Method and device for tracing poor quality complaint of wireless network

Abstract

The application relates to the technical field of mobile communication, and provides a wireless network quality complaint tracing method and device, which are executed by an intelligent agent deployed on a computing device. The method comprises the following steps: acquiring complaint information of a user, wherein the complaint information comprises a text description of network problems encountered by the user; analyzing the complaint information to obtain task content comprising a complaint problem type; generating a tool calling path and calling parameters of each tool on the tool calling path based on the task content and preset tool metadata, wherein the tool metadata comprises description information describing tool functions, input data specifications, output result specifications and tool dependency relationships; controlling each tool to execute according to the calling parameters of each tool along the tool calling path, and generating a root cause of the quality complaint according to analysis results of each tool on multi-source basic data related to the task content. The application can dynamically select a tool combination, and perform comprehensive root cause analysis based on multi-source basic data, so that the root cause of the quality complaint can be accurately located.

Description

Technical Field

[0001] This application relates to the field of mobile communication technology, and in particular to a method and apparatus for tracing the source of complaints about poor wireless network quality. Background Technology

[0002] In existing technologies, most solutions for handling poor wireless network quality complaints rely on a single data source, such as call text, location information, or deep packet inspection data. Furthermore, the processing employs a static tool call chain, meaning that the same fixed order of analysis tool calls is used regardless of the complaint content. This approach, combining a single data source with a static tool call chain, results in low diagnostic accuracy for complaints, makes it difficult to accurately pinpoint the root cause of poor quality complaints, and fails to meet the tracing needs of complex network scenarios. Summary of the Invention

[0003] This application provides a method and apparatus for tracing the source of complaints about poor wireless network quality, in order to solve the problems of low diagnostic accuracy of complaints and difficulty in accurately locating the root cause of poor quality complaints in the prior art.

[0004] This application provides a method for tracing the source of complaints about poor wireless network quality, which is executed by an intelligent agent deployed on a computing device. The method includes the following steps: Obtain user complaint information, which includes a textual description of the network problems encountered by the user; The complaint information is analyzed to obtain task content including the type of complaint issue; Based on the task content and the preset tool metadata, a tool call path and the call parameters of each tool on the tool call path are generated. The tool metadata includes descriptive information describing the tool's functions, input data specifications, output result specifications, and tool dependencies. Based on the calling parameters of each tool, the execution of each tool is controlled according to the tool calling path, and the root cause of the poor quality complaint is generated based on the analysis results of each tool on the multi-source basic data related to the task content.

[0005] According to the wireless network poor quality complaint tracing method provided in this application, the first tool on the tool call path is a data fusion tool, which is used to fuse the multi-source basic data.

[0006] According to the wireless network quality poor complaint tracing method provided in this application, based on the task content and preset tool metadata, a tool call path and call parameters for each tool on the tool call path are generated, including: Based on the task content, the preset knowledge base, and the preset tool metadata, a tool call path and the call parameters of each tool on the tool call path are generated. The preset knowledge base records historical task content and corresponding historical tool call paths.

[0007] According to the wireless network quality poor complaint tracing method provided in this application, after generating the root cause of the poor quality complaint based on the analysis results of multi-source basic data related to the task content by various tools, the method further includes: Based on the current thought chain formed by the agent, the agent's reflection mechanism is triggered, and the process jumps to execute the steps of generating tool call paths and call parameters of each tool on the tool call paths based on the task content and preset tool metadata, and generates tool call paths and call parameters that are different from the previous round.

[0008] According to the wireless network quality poor complaint tracing method provided in this application, for user connection cells extracted from multi-source basic data, the cell-related tools are invoked in the following three stages during each round of reflection: Filter all residential areas visited by the complaining user during the complaint period; All cells are sorted by cell risk to obtain the N candidate cells with the highest risk, where N is greater than or equal to 1; By combining the historical complaint records and spatial associations of the N candidate cells in the multi-source basic data, relevant cell-related tools are invoked.

[0009] According to the wireless network poor quality complaint tracing method provided in this application, the tool call path includes: risk assessment tool, complaint analysis tool and cell spatial analysis tool, and in the tool call path, the risk assessment tool is located before the complaint analysis tool and the cell spatial analysis tool, and the risk assessment tool is used to assess the cell risk index.

[0010] According to the wireless network quality poor complaint tracing method provided in this application, after generating the root cause of the poor quality complaint based on the analysis results of multi-source basic data related to the task content by various tools, the method further includes: Based on the root causes of the poor quality complaints, corresponding optimization suggestions are given, and a structured complaint analysis report is generated based on the root causes of the poor quality complaints and the optimization suggestions.

[0011] This application also provides a device for tracing the source of complaints about poor wireless network quality, which is executed by an intelligent agent deployed on a computing device, and includes the following modules: The complaint information acquisition module is used to acquire user complaint information, which includes a text description of the network problems encountered by the user. The complaint information parsing module is used to parse the complaint information to obtain task content including the type of complaint issue. The tool call generation module is used to generate a tool call path and call parameters for each tool on the tool call path based on the task content and preset tool metadata. The tool metadata includes descriptive information describing the tool function, input data specifications, output result specifications, and tool dependency relationships. The tool execution control module is used to control the execution of each tool according to the tool call path based on the calling parameters of each tool, and to generate the root cause of the poor quality complaint based on the analysis results of each tool on multi-source basic data related to the task content.

[0012] This application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the program to implement the wireless network quality poor complaint tracing method as described above.

[0013] This application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the wireless network poor quality complaint tracing method as described above.

[0014] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the wireless network poor quality complaint tracing method as described above.

[0015] The wireless network quality poor complaint tracing method and apparatus provided in this application parses user complaint information through an intelligent agent to obtain task content including the complaint problem type and multi-source basic data related to the complaint information. Based on the task content and preset tool metadata, it generates tool call paths and call parameters for each tool on the tool call path, realizing the dynamic generation of tool call paths according to task content. This allows for the dynamic selection of analysis tool combinations more suitable for the task content, improving the diagnostic accuracy of complaint problems, accurately locating the root cause of poor quality complaints, and meeting the tracing needs of complex network scenarios. During the execution of each tool in the control tool call path, each tool performs comprehensive analysis of multi-source basic data. Because it fully covers the multi-dimensional characteristics of network problems, it significantly improves the comprehensiveness and accuracy of network problem diagnosis and root cause location compared to the analysis results of single basic data. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1This is one of the flowcharts illustrating the method for tracing the source of complaints about poor wireless network quality provided in this application.

[0018] Figure 2 This is a schematic diagram illustrating the interaction between the intelligent agent, the tool, and the data storage and access layer in the wireless network poor quality complaint tracing method provided in this application.

[0019] Figure 3 This is the second flowchart illustrating the method for tracing the source of complaints about poor wireless network quality provided in this application.

[0020] Figure 4 This is a schematic diagram of the wireless network quality poor complaint tracing device provided in this application.

[0021] Figure 5 This is a schematic diagram of the structure of the electronic device provided in this application. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0023] The wireless network quality poor complaint tracing method of this application is executed by an intelligent agent deployed on a computing device, and the specific process is as follows: Figure 1 As shown, the procedure includes steps S110 to S140.

[0024] Step S110: Obtain user complaint information, which includes a text description of the network problem encountered by the user.

[0025] For example, when users encounter network problems, they typically log into a problem complaint website on their device and describe the problem in edited text, or they select a problem listed on the problem complaint website as their network problem. Regardless of the method, the user will receive a text description of the network problem. It is understandable that when a user submits an edited text description to the agent on the problem complaint website, the agent, in addition to obtaining the text description, can also retrieve user information such as the user's phone number and service plan (i.e., which services the user has subscribed to) from the problem complaint website. This user information, along with the text description, serves as the complaint information.

[0026] Step S120: Parse the complaint information to obtain task content including the type of complaint issue.

[0027] Specifically, the intelligent agent can invoke a large language model to perform semantic analysis on complaint information to determine whether it involves issues such as complaint hotspot identification, abnormal user cell access chain, spatial interference analysis, and dropped calls. Simultaneously, the agent can also parse the time and cell location of the user's problem from the complaint information. For example, if the complaint information is "A user with the number 135***0000 experienced a dropped call on the morning of March 24, 2026, in cell B of city A," then semantic analysis using the large language model will identify the complaint type as "dropped call." It's understandable that the large language model will also analyze key information such as the time ("March 24, 2026, 8:00 AM - 12:00 PM") and the cell ("City A, Cell B"). The key information derived from the semantic analysis of the large language model constitutes the task content.

[0028] Step S130: Based on the task content and the preset tool metadata, generate the tool call path and the call parameters of each tool on the tool call path.

[0029] An intelligent agent can maintain a such Figure 2 The tool microservice cluster includes: data fusion tool (data fusion / preprocessing tool), cell access chain tracing tool, complaint analysis tool, Top 3 time period tool, Bayesian evaluation tool, cell association analysis tool, cell spatial analysis tool, and report generation tool, etc.

[0030] Each tool has detailed tool metadata, which describes the tool's functionality, input data specifications, output result specifications, and tool dependencies. In other words, tool metadata is the foundation for an intelligent agent to dynamically schedule tools. Through standardized descriptions, the agent can understand the capabilities, inputs, outputs, and dependencies of each tool, enabling it to accurately match and combine tools after receiving the task content and generate a tool invocation path that matches the task content. Specifically, tool metadata includes: Tool ID and Name: The tool ID is a globally unique identifier for the tool to avoid name conflicts.

[0031] Tool Function Description: This describes the purpose of the tool, that is, what problems the tool can solve and what detection functions it performs. For example, a cell access chain tracing tool would be described as: analyzing the complete access path from the user's terminal to the core network and identifying performance anomalies at each node.

[0032] Input data specifications: These define the format, type, constraints, and source of the data that the tool needs to analyze (i.e., how the data is obtained, such as which datasets in the data storage and access layer it comes from).

[0033] Output specification: This defines the output content and corresponding output format of the tool.

[0034] Tool dependencies: These describe the dependencies between tools, i.e., which tools must be run before the current tool can be executed. For example, the cell access chain tracing tool depends on the Top 3 time period tool to output a precise time window.

[0035] In this step, the agent performs semantic matching and filtering based on the task content parsed in step S120 and the metadata of each tool, selects the tool that matches the task content, and then plans the path based on the tool dependencies between the selected tools, thereby generating the tool call path. It can be understood that when there are multiple complaint issue types in the task content, the agent can also analyze the causal relationships between the issues and plan the tool call path based on these causal relationships.

[0036] Meanwhile, the agent's understanding of the tool's input data specifications and output result specifications can generate calling parameters that are compatible with the tool (such as interpolation granularity, time window, and similarity threshold).

[0037] Step S140: Based on the calling parameters of each tool, control the execution of each tool according to the tool calling path, and generate the root cause of the quality complaint based on the analysis results of each tool on multi-source basic data related to the task content. That is, the agent controls each tool to execute in the order of the tool calling path, and counts the analysis results of each tool on multi-source basic data based on the calling parameters, thereby generating the root cause of the quality complaint.

[0038] In this step, such as Figure 2 As shown, based on the obtained task content, the intelligent agent can control various tools to read multi-source basic data related to the task content from the data storage and access layer. The multi-source basic data in this data storage and access layer includes: Extended Detail Record (xDR) data, Measurement Report (MR) data, cell log data, user location data, user cell access chain data, historical complaint records, cell geographic location data, and neighboring cell data, etc.

[0039] Continuing with the example above, having obtained the task details of a user's dropped call in City A, Cell B, the agent can control the tools in the cell's access path to read the corresponding xDR data, user cell access chain data, historical complaint records, and cell log data from the data storage and access layer. This data serves as the information the cell's tools need to analyze. The agent parses the source of the data to be analyzed in the tool's metadata and sends this source information to the cell's tools so they can access the relevant cell data. The agent also generates the root cause of the poor call quality complaint based on the analysis results from the cell's tools.

[0040] It should be noted that the intelligent agent can also automatically transfer and standardize the results between tools, continuously track the quality of intermediate results, and decide whether to continue, rerun, or backtrack the tool call.

[0041] In this embodiment, the intelligent agent is trained based on historical complaint records, historical multi-source basic data, historical tool calling schemes and corresponding historical root causes. The trained intelligent agent will form a thought chain for analyzing and processing complaint information as described in steps S110 to S140 above.

[0042] The wireless network quality poor complaint tracing method in this embodiment uses an intelligent agent to parse user complaint information to obtain task content including the complaint problem type and multi-source basic data related to the complaint information. Based on the task content and preset tool metadata, it generates tool call paths and call parameters for each tool on the tool call path. This realizes the dynamic generation of tool call paths according to task content, thereby dynamically selecting a combination of analysis tools more suitable for the task content, improving the diagnostic accuracy of complaint problems, accurately locating the root cause of poor quality complaints, and meeting the tracing needs of complex network scenarios. During the execution of each tool in the control tool call path, each tool performs comprehensive analysis of multi-source basic data. Because it fully covers the multi-dimensional characteristics of network problems, it significantly improves the comprehensiveness and accuracy of network problem diagnosis and root cause location compared to the analysis results of single basic data.

[0043] In some embodiments, the first tool on the tool call path is a data fusion tool, which is used to fuse the multi-source basic data. For example, the data fusion tool can unify multi-source basic data into the same time period, which is the time period involved in the complaint information. It can also align multi-source basic data by time, for example, by aligning xDR data, user cell access chain data, historical complaint records, and cell log data by time, so that each time point corresponds to one record for each of the following: xDR data, user cell access chain data, historical complaint records, and cell log data.

[0044] In this embodiment, a data fusion tool is used to fuse multi-source basic data, which facilitates comprehensive analysis by various tools based on the fused multi-source basic data and helps to pinpoint the root cause of poor quality complaints.

[0045] In some embodiments, step S130, based on the task content and preset tool metadata, generates a tool call path and call parameters for each tool on the tool call path, including: Based on the task content, the preset knowledge base, and the preset tool metadata, a tool call path and the call parameters of each tool on the tool call path are generated. The preset knowledge base records historical task content and corresponding historical tool call paths.

[0046] In this embodiment, since the preset knowledge base records the historical task content and the corresponding historical tool call path, and the historical tool call path has accurately located the root cause of the poor quality complaint in the historical task content, referring to the preset knowledge base can generate a tool call path that can more efficiently and accurately locate the root cause of the complaint problem when generating the tool call path and the tool call parameters.

[0047] In some embodiments, after generating the root cause of the quality complaint based on the analysis results of multi-source basic data related to the task content by each tool, the method further includes: Based on the current thought chain formed by the agent, the agent's reflection mechanism is triggered, and the process jumps to execute the steps of generating tool call paths and call parameters of each tool on the tool call paths based on the task content and preset tool metadata, and generates tool call paths and call parameters that are different from the previous round.

[0048] like Figure 3 As shown, the agent generates a tool invocation path based on the task content and preset tool metadata. For example, the tool invocation path is... Figure 3 The call paths ① to ⑥ are executed according to the tool call paths. After obtaining the root cause of the poor quality complaint by executing each tool according to the tool call path, the intelligent agent triggers the reflection mechanism based on the current thought chain.

[0049] Specifically, the agent's triggering of the reflection mechanism can include at least the following two situations: Scenario 1: The agent analyzes the confidence level of the root cause of the poor quality complaint. If the confidence level does not meet the confidence threshold, it indicates that the root cause of the poor quality complaint obtained by calling the tool this time is inaccurate, and a reflection mechanism needs to be triggered.

[0050] Scenario 2: The intelligent agent, considering the communication service scenario, may argue that in cases where multi-source basic data is missing, the tools used may provide inaccurate root causes for poor-quality complaints due to this lack of data, thus triggering a reflection mechanism. For example, if the analysis results of relevant tools in a community show significant interference, but the lack of data makes it impossible to determine whether the interference is internal or external, the intelligent agent can reflect on the missing neighboring community logs and trigger a route replanning behavior.

[0051] After the reflection mechanism is triggered, a new tool call path and call parameters are generated, different from the previous one, until the agent obtains what it considers the accurate root cause of the poor quality complaint.

[0052] In this embodiment, when the agent detects insufficient multi-source basic data or high uncertainty in the analysis results (i.e., low confidence), the agent automatically triggers a reflection mechanism and backtracks to the previous analysis stage. By adjusting the tool call path, optimizing parameters, or providing data supplementation suggestions, the accuracy, stability, and anti-interference ability of the root cause diagnosis of poor quality complaints are significantly improved.

[0053] In some embodiments, for user connection cells extracted from multi-source basic data, the invocation of cell-related tools is performed in the following three phases during each rethinking process: Phase 1: Filtering all the residential areas the complaining user passed through during the complaint period. Since most users describe their network problems in a broad sense, such as "yesterday my internet was slow," the description includes a wide time frame and no specific location of the residential area, the intelligent agent, upon receiving the complaint, first extracts all the residential areas the user passed through in the past 24 hours from multi-source basic data, i.e., all the residential areas the user's terminal connected to.

[0054] Phase Two: Sort all the communities by risk level to obtain the N candidate communities with the highest risk, where N is greater than or equal to 1. The specific value of N can be determined based on the community risk index. Specifically, risk assessment tools, such as Bayesian risk assessment tools, can be used to evaluate the risk index of all communities. A higher risk index indicates a greater probability that the community is the root cause of the poor quality complaints and should be analyzed first.

[0055] Phase 3: Combining the historical complaint records and spatial correlations of the N candidate cells in the multi-source basic data, call the cell-related tools, such as cell spatial analysis tools and cell correlation analysis tools.

[0056] In this embodiment, by executing the invocation of cell-related tools in three stages, cells with higher risk indices can be analyzed first, thereby improving the efficiency and accuracy of root cause analysis of poor quality complaints.

[0057] In some embodiments, the tool call path includes: a risk assessment tool, a complaint analysis tool, and a cell spatial analysis tool, and in the tool call path, the risk assessment tool is located before the complaint analysis tool and the cell spatial analysis tool. The risk assessment tool is used to assess the cell risk index, wherein the risk assessment tool may be a Bayesian risk assessment tool.

[0058] In this embodiment, the agent utilizes a built-in Bayesian risk assessment tool. This tool constructs a cell anomaly probability model based on historical complaint data. This model incorporates exponential decay and parameter adaptation mechanisms to continuously and dynamically track the cell's risk status. The agent can adjust the model's prior parameters according to real-time network conditions, making the cell risk assessment more closely reflect actual operational conditions and improving overall proactive early warning capabilities and decision-making foresight.

[0059] In some embodiments, after generating the root cause of the poor quality complaint based on the analysis results of multi-source basic data related to the task content by various tools, the method further includes: generating a structured complaint analysis report based on the root cause of the poor quality complaint. For example, the complaint analysis report includes at least three items: problem location, time period, and root cause prediction. This complaint analysis report can support multiple file formats such as JSON and PDF. Furthermore, different versions of the complaint analysis report can be generated for different groups (e.g., engineers, managers, and users) to achieve different management level displays.

[0060] Preferably, the intelligent agent can also provide corresponding optimization suggestions based on the root cause of the poor quality complaint, and write the optimization suggestions into the complaint analysis report to guide relevant personnel in handling the poor quality complaint.

[0061] The intelligent agent also writes the analysis reports into the data storage and access layer to support subsequent source tracing and intelligent agent iteration optimization. Meanwhile, to ensure user data security, the system uses SHA-256 encryption to anonymize all personally identifiable information in the analysis reports, ensuring compliance with data compliance and privacy protection requirements.

[0062] The following example, using the complaint information "A user with the SIM card number 135***0000 experienced a dropped call on the morning of March 24, 2026, in Community B, City A," illustrates the method for tracing the source of complaints regarding poor wireless network quality in this application, including: Step 1: The intelligent agent obtains the user's complaint information, namely, "the user with the number 135***0000 experienced a dropped call on the morning of March 24, 2026, in Community B of City A".

[0063] Step 2: The agent parses the complaint information to obtain task content including the complaint issue type, which is as follows: Complaint type: dropped call; Time range: 8:00-12:00 on March 24, 2026; Community: Community B in City A (with community-level accuracy).

[0064] Step 3: Based on the task content and the preset tool metadata, generate the tool call path and the call parameters of each tool on the tool call path.

[0065] Based on the task content, the agent determines that the task already has a specific time period and cell, and initially judges that the Top 3 time period tool and the cell access chain tracing tool can be skipped. Therefore, the following tool call path is generated: Use data fusion tools to fuse xDR data, user connection to the community's community log input, and historical complaint records, etc. Using Bayesian risk assessment tools: Preliminary quantification of the risk in problem areas; Use complaint analysis and community spatial analysis tools.

[0066] In this example, the agent trims unnecessary tools (Top 3 time period tools and cell access chain tracing tools) based on the task content, reducing redundant analysis and improving response efficiency.

[0067] Step 4: Based on the calling parameters of each tool, control the execution of each tool according to the tool calling path, and generate the root cause of the poor quality complaint based on the analysis results of multi-source basic data related to the task content by each tool. The execution results of each tool include the following: The intelligent agent invoked the data fusion tool to lock in and analyze multi-source basic data for the period from 08:00 to 12:00 on March 24, 2026.

[0068] The agent used a Bayesian risk assessment tool to evaluate the risk index of cell B as 92 points (high risk).

[0069] The agent determines that the user does not have a cross-cell situation, confirms that the skip cell access chain tracing tool strategy is effective, calls the complaint analysis tool, and finds that the keywords are concentrated in "dropped call" and "morning", and conducts complaint analysis based on these keywords.

[0070] The results of the cell spatial analysis tool indicate that there is a high level of coverage overlap between the neighboring cell C and the target cell B, which has a strong possibility of interference.

[0071] Step 4: Based on the results of the previous round, "Neighboring cell C and target cell B have a high level of coverage overlap area, which has a strong possibility of interference," the agent initiates a reflection mechanism and discovers that neighboring cell C lacks call drop logs for the past seven days. It is understandable that the agent's thought process includes the content that guides the discovery of "neighboring cell C lacking call drop logs for the past seven days."

[0072] After initiating the reflection mechanism, the agent regenerates the tool call path. For example, it adds a cell association analysis tool after the Bayesian risk assessment tool and supplements the dropped call logs of neighboring cell C for the past seven days. After executing the newly generated tool call path, if there are still cases of missing key data or insufficient confidence of the root cause, the reflection mechanism can be triggered iteratively in multiple rounds. If there are no missing key data and the confidence of the root cause meets the standard, a structured complaint analysis report is generated. The structured complaint analysis report is as follows: Anomaly location: Cell B; Time period: 2026-03-24, 10:00~10:06; Root cause speculation: Interference from neighboring cells + high load in this cell; Intelligent suggestions: Adjust power parameters, temporarily modify switching strategies, etc.

[0073] This example demonstrates how an intelligent agent can automatically tailor / plan analysis paths based on task content (whether it includes time, cell type, etc.), avoiding redundant calculations and improving efficiency. In different scenarios, the agent can also dynamically select tools, adjust parameters, and perform backtracking optimization based on data quality, showcasing the system's high level of intelligence and flexible adaptability.

[0074] The following describes the wireless network poor quality complaint tracing device provided in this application. The wireless network poor quality complaint tracing device described below and the wireless network poor quality complaint tracing method described above can be referred to in correspondence.

[0075] The wireless network quality poor complaint tracing device in this application embodiment is executed by an intelligent agent deployed on a computing device, such as... Figure 4 As shown, the device includes: The complaint information acquisition module 410 is used to acquire user complaint information, which includes a text description of the network problems encountered by the user.

[0076] The complaint information parsing module 420 is used to parse the complaint information to obtain task content including the type of complaint issue.

[0077] The tool call generation module 430 is used to generate a tool call path and call parameters for each tool on the tool call path based on the task content and preset tool metadata. The tool metadata includes descriptive information describing the tool's function, input data specifications, output result specifications, and tool dependencies.

[0078] The tool execution control module 440 is used to control the execution of each tool according to the tool call path based on the call parameters of each tool, and to generate the root cause of the poor quality complaint based on the analysis results of each tool on the multi-source basic data related to the task content.

[0079] In this embodiment of the wireless network quality poor complaint tracing device, the intelligent agent parses the user's complaint information to obtain task content including the type of complaint problem and multi-source basic data related to the complaint information. Based on the task content and preset tool metadata, a tool call path and the call parameters of each tool on the tool call path are generated. This realizes the dynamic generation of the tool call path according to the task content, thereby dynamically selecting a combination of analysis tools that is more suitable for the task content, improving the diagnostic accuracy of the complaint problem, accurately locating the root cause of the poor quality complaint, and meeting the tracing needs of complex network scenarios. During the execution of each tool in the control tool call path, each tool performs comprehensive analysis of multi-source basic data. Because it fully covers the multi-dimensional characteristics of network problems, compared with the analysis results of single basic data, it significantly improves the comprehensiveness and accuracy of network problem diagnosis and root cause location.

[0080] In some embodiments, the first tool on the tool call path is a data fusion tool, which is used to fuse the multi-source basic data.

[0081] In some embodiments, the tool call generation module 430 is specifically used to generate a tool call path and call parameters for each tool on the tool call path based on the task content, a preset knowledge base, and preset tool metadata. The preset knowledge base records historical task content and corresponding historical tool call paths.

[0082] In some embodiments, the wireless network quality poor complaint tracing device further includes: a reflection triggering module, which, after generating the root cause of the poor quality complaint based on the analysis results of the multi-source basic data related to the task content by each tool, triggers the agent reflection mechanism based on the agent's current thought chain, jumps to execute the tool call generation module 430, and generates a tool call path and call parameters different from the previous round.

[0083] In some embodiments, for user connection cells extracted from multi-source basic data, the invocation of cell-related tools is performed in the following three phases during each rethinking process: Filter all residential areas visited by the complaining user during the complaint period; All cells are sorted by cell risk to obtain the N candidate cells with the highest risk, where N is greater than or equal to 1; By combining the historical complaint records and spatial associations of the N candidate cells in the multi-source basic data, relevant cell-related tools are invoked.

[0084] In some embodiments, the tool call path includes: a risk assessment tool, a complaint analysis tool, and a cell spatial analysis tool, and in the tool call path, the risk assessment tool is located before the complaint analysis tool and the cell spatial analysis tool, and the risk assessment tool is used to assess the cell risk index.

[0085] In some embodiments, the wireless network quality poor complaint tracing device further includes an analysis report generation module, which is used to generate the root cause of the poor quality complaint based on the analysis results of the multi-source basic data related to the task content by each tool, provide corresponding optimization suggestions based on the root cause of the poor quality complaint, and generate a structured complaint analysis report based on the root cause of the poor quality complaint and the optimization suggestions.

[0086] Figure 5 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 5 As shown, the electronic device may include: a processor 510, a communications interface 520, a memory 530, and a communication bus 540, wherein the processor 510, the communications interface 520, and the memory 530 communicate with each other via the communication bus 540. The processor 510 can call logical instructions in the memory 530 to execute a method for tracing the source of complaints about poor wireless network quality, which is executed by an intelligent agent deployed on the computing device. This method includes: Obtain user complaint information, which includes a textual description of the network problems encountered by the user.

[0087] The complaint information is analyzed to obtain task content including the type of complaint.

[0088] Based on the task content and the preset tool metadata, a tool call path and the call parameters of each tool on the tool call path are generated. The tool metadata includes descriptive information about the tool's functions, input data specifications, output result specifications, and tool dependencies.

[0089] Based on the calling parameters of each tool, the execution of each tool is controlled according to the tool calling path, and the root cause of the poor quality complaint is generated based on the analysis results of each tool on the multi-source basic data related to the task content.

[0090] Furthermore, the logical instructions in the aforementioned memory 530 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0091] On the other hand, this application also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the wireless network quality poor complaint tracing method provided by the above methods, which is executed by an intelligent agent deployed on a computing device. The method includes: Obtain user complaint information, which includes a textual description of the network problems encountered by the user.

[0092] The complaint information is analyzed to obtain task content including the type of complaint.

[0093] Based on the task content and the preset tool metadata, a tool call path and the call parameters of each tool on the tool call path are generated. The tool metadata includes descriptive information about the tool's functions, input data specifications, output result specifications, and tool dependencies.

[0094] Based on the calling parameters of each tool, the execution of each tool is controlled according to the tool calling path, and the root cause of the poor quality complaint is generated based on the analysis results of each tool on the multi-source basic data related to the task content.

[0095] Furthermore, this application also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, is implemented to perform the wireless network quality poor complaint tracing method provided by the methods described above, executed by an intelligent agent deployed on a computing device, the method comprising: Obtain user complaint information, which includes a textual description of the network problems encountered by the user.

[0096] The complaint information is analyzed to obtain task content including the type of complaint.

[0097] Based on the task content and the preset tool metadata, a tool call path and the call parameters of each tool on the tool call path are generated. The tool metadata includes descriptive information about the tool's functions, input data specifications, output result specifications, and tool dependencies.

[0098] Based on the calling parameters of each tool, the execution of each tool is controlled according to the tool calling path, and the root cause of the poor quality complaint is generated based on the analysis results of each tool on the multi-source basic data related to the task content.

[0099] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The 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 the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0100] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0101] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for tracing the source of complaints about poor wireless network quality, characterized in that, The method, performed by an agent deployed on a computing device, includes: Obtain user complaint information, which includes a textual description of the network problems encountered by the user; The complaint information is analyzed to obtain task content including the type of complaint issue; Based on the task content and the preset tool metadata, a tool call path and the call parameters of each tool on the tool call path are generated. The tool metadata includes descriptive information describing the tool's functions, input data specifications, output result specifications, and tool dependencies. Based on the calling parameters of each tool, the execution of each tool is controlled according to the tool calling path, and the root cause of the poor quality complaint is generated based on the analysis results of each tool on the multi-source basic data related to the task content.

2. The method for tracing the source of complaints about poor wireless network quality according to claim 1, characterized in that, The first tool on the tool call path is the data fusion tool, which is used to fuse the multi-source basic data.

3. The method for tracing the source of complaints about poor wireless network quality according to claim 1, characterized in that, Based on the task content and preset tool metadata, generate tool call paths and call parameters for each tool on the tool call paths, including: Based on the task content, the preset knowledge base, and the preset tool metadata, a tool call path and the call parameters of each tool on the tool call path are generated. The preset knowledge base records historical task content and corresponding historical tool call paths.

4. The method for tracing the source of complaints about poor wireless network quality according to claim 1, characterized in that, After generating the root cause of the quality complaint based on the analysis results of multi-source basic data related to the task content from various tools, the process also includes: Based on the current thought chain formed by the agent, the agent's reflection mechanism is triggered, and the process jumps to execute the steps of generating tool call paths and call parameters of each tool on the tool call paths based on the task content and preset tool metadata, and generates tool call paths and call parameters that are different from the previous round.

5. The method for tracing the source of complaints about poor wireless network quality according to claim 4, characterized in that, For user connection cells extracted from multi-source basic data, the cell-related tools are invoked in the following three stages during each round of reflection: Filter all residential areas visited by the complaining user during the complaint period; All cells are sorted by cell risk to obtain the N candidate cells with the highest risk, where N is greater than or equal to 1; By combining the historical complaint records and spatial associations of the N candidate cells in the multi-source basic data, relevant cell-related tools are invoked.

6. The method for tracing the source of complaints about poor wireless network quality according to claim 1, characterized in that, The tool call path includes: risk assessment tool, complaint analysis tool and community spatial analysis tool. In the tool call path, the risk assessment tool is located before the complaint analysis tool and the community spatial analysis tool. The risk assessment tool is used to assess the community risk index.

7. The method for tracing the source of complaints about poor wireless network quality according to any one of claims 1 to 6, characterized in that, After generating the root cause of the quality complaint based on the analysis results of multi-source basic data related to the task content from various tools, the process also includes: Based on the root causes of the poor quality complaints, corresponding optimization suggestions are given, and a structured complaint analysis report is generated based on the root causes of the poor quality complaints and the optimization suggestions.

8. A device for tracing the source of complaints about poor wireless network quality, characterized in that, Performed by an agent deployed on a computing device, the device includes: The complaint information acquisition module is used to acquire user complaint information, which includes a text description of the network problems encountered by the user. The complaint information parsing module is used to parse the complaint information to obtain task content including the type of complaint issue. The tool call generation module is used to generate a tool call path and call parameters for each tool on the tool call path based on the task content and preset tool metadata. The tool metadata includes descriptive information describing the tool function, input data specifications, output result specifications, and tool dependency relationships. The tool execution control module is used to control the execution of each tool according to the tool call path based on the calling parameters of each tool, and to generate the root cause of the poor quality complaint based on the analysis results of each tool on multi-source basic data related to the task content.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the method for tracing the source of complaints about poor wireless network quality as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the method for tracing the source of complaints about poor wireless network quality as described in any one of claims 1 to 7.

11. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the method for tracing the source of complaints about poor wireless network quality as described in any one of claims 1 to 7.

Images