transaction operation of the first object can also be managed according to the target scoring result. The generation time of the target scoring result of the first object is obtained, and the identification information, target scoring result and generation time of the first object are stored. Each time the target scoring result is determined, it is stored once, which can facilitate the observation of the changes of a user over a period of time, so as to manage the transaction operation of the object in a timely manner. At the same time, if the target scoring result of the first object is greater than or equal to the scoring threshold, the transaction operation permission of the first object is closed. In addition, transaction permission management also includes downgrading, re-opening, etc. For the object, it can be to remind the first object of payment failure or collection failure, or to remind the first object to complete relevant account information, etc.
[0075] In an embodiment of the present application, when the data processing condition is triggered, the basic feature information and derived feature information of the first object are obtained from the database based on the identification information of the first object. The basic feature information is obtained by performing feature extraction processing on the real-time data of the first object, and the derived feature information is obtained by performing feature derivation processing on the historical data of the first object; since the basic feature information and the derived feature information are real-time data and historical data respectively, the embodiment of the present application can achieve timeliness. The basic feature information and the derived feature information are feature processed to obtain the target feature information of the first object. Processing is performed immediately after acquisition, rather than offline processing, which further reflects the timeliness of the embodiment of the present application. The N task processing models included in the task processing cluster are called to process the target feature information to obtain N standard scoring results, where N is an integer greater than 1. The N standard scoring results are integrated and processed to obtain the target scoring result of the first object, and the target scoring result is used to manage the transaction operation of the first object. In this process, since the target feature information is processed by the task processing cluster, each task processing model has a different division of labor, so the first object can be analyzed from multiple dimensions, thereby improving the accuracy of the classification result of the object. As can be seen from the above, the embodiment of the present application can ensure the timeliness of the object classification results by combining real-time data and historical data. At the same time, the target feature information is processed in multiple dimensions through the task processing cluster, thereby ensuring the accuracy of the object classification results.
[0076] The following is a detailed introduction to the construction process of the task processing cluster in the embodiment of the present application.
[0077] See also Figure 3 , is a flow chart of another data processing method disclosed in an embodiment of the present application. The data processing method can be executed by a computer device, which can be Figure 1Corresponding to the second server 102 shown, the data processing method may include but is not limited to the following steps S301-S305:
[0078] S301: When a data processing condition is triggered, basic feature information and derived feature information of the first object are obtained from a database based on identification information of the first object. The basic feature information is obtained by performing feature extraction processing on real-time data of the first object, and the derived feature information is obtained by performing feature derivation processing on historical data of the first object.
[0079] S302: Perform feature processing on the basic feature information and the derived feature information to obtain target feature information of the first object.
[0080] S303: Determine N task processing models using a preset model building method, where the preset model building method includes one or more of a priori method, hierarchical analysis method, machine learning method, and deep learning method, and generate a task processing cluster based on the N task processing models.
[0081] Step S303 mainly describes how to construct a task processing cluster. It can be understood that for different objects, the task processing models included in the task processing cluster can be different, and it can also be understood that for different objects, any task processing model can be flexibly added or changed.
[0082] When building each task processing model, you can use the following preset model building methods:
[0083] 1. A priori method. It is based on the experience of professional business experts to formulate corresponding effective rules. It can be understood that a rule is a model.
[0084] 2. Analytic Hierarchy Process (AHP). It is a research method that combines qualitative and quantitative methods to calculate decision weights for solving multi-objective complex problems. This method combines quantitative analysis with qualitative analysis, uses the decision maker's experience to judge the relative importance of the standards that can be achieved between the various measurement goals, and reasonably gives the weight of each standard of each decision-making plan, and uses the weight to find the order of advantages and disadvantages of each plan, which is more effectively applied to those topics that are difficult to solve with quantitative methods. Corresponding to this application, for example, the three models have similar outputs, but which model to use specifically also needs to measure the output and weight, and the model that is closer to the goal will be used.
[0085] 3. Machine learning methods. When using machine learning methods to train customer rating models, supervised learning models are generally used. For example, the lightGBM model. LightGBM (Light Gradient Boosting Machine) is a framework that implements the GBDT (Gradient Boosting Decision Tree) algorithm. It supports efficient parallel training and has the advantages of faster training speed, lower memory consumption, better accuracy, support for distributed and rapid processing of massive data, etc.
[0086] 4. Deep learning method: That is, the historical data and labels of the first object are used to train the initial neural network, and the network parameters of the initial neural network are adjusted through directional propagation to determine the target neural network.
[0087] Both machine learning methods and deep learning methods involve the field of artificial intelligence. Compared to machine learning methods, deep learning methods are relatively more complex. Machine learning methods generally use simpler and more linear algorithms, while deep learning methods use artificial neural networks, allowing for a higher level of complexity. Machine learning will require less data to make fairly accurate decisions. Since machine learning methods are generally simpler and require fewer parameters, models trained by machine learning methods can use smaller data sets. Deep learning methods use artificial neural networks to establish correlations and relationships with given data. Since each piece of data has different characteristics, deep learning methods generally require a large amount of data to accurately identify patterns in the data set.
[0088] Based on the above method, N task processing models are determined, and the idea of ensemble learning is used to generate a task processing cluster. The task processing cluster is then deployed to a computer device in real time so that the first object can be classified and processed in a timely manner. Among them, ensemble learning is a technology that uses multiple compatible learning algorithms / models to perform a single task. The main methods of ensemble learning can be classified into three categories: stacking, lifting and bagging. In an embodiment of the present application, any one of the above three categories can be used to determine the task processing set.
[0089] S304: calling N task processing models included in the task processing cluster to process the target feature information to obtain N standard scoring results, where N is an integer greater than 1.
[0090] S305: Integrate the N standard scoring results to obtain a target scoring result of the first object, where the target scoring result is used to manage the transaction operation of the first object.
[0091] Among them, step S304 and step S305 are the process of using the task processing cluster to process the target feature information of the first object to obtain the target scoring result, which can be as follows: Figure 4 As shown, the N task processing models may include an annotation registration processing model, a non-standard information processing model, and a comprehensive processing model, etc. Then, the target feature information of the first object is input into the task processing cluster, and N standard scoring results can be obtained, and one task processing model corresponds to one standard scoring result. For any object including the first object, the processing process of the task processing cluster can be as follows Figure 4 shown.
[0092] The embodiment of the present application introduces the process of constructing a task model cluster, and in this way, reflects the timeliness and accuracy of the data processing method of the present application.
[0093] See also Figure 5 , is a flowchart of another data processing method disclosed in an embodiment of the present application. For any object, the scoring and management of the object can be achieved through the following processing steps:
[0094] ① Message access: real-time access to the settlement object data of the new settled customer synchronized from the first server, including the identification information of the object.
[0095] ②Data splicing: also includes data query, querying basic feature information and derived feature information from the first database and the second database respectively according to the object identifier included in the settled object data, and splicing the queried basic feature information and derived feature information.
[0096] ③ Feature processing: Perform feature conversion on the spliced features to obtain target feature information.
[0097] ④Task processing cluster calculation: Input the target feature information into the N task processing models in the task processing cluster to score the object. The task processing model can be flexibly expanded or replaced according to needs. Then use the standard score calculation rules to obtain N standard scoring results corresponding to the N task processing models.
[0098] ⑤ Voting mechanism: For the N standard scoring results in the task processing cluster, the idea of ensemble learning is adopted, and the model results are summarized by means of a voting mechanism to finally obtain the target scoring result of the object.
[0099] ⑥Data post-processing: Real-time management and offline management can be performed based on the target scoring results. Real-time management can be the management of the transaction operations of the object. Offline management is to store the data for subsequent management of the object.
[0100] Therefore, based on the above description of the data processing method provided in the embodiment of the present application, it is possible to quickly and timely rate new objects in seconds, determine their classification, and manage the relevant transaction behaviors of the objects in a timely manner. At the same time, the task processing cluster is used to improve the accuracy of the object classification results, so that the quasi-real-time rating of the objects can be achieved.
[0101] The method of the embodiment of the present application is described in detail above. In order to facilitate better implementation of the above scheme of the embodiment of the present application, the device of the embodiment of the present application is provided below accordingly.
[0102] See also Figure 6 , Figure 6 It is a structural diagram of a data processing device provided in an embodiment of the present application. The data processing device can be set in the computer device provided in an embodiment of the present application. Figure 6 The data processing device shown may be a computer program (including program code) running in a computer device, and the data processing device may be used to execute Figure 2 , Figure 3 as well as Figure 5 Some or all of the steps in the method embodiment shown. Figure 6 , the data processing device may include the following units:
[0103] An acquisition unit 601 is used to acquire basic feature information and derived feature information of the first object from a database based on identification information of the first object when a data processing condition is triggered, wherein the basic feature information is obtained by performing feature extraction processing on real-time data of the first object, and the derived feature information is obtained by performing feature derivation processing on historical data of the first object;
[0104] The processing unit 602 is used to perform feature processing on the basic feature information and the derived feature information to obtain target feature information of the first object; call N task processing models included in the task processing cluster to process the target feature information to obtain N standard scoring results, where N is an integer greater than 1; integrate the N standard scoring results to obtain the target scoring result of the first object, and the target scoring result is used to manage the transaction operation of the first object.
[0105] In a possible implementation manner, the determining unit 603 is configured to:
[0106] When receiving the object entry data sent by the first server, determining that the data processing condition is triggered, and acquiring the identification information of the first object from the object entry data;
[0107] Alternatively, when a preset interval time is reached, it is determined that the data processing condition is triggered.
[0108] In a possible implementation, the database includes a first database and a second database, and when the acquisition unit 601 acquires the basic feature information and the derived feature information of the first object from the database based on the identification information of the first object, it is specifically configured to perform the following steps:
[0109] Based on the identification information of the first object, basic feature information of the first object is acquired from the first database through a real-time data query interface;
[0110] The derived feature information of the first object is acquired from the second database based on the identification information of the first object.
[0111] In a possible implementation, the acquisition unit 601 is further configured to acquire historical data of the first object, where the historical data includes account information and transaction information of the first object;
[0112] The processing unit 602 is also used to use a preset feature derivation method to perform feature derivation processing on the account information and transaction information of the first object to obtain derived feature information of the first object. The preset feature derivation method includes one or more of a data analysis model derivation method, an association derivation method, and a clustering derivation method.
[0113] In a possible implementation, the processing unit 602 calls N task processing models included in the task processing cluster to process the target feature information respectively, and when N standard scoring results are obtained, it is specifically used to perform the following steps:
[0114] Determining sub-feature information corresponding to each task processing model from the target feature information according to model types of the N task processing models included in the task processing cluster;
[0115] Calling each of the N task processing models to perform scoring processing on the corresponding sub-feature information to obtain N first scoring results;
[0116] The N first scoring results are converted into scores using a standard score calculation rule to obtain N standard scoring results.
[0117] In a possible implementation, the processing unit 602 converts the N first scoring results into scores by using the standard score calculation rule to obtain N standard scoring results, and is specifically configured to perform the following steps:
[0118] Dividing the N first scoring results into intervals according to the set M standard score segments to obtain M reference score segments, where M is an integer greater than or equal to 1;
[0119] For a scoring result to be processed, determining a first score segment in which the scoring result to be processed is located among the M reference score segments, wherein the scoring result to be processed is any one of the N first scoring results;
[0120] Determine a corresponding second score segment from the set M standard score segments according to the first score segment in which the to-be-processed scoring result is located;
[0121] The standard score calculation rule is used to calculate the lower limit value of the first score segment, the difference between the upper limit value and the lower limit value of the first score segment, the lower limit value of the second score segment, and the score result to be processed to obtain a standard score result of the score result to be processed.
[0122] In a possible implementation, the acquisition unit 601 is further configured to acquire a generation time of a target scoring result of the first object, and store the identification information of the first object, the target scoring result, and the generation time;
[0123] The processing unit 602 is further configured to close the transaction operation permission of the first object if the target scoring result of the first object is greater than or equal to a scoring threshold.
[0124] In one implementation, the processing unit 602 is further configured to perform the following steps:
[0125] Determine N task processing models using a preset model building method, wherein the preset model building method includes one or more of a priori method, hierarchical analysis method, machine learning method, and deep learning method;
[0126] A task processing cluster is generated based on the N task processing models.
[0127] According to another embodiment of the present application, Figure 6 The various units in the data processing device shown can be respectively or all combined into one or several other units to constitute, or some of the units (some) can also be split into multiple smaller units in function to constitute, which can achieve the same operation without affecting the realization of the technical effects of the embodiments of the present application. The above-mentioned units are divided based on logical functions. In practical applications, the function of a unit can also be implemented by multiple units, or the function of multiple units can be implemented by one unit. In other embodiments of the present application, the data processing device may also include other units. In practical applications, these functions can also be implemented with the assistance of other units, and can be implemented by the collaboration of multiple units.
[0128] According to another embodiment of the present application, the program can be executed by running on a general computing device such as a computer including a central processing unit (CPU), a random access memory medium (RAM), a read-only memory medium (ROM), and other processing elements and storage elements. Figure 2 , Figure 3 or Figure 5 A computer program (including program code) for each step involved in part or all of the method shown in the figure is used to construct a computer program (including program code) for each step involved in part or all of the method shown in the figure. Figure 6 The data processing device shown in the embodiment of the present application is implemented by the data processing device. The computer program can be recorded on a computer-readable storage medium, for example, and loaded into the above-mentioned computing device through the computer-readable storage medium and run therein.
[0129] In the embodiment of the present application, when the data processing condition is triggered, the basic feature information and derived feature information of the first object are obtained from the database based on the identification information of the first object. The basic feature information is obtained by performing feature extraction processing on the real-time data of the first object, and the derived feature information is obtained by performing feature derivation processing on the historical data of the first object; since the basic feature information and the derived feature information are real-time data and historical data respectively, the embodiment of the present application can achieve timeliness. The basic feature information and the derived feature information are feature processed to obtain the target feature information of the first object. Processing is performed immediately after acquisition, rather than offline processing, which further reflects the timeliness of the embodiment of the present application. The N task processing models included in the task processing cluster are called to process the target feature information to obtain N standard scoring results, where N is an integer greater than 1. The N standard scoring results are integrated and processed to obtain the target scoring result of the first object, and the target scoring result is used to manage the transaction operation of the first object. In this process, since the target feature information is processed by the task processing cluster, each task processing model has a different division of labor, so the first object can be analyzed from multiple dimensions, thereby improving the accuracy of the classification result of the object. As can be seen from the above, the embodiment of the present application can ensure the timeliness of the object classification results by combining real-time data and historical data. At the same time, by processing the target feature information in multiple dimensions through the task processing cluster, the accuracy of the object classification results is ensured, thereby achieving quasi-real-time rating of the object.
[0130] Based on the above method and device embodiments, the present application embodiment provides a computer device. Figure 7 , Figure 7 It is a structural diagram of a computer device provided in an embodiment of the present application. Figure 7The computer device shown includes at least a processor 701, an input interface 702, an output interface 703, and a computer-readable storage medium 704. The processor 701, the input interface 702, the output interface 703, and the computer-readable storage medium 704 may be connected via a bus or other means.
[0131] The computer-readable storage medium 704 may be stored in a memory of a computer device. The computer-readable storage medium 704 is used to store a computer program. The computer program includes computer instructions. The processor 701 is used to execute the program instructions stored in the computer-readable storage medium 704. The processor 701 (or CPU (Central Processing Unit)) is the computing core and control core of the computer device, which is suitable for implementing one or more computer instructions, and is specifically suitable for loading and executing one or more computer instructions to implement a corresponding method flow or a corresponding function.
[0132] The embodiment of the present application also provides a computer-readable storage medium (Memory), which is a memory device in a computer device for storing programs and data. It is understandable that the computer-readable storage medium here can include both built-in storage media in the computer device and, of course, extended storage media supported by the computer device. The computer-readable storage medium provides a storage space that stores the operating system of the computer device. In addition, one or more computer instructions suitable for being loaded and executed by the processor are also stored in the storage space. These computer instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory, or a non-volatile memory (Non-VolatileMemory), such as at least one disk storage; optionally, it can also be at least one computer-readable storage medium located away from the aforementioned processor.
[0133] In some embodiments, the processor 701 may load and execute one or more computer instructions stored in the computer-readable storage medium 704 to implement the above-mentioned Figure 2 , Figure 3 as well as Figure 5 In a specific implementation, the computer instructions in the computer-readable storage medium 704 are loaded by the processor 701 and execute the following steps:
[0134] When a data processing condition is triggered, basic feature information and derived feature information of the first object are obtained from a database based on identification information of the first object, wherein the basic feature information is obtained by performing feature extraction processing on real-time data of the first object, and the derived feature information is obtained by performing feature derivation processing on historical data of the first object;
[0135] Perform feature processing on the basic feature information and the derived feature information to obtain target feature information of the first object; call N task processing models included in the task processing cluster to process the target feature information to obtain N standard scoring results, where N is an integer greater than 1; integrate the N standard scoring results to obtain a target scoring result of the first object, and the target scoring result is used to manage the transaction operation of the first object.
[0136] In one possible implementation, the computer instructions in the computer-readable storage medium 704 are loaded by the processor 701 and are further used to perform the following steps:
[0137] When receiving the object entry data sent by the first server, determining that the data processing condition is triggered, and acquiring the identification information of the first object from the object entry data;
[0138] Alternatively, when a preset interval time is reached, it is determined that the data processing condition is triggered.
[0139] In a possible implementation, the database includes a first database and a second database, and when the computer instructions in the computer-readable storage medium 704 are loaded and executed by the processor 701 to obtain basic feature information and derived feature information of the first object from the database based on the identification information of the first object, the computer instructions are specifically used to perform the following steps:
[0140] Based on the identification information of the first object, basic feature information of the first object is acquired from the first database through a real-time data query interface;
[0141] The derived feature information of the first object is acquired from the second database based on the identification information of the first object.
[0142] In one possible implementation, the computer instructions in the computer-readable storage medium 704 are further used by the processor 701 to load and execute:
[0143] Acquire historical data of the first object, the historical data including account information and transaction information of the first object;
[0144] A preset feature derivation method is used to perform feature derivation processing on the account information and transaction information of the first object to obtain derived feature information of the first object. The preset feature derivation method includes one or more of a data analysis model derivation method, an association derivation method, and a clustering derivation method.
[0145] In a possible implementation, the computer instructions in the computer-readable storage medium 704 are loaded and executed by the processor 701 to call the N task processing models included in the task processing cluster to process the target feature information respectively, and when N standard scoring results are obtained, they are specifically used to perform the following steps:
[0146] Determining sub-feature information corresponding to each task processing model from the target feature information according to model types of the N task processing models included in the task processing cluster;
[0147] Calling each of the N task processing models to perform scoring processing on the corresponding sub-feature information to obtain N first scoring results;
[0148] The N first scoring results are converted into scores using a standard score calculation rule to obtain N standard scoring results.
[0149] In a possible implementation, the computer instructions in the computer-readable storage medium 704 are loaded and executed by the processor 701 to convert the N first scoring results into scores using the standard score calculation rule to obtain N standard scoring results, specifically for executing the following steps:
[0150] Dividing the N first scoring results into intervals according to the set M standard score segments to obtain M reference score segments, where M is an integer greater than or equal to 1;
[0151] For a scoring result to be processed, determining a first score segment in which the scoring result to be processed is located among the M reference score segments, wherein the scoring result to be processed is any one of the N first scoring results;
[0152] Determine a corresponding second score segment from the set M standard score segments according to the first score segment in which the to-be-processed scoring result is located;
[0153] The standard score calculation rule is used to calculate the lower limit value of the first score segment, the difference between the upper limit value and the lower limit value of the first score segment, the lower limit value of the second score segment, and the score result to be processed to obtain a standard score result of the score result to be processed.
[0154] In one possible implementation, the computer instructions in the computer-readable storage medium 704 are further used by the processor 701 to load and execute:
[0155] Acquire the generation time of the target scoring result of the first object, and store the identification information of the first object, the target scoring result, and the generation time;
[0156] If the target score result of the first object is greater than or equal to the score threshold, the transaction operation permission of the first object is closed.
[0157] In one implementation, the computer instructions in the computer-readable storage medium 704 are also used by the processor 701 to load and execute:
[0158] Determine N task processing models using a preset model building method, wherein the preset model building method includes one or more of a priori method, hierarchical analysis method, machine learning method, and deep learning method;
[0159] A task processing cluster is generated based on the N task processing models.
[0160] In the embodiment of the present application, when the data processing condition is triggered, the basic feature information and derived feature information of the first object are obtained from the database based on the identification information of the first object. The basic feature information is obtained by performing feature extraction processing on the real-time data of the first object, and the derived feature information is obtained by performing feature derivation processing on the historical data of the first object; since the basic feature information and the derived feature information are real-time data and historical data respectively, the embodiment of the present application can achieve timeliness. The basic feature information and the derived feature information are feature processed to obtain the target feature information of the first object. Processing is performed immediately after acquisition, rather than offline processing, which further reflects the timeliness of the embodiment of the present application. The N task processing models included in the task processing cluster are called to process the target feature information to obtain N standard scoring results, where N is an integer greater than 1. The N standard scoring results are integrated and processed to obtain the target scoring result of the first object, and the target scoring result is used to manage the transaction operation of the first object. In this process, since the target feature information is processed by the task processing cluster, each task processing model has a different division of labor, so the first object can be analyzed from multiple dimensions, thereby improving the accuracy of the classification result of the object. As can be seen from the above, the embodiment of the present application can ensure the timeliness of the object classification results by combining real-time data and historical data. At the same time, by processing the target feature information in multiple dimensions through the task processing cluster, the accuracy of the object classification results is ensured, thereby achieving quasi-real-time rating of the object.
[0161] According to one aspect of the present application, a computer program product or a computer program is provided, the computer program product or the computer program includes computer instructions, the computer instructions are stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the data processing method provided in the above-mentioned various optional modes.
[0162] It should be noted that, for the above-mentioned 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 the present application is not limited by the order of the actions described, because according to the present application, some steps may 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 the present application.
[0163] In the several embodiments provided in this application, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are only schematic, and the division of the modules described above is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed.
[0164] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.
Claims
1. A data processing method, characterized in that: The method comprises: When a data processing condition is triggered, basic feature information and derived feature information of the first object are obtained from a database based on identification information of the first object, wherein the basic feature information is obtained by performing feature extraction processing on real-time data of the first object, and the derived feature information is obtained by performing feature derivation processing on historical data of the first object; Performing feature processing on the basic feature information and the derived feature information to obtain target feature information of the first object; Calling N task processing models included in the task processing cluster to process the target feature information to obtain N standard scoring results, where N is an integer greater than 1; The N standard scoring results are integrated to obtain a target scoring result of the first object, and the target scoring result is used to manage the transaction operation of the first object.
2. The method according to claim 1, characterized in that The method further comprises: When receiving the object entry data sent by the first server, determining that the data processing condition is triggered, and acquiring the identification information of the first object from the object entry data; Alternatively, when a preset interval time is reached, it is determined that the data processing condition is triggered.
3. The method according to claim 1, characterized in that: The database includes a first database and a second database, and acquiring basic feature information and derived feature information of the first object from the database based on identification information of the first object includes: Based on the identification information of the first object, basic feature information of the first object is acquired from the first database through a real-time data query interface; The derived feature information of the first object is acquired from the second database based on the identification information of the first object.
4. The method according to any one of claims 1 to 3, characterized in that: The method further comprises: Acquire historical data of the first object, the historical data including account information and transaction information of the first object; A preset feature derivation method is used to perform feature derivation processing on the account information and transaction information of the first object to obtain derived feature information of the first object. The preset feature derivation method includes one or more of a data analysis model derivation method, an association derivation method, and a clustering derivation method.
5. The method according to claim 1, characterized in that The N task processing models included in the calling task processing cluster process the target feature information respectively to obtain N standard scoring results, including: Determining sub-feature information corresponding to each task processing model from the target feature information according to model types of the N task processing models included in the task processing cluster; Calling each of the N task processing models to perform scoring processing on the corresponding sub-feature information to obtain N first scoring results; The N first scoring results are converted into scores using a standard score calculation rule to obtain N standard scoring results.
6. The method according to claim 5, characterized in that The step of converting the N first scoring results into scores by using the standard score calculation rule to obtain N standard scoring results includes: Dividing the N first scoring results into intervals according to the set M standard score segments to obtain M reference score segments, where M is an integer greater than or equal to 1; For a scoring result to be processed, determining a first score segment in which the scoring result to be processed is located among the M reference score segments, wherein the scoring result to be processed is any one of the N first scoring results; Determine a corresponding second score segment from the set M standard score segments according to the first score segment in which the to-be-processed scoring result is located; The standard score calculation rule is used to calculate the lower limit value of the first score segment, the difference between the upper limit value and the lower limit value of the first score segment, the lower limit value of the second score segment, and the score result to be processed to obtain a standard score result of the score result to be processed.
7. The method according to any one of claims 1 to 3, characterized in that: The method further comprises: Acquire the generation time of the target scoring result of the first object, and store the identification information of the first object, the target scoring result, and the generation time; If the target score result of the first object is greater than or equal to the score threshold, the transaction operation permission of the first object is closed.
8. The method according to claim 1, characterized in that The method further comprises: Determine N task processing models using a preset model building method, wherein the preset model building method includes one or more of a priori method, hierarchical analysis method, machine learning method, and deep learning method; A task processing cluster is generated based on the N task processing models.
9. A data processing device, characterized in that: The device comprises: an acquisition unit, configured to acquire basic feature information and derived feature information of the first object from a database based on identification information of the first object when a data processing condition is triggered, wherein the basic feature information is obtained by performing feature extraction processing on real-time data of the first object, and the derived feature information is obtained by performing feature derivation processing on historical data of the first object; A processing unit, configured to perform feature processing on the basic feature information and the derived feature information to obtain target feature information of the first object; The processing unit is further used to call N task processing models included in the task processing cluster to process the target feature information to obtain N standard scoring results, where N is an integer greater than 1; The processing unit is further used to integrate the N standard scoring results to obtain a target scoring result of the first object, and the target scoring result is used to manage the transaction operation of the first object.
10. A computer device, characterized in that: The computer device comprises: a processor adapted to implement one or more computer programs; and, A computer storage medium storing one or more computer programs, wherein the one or more computer programs are suitable for being loaded by the processor and executing the data processing 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 one or more computer programs, and the one or more computer programs are suitable for being loaded by a processor and executing the data processing method according to any one of claims 1 to 8.
12. A computer program product, characterized in that The computer program product includes a computer program, which is stored in a computer-readable storage medium. A processor of a computer device reads the computer program from the computer-readable storage medium, and the processor executes the computer program, so that the computer device executes the data processing method according to any one of claims 1 to 8.
