Query processing method and device, storage medium and electronic equipment

Abstract

The invention discloses a query processing method and device, a storage medium and electronic equipment, and the method comprises the steps: obtaining a data conversion attribution information set for pre-troubleshooting of an information promotion transaction, receiving promotion conversion query information of a user side for the information promotion transaction, determining a target conversion query task route and a target conversion query task processing flow corresponding to the target conversion query task route by adopting a promotion conversion processing large model based on the promotion conversion query information and the data conversion attribution information set, and executing the target conversion query task processing flow based on the data conversion attribution information set to obtain a promotion conversion query reply, and outputting the promotion conversion query reply to the user side.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a query processing method, device, storage medium and electronic device. Background Art

[0002] With the rapid development of the information promotion industry, in the process of information promotion, by setting a variety of conversion goals (such as registration, ordering, purchase, etc.), the key behaviors of users in the information promotion conversion path are tracked. Information promotion conversion: Conversion can be understood as a specific user action (such as online purchase or mobile phone call) that is valuable to a certain transaction (such as product promotion, information diversion). When a user interacts with the promotional information delivered (for example, clicks on a text ad or watches a video ad), and then performs a specific user action set above, it is counted as an information promotion conversion (or data conversion).

[0003] However, the data complexity in the information promotion delivery environment continues to increase, including multi-dimensional data sources such as user behavior data, traffic data, anti-fraud data, and conversion configuration. The efficient integration and analysis of these data has become the key to improving data conversion. In this process, information promotion attribution analysis is often involved. Information promotion attribution analysis refers to the process of tracing the user's final conversion behavior to a certain information promotion or traffic source. Attribution analysis can help optimize delivery strategies and save delivery resources. Summary of the invention

[0004] This specification provides a query processing method, device, storage medium and electronic device, and the technical solution is as follows:

[0005] In a first aspect, this specification provides a query processing method, the method comprising:

[0006] Obtain data conversion attribution information set for pre-screening of information promotion affairs;

[0007] Receive promotion conversion query information from the user end for the information promotion transaction, determine a target conversion query task route and a target conversion query task processing flow corresponding to the target conversion query task route using a promotion conversion processing large model based on the promotion conversion query information and the data conversion attribution information set, and execute the target conversion query task processing flow based on the data conversion attribution information set to obtain a promotion conversion query reply;

[0008] The promotion conversion query reply is output to the user terminal.

[0009] In a second aspect, this specification provides a query processing device, the device comprising:

[0010] An acquisition module is used to obtain a data conversion attribution information set for pre-screening of information promotion affairs;

[0011] A processing module, for receiving promotion conversion query information from a user terminal for the information promotion transaction, determining a target conversion query task route and a target conversion query task processing flow corresponding to the target conversion query task route using a promotion conversion processing large model based on the promotion conversion query information and the data conversion attribution information set, and executing the target conversion query task processing flow based on the data conversion attribution information set to obtain a promotion conversion query reply;

[0012] A reply module is used to output the promotion conversion query reply to the user terminal.

[0013] In a third aspect, the present specification provides a computer storage medium, wherein the computer storage medium stores at least one instruction, wherein the instruction is suitable for being loaded by a processor and executing the method steps of one or more embodiments of the present specification.

[0014] In a fourth aspect, the present specification provides a computer program product, wherein the computer program product stores at least one instruction, wherein the instruction is suitable for being loaded by a processor and executing the method steps of one or more embodiments of the present specification.

[0015] In a fifth aspect, the present specification provides an electronic device, which may include: a processor and a memory; wherein the memory stores a computer program, and the computer program is suitable for being loaded by the processor and executing the method steps of one or more embodiments of the present specification.

[0016] The beneficial effects brought by the technical solutions provided by some embodiments of this specification include at least:

[0017] In one or more embodiments of the present specification, the service platform integrates multi-source data in advance in the pre-troubleshooting stage of information promotion affairs to generate a data conversion attribution information set, thereby effectively reducing unnecessary repeated calculations in subsequent query processes. After the user submits the promotion conversion query information, the service platform automatically generates a target conversion query task route based on the task routing decision-making ability of the promotion conversion processing large model to select the most appropriate conversion query task processing flow, and conducts a comprehensive analysis of the causes of conversion failure from multiple dimensions such as traffic data, anti-cheating data, and conversion configuration based on the data conversion attribution information set to ensure the accuracy and efficiency of the troubleshooting process; the service platform outputs the troubleshooting results in a user-friendly form based on the promotion conversion processing large model, which lowers the threshold for understanding technical information, reduces the time and cost of manual troubleshooting, and realizes automated closed-loop processing from pre-troubleshooting to multiple rounds of queries to accurate replies, reducing the necessity of manual intervention, and significantly improving the intelligence, automation, and precision of conversion problem troubleshooting, thereby helping the platform to quickly locate the causes of abnormal promotion effects and provide solutions, thereby improving the optimization effect of information promotion. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in this specification or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 It is a scenario diagram of a query processing system provided in this specification;

[0020] Figure 2 It is a flowchart of a query processing method provided in this specification;

[0021] Figure 3 It is a flowchart of determining a data conversion attribution information set provided in this specification;

[0022] Figure 4 This is a schematic diagram of a pre-check scenario provided in this manual;

[0023] Figure 5 It is a flow chart of a large model processing provided in this manual;

[0024] Figure 6 This is a schematic diagram of a route generation process provided in this manual;

[0025] Figure 7 This is a flowchart of a target conversion query task processing flow provided in this specification;

[0026] Figure 8 This is a flowchart of a conversion query task analysis provided in this specification;

[0027] Fig. 9 It is a structural schematic diagram of a query processing device provided in this specification;

[0028] Fig.10 It is a structural schematic diagram of an electronic device provided in this manual. DETAILED DESCRIPTION

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

[0030] In the description of this specification, it should be understood that the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In the description of this specification, it should be noted that, unless otherwise clearly specified and limited, "including" and "having" and any of their variations 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 limited to the listed steps or units, but optionally also includes steps or units that are not listed, or optionally also includes other steps or units inherent to these processes, methods, products or devices. For those of ordinary skill in the art, the specific meanings of the above terms in this specification can be understood in specific circumstances. In addition, in the description of this specification, unless otherwise specified, "multiple" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that there can be three relationships, for example, A and / or B, which can represent: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the associated objects before and after are an "or" relationship.

[0031] In related technologies, the current attribution analysis process is based on data conversion problem tickets. The troubleshooting of conversion problem tickets requires the intervention of data experts and the entire query link is very time-consuming. Since data experts cannot directly locate the specific link with the problem based on the ticket, they can only check each item from top to bottom according to the information promotion conversion link. In extreme cases, a single troubleshooting ticket is very time-consuming. The main reasons for the inefficiency of traditional attribution analysis of data conversion problems are as follows:

[0032] 1. Low troubleshooting efficiency: When receiving a work order, data experts will query multiple data tables to troubleshoot problems, such as conversion result tables, traffic log tables, conversion unit configuration tables, anti-cheating tables, etc., and locate problems based on the logical connection of multiple tables. The troubleshooting efficiency is low, resulting in low work order processing efficiency.

[0033] 2. Lack of standard troubleshooting: The accuracy of locating the cause of the conversion problem is limited by the subjective logical judgment of the current on-duty data experts, and there is a lack of a standard troubleshooting process. At the same time, the reasons for attribution failure are not uniformly explained to the outside world, which can easily cause difficulties for customers to understand.

[0034] 3. High threshold for use: personnel may need to have certain data processing capabilities

[0035] These factors make it difficult for the platform to promptly and accurately identify the root cause of the problem when conversion problems arise, thus affecting the efficiency of optimizing information promotion effects.

[0036] The present specification is described in detail below with reference to specific embodiments.

[0037] See also Figure 1 , is a scenario diagram of a query processing system provided in this specification. Figure 1 As shown, the query processing system may include at least a client cluster and a service platform 100 .

[0038] The client cluster may include at least one client, such as Figure 1 As shown, it specifically includes client 1 corresponding to user 1, client 2 corresponding to user 2, ..., client n corresponding to user n, where n is an integer greater than 0.

[0039] Each client in the client cluster may be an electronic device with communication function, including but not limited to: wearable device, handheld device, personal computer, tablet computer, vehicle-mounted device, smart phone, computing device or other processing device connected to wireless modem, etc. Electronic devices may be called different names in different networks, such as: user equipment, access terminal, user unit, user station, mobile station, mobile station, remote station, remote terminal, mobile device, user terminal, electronic device, wireless communication device, user agent or user device, cellular phone, cordless phone, personal digital assistant (PDA), electronic device in 5G network or future evolution network, etc.

[0040] The service platform 100 can be a separate server device, such as a rack-mounted, blade, tower, or cabinet-mounted server device, or a workstation, mainframe computer, or other hardware device with strong computing power; it can also be a server cluster composed of multiple servers, and the servers in the service cluster can be composed in a symmetrical manner, wherein each server has equivalent functions and status in the transaction link, and each server can provide services to the outside independently, and the independent service can be understood as not requiring the assistance of other servers.

[0041] In one or more embodiments of the present specification, the service platform 100 may establish a communication connection with at least one client in the client cluster, and complete data interaction during the query processing, such as online transaction data interaction, based on the communication connection.

[0042] It should be noted that the service platform 100 and at least one client in the client cluster establish a communication connection through a network for interactive communication, wherein the network can be a wireless network or a wired network, the wireless network includes but is not limited to a cellular network, a wireless local area network, an infrared network or a Bluetooth network, and the wired network includes but is not limited to Ethernet, a universal serial bus (USB) or a controller local area network. In one or more embodiments of the specification, technologies and / or formats including Hyper Text Mark-up Language (HTML), Extensible Markup Language (XML), etc. are used to represent data (such as a target compressed package) exchanged through a network. In addition, conventional encryption technologies such as Secure Socket Layer (SSL), Transport Layer Security (TLS), Virtual Private Network (VPN), Internet Protocol Security (IPsec), etc. can also be used to encrypt all or some links. In other embodiments, customized and / or dedicated data communication technologies can also be used to replace or supplement the above data communication technologies.

[0043] The query processing system embodiment provided in this specification and the query processing method in one or more embodiments belong to the same concept. The execution subject corresponding to the query processing method involved in one or more embodiments of the specification can be the above-mentioned service platform 100; the execution subject corresponding to the query processing method involved in one or more embodiments of the specification can also be the electronic device corresponding to the client, which is determined based on the actual application environment. The implementation process of the query processing system embodiment can be detailed in the following method embodiment, which will not be repeated here.

[0044] based on Figure 1 The scenario diagram is shown, and the query processing method provided by one or more embodiments of this specification is introduced in detail below.

[0045] See also Figure 2 , is a flowchart of a query processing method for one or more embodiments of the present specification, which can be implemented by a computer program and can be run on a query processing device based on the von Neumann architecture. The computer program can be integrated into an application or run as an independent tool application. The query processing device can be a service platform.

[0046] Specifically, the query processing method includes:

[0047] S102: Obtaining a data conversion attribution information set for pre-screening of information promotion affairs;

[0048] Information promotion transactions: refers to the promotion activities conducted by the information promotion end (such as advertisers) on the transaction platform, including various conversion goals of advertising, such as registration, ordering, purchase and other user behaviors.

[0049] Pre-check: refers to the process of troubleshooting conversion issues on the data involved in the conversion event before the user (such as the user on the promotion side or the user on the service side) queries the system, so as to locate the reasons that may lead to the failure of the conversion event in advance.

[0050] The data conversion attribution information set can be understood as a standardized data set for attribution analysis integrated in the pre-screening stage. This information set contains data sets from multiple (promotion conversion) data sources and annotated conversion attribution annotation information, mainly but not limited to:

[0051] Raw data: Raw conversion data that has not been successfully attributed.

[0052] Traffic data: behavioral data such as ad display, clicks, and visits.

[0053] Anti-fraud data: system detection data on abnormal traffic and fraudulent behavior.

[0054] Attribution result data: the output of the platform’s attribution model.

[0055] Configuration data: conversion configuration data of the information promotion end (such as the tracking code of the conversion target page, return parameters, etc.).

[0056] Schematically, unattributed conversion data and data related to attribution analysis are extracted from multiple promotion conversion data sources, and the data from these data sources are integrated and preprocessed to generate a data conversion attribution information set. In the pre-screening stage, the system automatically analyzes these data based on preset attribution rules, anti-cheating models, traffic matching rules, etc., and labels the unattributed data corresponding to each information conversion event with possible conversion attribution problem labels (that is, conversion attribution annotation information). The conversion attribution annotation information corresponding to each information conversion event and the conversion attribution screening data set constitute the pre-screening data conversion attribution information set.

[0057] Example: Assume that the conversion rate of "successful order placement" for an advertiser's advertising campaign is abnormally low. The system will extract information from the following data sources in advance:

[0058]

[0059] Pre-check results (data conversion attribution information set):

[0060] The conversion ID of the conversion data corresponding to the information conversion event: C_001, and the attribution failure reason label: time window exceeded, configuration error;

[0061] It can be understood that the pre-check data conversion attribution information set is generated in real time. In the preprocessing stage of the platform system, the data conversion attribution information set is analyzed and generated in advance to provide data support for subsequent user queries and problem troubleshooting. The automation of pre-checking greatly shortens the time of manual troubleshooting, provides data support for subsequent large model query responses, and improves troubleshooting efficiency.

[0062] S104: receiving promotion conversion query information from the user end for the information promotion transaction, determining a target conversion query task route and a target conversion query task processing flow corresponding to the target conversion query task route using a promotion conversion processing large model based on the promotion conversion query information and the data conversion attribution information set, and executing the target conversion query task processing flow based on the data conversion attribution information set to obtain a promotion conversion query reply;

[0063] Promotion conversion query information: The query request for a certain information conversion event entered by the user in the intelligent agent service associated with the promotion conversion processing big model (a service entity composed of a series of custom actions, which has the ability to use the big model capabilities to assist a class of users to complete complex tasks), contains information such as the advertiser's advertising ID and conversion event type.

[0064] Information conversion events can be understood as the specific conversion behavior definition for a certain information delivery data;

[0065] Large model for promotion conversion processing: A large model for promotion conversion processing is obtained after scene adaptation using sample data based on a large language model (such as GPT) in a promotion conversion scenario. It is used to analyze the user's query content and combine it with the system's data conversion attribution information set to dynamically generate conversion query task routing and then select the most suitable task processing flow to generate the troubleshooting results required by the user.

[0066] Target conversion query task routing: refers to the task processing path information selected by the system based on the user's query requirements by promoting the conversion processing model, that is, which data sources the system needs to extract data from, which troubleshooting logic to execute, and which analysis results to generate.

[0067] Target conversion query task processing flow: A standardized troubleshooting process planned based on the target conversion query task routing, including a series of automated data analysis steps, such as: checking whether the traffic data matches, checking whether the anti-cheating data passes, checking whether the conversion configuration is correct, etc.

[0068] Schematically, the user inputs promotion conversion query information to the intelligent agent service associated with the promotion conversion processing big model, such as "Why does the conversion event 'order successful' of advertising ID 12345 have no conversion attribution?" The promotion conversion processing big model parses the user's promotion conversion query information input into intent and slot information to obtain the user's conversion query intention, for example: Intent: query the reason for conversion failure, slot information: advertising ID = 12345, conversion event type = order successful, the promotion conversion processing big model then automatically selects the target conversion query task route based on the user's conversion query intention and the pre-checked data conversion attribution information set, such as: checking traffic matching, checking anti-cheating results, and checking conversion page configuration. The promotion conversion processing big model calls resources to execute each step in the task processing flow, gradually investigates the cause of the problem and generates investigation results. The promotion conversion processing big model generates a user-friendly promotion conversion query response based on the investigation results and feeds back to the user in the form of natural language.

[0069] For example, a query submitted by a user: "Why is there no attribution for the conversion event 'Order Success' for ad ID 12345?"

[0070] The task processing flow selected by the large model for generalization and transformation processing is as follows:

[0071] Check traffic data steps: Is there any ad click record? Check result: There is an ad click record

[0072] Check anti-fraud data step: Is the click marked as fraudulent traffic? Check result: Not marked as fraudulent traffic

[0073] Check the conversion configuration steps: Is the tracking code deployed on the target page? Check result: The tracking code is missing;

[0074] The query response generated by the promotion conversion processing model by executing the task processing flow: "The conversion event 'Order Success' of ad ID 12345 was not successfully attributed because the target page lacks the conversion tracking code. Please deploy the complete tracking code in the target page."

[0075] S106: Outputting the promotion conversion query reply to the user terminal.

[0076] Optionally, the system sends the promotion conversion query reply to the user end, and the user can view the troubleshooting report on the interactive interface;

[0077] In one or more embodiments of the present specification, the service platform integrates multi-source data in advance in the pre-troubleshooting stage of information promotion affairs to generate a data conversion attribution information set, thereby effectively reducing unnecessary repeated calculations in subsequent query processes. After the user submits the promotion conversion query information, the service platform automatically generates a target conversion query task route based on the task routing decision-making ability of the promotion conversion processing large model to select the most appropriate conversion query task processing flow, and conducts a comprehensive analysis of the causes of conversion failure from multiple dimensions such as traffic data, anti-cheating data, and conversion configuration based on the data conversion attribution information set to ensure the accuracy and efficiency of the troubleshooting process; the service platform outputs the troubleshooting results in a user-friendly form based on the promotion conversion processing large model, which lowers the threshold for understanding technical information, reduces the time and cost of manual troubleshooting, and realizes automated closed-loop processing from pre-troubleshooting to multiple rounds of queries to accurate replies, reducing the necessity of manual intervention, and significantly improving the intelligence, automation, and precision of conversion problem troubleshooting, thereby helping the platform to quickly locate the causes of abnormal promotion effects and provide solutions, thereby improving the optimization effect of information promotion.

[0078] See also Figure 3 , Figure 3 This is a flowchart of determining a data conversion attribution information set proposed in one or more embodiments of this specification. For the specific implementation of obtaining the data conversion attribution information set for pre-screening of information promotion affairs, reference may be made to the following implementations:

[0079] S202: Acquire multiple promotion conversion data sources in the information promotion affairs, and regularize the data structure of each of the promotion conversion data sources to obtain a conversion attribution investigation data set;

[0080] Multiple promotion conversion data sources: Multiple data sources, including various data involved in the entire chain of information promotion delivery. For example: raw material data, traffic data, anti-cheating data, attribution result data, configuration data.

[0081] Data structure regularization: It can be understood as converting unstructured data or semi-structured data from different data sources into unified, standardized structured data to provide data support for subsequent data analysis and problem troubleshooting.

[0082] Schematically, the system extracts data sets related to promotion conversion from multiple data sources, cleans and deduplicates the extracted data sets to ensure the integrity and accuracy of the data, formats the data from different sources involved in the conversion events, and unifies the field names and data structures. For example, standardized field names such as "advertising ID", "conversion timestamp", and "conversion status" are uniformly used. The sorted data set contains all the information related to each conversion event and becomes the conversion attribution troubleshooting data set for subsequent analysis. The conversion attribution troubleshooting data set can also be called the conversion attribution troubleshooting intermediate data (set). Through some pre-built common underlying data and standardized attribution problem troubleshooting links, the conversion attribution troubleshooting intermediate data (set) is finally located to the specific reasons for the conversion roughness involved in each conversion event.

[0083] In a feasible implementation manner, the data structure regularization of each of the promotion conversion data sources to obtain a conversion attribution investigation data set may be performed by referring to the following implementation manner:

[0084] A plurality of key attribution fields for the information conversion event are extracted from each of the promotion conversion data sources, and a preset data set structure is used to regularize the data structure based on the key attribution fields to obtain a conversion attribution screening data set.

[0085] Key attribution fields refer to the core field information extracted from multiple promotion conversion data sources that supports attribution analysis and is used to build a unified data set structure. These key attribution fields cover the full-link information required for conversion attribution, such as: Conversion event information field type: the key attribution fields involved can be conversion ID and conversion time; Ad click information type: the key fields involved can be click ID, click time, and source platform; Cheating information type: the key fields involved can be cheating tags and abnormal reasons; Configuration status information type: the key fields involved can be tracking code status and return parameter status.

[0086] Schematically, the system extracts the key attribution fields for each conversion event from multiple promotion conversion data sources. Optionally, the extracted key attribution fields are filtered and classified according to the type and purpose of the field; the data fields from different sources are standardized to unify the field names and data types. For example: the click_id in the traffic data and the clk_id in the raw material data are unified as the click ID, and data cleaning is performed at the same time to remove invalid records and correct abnormal data; then the extracted key fields are integrated into a standardized conversion attribution troubleshooting data set according to the preset data set structure. The preset data set structure can be in the form of a data set table; the generated troubleshooting data set is then saved in a standardized structure for subsequent attribution analysis and troubleshooting.

[0087] It can be understood that through the above method, standardized integration of multi-source data is achieved, and data from different sources are unified into a standardized data set structure to provide high-quality data support for subsequent analysis. Automated field extraction and regularization reduce the need for manual operations and greatly improve data processing speed. The regularized data set can reflect complete conversion chain information, help accurately locate attribution problems, and improve advertising optimization effects.

[0088] S204: Perform conversion status attribution processing based on the conversion attribution investigation data to obtain conversion attribution annotation information of each information conversion event;

[0089] Conversion status attribution processing means that the system analyzes the raw data involved in the conversion event through the attribution model or attribution problem troubleshooting link, determines its "conversion status", and marks the conversion attribution information of the "conversion status".

[0090] "Conversion Status" refers to the transaction completion status of each (information) conversion event in the entire information promotion delivery link, including but not limited to: successful conversion status (the promotion covered user completed the conversion target set by the (information) conversion event (such as registration, order placement, purchase), unsuccessful conversion status (the promotion covered user did not complete the conversion target set by the (information) conversion event (such as registration, order placement, purchase), unsuccessful attribution status (the promotion covered user completed the conversion target, but the system was unable to associate the conversion behavior with the information promotion click data), etc.;

[0091] The conversion attribution annotation information is a problem cause label for each conversion event, describing the specific cause of the conversion status (successful conversion, unsuccessful conversion, unsuccessful attribution) of the conversion event, such as: time window limit exceeded, traffic data matching failed, configuration error, etc.

[0092] In a feasible implementation, the data of each conversion event is read from the conversion attribution troubleshooting data set, and the "conversion status" is determined according to the user's behavior trajectory and the conversion target configuration of the conversion event. Then, the attribution rule check is performed on the "conversion status" to obtain the inspection result, such as checking the time window rule (whether the conversion occurs within the effective time range after the advertisement is clicked), checking the traffic data matching rule (whether there is a click record that matches the conversion event), checking the anti-cheating rule (whether the user is marked as cheating traffic), checking the configuration item status (whether the target page has deployed the tracking code, whether the return parameters are correct), etc.; based on the inspection result, the conversion attribution annotation information of the conversion event is generated, and the data corresponding to the conversion event is annotated with the conversion attribution annotation information in the conversion attribution troubleshooting data set.

[0093] Furthermore, the implementation of conversion status attribution processing enables the system to more accurately analyze the actual business status of each conversion event and generate clear annotation information for events that are not successfully attributed, helping advertisers to quickly locate the cause of the problem and improve the efficiency of optimizing advertising conversion effects.

[0094] In a feasible implementation manner, the conversion status attribution processing based on the conversion attribution screening data is performed to obtain the conversion attribution annotation information of each information conversion event, and reference may be made to the following implementation manner:

[0095] A2: using the attribution problem pre-checking link to perform conversion attribution detection processing on the conversion attribution checking data set to obtain multiple pre-checking link item record results;

[0096] Attribution problem pre-check link: It is a set of standardized rules for analyzing conversion event attribution problems. It performs attribution detection tasks step by step according to the link items to analyze the potential causes of conversion failure. The pre-check link is defined in a modular way. Each link item corresponds to a specific detection task, which is connected in series in a step-by-step manner to perform preset attribution problem troubleshooting tasks such as attribution rules, configuration detection, and cheating screening.

[0097] For example:

[0098] Time window detection link item: The detection content is to determine whether the conversion time is within the effective time range after the click;

[0099] Traffic matching detection link item: The detection content is to check whether the conversion can match the click behavior;

[0100] Configure the detection link item: The detection content is to check whether the tracking code of the advertiser's main page is deployed correctly;

[0101] Exemplarily, the pre-check link is initialized first, which can be a preset attribution problem pre-check link loaded by the system, and each link item corresponds to an independent attribution detection module. Link example: time window detection → traffic matching detection → configuration detection → anti-cheating screening.

[0102] Then, link detection is performed item by item according to the pre-checked links. The input is the conversion attribution check data set. The processing process is to detect each conversion event through each link item one by one and generate the corresponding detection results. The output is the link item record result of each conversion event. After each link item detection is completed, the detection result is recorded as a link item result set, such as whether the detection is passed and the reason for the detection failure;

[0103] A4: Generate conversion attribution annotation information corresponding to each information conversion event based on the pre-check link item record result.

[0104] Indicatively, the link item record results are integrated first, the link item record results of each conversion event are summarized, and the link items that failed the test and their reasons are analyzed; then the attribution annotation information is generated, that is, the system generates problem labels based on the failed link items, such as "configuration error", "time window exceeded", etc. You can also attach solution suggestions to each problem label, such as "check the tracking code deployment of the target page". Then output the annotation information, which includes but is not limited to: conversion status (successful attribution, unattributed, abnormal attribution), specific problem labels for the reasons for unattribution, and solution suggestions for each problem label.

[0105] In schematic form, through the modular detection process of the pre-check link, the system can quickly locate the specific cause of the conversion failure and provide the platform with accurate problem analysis. The annotation information can be supplemented with solution suggestions to help the platform quickly fix the problem and optimize the conversion effect of information delivery. At the same time, the link-based detection logic reduces invalid full-volume analysis, improves the troubleshooting efficiency, and provides high-quality data support for subsequent attribution processing.

[0106] For example, please refer to Figure 4 , Figure 4 This is a schematic diagram of a pre-screening scenario. Figure 4 The overall process of transformation problem identification is described, covering the entire process from extraction of underlying data sources to identification and annotation of transformation problems. Figure 4 The bottom layer is the data source layer corresponding to various promotion conversion data sources, including various data sources related to advertising promotion and user conversion. These data are the basis for the entire problem troubleshooting and attribution analysis.

[0107] Promotion conversion data sources such as:

[0108] ZFB_TRADE: Payment transaction data, used to identify user behavior in the transaction link.

[0109] LEAVE_INFO: User exit information, recording the user's stay and exit behavior on the target page.

[0110] MEMBER_CARD: Membership card data, used to identify member-related conversion behaviors.

[0111] BENEFIT_EXCHANGE: Benefit exchange record, used to track the user's behavior of receiving coupons or other benefits.

[0112] Traffic data: the user’s behavioral path from clicking on an ad to visiting the target page.

[0113] Anti-fraud data: Data that detects and flags invalid traffic or abnormal behavior.

[0114] Based on multiple promotion conversion data sources, the system integrates and regularizes multiple data sources to generate conversion attribution troubleshooting intermediate data, that is, conversion attribution troubleshooting data set, by executing S202. This intermediate data set is the core data support for subsequent attribution analysis and problem annotation. The system locates conversion attribution problems by executing S204. After generating intermediate data, the system locates attribution problems for conversion events based on standardized attribution problem troubleshooting links to help identify the reasons for the failure of each conversion event. The attribution problem troubleshooting link may include a standard attribution problem location link and a custom attribution problem location link;

[0115] The standard attribution problem location link contains a series of preset detection rules, which are used to automatically check whether the conversion event meets the attribution conditions. The detection content may include:

[0116] Click attribution rule targeting: Check whether the conversion event matches the ad click record (such as a click within a time window).

[0117] Raw material rule configuration positioning: Check whether the conversion event meets the preset attribution rules (such as conversion behavior path).

[0118] Traffic anomaly data detection: Mark abnormal traffic and eliminate cheating or invalid clicks.

[0119] The custom attribution problem location link includes information delivery end that can configure personalized attribution detection rules according to specific transaction requirements. Example: Customized analysis of attribution rules for certain special conversion events (such as large orders).

[0120] Furthermore, based on the link record results of the attribution problem location, the system generates problem annotation information for each conversion event. This annotation information is used to identify the reasons for the conversion failure and provide optimization suggestions. The annotation process can be:

[0121] 1. Integrate link results: Summarize the items that failed the link detection for each conversion event and analyze the reasons for failure.

[0122] 2. Generate annotation information: The system generates problem labels based on the link items that failed detection, such as: time window limit exceeded, configuration error (such as missing tracking code of the target page), traffic matching failure; and additional solution suggestions, such as: adjusting the advertising delivery time window, checking the deployment of tracking code of the target page;

[0123] S206: Generate a pre-screened data conversion attribution information set based on the conversion attribution annotation information and the conversion attribution screening data set.

[0124] Pre-check data conversion attribution information set: refers to the complete and standardized attribution data set generated by the system in the pre-check stage, including the attribution status, annotation information, problem causes, etc. of each conversion event.

[0125] Indicatively, the system merges the conversion attribution troubleshooting data set with the annotation information to form a complete pre-troubleshooting information set and generates pre-troubleshooting results. The attribution results, problem causes, solution suggestions, and other information of each conversion event are included in the pre-troubleshooting information set.

[0126] In one or more embodiments of the present specification, a full-link automated processing system is constructed from multi-source data extraction to conversion attribution problem annotation. First, through S202, the data structure of various promotion conversion data sources is regularized to generate a standardized conversion attribution troubleshooting data set, providing high-quality data support for subsequent analysis; then in S204, through the preset attribution problem troubleshooting link, the attribution status of the conversion event in terms of time window, traffic matching, configuration correctness, etc. is gradually detected, and the detection results are recorded; finally, S206 generates annotation information for each conversion event based on the detection record, clarifies the reasons for non-attribution or failure, and adds optimization suggestions. On the whole, this process greatly improves the efficiency and accuracy of the analysis of conversion problems, reduces manual intervention, and provides technical support for users of information promotion and delivery to quickly troubleshoot conversion problems and optimize delivery strategies.

[0127] See also Figure 5 , Figure 5 This is a flowchart of a large model processing proposed in one or more embodiments of this specification. For the specific implementation of the determination of the target conversion query task route and the target conversion query task processing flow corresponding to the target conversion query task route by using the promotion conversion processing large model based on the promotion conversion query information and the data conversion attribution information set, reference may be made to the following implementations:

[0128] S3002: Inputting the promotion conversion query information and the data conversion attribution information set into the promotion conversion processing model;

[0129] S3004: understanding the intent of the promotion conversion query information through the promotion conversion processing large model to obtain the user conversion query intent, and generating a target conversion query task route based on the user conversion query intent and the data conversion attribution information set;

[0130] Schematically, the user inputs promotion conversion query information to the intelligent agent service associated with the promotion conversion processing big model, such as "Why does the conversion event 'order success' of advertisement ID: 12345 have no conversion attribution?" The promotion conversion processing big model parses the user's promotion conversion query information input into intent and slot information to obtain the user's conversion query intent, for example: intent: query the reason for conversion failure, slot information: advertisement ID = 12345, conversion event type = order success, and the subsequent promotion conversion processing big model automatically generates or determines the target conversion query task route based on the user's conversion query intent and the pre-screened data conversion attribution information set;

[0131] Optional, see Figure 6 , Figure 6 The present invention is a flowchart of route generation. Specifically, the target conversion query task route generated based on the user conversion query intention and the data conversion attribution information set may be generated by referring to the following method:

[0132] S4002: Determine the conversion query task type based on the user conversion query intention, and determine the conversion attribution annotation information and conversion attribution screening data set for the information conversion event in the data conversion attribution information set;

[0133] User conversion query intent: The goals and needs expressed by users during query, extracted through the natural language understanding (NLU) technology of the big model. For example: querying the reasons for conversion failure of an advertising ID, querying the attribution status of a specific conversion event, and analyzing the attribution statistics of a certain advertising campaign.

[0134] Conversion query task type: The task type is a query target category generated based on user intent, which is used to guide the subsequent task routing generation. For example: Failure cause troubleshooting task: Analyze the failure cause of a certain conversion event. Attribution status query task: Check the attribution success rate and statistics of a certain advertising campaign. Problem optimization suggestion task: Generate optimization suggestions for attribution problems.

[0135] Conversion attribution labeling information: Conversion attribution labeling information is the attribution analysis result of a single conversion event, identifying the reasons for attribution failure and optimization suggestions. For example: Problem label: time window exceeded, traffic matching failed, configuration error.

[0136] Schematically, the system extracts keywords and context information from the natural language query input by the user, and the promotion conversion processing large model parses the keywords and context information through natural language understanding (NLU) to obtain query intent and slot information. The user's query target is determined based on the query intent and slot information, and then the query target is mapped to a predefined conversion query task type. Furthermore, the system parses the conversion attribution annotation information and the conversion attribution troubleshooting data set for the information conversion event corresponding to the user's conversion query intent from the data conversion attribution information set, and extracts the attribution annotation information that matches the conversion event corresponding to the user's conversion query intent. The annotation information is used to provide failure reason labels related to the current query and even references to solution suggestions.

[0137] S4004: Determine multiple data source selection information and data conversion attribution environment information for the conversion attribution screening data set based on the conversion attribution annotation information;

[0138] Information from multiple data sources: Select appropriate data sources based on conversion attribution annotation information, including but not limited to: Time-related data: click time and conversion time in traffic data, Configuration-related data: tracking code status and return parameter status in configuration data. Traffic validity data: cheating tags in anti-cheating data.

[0139] Data conversion attribution environment information: describes the contextual environment related to the current attribution analysis target corresponding to the user's conversion query intent, and is used to guide the execution of the troubleshooting logic. Examples include but are not limited to: attribution rule environment: such as time window rules and traffic matching logic; platform feature environment: such as the attribution logic of a specific advertising platform.

[0140] Indicatively, the promotion conversion processing model parses the attribution annotation information, extracts the problem labels and even the suggestion information, and preliminarily determines the problem categories that need to be investigated based on the content of the annotation information. For example, if the annotation information contains "time window exceeded", time-related data needs to be selected. If the annotation information contains "configuration error", configuration-related data needs to be selected.

[0141] According to the question category, query the data source that matches the question category label, so as to extract the target data source information and generate multiple data source selection information. For example, for time-related questions, select the click time and conversion time in the traffic data. For configuration-related questions, select the tracking code status and return parameter status in the configuration data.

[0142] Furthermore, the promotion conversion processing model analyzes the attribution rule environment and the platform feature environment. Analyzing the attribution rule environment means checking the context information of the attribution rule, such as the time window length and traffic matching conditions. Analyzing the platform feature environment can be understood as checking whether the current advertising platform has special attribution rules or logic. For example, some advertising platforms may have customized click matching window rules.

[0143] The data conversion attribution environment information is generated by combining the attribution rule environment analysis results and the platform characteristic environment analysis results.

[0144] It is understandable that by parsing the content of the annotation information, the data source fields that need to be extracted are automatically determined to ensure the efficiency and pertinence of task execution, and the data source fields required for the task are dynamically matched according to the attribution annotation information to avoid redundant calculations; the attribution rules and platform characteristic environment information are dynamically analyzed to support personalized attribution problem troubleshooting logic and improve the accuracy of the problem troubleshooting logic. The data source selection and environment information are closely integrated to support the optimization of complex attribution task processing flows, significantly improving the efficiency and pertinence of conversion problem troubleshooting.

[0145] S4006: Determine task processing flow allocation logic and task routing configuration prompt based on the conversion query task type, the multiple data source selection information and the data conversion attribution environment information;

[0146] The task processing flow allocation logic describes how to decompose the query task into a series of specific detection tasks and determine the execution order and dependencies of each task. For example, the time window detection is executed first. The flow matching detection is executed after the time window detection passes.

[0147] Task routing configuration hints are auxiliary information for task processing flows, which are used to dynamically adjust the order or logic of task flows, for example, to prioritize certain detection tasks for specific problems. Optimize the execution mode of detection tasks based on data environment information.

[0148] Schematically, the generalized conversion processing large model analyzes the conversion query task type, the multiple data source selection information and the data conversion attribution environment information, determines the subtasks in the task processing flow according to the conversion query task type, and generates a list of detection tasks that need to be executed according to the subtasks. For example: the failure cause troubleshooting task may include: time window detection, traffic matching detection, configuration status detection, anti-cheating detection; then, determine the execution order of the subtasks in the detection task list, the execution order of the subtasks in the task list determines the task processing flow allocation logic, for example, it can be based on the priority to adjust the subtask execution order, perform task dependency analysis to determine the execution order; then optimize the execution allocation logic of the task processing flow according to the data conversion attribution environment information;

[0149] Meanwhile, the task routing configuration prompt is an auxiliary description prompt for the task processing flow, which is used to dynamically adjust the execution logic of the task flow. For example: prompting that some detection tasks can be executed in parallel, and specifying specific parameters (such as time window length, anti-cheating threshold) for some detection tasks.

[0150] S4008: Generate a target transformed query task route based on the task processing flow allocation logic and the task routing configuration prompt.

[0151] The target transformed query task route is the generated complete query task path, which includes all subtasks of the execution task flow and their order, parameters, and logical configurations.

[0152] Schematically, integrate the task processing flow allocation logic and the routing configuration prompt. By parsing the task processing flow allocation logic, arrange each detection task in sequence according to the allocation logic. Then apply the routing configuration prompt to generate the target transformed query task route, that is, adjust the task execution method in combination with the prompt information, such as executing some tasks in parallel and dynamically adjusting task parameters. Then generate the target transformed query task route.

[0153] In this specification, a dynamic task processing framework from user intention parsing to task route generation is constructed. First, S4002 accurately extracts the user's query intention, matches relevant annotation information and troubleshooting data sets, and provides data support for task planning; S4004 intelligently analyzes the annotation information, dynamically selects multiple data sources and environmental information to ensure the pertinence and context adaptation of the troubleshooting task; S4006 generates an optimized task processing flow allocation logic based on the task type and environmental information, clarifies the execution order and supports parallel tasks; finally, S4008 combines the allocation logic and the configuration prompt to generate a complete target task route, covering all key detection steps. The overall process realizes an efficient, accurate, and intelligent transformation problem troubleshooting path planning, significantly improves the troubleshooting efficiency and user experience, and provides technical support for advertisers to quickly locate transformation problems.

[0154] S3006: Determine the target transformed query task processing flow of the target transformed query task route through the promotion transformation processing large model.

[0155] The target transformed query task processing flow is a series of specific detection steps and execution logics corresponding to the task route, which clarifies the input data, execution rules, and output results of the task. For example: Input: click time, conversion time. Execution rule: Determine whether the conversion time is within 30 minutes after the click time. Output: Whether the time window is valid.

[0156] Schematically, the promotion transformation processing large model parses the target transformed query task route to determine the detection tasks that need to be executed. For example: the route includes time window detection, traffic matching detection, and configuration detection.

[0157] Then, a specific target conversion query task processing flow is generated, and specific execution logic is generated for each detection task according to the task type and input data.

[0158] At the same time, the target conversion query task processing flow can be optimized to obtain an optimized target conversion query task processing flow, which can prioritize tasks with low dependencies, such as time window detection. Independent tasks can be executed in parallel, such as configuration detection and anti-cheating detection. It can be a dynamic adjustment of parameters: based on the attribution environment information, dynamically adjust the parameters of the detection task, such as the length of the time window rule.

[0159] Example: Target Task Processing Flow

[0160] 1) Input information:

[0161] Target task routing: time window detection → traffic matching detection → configuration detection.

[0162] Data source: Click time: 2025-01-15 10:00. Conversion time: 2025-01-15 10:45. Tracking code status: Missing.

[0163] 2) Generate processing flow:

[0164] Time window detection: Input: click time (2025-01-15 10:00), conversion time (2025-01-1510:45). Rule: The time window is 30 minutes. Output: The time window is invalid.

[0165] Traffic matching detection: Input: click ID, conversion ID. Rule: Click records must match conversion records. Output: Not executed (time window detection failed).

[0166] Configuration check: Input: Tracking code status. Rule: The target page must have the tracking code deployed. Output: Configuration error.

[0167] S3008: Execute the target conversion query task processing flow based on the data conversion attribution information set to obtain a promotion conversion query response.

[0168] The promotion conversion query response is a response generated based on the execution results of the processing flow, including the troubleshooting results of the conversion problem and optimization suggestions. For example:

[0169] Cause of the problem: time window exceeded, configuration error.

[0170] Solution: Adjust the time window rules and check the tracking code deployment of the target page.

[0171] In schematic form, the data required for the task, such as click time, conversion time, and configuration status, is extracted from the data conversion attribution information set, and the task flow is executed step by step. The detection task is executed step by step according to the allocation logic of the processing flow, and the execution result of each task is recorded. The execution results of all tasks are integrated into a complete troubleshooting report, including the pass status and failure reason of each task. At the same time, optimization suggestions can also be generated, that is, targeted suggestions are generated based on the detection results to help users quickly solve problems.

[0172] In one or more embodiments of this specification, a complete closed loop from target task routing to task flow execution and query response generation is achieved through the above method. The task flow is dynamically generated according to user queries and data environment to ensure the optimal path. By gradually executing the task flow, the root cause of the conversion problem can be accurately located, and detailed troubleshooting reports and optimization suggestions can be provided to help users quickly solve conversion problems. The processing efficiency and user experience of conversion query tasks are greatly improved, and an accurate and fast problem diagnosis tool is provided for information delivery conversion.

[0173] Optional, see Figure 7 , Figure 7 The present invention is a flowchart of a target conversion query task processing flow. Specifically, the target conversion query task processing flow is executed based on the data conversion attribution information set to obtain a promotion conversion query response. The following method can be referred to:

[0174] S5002: parsing the task route configuration prompt in the target conversion query task route, and selecting a target call data source from the data conversion attribution information set based on the task route configuration prompt;

[0175] According to some embodiments, the task routing configuration prompt can be understood as auxiliary information for dynamically optimizing task execution, including data source selection, execution order adjustment, task parameter setting, etc. For example: Data source selection prompt: traffic data (click time) and configuration data (tracking code status) are required. Sequence prompt: check the time window first.

[0176] Target call data source: A specific data set for task execution extracted from the data conversion attribution information set according to the configuration prompt. For example: click time, conversion time (traffic data), tracking code status (configuration data).

[0177] The promotion conversion processing model extracts the task routing configuration prompts from the target conversion query task routing, parses the task routing configuration prompts in the target conversion query task routing, determines the required target data source, and determines the priority and dependency of data extraction. According to the task routing configuration prompts, the fields required for the task are extracted from the data conversion attribution information set. Specifically, the priority fields can be extracted: fields with high task dependency are selected first, for example: time window detection prioritizes click time and conversion time, and configuration detection selects tracking code status and return parameter status.

[0178] S5004: importing the data conversion attribution information set and the target call data source into the target conversion query task processing flow, and performing conversion query task parsing through the target conversion query task processing flow to obtain promotion conversion query results;

[0179] Conversion query task processing flow: The specific execution path corresponding to the task routing defines the input, rules, and output of each detection task. For example: Time window detection: input click time and conversion time, and output whether the time window is valid.

[0180] Promotion conversion query results: The analysis results after the task flow is executed, including a summary of the pass status and failure reasons of each detection task.

[0181] Schematically, the data conversion attribution information set and the target call data source are imported into the target conversion query task processing flow, the data conversion attribution information set provides the global background information of the task flow execution, and the target call data source provides the specific input data of each detection task in the task flow. The detection tasks are executed in sequence according to the task flow allocation logic, and the promotion conversion query results of each task are recorded, including the passing status and the failure reason.

[0182] S5006: Based on the promotion conversion query result, return result packaging is performed to obtain a promotion conversion query reply.

[0183] Return result encapsulation processing: Convert the promotion conversion query results into user-friendly response content, including the cause of the problem and solution suggestions.

[0184] Promotion conversion query response: The query feedback finally provided to users, including troubleshooting results and optimization plans.

[0185] In one or more embodiments of this specification, through the implementation of S5002-S5008, a complete closed loop from target data source extraction to task flow execution, and then to result generation and optimization suggestion output is completed, and the data required for the task is efficiently extracted through dynamic parsing configuration prompts. The task flow is executed step by step, and clear troubleshooting results can be output. At the same time, the detection results are encapsulated, targeted optimization suggestions are provided, and user experience and problem solving efficiency are improved. The above process significantly enhances the automation and intelligence level of advertising conversion problem troubleshooting.

[0186] Optional, see Figure 8 , Figure 8 It is a flow chart of conversion query task analysis, specifically implementing the importing of the data conversion attribution information set and the target call data source into the target conversion query task processing flow, and performing conversion query task analysis through the target conversion query task processing flow to obtain promotion conversion query results, which can refer to the following methods:

[0187] S6002: Importing the data conversion attribution information set and the target call data source into the target conversion query task processing flow, and performing task information parsing through the target conversion query task processing flow to obtain tracking missing task information for the dialog state tracking mechanism;

[0188] Dialog State Tracking (DST): The large model uses multiple rounds of dialogue to dynamically track user intent, context, and unfinished task information to ensure the integrity of task information in multiple rounds of interaction.

[0189] Tracking missing task information: refers to the information required for detection tasks that cannot be performed in the current task flow due to insufficient input data, or the information required for detection tasks that cannot be performed in the current task flow because the current target query intent is too large and the required query scope needs to be further refined. For example, the user did not explicitly provide an advertising ID or a time range for the query.

[0190] Indicatively, the data conversion attribution information set and the target call data source are imported into the task processing flow. According to the requirements of the task processing flow, identify whether there are insufficient input data, missing context information, narrowing the query intent, and other task requirements, and then check whether each detection task has complete input data. For example, time window detection requires click time and conversion time, and traffic matching detection requires click ID and conversion ID. If there are unmet data requirements in the target conversion query task processing flow, the promotion conversion processing model marks this information as "tracking missing task information."

[0191] Example: Input data: Data conversion attribution information set: including click time and configuration status. Call data source: click time (2025-01-15 10:00), configuration status (missing).

[0192] Task flow requirements: Time window detection: click time and conversion time are required. Traffic matching detection: click ID and conversion ID are required.

[0193] Parsing results: Tracking missing task information: conversion time (time window detection dependency). Conversion ID (traffic matching detection dependency).

[0194] S6004: generating at least one round of information guidance inquiry based on the tracking missing task information, outputting the information guidance inquiry to the target user through the promotion conversion processing large model, and receiving user supplementary context information for the information guidance inquiry;

[0195] Information-guided inquiries: Automatically generate supplementary inquiries for users based on the missing information in the task flow. For example: "Please provide the conversion event time range of ad ID 12345", "Conversion ID is not provided, please provide the corresponding conversion ID";

[0196] User-supplemented contextual information: additional information supplemented by users through query interaction, which is used to improve the input data of the task flow.

[0197] Indicatively, the missing task information is parsed to generate query content item by item. The promotion conversion processing model can be used to generate targeted information to guide the query for the missing task information, ensuring that the query content is clear and easy to understand. The system displays the query content to the user through the dialogue interface. Then the user input is received, and the user completes the input data by supplementing the context information, such as providing a conversion ID or time range.

[0198] S6006: Based on the user's supplementary context information, conversion query task processing is performed through the target conversion query task processing flow to obtain a promotion conversion query result.

[0199] Schematically, the promotion conversion processing model updates the task input by integrating the user's supplementary context information, integrating the user's supplementary context information with the original input data. At the same time, the task flow is re-executed to ensure the integrity of the task flow after the missing information is supplemented. Then, the detection tasks are executed step by step based on the task flow, that is, the detection tasks are executed in sequence according to the task allocation logic, and the execution status and output results of each task are recorded to obtain the promotion conversion query results.

[0200] In one or more embodiments of the present specification, closed-loop processing of dynamic information tracking, user interaction supplementation, and complete task flow execution of conversion query tasks is achieved through the above-mentioned method. First, S6002 accurately locates missing information by parsing task flow dependencies to ensure that task execution requirements are fully identified; then, S6004 dynamically generates personalized information-guided inquiries, and efficiently supplements contextual data in combination with user interactions; finally, S6006 integrates the supplementary information to re-execute the task flow to generate accurate query results and problem analysis. The overall process has the characteristics of efficient information tracking, intelligent user guidance, and dynamic task optimization, which greatly improves the troubleshooting efficiency and user experience of complex conversion problems, while reducing processing interruptions caused by missing information, and provides accurate and actionable optimization suggestions for promotional information delivery conversion.

[0201] The following will be combined Fig. 9 , the query processing device provided in this specification is introduced in detail. It should be noted that, Fig. 9 The query processing device shown is used to execute the Figures 1 to 8 For the convenience of explanation, only the parts related to this specification are shown. For the specific technical details not disclosed, please refer to this specification. Figures 1 to 8 The embodiment shown.

[0202] See also Fig. 9 , which shows a schematic diagram of the structure of the query processing device of this specification. The query processing device 1 can be implemented as all or part of the user electronic device through software, hardware or a combination of both. According to some embodiments, the query processing device 1 includes an acquisition module 11, a processing module 12 and a reply module 13, which are specifically used to:

[0203] An acquisition module 11 is used to acquire a data conversion attribution information set for pre-screening of information promotion affairs;

[0204] The processing module 12 is used to receive the promotion conversion query information of the user end for the information promotion transaction, determine the target conversion query task route and the target conversion query task processing flow corresponding to the target conversion query task route by using the promotion conversion processing large model based on the promotion conversion query information and the data conversion attribution information set, and execute the target conversion query task processing flow based on the data conversion attribution information set to obtain a promotion conversion query reply;

[0205] The reply module 13 is used to output the promotion conversion query reply to the user terminal.

[0206] In a feasible implementation manner, obtaining a data conversion attribution information set for pre-screening of information promotion affairs includes:

[0207] Acquire multiple promotion conversion data sources in information promotion affairs, and regularize the data structure of each of the promotion conversion data sources to obtain a conversion attribution investigation data set;

[0208] Perform conversion status attribution processing based on the conversion attribution investigation data to obtain conversion attribution annotation information for each of the information conversion events;

[0209] A pre-screened data conversion attribution information set is generated based on the conversion attribution annotation information and the conversion attribution screening data set.

[0210] In a feasible implementation manner, the step of regularizing the data structure of each of the promotion conversion data sources to obtain a conversion attribution investigation data set includes:

[0211] A plurality of key attribution fields for the information conversion event are extracted from each of the promotion conversion data sources, and a preset data set structure is used to regularize the data structure based on the key attribution fields to obtain a conversion attribution screening data set.

[0212] In a feasible implementation manner, the conversion status attribution processing based on the conversion attribution screening data is performed to obtain the conversion attribution annotation information of each information conversion event, including:

[0213] Using the attribution problem pre-checking link to perform conversion attribution detection processing on the conversion attribution checking data set to obtain multiple pre-checking link item record results;

[0214] Based on the pre-check link item record results, conversion attribution annotation information corresponding to each of the information conversion events is generated.

[0215] In a feasible implementation, the method of determining a target conversion query task route and a target conversion query task processing flow corresponding to the target conversion query task route by using a promotion conversion processing large model based on the promotion conversion query information and the data conversion attribution information set includes:

[0216] Input the promotion conversion query information and the data conversion attribution information set into the promotion conversion processing model,

[0217] The promotion conversion query information is understood by the promotion conversion processing large model to obtain the user conversion query intention, and a target conversion query task route is generated based on the user conversion query intention and the data conversion attribution information set;

[0218] The target conversion query task processing flow of the target conversion query task routing is determined by promoting the conversion processing large model.

[0219] In a feasible implementation manner, the generating of a target conversion query task route based on the user conversion query intention and the conversion attribution annotation information in the promotion conversion query information includes:

[0220] Determine the conversion query task type based on the user conversion query intention, and determine the conversion attribution annotation information and conversion attribution screening data set for the information conversion event in the data conversion attribution information set;

[0221] Determine, based on the conversion attribution annotation information, multiple data source selection information and data conversion attribution environment information for the conversion attribution screening data set;

[0222] Determine the task processing flow allocation logic and task routing configuration prompt based on the conversion query task type, the multiple data source selection information and the data conversion attribution environment information;

[0223] A target conversion query task route is generated based on the task processing flow allocation logic and the task route configuration prompt.

[0224] In a feasible implementation manner, the step of executing the target conversion query task processing flow based on the data conversion attribution information set to obtain a promotion conversion query response includes:

[0225] Parsing the task route configuration prompt in the target conversion query task route, and selecting a target call data source from the data conversion attribution information set based on the task route configuration prompt;

[0226] Importing the data conversion attribution information set and the target call data source into the target conversion query task processing flow, and performing conversion query task analysis through the target conversion query task processing flow to obtain promotion conversion query results;

[0227] The return result is packaged based on the promotion conversion query result to obtain a promotion conversion query reply.

[0228] In a feasible implementation manner, the data conversion attribution information set and the target call data source are imported into the target conversion query task processing flow, and the conversion query task is parsed through the target conversion query task processing flow to obtain the promotion conversion query result, including:

[0229] Importing the data conversion attribution information set and the target call data source into the target conversion query task processing flow, and performing task information parsing through the target conversion query task processing flow to obtain tracking missing task information for the dialog state tracking mechanism;

[0230] Generate at least one round of information-guiding inquiries based on the tracking missing task information, output the information-guiding inquiries to the target user through the promotion conversion processing model, and receive user supplementary context information for the information-guiding inquiries;

[0231] Based on the user's supplementary context information, conversion query task processing is performed through the target conversion query task processing flow to obtain a promotion conversion query result.

[0232] It should be noted that the query processing device provided in the above embodiment only uses the division of the above functional modules as an example when executing the query processing method. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the query processing device provided in the above embodiment and the query processing method embodiment belong to the same concept, and the implementation process thereof is detailed in the method embodiment, which will not be repeated here.

[0233] The above serial numbers in this specification are for description only and do not represent the advantages or disadvantages of the embodiments.

[0234] In this specification, the service platform integrates multi-source data in advance to generate a data conversion attribution information set in the pre-screening stage of information promotion affairs, thereby effectively reducing unnecessary repeated calculations in the subsequent query process. After the user submits the promotion conversion query information, the service platform automatically generates a target conversion query task route based on the task routing decision-making ability of the promotion conversion processing large model to select the most appropriate conversion query task processing flow, and conducts a comprehensive analysis of the reasons for conversion failure from multiple dimensions such as traffic data, anti-cheating data, and conversion configuration based on the data conversion attribution information set to ensure the accuracy and efficiency of the screening process; the service platform outputs the screening results in a user-friendly form based on the promotion conversion processing large model, which lowers the threshold for understanding technical information, reduces the time and cost of manual screening, and realizes automated closed-loop processing from pre-screening to multiple rounds of queries to accurate replies, reducing the necessity of manual intervention, and significantly improving the intelligence, automation, and precision of conversion problem screening, thereby helping the platform to quickly locate the causes of abnormal promotion effects and provide solutions, thereby improving the optimization effect of information promotion.

[0235] The present specification also provides a computer storage medium, which can store multiple instructions, which are suitable for being loaded and executed by a processor as described above. Figures 1 to 8 The query processing method of the embodiment shown in the figure can be found in the specific execution process. Figures 1 to 8 The specific description of the illustrated embodiment will not be repeated here.

[0236] The present specification also provides a computer program product, which stores at least one instruction, and the at least one instruction is loaded and executed by the processor as described above. Figures 1 to 8 The query processing method of the embodiment shown in the figure can be found in the specific execution process. Figures 1 to 8 The specific description of the illustrated embodiment will not be repeated here.

[0237] Please refer to Fig.10 , is a block diagram of a structure of an electronic device provided in an embodiment of this specification. The electronic device in this specification may include one or more of the following components: a processor 1010, a memory 1020, an input device 1030, an output device 1040, and a bus 1050. The processor 1010, the memory 1020, the input device 1030, and the output device 1040 may be connected via a bus 1050.

[0238] The processor 1010 may include one or more processing cores. The processor 1010 uses various interfaces and lines to connect various parts of the entire electronic device, and executes various functions of the electronic device and processes data by running or executing instructions, programs, code sets or instruction sets stored in the memory 1020, and calling data stored in the memory 1020. Optionally, the processor 1010 can be implemented in at least one hardware form of digital signal processing (DSP), field-programmable gate array (FPGA), and programmable logic array (PLA). The processor 1010 can integrate one or a combination of a central processing unit (CPU), a graphics processing unit (GPU), and a modem. Among them, the CPU mainly processes the operating system, user interface, and application programs; the GPU is responsible for rendering and drawing display content; and the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 1010, but may be implemented separately through a communication chip.

[0239] The memory 1020 may include a random access memory (RAM) or a read-only memory (ROM). Optionally, the memory 1020 includes a non-transitory computer-readable storage medium. The memory 1020 may be used to store instructions, programs, codes, code sets, or instruction sets.

[0240] The input device 1030 is used to receive input instructions or data, and the input device 1030 includes but is not limited to a keyboard, a mouse, a camera, a microphone or a touch device. The output device 1040 is used to output instructions or data, and the output device 1040 includes but is not limited to a display device and a speaker. In the embodiment of this specification, the input device 1030 can be a temperature sensor for obtaining the operating temperature of the electronic device. The output device 1040 can be a speaker for outputting audio signals.

[0241] In addition, those skilled in the art will appreciate that the structure of the electronic device shown in the above drawings does not constitute a limitation on the electronic device, and the electronic device may include more or fewer components than shown, or combine certain components, or arrange the components differently. For example, the electronic device also includes a radio frequency circuit, an input unit, a sensor, an audio circuit, a wireless fidelity (WIFI) module, a power supply, a Bluetooth module and other components, which will not be described in detail here.

[0242] In the embodiments of this specification, the execution subject of each step may be the electronic device described above. Optionally, the execution subject of each step is the operating system of the electronic device. The operating system may be an Android system, an IOS system, or other operating systems, which is not limited in the embodiments of this specification.

[0243] exist Fig.10 In the electronic device, the processor 1010 can be used to call the program stored in the memory 1020 and execute it to implement the query processing method as described in the various method embodiments of this specification.

[0244] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0245] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, stored data, displayed data, etc.) and signals involved in the embodiments of this specification are all authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions. For example, the promotion conversion query information and promotion conversion query replies involved in this specification are all obtained with full authorization.

[0246] The above disclosure is only the preferred embodiment of this specification, which certainly cannot be used to limit the scope of rights of this specification. Therefore, equivalent changes made according to the claims of this specification are still within the scope covered by this specification.

Claims

1. A query processing method, the method comprising: Acquire a data conversion attribution information set for pre-screening of information promotion affairs, and receive promotion conversion query information from a user terminal for the information promotion affairs; Based on the promotion conversion query information and the data conversion attribution information set, a promotion conversion processing large model is used to determine a target conversion query task route and a target conversion query task processing flow corresponding to the target conversion query task route, and based on the data conversion attribution information set, the target conversion query task processing flow is executed to obtain a promotion conversion query reply; The promotion conversion query reply is output to the user terminal.

2. According to the method of claim 1, the step of obtaining a data conversion attribution information set for pre-screening of information promotion affairs comprises: Acquire multiple promotion conversion data sources in information promotion affairs, and regularize the data structure of each of the promotion conversion data sources to obtain a conversion attribution investigation data set; Perform conversion status attribution processing based on the conversion attribution investigation data to obtain conversion attribution annotation information for each of the information conversion events; A pre-screened data conversion attribution information set is generated based on the conversion attribution annotation information and the conversion attribution screening data set.

3. According to the method of claim 2, the step of regularizing the data structure of each of the promotion conversion data sources to obtain a conversion attribution investigation data set comprises: A plurality of key attribution fields for the information conversion event are extracted from each of the promotion conversion data sources, and a preset data set structure is used to regularize the data structure based on the key attribution fields to obtain a conversion attribution screening data set.

4. The method according to claim 2, wherein the conversion status attribution processing is performed based on the conversion attribution screening data to obtain conversion attribution annotation information of each information conversion event, including: Using the attribution problem pre-checking link to perform conversion attribution detection processing on the conversion attribution checking data set to obtain multiple pre-checking link item record results; Based on the pre-check link item record results, conversion attribution annotation information corresponding to each of the information conversion events is generated.

5. The method according to claim 1, wherein the step of determining a target conversion query task route and a target conversion query task processing flow corresponding to the target conversion query task route by using a promotion conversion processing large model based on the promotion conversion query information and the data conversion attribution information set comprises: Input the promotion conversion query information and the data conversion attribution information set into the promotion conversion processing model, The promotion conversion query information is understood by the promotion conversion processing large model to obtain the user conversion query intention, and a target conversion query task route is generated based on the user conversion query intention and the data conversion attribution information set; The target conversion query task processing flow of the target conversion query task routing is determined by promoting the conversion processing large model.

6. The method according to claim 5, wherein generating a target conversion query task route based on the user conversion query intention and the conversion attribution annotation information in the promotion conversion query information comprises: Determine the conversion query task type based on the user conversion query intention, and determine the conversion attribution annotation information and conversion attribution screening data set for the information conversion event in the data conversion attribution information set; Determine, based on the conversion attribution annotation information, multiple data source selection information and data conversion attribution environment information for the conversion attribution screening data set; Determine the task processing flow allocation logic and task routing configuration prompt based on the conversion query task type, the multiple data source selection information and the data conversion attribution environment information; A target conversion query task route is generated based on the task processing flow allocation logic and the task route configuration prompt.

7. According to the method of claim 6 or 1, executing the target conversion query task processing flow based on the data conversion attribution information set to obtain a promotion conversion query response comprises: Parsing the task route configuration prompt in the target conversion query task route, and selecting a target call data source from the data conversion attribution information set based on the task route configuration prompt; Importing the data conversion attribution information set and the target call data source into the target conversion query task processing flow, and performing conversion query task analysis through the target conversion query task processing flow to obtain promotion conversion query results; The return result is packaged based on the promotion conversion query result to obtain a promotion conversion query reply.

8. The method according to claim 7, wherein the step of importing the data conversion attribution information set and the target call data source into the target conversion query task processing flow, and performing conversion query task parsing through the target conversion query task processing flow to obtain a promotion conversion query result, comprises: Importing the data conversion attribution information set and the target call data source into the target conversion query task processing flow, and performing task information parsing through the target conversion query task processing flow to obtain tracking missing task information for the dialog state tracking mechanism; Generate at least one round of information-guiding inquiries based on the tracking missing task information, output the information-guiding inquiries to the target user through the promotion conversion processing model, and receive user supplementary context information for the information-guiding inquiries; Based on the user's supplementary context information, conversion query task processing is performed through the target conversion query task processing flow to obtain a promotion conversion query result.

9. A query processing device, comprising: An acquisition module is used to acquire a data conversion attribution information set pre-screened for information promotion affairs, and receive promotion conversion query information from a user terminal for the information promotion affairs; A processing module, for determining a target conversion query task route and a target conversion query task processing flow corresponding to the target conversion query task route by using a promotion conversion processing large model based on the promotion conversion query information and the data conversion attribution information set, and executing the target conversion query task processing flow based on the data conversion attribution information set to obtain a promotion conversion query reply; A reply module is used to output the promotion conversion query reply to the user terminal.

10. A computer storage medium storing a plurality of instructions, wherein the instructions are suitable for being loaded by a processor and executing the method steps according to any one of claims 1 to 8.

11. A computer program product, the computer program product storing at least one instruction, wherein the at least one instruction is loaded by a processor and executes the method steps according to any one of claims 1 to 8.

12. An electronic device comprising: A processor and a memory; wherein the memory stores a computer program, and the computer program is suitable for being loaded by the processor and executing the method steps as claimed in any one of claims 1 to 8.

Images