Business processing method and related device

Abstract

The invention provides a service processing method and a related device. The embodiment of the invention can be applied to various scenes such as artificial intelligence. The method comprises the steps that firstly, inquiry information of a service is acquired, and the inquiry information of the service carries identification information of the service; then, K pieces of log data corresponding to the business are searched according to the identification information of the business, and the K pieces of log data correspond to the K log sources; then, based on the inquiry information of the business, performing data statistics on the K log data, and generating a statistical result; and finally, reply information is generated based on a statistical result. According to the method provided by the embodiment of the invention, the service problem can be quickly positioned and traced by obtaining and counting the related log data, so that the efficiency of solving the problem is improved, and the service delay is reduced; the corresponding K pieces of log data are searched according to the identification information of the service, it can be ensured that key information related to the specific service is obtained, and the accuracy of problem analysis is improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a business processing method and related devices. Background Art

[0002] Palm swiping service is an identity authentication or payment method based on palm feature recognition technology. It uses biometric technology to confirm personal identity or complete payment transactions by scanning or reading the characteristic information of the palm (such as palm prints, veins, etc.).

[0003] With the rapid development of the Shuazhang business, in order to improve business stability and performance, it is necessary to accurately obtain and locate the root causes of Shuazhang business problems. In related technologies, experts or developers in various business fields usually use their own experience to find relevant data information in the Shuazhang business process and summarize the root causes of Shuazhang business problems. However, due to the fragmentation of logs, diverse data sources, and complex business environments, tracing and locating business problems becomes challenging, resulting in inefficient problem solving, which may delay the normal operation of the business and cause inconvenience to users. Summary of the Invention

[0004] The embodiments of the present application provide a business processing method and related devices, which can accurately trace and locate business problems in a business environment with scattered logs, diverse data sources, and complex business environments, and improve the reliability of response information to business problems.

[0005] One aspect of the present application provides a service processing method, including:

[0006] Obtaining query information of the service, wherein the query information of the service carries identification information of the service;

[0007] Search for K log data corresponding to the business based on the business identification information, where the K log data correspond to K log sources, and K is an integer greater than or equal to 1;

[0008] Based on the business query information, perform data statistics on K log data and generate statistical results;

[0009] Generate reply information based on the statistical results.

[0010] Another aspect of the present application provides a business processing device, including: an inquiry information acquisition module, a log data search module, a log data statistics module, and a reply information generation module; specifically:

[0011] An inquiry information acquisition module is used to obtain inquiry information of a service, wherein the inquiry information of the service carries identification information of the service;

[0012] A log data search module is used to search for K log data corresponding to a business according to the identification information of the business, wherein the K log data correspond to K log sources, and K is an integer greater than or equal to 1;

[0013] The log data statistics module is used to perform data statistics on K log data based on business query information and generate statistical results;

[0014] The reply information generation module is used to generate reply information based on the statistical results.

[0015] In another implementation of the embodiment of the present application, the log data statistics module is further configured to:

[0016] The business question-answering model is used to perform semantic analysis on the query information and generate semantic analysis results. The business question-answering model is trained based on sample knowledge data corresponding to the business.

[0017] According to the semantic analysis results, determine the data statistics method for aggregating K log data;

[0018] Perform data statistics on K log data based on a data statistics method to generate statistical results.

[0019] In another implementation of the embodiment of the present application, the business processing device further includes: a business question and answer model training module; specifically: the business question and answer model training module is used to:

[0020] Obtain sample knowledge data corresponding to the business, P sample business problems, and P sample business strategy labels corresponding to the P sample business problems, where P is an integer greater than or equal to 1;

[0021] Input sample knowledge data and P sample business questions into the basic question-answering model;

[0022] Through the basic question-answering model, P sample business questions are analyzed and P sample business reasoning information is generated based on the sample knowledge data;

[0023] The basic question-answering model is trained based on sample knowledge data, P sample business questions, P sample business strategy labels, and P sample business reasoning information to adjust the parameters of the basic question-answering model and obtain a business question-answering model.

[0024] In another implementation of the embodiment of the present application, the business question-answering model training module is further used to:

[0025] Encode the sample knowledge data and P sample business problems respectively to generate sample knowledge encoding data and P sample business problem encoding data;

[0026] Calculate the P probability distributions of P sample business problem encoding data for the sample knowledge encoding data respectively;

[0027] Determine sample knowledge data corresponding to each sample business problem among P sample business problems according to P probability distributions;

[0028] According to the sample knowledge data corresponding to each of the P sample business problems, P pieces of sample business reasoning information corresponding to the P sample business problems are generated.

[0029] In another implementation of the embodiment of the present application, the business question-answering model training module is further used to:

[0030] Parse the sample knowledge data to generate entity information of the sample knowledge data;

[0031] and, analyzing P sample business problems, storing the P sample business problems in a relational database structure, and generating P relational data tables;

[0032] Establish the corresponding relationship between the entity information of the sample knowledge data and P relational data tables, and store the corresponding relationship.

[0033] In another implementation of the embodiment of the present application, the business question-answering model training module is further used to:

[0034] Calculate the sample business reasoning loss value based on the sample knowledge data, P sample business problems, P sample business strategy labels, and P sample business reasoning information;

[0035] The parameters of the basic question-answering model are adjusted based on the business reasoning loss value to obtain a business question-answering model.

[0036] In another implementation of the embodiment of the present application, the reply information generation module is further configured to:

[0037] Determine the corpus text information corresponding to the business through the business question-answering model;

[0038] The statistical results are subjected to corpus generation based on the corpus text information to obtain the response information.

[0039] In another implementation of the embodiment of the present application, the service processing device further includes a log data generation module; specifically, the log data generation module is configured to:

[0040] Receive business request information corresponding to the business;

[0041] Obtaining one or more of time information, interface information, and terminal information corresponding to the received service request information;

[0042] generating identification information according to one or more of time information, interface information, and terminal information;

[0043] Log data corresponding to the business request information is generated according to the business request information and the identification information, and the log data corresponding to the business request information is stored, wherein the log source of the log data corresponding to the business request information is the business request service.

[0044] In another implementation of the embodiment of the present application, the log data generation module is further configured to:

[0045] Parse the service request information to obtain the identity information carried in the service request information;

[0046] Verify the identity information and generate an identity authentication result;

[0047] According to the identity authentication result and the identification information, log data corresponding to the identity authentication is generated and stored, wherein the log source of the log data corresponding to the identity authentication is the identity authentication service.

[0048] In another implementation of the embodiment of the present application, the log data generation module is further configured to:

[0049] If the identity verification result indicates that the identity information verification is successful, the account information corresponding to the identity information is obtained;

[0050] Parse the service request information to obtain the consumption request information carried in the service request information;

[0051] Generate target behavior results based on consumption request information and account information;

[0052] According to the deduction result and the identification information, log data corresponding to the target behavior is generated and stored, wherein the log source of the log data corresponding to the target behavior is the target behavior service.

[0053] In another implementation of the embodiment of the present application, the service processing device further includes a query information generation module; specifically, the query information generation module is configured to:

[0054] receiving service alarm information, wherein the service alarm information carries identification information of the service;

[0055] Analyze business alarm information to obtain business identification information;

[0056] Generate business query information based on business identification information and business alarm information.

[0057] Another aspect of the present application provides a computer device, comprising:

[0058] memories, transceivers, processors, and bus systems;

[0059] Wherein, the memory is used to store programs;

[0060] The processor is used to execute the program in the memory, including executing the above-mentioned methods;

[0061] The bus system is used to connect the memory and the processor so that the memory and the processor can communicate with each other.

[0062] Another aspect of the present application provides a computer-readable storage medium, wherein instructions are stored in the computer-readable storage medium. When the computer-readable storage medium is run on a computer, the computer is enabled to execute the above-mentioned methods.

[0063] Another aspect of the present application provides a computer program product or computer program, which includes computer instructions 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 performs the methods provided by the above aspects.

[0064] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages:

[0065] The present application provides a business processing method and related devices, which include: obtaining query information carrying identification information of the business, thereby searching for multiple log data of different log sources corresponding to the business based on the identification information of the business, and then performing data statistics on all log data based on the query information of the business to obtain statistical results; finally, generating reply information corresponding to the query information based on the statistical results. The business processing method provided by the embodiment of the present application can quickly locate and trace business problems by automatically obtaining and counting relevant log data, avoiding the tedious work of manually searching and analyzing a large number of scattered logs, thereby improving the efficiency of problem solving and reducing business delays; searching for the corresponding K log data based on the identification information of the business can ensure that key information related to the specific business is obtained, thereby improving the accuracy of problem analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 A schematic diagram of the architecture of a business processing system provided in one embodiment of the present application;

[0067] Figure 2 A flowchart of a business processing method provided in one embodiment of the present application;

[0068] Figure 3 A flowchart of a business processing method provided in another embodiment of the present application;

[0069] Figure 4 A flowchart of a business processing method provided in another embodiment of the present application;

[0070] Figure 5 A flowchart of a business processing method provided in another embodiment of the present application;

[0071] Figure 6 A flowchart of a business processing method provided in another embodiment of the present application;

[0072] Figure 7 A flowchart of a business processing method provided in another embodiment of the present application;

[0073] Figure 8 A flowchart of a business processing method provided in another embodiment of the present application;

[0074] Figure 9 A schematic diagram of training a business question-answering model provided in an embodiment of the present application;

[0075] Figure 10 A flowchart of the palm-swiping service provided in one embodiment of the present application;

[0076] Figure 11 A schematic diagram of aggregating multiple log sources provided in an embodiment of the present application;

[0077] Figure 12 A schematic diagram of a method for handling palm-swiping service issues provided in one embodiment of the present application;

[0078] Figure 13 A flowchart of a business processing method provided in another embodiment of the present application;

[0079] Figure 14 A flowchart of a business processing method provided in another embodiment of the present application;

[0080] Figure 15 A flowchart of a business processing method provided in another embodiment of the present application;

[0081] Figure 16 A flowchart of a business processing method provided in another embodiment of the present application;

[0082] Figure 17 A schematic diagram of the structure of a service processing device provided in one embodiment of the present application;

[0083] Figure 18 A schematic diagram of the server structure provided for an embodiment of the present application. DETAILED DESCRIPTION

[0084] The embodiment of the present application provides a business processing method that can quickly locate and trace business problems by automatically acquiring and counting relevant log data, thereby improving the efficiency of problem solving and reducing business delays; searching for corresponding K log data based on the identification information of the business can ensure that key information related to the specific business is obtained, thereby improving the accuracy of problem analysis.

[0085] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the numbers used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can, for example, be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "corresponding to" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0086] In the embodiments of the present application, the term "module" or "unit" refers to a computer program or a part of a computer program that has 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 circuits or memories) 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 part of an overall module or unit that includes the function of the module or unit.

[0087] To facilitate understanding of the technical solutions provided in the embodiments of the present application, some key terms used in the embodiments of the present application are explained here:

[0088] Palm scanning services refer to business activities involving the use of palms, including payment and access control. Palm scanning services are a biometric technology that uses palm lines to complete payment or other identity verification. Palm scanning services leverage the development of biometric technology to bring greater convenience and security to people's lives and work. Palm scanning services are of great significance to user convenience and security.

[0089] Business issues: Exceptions, errors, or requirements that arise in the Shuazhang business require problem location and resolution to improve business efficiency and user experience.

[0090] Source tracing and location: Obtaining and locating the root cause and source of the problem. In the embodiment of this application, it refers to using a question-answering model (e.g., a GPT model) to trace and locate the source of the problem in the brush palm business in order to accurately solve the problem.

[0091] GPT (Generative Pretrained Transformer) model: the full name is Generative Pretrained Transformer, which is a pre-trained language model based on the Transformer architecture. Through pre-training of a large amount of text data, it can realize tasks such as generation, classification, and translation of natural language text. The core idea of ​​the GPT model is to use the pre-training model to learn the statistical laws and semantic representation of the language, so as to improve the performance of the model in different tasks. During the pre-training process, the GPT model will learn the grammar, semantics and contextual information of the language, which can be used to generate natural language text, answer questions, perform translation and other tasks. The GPT model has the ability to understand and generate natural language text, and is used to analyze the business problems of the brush palm in the embodiment of this application.

[0092] Data aggregation: The process of merging and integrating data from different sources to enable deeper data analysis and processing. In the present embodiment, data aggregation is used to integrate logs and related data generated by multiple terminals to achieve comprehensive data analysis.

[0093] Business optimization: The process of improving and optimizing the Shuazhang service by tracing and locating problems. Business optimization aims to improve business efficiency, user satisfaction, and system performance.

[0094] The embodiments of the present application relate to artificial intelligence (AI) and large models, and are designed based on speech technology (Speech Technology) and natural language processing (NLP) in artificial intelligence.

[0095] Artificial Intelligence (AI) refers to the theories, methods, techniques, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, to perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results. In other words, AI is a comprehensive technology within computer science that seeks to understand the essence of intelligence and produce new intelligent machines that can respond in a manner similar to human intelligence. AI also studies the design principles and implementation methods of various intelligent machines, enabling them to possess the capabilities of perception, reasoning, and decision-making.

[0096] Artificial intelligence (AI) technology is a comprehensive discipline encompassing a wide range of fields, encompassing both hardware and software technologies. Foundational AI technologies generally include sensors, specialized AI chips, cloud computing, distributed storage, big data processing, operating / interaction systems, and mechatronics. AI software technologies primarily encompass computer vision, speech processing, natural language processing, and machine learning / deep learning.

[0097] Key technologies in speech technology include automatic speech recognition (ASR), text-to-speech (TTS), and voiceprint recognition. Enabling computers to hear, see, speak, and feel is the future direction of human-computer interaction, with speech becoming one of the most promising methods of human-computer interaction.

[0098] Natural language processing (NLP) is a key area of ​​research in computer science and artificial intelligence. It studies the theories and methods that enable effective communication between humans and computers using natural language. Natural language processing (NLP) integrates linguistics, computer science, and mathematics. Therefore, research in this field involves natural language—the language we use in everyday life—and is closely linked to the study of linguistics. Natural language processing technologies typically include text processing, semantic understanding, machine translation, robotic question answering, and knowledge graphs.

[0099] A pre-training model, also known as a cornerstone model or a large model, refers to a deep neural network (DNN) with large parameters. It is trained on massive amounts of unlabeled data. Leveraging the function approximation capabilities of large-parameter DNNs, the pretrained machine learning (PTM) extracts common features from the data. Through techniques such as fine tuning, efficient parameter fine tuning (PEFT), and prompt-tuning, it is then adapted for downstream tasks. Therefore, pre-trained models can achieve ideal results in few-shot or zero-shot scenarios. Based on the data modality processed, PTMs can be categorized into language models (ELMO, BERT, GPT), vision models (swin-transformer, ViT, V-MOE), speech models (VALL-E), and multimodal models (ViBERT, CLIP, Flamingo, Gato). Multimodal models refer to models that establish feature representations from two or more data modalities. Pre-trained models are important tools for outputting artificial intelligence generated content (AIGC) and can also serve as a universal interface for connecting multiple task-specific models.

[0100] With the rapid development of the ShuaiZhang business, in order to improve business stability and performance, it is necessary to accurately obtain and locate the root causes of ShuaiZhang business issues. However, due to the fragmented logs, diverse data sources, and complex business environment, tracing and locating business issues has become challenging. Traditional methods usually rely on experts or developers in various business fields to find relevant data information in the ShuaiZhang business process based on their own experience and summarize the root causes of ShuaiZhang business issues, which often cannot effectively address these issues.

[0101] The embodiment of the present application provides a business processing method, which uses a question-answering model to trace and locate business problems in the palm-swiping business. Specifically, the question-answering model is used to analyze the aggregated log data in the palm-swiping business, and by confirming the business problem, finding the log data generated by multiple log sources, and analyzing the cause of the business problem, an analysis conclusion is generated and relevant data is provided. The method provided in the embodiment of the present application can analyze business problems, determine the root cause of the problem, and provide solutions and optimization suggestions. The method provided in the embodiment of the present application can more accurately locate palm-swiping business problems, improve the efficiency and accuracy of problem solving, and thus improve the stability and performance of the business.

[0102] For easier understanding, see Figure 1 , Figure 1 This is an application environment diagram of the business processing method in the embodiment of the present application, such as Figure 1 As shown, the business processing method in the embodiment of the present application is applied to a business processing system. The business processing system includes: a server and a terminal device; wherein the server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (Content Delivery Network, CDN), and big data and artificial intelligence platforms. The terminal can be a smart phone, tablet computer, laptop computer, desktop computer, smart speaker, smart watch, etc., but is not limited to this. The terminal and the server can be directly or indirectly connected by wired or wireless communication, and the embodiment of the present application is not limited here.

[0103] The server first obtains the business query information, where the business query information carries the business identification information; then, the server searches for K log data corresponding to the business based on the business identification information, where the K log data correspond to K log sources; then, based on the business query information, the server performs data statistics on the K log data and generates statistical results; finally, the server generates reply information based on the statistical results.

[0104] The following describes the business processing method in this application from the perspective of the server. Figure 2 The service processing method provided in the embodiment of the present application includes: Step S110 to Step S140. Specifically:

[0105] S110: Obtain business inquiry information.

[0106] The service inquiry information carries the identification information of the service.

[0107] It can be understood that a service is a business performed through an application. The service performed through an application refers to the process involved in executing the service, which can be implemented through the application. In actual applications, services can include palm-swiping services, payment services, search services, and resource transfer services.

[0108] Inquiry information is used to inquire about business-related knowledge; for example, knowledge inquiry information is used to indicate: what is a certain business (for example, what is palm payment), what is the cause of a problem with a certain business (for example, what is the reason for palm payment failure), etc.

[0109] The inquiry information may be at least one of text, picture, and audio; for example, the inquiry information may be a question expressed in text, a question expressed in both text and picture, or a question expressed in audio.

[0110] The identification information of a business is a full-process ID (also called Graph ID), which is usually generated at the entrance of the service of the business and transmitted based on the framework context. The identification information is used to identify and concatenate requests during log printing and data recording, and can be passed between multiple interfaces. When crossing different services or interfaces, the identification information is passed through in the background framework context, and the requests processed by the code corresponding to different sequence diagrams can be concatenated by the identification information, thereby realizing the transmission of log data, context, etc. By injecting identification information when calling the interface or entity method, it can be obtained and located in the entire process corresponding to the business. For example, when locating rejection, it is necessary to find the data on the entire link of the current recognition (images, scores, quality judgments, verification methods, and calls from other related ends).

[0111] The identification information may be a random string generated by one or more of request information, interface information, machine IP, thread ID, timestamp, incremental ID, and other random numbers.

[0112] For example, identification information is generated at the entry point of the service request and injected into the context of the background framework for full transparent transmission. The identification information can be composed of one or more information including request information, interface information, machine IP, thread ID, timestamp, incremental ID, and other random numbers. It should be emphasized that the reporting of identification information is carried out in the form of use cases, so it can be obtained across use cases through identification information. The purpose of obtaining use cases is to sort out the entire call chain to determine the entities on the entire call chain, so that subsequent modules can obtain entities. In the scenario of swiping palm payment, the entity mainly refers to the entity of the swiping palm business to determine the requests of multiple device ends. When calling the interface or the entity method, the identification information is injected into the corresponding code. In this way, the content of multiple device ends can be obtained and located in the swiping palm business. The call across the device ends is obtained through the identification information. By connecting the identification information in series, it can be obtained and analyzed across the device ends in the swiping palm business process.

[0113] S120 : Search for K log data corresponding to the business according to the identification information of the business.

[0114] K log data correspond to K log sources, and K is an integer greater than or equal to 1.

[0115] It's understandable that the use of identification information allows for full-link acquisition and location within complex business processes, particularly for requests from multiple devices involved in a palm-swipe payment system, providing accurate log data and contextual information. Data related to a palm-swipe action can be acquired and located using identification information. As a unique identifier, identification information accurately locates logs and data related to palm-swipe actions.

[0116] S130 : Based on the business query information, perform data statistics on the K log data and generate statistical results.

[0117] It is understood that, based on the contextual information of the use case, relevant data is contextually analyzed and correlated. This allows for the merging and aggregation of multiple logs and data generated by the same use case. Question-answering models (such as the GPT model) are used to understand business inquiry information and aggregate the palm-swiping data based on the model's data requirements. Through analysis and processing by the GPT model, key information, features, and statistical results can be extracted from the palm-swiping data.

[0118] S140. Generate reply information based on the statistical results.

[0119] It can be understood that the inquiry information is used to ask business-related questions. According to the business identification information carried in the inquiry information, multiple log data corresponding to multiple different log sources corresponding to the business are searched, and all log data are integrated to generate reply information. Therefore, generating corresponding reply information based on the business inquiry information can be achieved through question and answer, that is, when the server obtains the business inquiry information, it replies with the integrated statistical results, and the statistical results are used to reply to the business inquiry information.

[0120] Exemplarily, when the query information of a business is text, the server obtains the text, processes the text through the question-answering model, and obtains the knowledge data of the business; for example, the query information is: "What is the reason for the failure of the palm payment business in the past week?" The question-answering model obtains the palm payment data of the past week based on the time information, and determines the business that failed the palm payment from it. Assuming that there are M businesses that failed the palm payment, M identification information corresponding to these M failed palm payment businesses is obtained, and all log data corresponding to the M failed palm payment businesses (including payment request logs, identity authentication logs, and payment deduction logs, etc.) are determined through the M identification information. , and conduct data statistics on all log data, and find that among the M failed palm payment businesses, there are X failures due to payment request interface errors, Y failures due to identity verification failures, and Z failures due to insufficient account balance for deduction. Based on the statistical results, a reply message is generated: "There were M failed palm payment businesses in the past week, among which X failures due to payment request interface errors accounted for x%, Y failures due to identity verification failures accounted for y%, and Z failures due to insufficient account balance for deduction, accounting for z%." Furthermore, the reply message may include the root cause, possible solutions, or recommended optimization measures corresponding to the business inquiry information.

[0121] The business processing method provided in the embodiment of the present application can quickly locate and trace business problems by automatically acquiring and counting relevant log data, avoiding the tedious work of manually searching and analyzing a large number of scattered logs, thereby improving the efficiency of problem solving and reducing business delays; searching for the corresponding K log data according to the identification information of the business can ensure that key information related to the specific business is obtained, thereby improving the accuracy of problem analysis.

[0122] In this application Figure 2 In an optional embodiment of the business processing method provided in the corresponding embodiment, please refer to Figure 3 , step S130 further includes sub-steps S131 to S133. Specifically:

[0123] S131. Perform semantic analysis on the query information through the business question-answering model to generate a semantic analysis result.

[0124] Among them, the business question-answering model is trained based on sample knowledge data corresponding to the business.

[0125] It can be understood that the business question-answering model is obtained by training the question-answering model based on sample knowledge data corresponding to the business, which means that the business question-answering model has learned business-related knowledge and patterns and can understand and analyze the meaning of inquiry information.

[0126] When performing semantic analysis on a query, the system first needs to analyze the vocabulary within the query, including identifying words, phrases, and keywords. Furthermore, it analyzes the sentence structure of the query to determine the subject, predicate, and object of the sentence. Analyzing the vocabulary within the query helps understand the basic structure and topic of the question; analyzing the sentence structure of the query helps understand the logical relationships and semantic structure of the sentence. Next, based on the results of lexical and syntactic analysis, it attempts to understand the semantic meaning of the query, which may include identifying the intent, topic, and entities of the question. The query is then mapped to existing business knowledge and domain models to identify concepts, rules, and relationships relevant to the question. Finally, based on the results of semantic understanding and knowledge mapping, a semantic analysis result is generated. Through semantic analysis, the system can better understand the user's question and provide meaningful information for subsequent processing and response. For example, semantic analysis results can be used to determine the question type (such as query, command, request), extract key information, and identify relevant entities. This allows for more accurate data aggregation and statistical operations in subsequent steps based on the semantic analysis results.

[0127] S132: Determine a data statistics method for aggregating the K log data based on the semantic analysis result.

[0128] It is understandable that based on the results of semantic analysis, a data statistical method for aggregating the K log data is determined. The semantic analysis results can provide key information and intent about the query information, and based on this information, it can be determined how to effectively aggregate and count the K log data. For example, it is possible to determine the statistical indicators to be calculated (such as average, total, percentage, etc.), or to determine how to associate different log data for comprehensive analysis. At the same time, it is also necessary to consider the characteristics of the K log data, such as the distribution, time sequence, and correlation of the data. These characteristics may affect the selection and application of statistical methods and determine the granularity of data aggregation, such as whether to aggregate the entire data set or group and aggregate according to time range, category, or other dimensions. By carefully considering the semantic analysis results and data characteristics and selecting appropriate data statistical methods, the K log data can be aggregated and analyzed more effectively, thereby providing meaningful statistical results.

[0129] S133: Perform data statistics on the K log data based on a data statistics method to generate statistical results.

[0130] It is understood that, based on the determined statistical method, actual statistical operations are performed on the K log data to generate statistical results. This step involves specific processing of the K log data, calculating and summarizing relevant statistical indicators according to the previously determined statistical method, and generating the results.

[0131] The method provided in the embodiments of this application uses semantic analysis of the question-answering model to determine an appropriate data statistics method and perform statistics on log data to generate meaningful statistical results. This can help better understand and answer business-related questions and provide more accurate and useful information.

[0132] In this application Figure 3 In an optional embodiment of the business processing method provided in the corresponding embodiment, please refer to Figure 4 , before step S131, it also includes steps S210 to S240. Specifically:

[0133] S210 : Obtain sample knowledge data corresponding to the business, P sample business problems, and P sample business strategy labels corresponding to the P sample business problems.

[0134] Wherein, P is an integer greater than or equal to 1.

[0135] It's understood that the sample knowledge data corresponding to a business refers to key concepts in the domain knowledge of that business. For example, in the field of palm-swipe payment, sample knowledge data includes misidentification, rejection, payment timeout, and device network terminal. Sample business problems refer to specific business-related questions or scenarios. These problems help the model understand and learn various situations that may be encountered in real-world business situations. For example, in the field of palm-swipe payment, sample business problems include: how to improve palm-swipe payment recognition accuracy, what are the causes of misidentification during the payment process, etc. Sample business strategy tags refer to the strategies or solutions corresponding to the sample business problems. They provide the expected answer or handling method for each sample business problem. For example, in the field of palm-swipe payment, sample business strategy tags include: strategies for improving recognition accuracy: for example, recommending users to keep their palms clean, operate the palm-swipe device correctly, and optimize the recognition algorithm; analysis of the causes of misidentification: possible causes include wet palms, environmental interference, and device malfunction; payment timeout handling strategies: prompting users to check their network connection, re-initiate payment, and providing refund channels; and solutions to device network terminal problems: checking device status, repairing the network connection, replacing the device, or seeking technical support.

[0136] S220: Input the sample knowledge data and P sample business questions into the basic question-answering model.

[0137] It's understandable that a basic question-answering model is a model capable of answering questions. It's typically a pre-trained model that analyzes and processes input questions and attempts to provide an answer. The basic question-answering model is a general-purpose model based on natural language processing technology, capable of understanding and processing questions expressed in human language. It typically uses deep learning algorithms, such as neural networks, to learn language patterns and regularities. Basic question-answering models can be trained through unsupervised learning on large amounts of text data, thereby learning the semantics, grammar, and logical relationships of language. When sample knowledge data and sample business questions are input into the basic question-answering model, the model attempts to understand the meaning of the question and generates an answer using the information in the sample knowledge data. The model's answer may be based on existing knowledge and patterns, or derived through analysis and reasoning of the question. By inputting sample data into the basic question-answering model, the model's capabilities can be leveraged to process and answer business-related questions. However, because the basic question-answering model is general-purpose, it may not fully understand the details and characteristics of a specific business. Therefore, in step S240, the basic question-answering model is further trained and adjusted based on sample business questions, sample business policy labels, and sample business reasoning information to better adapt it to specific business needs. The performance and accuracy of the basic question-answering model depend on the quality and quantity of its training data, as well as the model architecture and parameter settings. In practical applications, the basic question-answering model needs to be continuously optimized and improved to improve the accuracy and reliability of its answers.

[0138] S230: Analyze P sample business questions through the basic question-answering model, and generate P sample business reasoning information based on the sample knowledge data.

[0139] It is understandable that the basic question-answering model will analyze the P sample business questions that are input. The basic question-answering model will use its learned language knowledge and patterns to try to understand the intention and meaning of the question. Based on the sample knowledge data, the basic question-answering model will generate P sample business reasoning information, which is the result of the model's reasoning and inference based on the question and sample knowledge data. The reasoning information may include answers, suggestions, explanations or other relevant information related to the question. For example, if a sample business question is "How to improve the recognition accuracy of palm-swipe payment", the basic question-answering model may generate the following sample business reasoning information based on the relevant information in the sample knowledge data: Keep your palms clean and dry, ensure that the palm-swipe device works properly, and use more advanced recognition technology. The generated sample business reasoning information can help us better understand the problem and provide valuable reference for subsequent steps. It should be noted that the quality and accuracy of sample business reasoning information depends on the performance of the basic question-answering model and the quality and completeness of the sample knowledge data. To improve the quality of sample business reasoning information, we can take the following measures: Optimize the model: Improve the performance and accuracy of the model by adjusting the model's parameters, architecture, or training data; Increase sample knowledge data: Provide more comprehensive, accurate, and rich sample knowledge data to help the model better understand business problems; Manual review: Manually review and correct the generated sample business reasoning information to ensure its accuracy and reliability. In this way, through step S230, the capabilities of the basic question-answering model can be utilized to generate reasoning information related to the sample business problem, providing a useful reference for subsequent training and optimization.

[0140] S240. Train the basic question-answering model based on the sample knowledge data, P sample business questions, P sample business strategy labels, and P sample business reasoning information to adjust the parameters of the basic question-answering model and obtain a business question-answering model.

[0141] As can be understood, the basic question-answering model is trained based on sample knowledge data, P sample business questions, P sample business policy labels, and P sample business reasoning information. The purpose of training is to adjust the model's parameters so that the model can better understand business questions and generate reasoning information consistent with the policy labels. First, the sample knowledge data provides important concepts and knowledge about the business domain, helping the model better understand the context and relevant information of the question. The sample business questions represent various situations that may arise in real business situations, allowing the model to learn the characteristics and handling methods of different problems. The sample business policy labels provide the model with the expected answer or solution for each question. These policy labels tell the model what action to take or what advice to give in specific situations. The sample business reasoning information demonstrates the model's ability to analyze and understand the sample business questions. By analyzing this reasoning information, we can understand the model's thinking process and inference methods, thereby identifying potential problems or shortcomings. During the training process, supervised learning methods can be used to compare the sample business questions, sample business policy labels, and sample business reasoning information with the model's output. Based on the comparison results, the model's parameters are adjusted to optimize model performance. Specifically, techniques such as the backpropagation algorithm may be used to update the weights and biases of the model based on feedback from the error or loss function, so that the output of the model is closer to the desired answer or policy label. Such a training process can continuously improve the performance of the model and make it better adapted to specific business needs. Through the training of step S240, an optimized and adjusted business question-answering model is obtained. The model has higher accuracy and reliability in understanding and answering business questions, and can better support the operation of actual business. It should be noted that the training process may require multiple iterations, and different parameters need to be tested and adjusted to find the optimal model configuration. In addition, in order to improve the generalization ability of the model, data enhancement, regularization and other technologies can also be used to prevent overfitting.

[0142] The method provided in the embodiment of the present application trains and optimizes the basic question-answering model by utilizing sample data and business questions, thereby obtaining a question-answering model specifically for the business. The business question-answering model can better adapt to the needs and characteristics of specific businesses and improve the accuracy and reliability of answering questions.

[0143] In this application Figure 4 In an optional embodiment of the business processing method provided in the corresponding embodiment, please refer to Figure 5 , step S230 further includes sub-steps S231 to S234. Specifically:

[0144] S231. Encode the sample knowledge data and P sample business problems respectively to generate sample knowledge encoding data and P sample business problem encoding data.

[0145] It can be understood that the sample knowledge data and P sample business questions are encoded. The purpose of encoding is to convert the data into a form that the model can understand and process. This can be achieved through various encoding methods, such as word vector encoding and semantic encoding. Through encoding, the sample knowledge data and sample business questions are represented as digital vectors, allowing the model to perform calculations and analysis. Encoding is the process of converting raw data into a representation suitable for model processing. For sample knowledge data, encoding helps the model better understand and process this data. Through encoding, we can convert different types of sample knowledge data, such as text, images, and audio, into digital vectors or other encoded representations. The specific encoding method can be selected based on the model and task being used. Common encoding methods include word vector encoding, semantic encoding, and image feature encoding. For example, in natural language processing tasks, word vector encoding can be used to represent words as vectors so that the model can capture semantic relationships between words. The P sample business questions also require encoding. The purpose of encoding is to convert the questions into a form that the model can understand so that the model can analyze and reason about them. The encoding method can be similar to that used for the sample knowledge data. Through the encoding process of step S231, we can convert the sample knowledge data and sample business problems into a form that can be processed by the model, which provides a basis for subsequent model analysis and reasoning.

[0146] S232. Calculate P probability distributions of P sample business problem encoding data with respect to the sample knowledge encoding data respectively.

[0147] It can be understood that the P probability distributions calculated for each of the P sample business problem encoding data with respect to the sample knowledge encoding data represent the degree of association between each sample business problem and the sample knowledge encoding data. Specifically, the probability distributions can be calculated using the following methods: Similarity measurement: Various similarity measurement methods can be used to compare the sample business problem encoding data and the sample knowledge encoding data. For example, cosine similarity or Euclidean distance between vectors can be used to measure the similarity between two encoding data. Probabilistic model: A probabilistic model can be used to calculate the probability distribution of the sample business problem encoding data with respect to the sample knowledge encoding data. For example, a probabilistic model such as a naive Bayesian model or a conditional random field can be used to predict the probability distribution based on the sample business problem encoding data and the sample knowledge encoding data. Deep learning model: Within the deep learning framework, a neural network model can be used to automatically learn the mapping relationship between the sample business problem encoding data and the sample knowledge encoding data and calculate the corresponding probability distribution. Through the calculated P probability distributions, each sample business problem has a measure of the degree of association with the sample knowledge encoding data. A higher probability value indicates a greater correlation between the sample business problem and the corresponding sample knowledge encoding data, while a lower probability value indicates a weaker correlation. These probability distributions can be used in subsequent decision-making and reasoning processes. For example, when generating reasoning information based on sample business problems, more relevant sample knowledge encoding data can be selected based on the probability distribution to improve the accuracy and reliability of reasoning.

[0148] S233. Determine sample knowledge data corresponding to each of the P sample business problems according to the P probability distributions.

[0149] It can be understood that determining the sample knowledge data corresponding to each of the P sample business questions based on P probability distributions involves determining which sample knowledge data are associated with each sample business question based on the probability distributions. A higher probability value indicates a greater relevance between the sample business question and the corresponding sample knowledge data. Specifically, each sample business question corresponds to a probability distribution that represents the degree of relevance between the sample business question and the sample knowledge data. A higher probability value indicates a stronger relevance between the sample business question and a particular sample knowledge data, while a lower probability value indicates a weaker relevance. Based on the probability distributions, the following method can be used to determine the sample knowledge data corresponding to each sample business question: Select the sample knowledge data with the highest probability value. This method can be directly selected as the answer to the corresponding sample business question. This method is simple and intuitive, but it may overlook other sample knowledge data with lower probability values ​​that still have a certain relevance. Considering multiple sample knowledge data with higher probability values: In addition to selecting the sample knowledge data with the highest probability value, other sample knowledge data with higher probability values ​​can also be considered. This can be achieved by setting a threshold or employing some weighting strategy. Combining prior knowledge and contextual information: When determining the sample knowledge data corresponding to a sample business problem, a comprehensive judgment can be made based on prior knowledge and contextual information. For example, more appropriate sample knowledge data can be selected based on knowledge of the business domain and the semantics of the problem. Iterative optimization: In practical applications, iterative optimization can be used to continuously adjust and improve the sample knowledge data selection strategy to improve the accuracy and reliability of decision-making. By determining the sample knowledge data corresponding to each sample business problem based on a probability distribution, the sample business problem can be matched with the relevant sample knowledge data.

[0150] S234 . Generate P pieces of sample business reasoning information corresponding to the P sample business problems based on the sample knowledge data corresponding to each of the P sample business problems.

[0151] It can be understood that based on the sample knowledge data corresponding to each of the P sample business problems, P sample business reasoning information corresponding to the P sample business problems is generated. Based on the correspondence between the sample business problems and the sample knowledge data, the model can generate corresponding sample business reasoning information. This reasoning information can be based on analysis and inference of the sample knowledge data, providing possible answers or suggestions for each sample business problem. First, each sample business problem is associated with specific sample knowledge data. The sample knowledge data can be knowledge, rules, patterns, or other related information about a certain domain. Through the previous steps, we have identified the sample knowledge data corresponding to each sample business problem. Next, reasoning and analysis are performed based on the sample knowledge data. The reasoning process may involve interpretation, application, and expansion of the sample knowledge data. For example, if the sample knowledge data is rules or patterns, we can use these rules or patterns to derive conclusions or suggestions related to the sample business problem. Various techniques and methods may be used in the reasoning process, such as rule-based reasoning, case-based reasoning, and machine learning models. The specific reasoning method will depend on the characteristics of the sample knowledge data and the requirements of the task. Generating sample business reasoning information requires expressing the reasoning results in some form, which can be a textual description, numerical results, charts, or other suitable representations. Sample business reasoning information should be able to clearly convey the reasoning results and related recommendations.

[0152] The method provided in the embodiment of the present application can process sample knowledge data and sample business problems more carefully, and generate sample business reasoning information based on the relationship between them. Such processing method can improve the accuracy and reliability of the model, enabling it to better understand and answer business problems.

[0153] In this application Figure 4 In an optional embodiment of the business processing method provided in the corresponding embodiment, please refer to Figure 6 , after step S220, further steps S250 to S270 are included. Specifically:

[0154] S250: Analyze the sample knowledge data to generate entity information of the sample knowledge data.

[0155] It's understandable that the sample knowledge data is broken down into more specific entity information. By parsing the sample knowledge data, key entities, attributes, and relationships can be extracted. Entity information can include the entity's name, type, and attributes. This step helps better understand and organize the sample knowledge data, preparing it for subsequent processing and analysis.

[0156] S260: Analyze P sample business problems, store the P sample business problems in a relational database structure, and generate P relational data tables.

[0157] It's understandable that the P sample business problems are converted into a relational database structure for storage. A relational database is a common data storage and management method that organizes data in a tabular format and represents relationships between entities by linking different tables. Storing the sample business problems in relational tables allows for more efficient data management and query, facilitating subsequent analysis and processing.

[0158] S270: Establish a correspondence between the entity information of the sample knowledge data and P relational data tables, and store the correspondence.

[0159] It is understood that a mapping relationship is established between the entity information of the sample knowledge data and the relational data table. By establishing this mapping relationship, the entity information can be associated with a specific data table, making it easier to query and use it when needed. Storing this mapping relationship ensures that relevant data can be found quickly and accurately in subsequent processing.

[0160] The method provided in the embodiment of the present application can organize and manage data more effectively and improve data availability and query efficiency by parsing sample knowledge data and storing it in the structure of a relational database; by establishing a correspondence between the entity information of the sample knowledge data and the relational data table, it can ensure the consistency and accuracy of the data and avoid data errors and inconsistencies.

[0161] In this application Figure 4 In an optional embodiment of the business processing method provided in the corresponding embodiment, please refer to Figure 7 , step S240 further includes sub-steps S241 to S242. Specifically:

[0162] S241. Calculate the sample business reasoning loss value based on the sample knowledge data, P sample business problems, P sample business strategy labels, and P sample business reasoning information.

[0163] As you can understand, the sample business reasoning loss is calculated based on the sample knowledge data, sample business questions, sample business policy labels, and sample business reasoning information. The sample business reasoning loss measures the error or inaccuracy of the model when reasoning about the sample business question. Typically, loss functions such as mean squared error (MSE) and cross-entropy are used to calculate the loss. The purpose of calculating the loss is to evaluate model performance and provide a basis for subsequent parameter adjustments.

[0164] S242. Adjust the parameters of the basic question-answering model based on the business reasoning loss value to obtain a business question-answering model.

[0165] It's understandable that based on the calculated inference loss for a sample business, we can adjust the parameters of the basic question-answering model. The goal of parameter adjustment is to optimize the model's performance, enabling it to more accurately generate answers when reasoning about business questions. Common parameter adjustment methods include optimization algorithms such as gradient descent and stochastic gradient descent. By continuously adjusting parameters, the model can gradually learn better feature representations and inference patterns, thereby improving its performance.

[0166] The method provided in the embodiment of the present application calculates the sample business reasoning loss value and adjusts the parameters, so that the model can better adapt to the characteristics of the business problem and improve the accuracy of reasoning. Parameter adjustment can help the model find a more optimal parameter configuration, thereby improving the performance and generalization ability of the model; according to specific business scenarios and needs, by adjusting the parameters of the model, the model can be better adapted to different business problems and data distributions; through continuous parameter adjustment and optimization, the stability and reliability of the model can be improved, and errors and uncertainties can be reduced.

[0167] In this application Figure 3 In an optional embodiment of the business processing method provided in the corresponding embodiment, please refer to Figure 8 , step S140 further includes sub-steps S141 to S142. Specifically:

[0168] S141. Determine the corpus text information corresponding to the business through the business question-answering model.

[0169] It is understandable that the business question answering model will use its pre-learned knowledge and patterns to determine the corpus text information related to the business based on the input business question. This corpus text information may include common expressions, keywords, phrases, etc. related to the business question. The process of determining the corpus text information can help the model better understand the business question.

[0170] S142. Generate corpus based on the statistical results based on the corpus text information to obtain response information.

[0171] It is understood that based on the corpus text information determined in step S141, the model processes and generates statistical results. By using this corpus text information, the model can generate more expressive and targeted responses. The corpus generation process can make responses more natural and easy to understand, improving communication effectiveness with users.

[0172] The method provided in the embodiment of the present application can make the reply more in line with business reality and provide more accurate and useful information by determining corpus text information related to the business and generating a reply based on it; using corpus text information for generation can make the reply more natural and fluent, and easier for users to understand and accept; according to different business problems, the relevant corpus text information can be flexibly determined, so that the model can better adapt to various business scenarios and user needs; this generation method based on corpus text information demonstrates the intelligent processing capabilities of the model and can better simulate human language expression and thinking.

[0173] For ease of understanding, an embodiment of the present application provides a business processing method, which is used to obtain and locate the root cause of the problem of the Shua Zhang business, and implements the business processing method through a business question and answer model. The business question and answer model uses a question and answer model (such as a GPT model) to analyze all log data in the aggregated Shua Zhang business process and the inquiry information of the current business, and generates analysis conclusions and provides relevant data by confirming the questions about the business in the inquiry information, finding the business data generated by multiple log sources, and analyzing the causes of business problems. The business question and answer model can use the GPT model to analyze business problems, determine the root cause of the problem, and provide solutions and optimization suggestions. Through the business question and answer model, problems can be located more accurately, and the efficiency and accuracy of problem solving can be improved, thereby improving the stability and performance of the business.

[0174] Specifically:

[0175] First, train the business question-answering model. The business question-answering model is pre-trained using a question-answering model (such as the GPT model) to learn domain knowledge of palm-swipe payment scenarios and master important concepts in the palm-swipe payment field, such as misidentification, rejection, payment timeout, device network terminal, etc., and understand their meaning. In addition, the business question-answering model also needs to learn the entire business logic related to these concepts. It can learn and train business problems and corresponding solutions. For example, in the event of rejection, it clearly defines which data, equipment, on-site photos, user verification methods, and other logic need to be analyzed. Please refer to Figure 9 , Figure 9 A schematic diagram of training a business question-answering model provided in an embodiment of the present application.

[0176] The input to the business question-answering model primarily consists of domain knowledge, business problems, and response strategies. After pre-training, the business question-answering model can locate and judge business problems based on learned logic and identify the additional knowledge required to determine further content. The business question-answering model deeply understands and encodes the input domain knowledge and associates it with the pre-trained question-answering model. Through training, the business question-answering model can grasp the characteristics and patterns of business problems and provide corresponding solutions based on the problem description.

[0177] The input data used for training (including domain knowledge, business problems, and response strategies) can be domain knowledge stored in key-value (KV) or JSON format, SQL syntax, and the relationship between entities and SQL tables. For example, in the existing system, the attribute correspondence between entities and existing database tables (such as SQL tables), such as the SQL table corresponding to accounts. Here, you need to tell the model: the name of the SQL table corresponding to the entity account, the name of the field corresponding to the account attribute A, etc. The trained business question-and-answer model can output its understanding of a specific business item and logic, and can output the SQL for creating a table or searching for a specific entity.

[0178] The business question-answering model is pre-trained based on the question-answering model, giving it conversational capabilities. Through conversational interaction with the business question-answering model, business problems can be solved and located. The business question-answering model understands the given business problem description and, based on pre-trained knowledge and logic, analyzes and infers, providing a solution to the problem. The business question-answering model also has self-learning capabilities, continuously enriching its knowledge base through conversations with users. Based on the complexity of the problem and domain requirements, it can identify the additional knowledge and information needed to solve the problem.

[0179] The pre-training and learning capabilities of the business question-answering model make it an intelligent business problem-solving model. Based on existing domain knowledge and logic, the business question-answering model can interact with users through conversations to solve and locate complex business problems. Furthermore, the business question-answering model continuously accumulates and updates knowledge, improving its understanding and problem-solving capabilities, thereby providing more accurate and efficient solutions.

[0180] Next, generate the business identification information. Figure 10As shown, when a user swipes their palm through a palm-swiping terminal, identification information of the palm-swiping service is generated at the entrance of the palm-swiping system. Preferably, the identification information can be a full-process dyeing ID (also called Graph ID), which is transparently transmitted based on the framework context. The identification information is used to identify and concatenate requests when printing logs and recording data, and can be transmitted between multiple interfaces. When crossing different services or interfaces, the identification information is transparently transmitted in the background framework context, and the requests processed by the codes corresponding to different sequence diagrams can be concatenated by the identification information, thereby realizing the transmission of log data, context, etc. The identification information can be a random string generated by one or more of the request information, interface information, machine IP, thread ID, timestamp, incremental ID, and other random numbers. By injecting identification information when calling the interface or entity method, it can be obtained and located in the entire process corresponding to the business. For example, when locating rejection, it is necessary to find the data on the entire link of the current identification (pictures, scores, quality judgments, verification methods, and calls from other associated terminals).

[0181] For example, a coloring ID is generated at the entry point of a service-oriented architecture (SOA) microservice request and injected into the context of the backend framework for full transparency. The coloring ID is a random string generated from the request information, interface information, machine IP, thread ID, timestamp, incrementing ID, and other random numbers. It is important to emphasize that the coloring ID is reported according to the use case, so it can be obtained across use cases using the coloring ID. The purpose of obtaining use cases is to sort out the entire call chain to identify the entities in the entire call chain, so that the entities can be obtained later. Here, the entity mainly refers to the entity of the brush palm service to identify requests from multiple devices. Specifically: the request information, interface information, machine IP, thread ID, timestamp, and incrementing ID corresponding to the brush palm request are obtained. Based on the request information, interface information, machine IP, thread ID, timestamp, and incrementing ID, a coloring ID is generated. This coloring ID can be obtained and located in the subsequent call chain. The generated coloring ID is injected into the context of the backend framework to achieve full transparency. By injecting it into the framework context, the coloring ID can be transmitted and used throughout the entire business process. During the entire transparent transmission process, the log data with the dye ID will be reported. Preferably, the reporting of the dye ID should be carried out in the form of a use case so that it can correspond to the sequence diagram in the subsequent acquisition. When the interface is called or the entity method is called, the dye ID is injected into the corresponding code to obtain and locate the content of multiple device ends in the palm-swiping business. The dye ID is used to obtain calls across device ends. By connecting the dye IDs in series, acquisition and analysis can be performed across device ends in the business process.

[0182] This embodiment of the application implements the generation and injection of full-process dye IDs, as well as related log reporting and business acquisition capabilities. By using dye IDs, full-link acquisition and location can be achieved within complex business processes, particularly for requests from multiple devices involved in the palm-of-hand payment system, providing accurate data and contextual information. This allows the system to better locate problems, troubleshoot, and optimize performance.

[0183] Then, the log data of the business is obtained based on the identification information of the business, and the obtained log data is aggregated to obtain statistical results. Figure 11 As shown, the business question-answering model is used to understand business problems and identify data needs, and then the logs and data involved are classified, aggregated, counted and processed (such as error code escape, etc.), thereby realizing the aggregation of multi-terminal and multi-log source swiping data. Since the data of the same use case are related, they can be obtained through the dyeing ID and contextual analysis can be performed. For image data generated on the device side, single-device aggregation analysis is usually required to determine the imaging quality of the device, etc. Specifically, the relevant data of a swiping behavior is obtained and located through the full dyeing ID. The dyeing ID is used as a unique identifier to accurately find the log data involving the swiping behavior. According to the context information of the use case in the log data, the relevant data is contextually analyzed and associated. In this way, multiple log data generated by the same use case can be merged and aggregated. The business question-answering model is used to understand business problems, and the swiping data is aggregated according to the model's data needs. Through the analysis and processing of the business question-answering model, key information, features and statistical results in the swiping data can be extracted.

[0184] By combining end-to-end ID and business question-and-answer models, we can aggregate and analyze palm-swiping data. We correlate logs and data from different terminals, extract key information, and perform statistical analysis and processing. This allows for comprehensive analysis of palm-swiping data from multiple log sources, providing valuable information for subsequent problem identification, business optimization, and decision-making.

[0185] Finally, the response information is generated based on the statistical results. Figure 12As shown, the business question-and-answer model is used to trace and locate issues in inquiries from the Shuazhang service. This model analyzes the current business issue in aggregated logs, analyzes the issue based on the log data, generates conclusions, and provides relevant data. Specifically, the current business issue must be clearly identified. This can be determined using pre-aggregated logs and the issue description. For Shuazhang service issues, such as transaction failures and abnormal behavior, the module must accurately identify and confirm the issue type. To accurately analyze the cause of the issue, the module must source and collect relevant business data from multiple log sources. This data can come from different devices, business processes, and different aspects. By collecting and integrating this data, a more comprehensive context for the business issue is obtained. Based on the aggregated logs and collected business data, the module uses the business question-and-answer model to conduct in-depth analysis of the issue. The business question-and-answer model understands the issue description and relevant data, identifying possible causes and influencing factors. By analyzing the cause of the issue, the module can provide explanations and inferences. After completing the analysis of the business issue, the module generates conclusions and solutions. These conclusions can include the root cause of the issue, possible solutions, or recommended optimization measures. At the same time, the module can also provide relevant data and statistical results to support the implementation of business decisions and problem solving.

[0186] The method provided in the embodiment of the present application uses a business question and answer model to analyze business problems, so as to understand the context, causes and influencing factors of the problem. The business question and answer model can make in-depth inferences and interpretations based on the problem description and data, and provide support for problem location. By analyzing multiple log sources and related data, the business question and answer model can determine the root cause of the business problem, which helps to clarify the nature of the problem and provide direction for problem solving. The business question and answer model generates analysis conclusions and solutions to help the business team understand and solve problems. Conclusions and solutions can include the causes of the problem, influencing factors, optimization suggestions, etc., thereby providing guidance for business decision-making and problem solving. The business question and answer model uses the powerful analytical capabilities of the GPT model, combined with aggregated logs and related data, to provide support for tracing and locating the business problems of Shuaizhang. By deeply analyzing the causes of business problems, it can help the business team better understand the problem, and provide solutions and optimization suggestions to improve the stability and performance of the business.

[0187] The method provided in the embodiment of the present application applies the GPT model to the business question-and-answer model, traces and locates the business problems of Shuazhang through the business question-and-answer model, obtains the business data generated by multiple log sources by analyzing the aggregated logs and data, and uses the business question-and-answer model to analyze the business problems. This method, which combines natural language processing and data analysis, provides new ideas and tools for locating and solving business problems of Shuazhang. However, the implementation of this technical solution faces difficulties such as log dispersion, data association and context analysis. It is necessary to overcome the complexity of data processing and model application to ensure the accuracy and reliability of traceability and positioning.

[0188] In this application Figure 2 In an optional embodiment of the business processing method provided in the corresponding embodiment, please refer to Figure 13 The business processing method further includes steps S310 to S340. Specifically:

[0189] S310: Receive service request information corresponding to the service.

[0190] It can be understood that in the palm-swipe payment scenario, the service is palm-swipe payment. When the user swipes his palm through the terminal, corresponding service request information is generated, which includes the palm-swipe image collected when the user swipes his palm.

[0191] S320: Obtain one or more of time information, interface information, and terminal information corresponding to the received service request information.

[0192] It can be understood that when the user swipes his palm through the terminal, the corresponding time information, interface information and terminal information are obtained. The time information can be the timestamp corresponding to the generated business request information, the interface information is the system interface corresponding to the palm swiping system, and the terminal information is the device identification information corresponding to the terminal device used by the user when swiping his palm.

[0193] S330: Generate identification information according to one or more of the time information, the interface information, and the terminal information.

[0194] It can be understood that the identification information is generated based on one or more of the time information, interface information and terminal information. Preferably, the identification information can be a dye ID; the identification information can also be a random string generated by one or more of the request information, interface information, machine IP, thread ID, timestamp, incremental ID and other random numbers.

[0195] S340: Generate log data corresponding to the service request information according to the service request information and the identification information, and store the log data corresponding to the service request information.

[0196] The log source of the log data corresponding to the business request information is the business request service.

[0197] It is understood that based on the service request information and identification information, the system will generate corresponding log data. This log data may include detailed information about the service request, processing process, results, etc., to facilitate subsequent analysis and troubleshooting. The generated log data is stored for subsequent query and analysis. The storage method may be to write the log data to a database, file, or other storage medium.

[0198] The method provided in the embodiments of this application records the relevant information of each service request, forming valuable log data. This log data can be used to obtain the operating status of the service, identify potential problems, optimize business processes, etc. At the same time, by using identification information, different service requests can be easily distinguished and retrieved.

[0199] In this application Figure 9 In an optional embodiment of the business processing method provided in the corresponding embodiment, please refer to Figure 14 , after step S340, further steps S410 to S430 are included. Specifically:

[0200] S410: Analyze the service request information to obtain the identity information carried in the service request information.

[0201] It is understandable that the image collected when the user swipes his palm in the service request information is parsed to obtain the corresponding identity information.

[0202] S420: Verify the identity information and generate an identity authentication result.

[0203] It is understood that the identity information is verified and an authentication result is generated. For example, it is compared with known identity information in a database, or verified using an encryption algorithm. The purpose of verification is to ensure the authenticity and validity of the identity information.

[0204] S430: Generate log data corresponding to the identity authentication according to the identity authentication result and the identification information, and store the log data corresponding to the identity authentication.

[0205] The log source of the log data corresponding to identity authentication is the identity authentication service.

[0206] It is understandable that the log data corresponding to the identity authentication includes the result of the identity authentication (such as success or failure) and related identification information. These log data are then stored for subsequent query and analysis. In the scenario of swiping the palm to pay, steps S420 and S430 can ensure that the payment operation is initiated by a legitimate user and record the result of the identity authentication. For example, if the identity authentication result is successful, the system may record relevant information for subsequent auditing and tracing. If the identity authentication fails, the system may trigger the corresponding error handling logic, such as requiring the user to re-authenticate.

[0207] The method provided in the embodiment of the present application injects identification information into the log data corresponding to the identity authentication during the identity authentication process, laying the foundation for subsequently obtaining the log data corresponding to the business through the identification information.

[0208] In this application Figure 10 In an optional embodiment of the business processing method provided in the corresponding embodiment, please refer to Figure 15 , after step S430, further steps S510 to S540 are included. Specifically:

[0209] S510: If the identity authentication result indicates that the identity information verification is successful, then obtain the account information corresponding to the identity information.

[0210] It is understandable that if the identity authentication result indicates that the identity information verification is passed, that is, the user's identity is legal, the balance information of the account corresponding to the identity information is obtained.

[0211] S520: Parse the service request information to obtain consumption request information carried in the service request information.

[0212] It is understandable that the consumption request information may be the consumption amount, and the consumption amount in the consumption request can be obtained by parsing the service request information.

[0213] S530: Generate target behavior results based on the consumption request information and account information.

[0214] It can be understood that the target behavior result is generated based on the consumption request information and account information. When the balance number displayed by the account balance information is greater than or equal to the consumption amount, the payment is successful. When the balance number displayed by the account balance information is less than the consumption amount, the payment fails.

[0215] S540: Generate log data corresponding to the target behavior based on the deduction result and the identification information, and store the log data corresponding to the target behavior.

[0216] The log source of the log data corresponding to the target behavior is the target behavior service.

[0217] It is understandable that based on the generated target behavior results and identification information, log data corresponding to the target behavior is generated and stored. Similar to the previous steps, the log data here will record the results of the target behavior for subsequent query and analysis. In the scenario of swiping the palm to pay, these steps can ensure that only legitimate users who have been authenticated can perform consumption operations, and record related consumption requests and results. For example, if the consumption request information is legal and the account balance is sufficient, the system may generate a target behavior result of a successful transaction and record the corresponding log data. If there is a problem with the consumption request information (such as insufficient balance), a transaction failure result may be generated and the corresponding log may be recorded.

[0218] The method provided in the embodiment of the present application injects identification information into the log data corresponding to the transaction deduction during the identity verification process, laying the foundation for subsequently obtaining the log data corresponding to the business through the identification information.

[0219] In this application Figure 2 In an optional embodiment of the business processing method provided in the corresponding embodiment, please refer to Figure 16 , before step S110, it also includes steps S610 to S630. Specifically:

[0220] S610: Receive service alarm information.

[0221] The service alarm information carries the identification information of the service.

[0222] It is understandable that when there is alarm information in the reported log data, identification information of the service corresponding to the alarm information is determined.

[0223] S620: Analyze the service alarm information to obtain service identification information.

[0224] It is understood that by understanding the format or protocol of the alarm information and identifying specific fields or tags, the system can determine the specific business related to the alarm through parsing.

[0225] S630: Generate service query information based on the service identification information and service alarm information.

[0226] It is understandable that, after the identification information of the abnormal business is determined, inquiry information for the business is generated to inquire about the reason why the abnormality occurs in the business.

[0227] The method provided in the embodiment of the present application generates query information for the service based on the service identification information and the original service alarm information. The query information can be used to further understand the service situation, determine the severity of the problem, or obtain more details.

[0228] The following is a detailed description of the service processing device in this application. Figure 17 . Figure 17 This is a schematic diagram of an embodiment of a business processing device 10 in an embodiment of the present application. The business processing device 10 includes: an inquiry information acquisition module 110, a log data search module 120, a log data statistics module 130 and a reply information generation module 140; specifically:

[0229] The query information acquisition module 110 is used to obtain query information of a service, wherein the query information of the service carries identification information of the service;

[0230] The log data search module 120 is used to search for K log data corresponding to the business according to the identification information of the business, wherein the K log data correspond to K log sources, and K is an integer greater than or equal to 1;

[0231] The log data statistics module 130 is used to perform data statistics on K log data based on the business query information and generate statistical results;

[0232] The reply information generating module 140 is configured to generate reply information based on the statistical results.

[0233] The business processing device provided in the embodiment of the present application can quickly locate and trace business problems by automatically acquiring and counting relevant log data, avoiding the tedious work of manually searching and analyzing a large number of scattered logs, thereby improving the efficiency of problem solving and reducing business delays; searching for the corresponding K log data according to the identification information of the business can ensure that key information related to the specific business is obtained, thereby improving the accuracy of problem analysis.

[0234] In this application Figure 17 In an optional embodiment of the service processing device provided by the corresponding embodiment, the log data statistics module is further configured to:

[0235] The business question-answering model is used to perform semantic analysis on the query information and generate semantic analysis results. The business question-answering model is trained based on sample knowledge data corresponding to the business.

[0236] According to the semantic analysis results, determine the data statistics method for aggregating K log data;

[0237] Perform data statistics on K log data based on a data statistics method to generate statistical results.

[0238] The device provided in the embodiment of the present application determines an appropriate data statistics method through semantic analysis of the question-answering model and performs statistics on log data to generate meaningful statistical results. This can help better understand and answer business-related questions and provide more accurate and useful information.

[0239] In this application Figure 17 In an optional embodiment of the business processing device provided by the corresponding embodiment, the business processing device further includes: a business question and answer model training module; specifically: the business question and answer model training module is used to:

[0240] Obtain sample knowledge data corresponding to the business, P sample business problems, and P sample business strategy labels corresponding to the P sample business problems, where P is an integer greater than or equal to 1;

[0241] Input sample knowledge data and P sample business questions into the basic question-answering model;

[0242] Through the basic question-answering model, P sample business questions are analyzed and P sample business reasoning information is generated based on the sample knowledge data;

[0243] The basic question-answering model is trained based on sample knowledge data, P sample business questions, P sample business strategy labels, and P sample business reasoning information to adjust the parameters of the basic question-answering model and obtain a business question-answering model.

[0244] The device provided in the embodiment of the present application trains and optimizes the basic question-answering model by utilizing sample data and business questions, thereby obtaining a question-answering model specifically for the business. The business question-answering model can better adapt to the needs and characteristics of specific businesses and improve the accuracy and reliability of answering questions.

[0245] In this application Figure 17 In an optional embodiment of the business processing device provided by the corresponding embodiment, the business question-answering model training module is further configured to:

[0246] Encode the sample knowledge data and P sample business problems respectively to generate sample knowledge encoding data and P sample business problem encoding data;

[0247] Calculate the P probability distributions of P sample business problem encoding data for the sample knowledge encoding data respectively;

[0248] Determine sample knowledge data corresponding to each sample business problem among P sample business problems according to P probability distributions;

[0249] According to the sample knowledge data corresponding to each of the P sample business problems, P pieces of sample business reasoning information corresponding to the P sample business problems are generated.

[0250] The device provided in the embodiment of the present application can process sample knowledge data and sample business problems more carefully, and generate sample business reasoning information based on the relationship between them. Such processing method can improve the accuracy and reliability of the model, enabling it to better understand and answer business problems.

[0251] In this application Figure 17 In an optional embodiment of the business processing device provided by the corresponding embodiment, the business question-answering model training module is further configured to:

[0252] Parse the sample knowledge data to generate entity information of the sample knowledge data;

[0253] and, analyzing P sample business problems, storing the P sample business problems in a relational database structure, and generating P relational data tables;

[0254] Establish the corresponding relationship between the entity information of the sample knowledge data and P relational data tables, and store the corresponding relationship.

[0255] The device provided in the embodiment of the present application can organize and manage data more effectively and improve data availability and query efficiency by parsing sample knowledge data and storing it in the structure of a relational database; by establishing a correspondence between the entity information of the sample knowledge data and the relational data table, it can ensure the consistency and accuracy of the data and avoid data errors and inconsistencies.

[0256] In this application Figure 17 In an optional embodiment of the business processing device provided by the corresponding embodiment, the business question-answering model training module is further configured to:

[0257] Calculate the sample business reasoning loss value based on the sample knowledge data, P sample business problems, P sample business strategy labels, and P sample business reasoning information;

[0258] The parameters of the basic question-answering model are adjusted based on the business reasoning loss value to obtain a business question-answering model.

[0259] The device provided in the embodiment of the present application calculates the sample business inference loss value and adjusts the parameters, so that the model can better adapt to the characteristics of the business problem and improve the accuracy of reasoning. Parameter adjustment can help the model find a more optimal parameter configuration, thereby improving the performance and generalization ability of the model; according to specific business scenarios and needs, by adjusting the parameters of the model, the model can be better adapted to different business problems and data distributions; through continuous parameter adjustment and optimization, the stability and reliability of the model can be improved, and errors and uncertainties can be reduced.

[0260] In this application Figure 17 In an optional embodiment of the service processing device provided by the corresponding embodiment, the reply information generating module is further configured to:

[0261] Determine the corpus text information corresponding to the business through the business question-answering model;

[0262] The statistical results are subjected to corpus generation based on the corpus text information to obtain the response information.

[0263] The device provided in the embodiment of the present application can make the reply more in line with the business reality and provide more accurate and useful information by determining the corpus text information related to the business and generating a reply based on it; using the corpus text information for generation can make the reply more natural and fluent, and easier for users to understand and accept; according to different business problems, the relevant corpus text information can be flexibly determined, so that the model can better adapt to various business scenarios and user needs; this generation method based on corpus text information demonstrates the intelligent processing capability of the model and can better simulate human language expression and thinking.

[0264] In this application Figure 17 In an optional embodiment of the service processing device provided by the corresponding embodiment, the service processing device further includes a log data generation module; specifically, the log data generation module is configured to:

[0265] Receive business request information corresponding to the business;

[0266] Obtaining one or more of time information, interface information, and terminal information corresponding to the received service request information;

[0267] generating identification information according to one or more of time information, interface information, and terminal information;

[0268] Log data corresponding to the business request information is generated according to the business request information and the identification information, and the log data corresponding to the business request information is stored, wherein the log source of the log data corresponding to the business request information is the business request service.

[0269] The device provided in the embodiments of the present application records the relevant information of each service request, forming valuable log data. This log data can be used to obtain the operating status of the service, identify potential problems, optimize business processes, and so on. Furthermore, by using identification information, different service requests can be easily distinguished and retrieved.

[0270] In this application Figure 17 In an optional embodiment of the business processing device provided by the corresponding embodiment, the log data generating module is further configured to:

[0271] Parse the service request information to obtain the identity information carried in the service request information;

[0272] Verify the identity information and generate an identity authentication result;

[0273] According to the identity authentication result and the identification information, log data corresponding to the identity authentication is generated and stored, wherein the log source of the log data corresponding to the identity authentication is the identity authentication service.

[0274] The device provided in the embodiment of the present application injects identification information into the log data corresponding to the identity authentication during the identity authentication process, laying the foundation for subsequently obtaining the log data corresponding to the business through the identification information.

[0275] In this application Figure 17 In an optional embodiment of the business processing device provided by the corresponding embodiment, the log data generating module is further configured to:

[0276] If the identity verification result indicates that the identity information verification is successful, the account information corresponding to the identity information is obtained;

[0277] Parse the service request information to obtain the consumption request information carried in the service request information;

[0278] Generate target behavior results based on consumption request information and account information;

[0279] According to the deduction result and the identification information, log data corresponding to the target behavior is generated and stored, wherein the log source of the log data corresponding to the target behavior is the target behavior service.

[0280] The device provided in the embodiment of the present application injects identification information into the log data corresponding to the transaction deduction during the identity verification process, laying the foundation for subsequently obtaining the log data corresponding to the business through the identification information.

[0281] In this application Figure 17In an optional embodiment of the service processing device provided by the corresponding embodiment, the service processing device further includes a query information generation module; specifically, the query information generation module is configured to:

[0282] receiving service alarm information, wherein the service alarm information carries identification information of the service;

[0283] Analyze business alarm information to obtain business identification information;

[0284] Generate business query information based on business identification information and business alarm information.

[0285] The apparatus provided in the embodiment of the present application generates query information for the service based on the service identification information and the original service alarm information. The query information can be used to further understand the service situation, determine the severity of the problem, or obtain more details.

[0286] In this application Figure 17 In an optional embodiment of the service processing device provided by the corresponding embodiment,

[0287] Figure 18 : This is a schematic diagram of a server structure provided in an embodiment of the present application. The server 300 may have relatively large differences due to different configurations or performances, and may include one or more central processing units (CPUs) 322 (for example, one or more processors) and memories 332, and one or more storage media 330 (for example, one or more massive storage devices) for storing application programs 342 or data 344. Among them, the memories 332 and the storage media 330 may be temporary storage or permanent storage. The program stored in the storage medium 330 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the server. Furthermore, the central processing unit 322 may be configured to communicate with the storage medium 330 to execute a series of instruction operations in the storage medium 330 on the server 300.

[0288] The server 300 may also include one or more power supplies 326, one or more wired or wireless network interfaces 350, one or more input and output interfaces 358, and / or one or more operating systems 341, such as Windows Server 2003 or Windows Server 2003R. TM , Mac OS X TM , Unix TM , Linux TM , FreeBSD TM etc.

[0289] The steps performed by the server in the above embodiment can be based on the Figure 15 The server structure shown.

[0290] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0291] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interface, device or unit, which can be electrical, mechanical or other forms.

[0292] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0293] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0294] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, 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, which is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0295] 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, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A business processing method, characterized in that: include: Obtaining query information of a service, wherein the query information of the service carries identification information of the service; Searching for K log data corresponding to the business according to the identification information of the business, wherein the K log data correspond to K log sources, and K is an integer greater than or equal to 1; Based on the query information of the business, performing data statistics on the K log data to generate statistical results; Reply information is generated based on the statistical results.

2. The service processing method according to claim 1, wherein: The performing of data statistics on the K log data based on the query information of the business to generate statistical results includes: Performing semantic analysis on the query information using a business question-answering model to generate a semantic analysis result, wherein the business question-answering model is trained based on sample knowledge data corresponding to the business; Determine a data statistics method for performing data aggregation on the K log data according to the semantic analysis result; Perform data statistics on the K log data based on the data statistics method to generate statistical results.

3. The service processing method according to claim 2, wherein: Before performing semantic analysis on the query information using the question-answer model and generating a semantic analysis result, the method further includes: Obtaining sample knowledge data corresponding to the business, P sample business problems, and P sample business strategy labels corresponding to the P sample business problems, where P is an integer greater than or equal to 1; Inputting the sample knowledge data and the P sample business questions into the basic question-answering model; Analyze the P sample business questions through the basic question-answering model, and generate P sample business reasoning information based on the sample knowledge data; The basic question-answering model is trained according to the sample knowledge data, P sample business questions, the P sample business strategy labels and the P sample business reasoning information to adjust the parameters of the basic question-answering model to obtain the business question-answering model.

4. The service processing method according to claim 3, wherein: The analyzing the P sample business problems and generating P sample business reasoning information based on the sample knowledge data includes: Encoding the sample knowledge data and the P sample business problems respectively to generate sample knowledge encoding data and P sample business problem encoding data; Calculate P probability distributions of the P sample business problem coded data with respect to the sample knowledge coded data respectively; Determining sample knowledge data corresponding to each of the P sample business problems according to the P probability distributions; According to the sample knowledge data corresponding to each sample business problem in the P sample business problems, P pieces of sample business reasoning information corresponding to the P sample business problems are generated.

5. The business processing method according to claim 3, wherein: After inputting the sample knowledge data and the P sample business questions into the basic question-answering model, the method further includes: Parsing the sample knowledge data to generate entity information of the sample knowledge data; and, analyzing the P sample business problems, storing the P sample business problems in a relational database structure, and generating P relational data tables; A correspondence between the entity information of the sample knowledge data and the P relational data tables is established, and the correspondence is stored.

6. The service processing method according to claim 3, wherein: The training of the basic question-answering model based on the sample knowledge data, the P sample business questions, the P sample business policy labels, and the P sample business reasoning information to adjust the parameters of the basic question-answering model to obtain the business question-answering model includes: Calculating a sample business reasoning loss value based on the sample knowledge data, the P sample business problems, the P sample business strategy labels, and the P sample business reasoning information; The parameters of the basic question-answering model are adjusted based on the business reasoning loss value to obtain the business question-answering model.

7. The service processing method according to claim 2, wherein: The generating of reply information based on the statistical result includes: Determine the corpus text information corresponding to the business through the business question-answering model; Corpus generation is performed on the statistical results based on the corpus text information to obtain reply information.

8. The business processing method according to claim 1, wherein: The method further comprises: Receiving service request information corresponding to the service; Obtaining one or more of time information, interface information, and terminal information corresponding to receiving the service request information; generating the identification information according to one or more of the time information, the interface information, and the terminal information; Log data corresponding to the business request information is generated according to the business request information and the identification information, and the log data corresponding to the business request information is stored, wherein a log source of the log data corresponding to the business request information is a business request service.

9. The service processing method according to claim 8, wherein: After establishing the association relationship between the identification information and the service, the method further includes: Parsing the service request information to obtain the identity information carried in the service request information; Verifying the identity information and generating an identity verification result; Log data corresponding to the identity authentication is generated according to the identity authentication result and the identification information, and the log data corresponding to the identity authentication is stored, wherein a log source of the log data corresponding to the identity authentication is an identity authentication service.

10. The service processing method according to claim 8, wherein: After verifying the identity information and generating an identity authentication result, the method further includes: If the identity verification result indicates that the identity information is verified successfully, obtaining the account information corresponding to the identity information; Parsing the service request information to obtain consumption request information carried in the service request information; generating a target behavior result according to the consumption request information and the account information; Log data corresponding to the target behavior is generated according to the deduction result and the identification information, and the log data corresponding to the target behavior is stored, wherein a log source of the log data corresponding to the target behavior is a target behavior service.

11. The service processing method according to claim 1, wherein: Before obtaining the query information of the business, the following steps are also included: receiving service alarm information, wherein the service alarm information carries identification information of the service; Parsing the service alarm information to obtain identification information of the service; Generate query information of the service according to the identification information of the service and the service alarm information.

12. A business processing device, characterized in that: include: An inquiry information acquisition module, configured to acquire inquiry information of a service, wherein the inquiry information of the service carries identification information of the service; a log data search module, configured to search for K log data corresponding to the business according to the identification information of the business, wherein the K log data correspond to K log sources, and K is an integer greater than or equal to 1; A log data statistics module, configured to perform data statistics on the K log data based on the query information of the business and generate statistical results; A reply information generating module is used to generate reply information based on the statistical results.

13. A computer device, characterized in that: include: Memory, processor, and bus system; Wherein, the memory is used to store programs; The processor is configured to execute the program in the memory, including executing the service processing method according to any one of claims 1 to 11; The bus system is used to connect the memory and the processor so that the memory and the processor can communicate with each other.

14. A computer-readable storage medium comprising instructions, characterized in that: When the method is executed on a computer, the computer is enabled to execute the business processing method according to any one of claims 1 to 11.

15. A computer program product comprising a computer program, characterized in that The computer program is executed by a processor to implement the service processing method according to any one of claims 1 to 11.

Images