Conference data analysis method and device and electronic equipment

Abstract

The invention relates to the technical field of data analysis, in particular to a conference data analysis method and device and electronic equipment. The method comprises the following steps: receiving user question information; performing intention analysis on the user question information through the target large language model to obtain an intention analysis result, and determining to-be-counted conference information based on the intention analysis result; determining at least one conference data statistical model to be called based on the conference information to be counted; and calling at least one conference data statistical model to perform data retrieval to obtain initial conference data, and performing conference data analysis on the initial conference data through the target large language model to generate conference statistical data and a conference analysis result. The conference data retrieval and analysis can be carried out through the big language model, the conference statistical data can be obtained, the conference data analysis can be carried out by means of the target big language model, the result containing the conference abnormal reason and the conference optimization suggestion is generated, and more valuable reference is provided for conference decision making.

Description

Technical Field

[0001] This disclosure relates to the field of data analysis technology, and in particular to a method, apparatus and electronic device for analyzing conference data. Background Technology

[0002] Traditional conference data analysis systems typically rely on predefined statistical reports and fixed query conditions, presenting results through database queries combined with visual charts. These systems require the pre-development of numerous data interfaces and front-end interfaces, with each statistical requirement necessitating independent design, development, and deployment processes, resulting in poor scalability and long development cycles. Furthermore, these systems only provide data visualization capabilities, lacking the ability for in-depth data analysis and decision-making recommendations.

[0003] Therefore, how to achieve intelligent data analysis and decision support in the field of conference data analysis has become an urgent problem to be solved. Summary of the Invention

[0004] To address the aforementioned technical problems, this disclosure provides a method, apparatus, and electronic device for analyzing conference data.

[0005] Firstly, this disclosure provides a method for analyzing conference data, including: Receive user questions; the questions are requests for analysis of meeting data, that is, the questions are related to requests for analysis of meeting data. The user's question information is analyzed using a target large language model to obtain intent analysis results, and meeting information to be statistically analyzed is determined based on the intent analysis results; the meeting information to be statistically analyzed includes at least one of meeting statistical indicators, meeting statistical dimensions, and meeting statistical parameters. Based on the meeting information to be statistically analyzed, at least one meeting data statistical model to be invoked is determined; At least one of the aforementioned meeting data statistical models is invoked to retrieve initial meeting data, and the initial meeting data is analyzed using a target large language model to generate meeting statistics and meeting analysis results; the meeting analysis results include the causes of meeting anomalies and meeting optimization suggestions.

[0006] Secondly, this disclosure provides a device for analyzing conference data, including: A receiving unit is used to receive user query information; the query information is an analysis request for meeting data. The analysis unit is used to perform intent analysis on the user's question information through a target large language model, obtain intent analysis results, and determine the meeting information to be statistically analyzed based on the intent analysis results; the meeting information to be statistically analyzed includes at least one of meeting statistical indicators, meeting statistical dimensions, and meeting statistical parameters; The determining unit is used to determine at least one meeting data statistical model to be invoked based on the meeting information to be statistically analyzed. The calling unit is used to call at least one of the meeting data statistical models to retrieve data, obtain initial meeting data, and perform meeting data analysis on the initial meeting data through a target large language model to generate meeting statistics and meeting analysis results; the meeting analysis results include the reasons for meeting anomalies and meeting optimization suggestions.

[0007] Thirdly, the present invention provides an electronic device, comprising: a memory and a processor, wherein the memory is used to store a computer program; and the processor is used to cause the electronic device to implement the conference data analysis method provided in any of the first aspects when executing the computer program.

[0008] Fourthly, the present invention provides a computer-readable storage medium comprising: storing a computer program on the computer-readable storage medium, the computer program being executed by a controller as a method for analyzing conference data as provided in any of the first aspects.

[0009] Fifthly, the present invention provides a computer program product that, when run on a computer, causes the computer to perform a method for analyzing conference data as provided in any of the first aspects.

[0010] It should be noted that the aforementioned computer instructions may be stored, in whole or in part, on the first computer-readable storage medium. The first computer-readable storage medium may be packaged together with the controller of the conference data analysis device, or it may be packaged separately from the controller of the conference data analysis device; this disclosure does not limit this. The descriptions of the second, third, fourth, and fifth aspects of this disclosure can be referenced to the detailed description of the first aspect; and the beneficial effects of the descriptions of the second, third, fourth, and fifth aspects can be referenced to the analysis of the beneficial effects of the first aspect, and will not be repeated here.

[0011] In this disclosure, the name of the aforementioned meeting data analysis device does not limit the device or functional module itself. In actual implementation, these devices or functional modules may appear under other names. As long as the function of each device or functional module is similar to that of this disclosure, it falls within the scope of this disclosure and its equivalents.

[0012] These or other aspects of this disclosure will become more readily apparent in the following description.

[0013] The technical solution provided in this disclosure has the following advantages compared with the prior art: The meeting data analysis method provided in this application is as follows: First, a meeting data analysis request submitted by a user in natural language is received. Then, the user's intent is parsed using a target large language model, and structured query elements are extracted. Next, a pre-set statistical model is matched to these elements, and its interface is called to retrieve the data. Finally, the large language model is used to perform in-depth analysis and reasoning on the data to obtain the meeting analysis results. Compared with existing technologies, traditional systems require the separate development of interfaces for each analysis requirement, resulting in fixed interaction methods and a lack of intelligent analysis capabilities. This application uses a large language model for meeting data retrieval and analysis, which not only obtains meeting statistics but also uses the target large language model to analyze meeting data, generating results that include reasons for meeting anomalies and meeting optimization suggestions, providing more valuable references for meeting decisions. Attached Figure Description

[0014] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0015] To more clearly illustrate the technical solutions in the embodiments of this disclosure or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 A flowchart illustrating a method for analyzing meeting data provided in this embodiment of the disclosure; Figure 2 A flowchart illustrating yet another method for analyzing conference data provided in this disclosure embodiment; Figure 3 A schematic flowchart of a meeting data analysis device provided in this embodiment of the present disclosure; Figure 4 This is a schematic block diagram of an electronic device provided in an embodiment of the present disclosure. Detailed Implementation

[0017] To better understand the above-mentioned objectives, features, and advantages of this disclosure, the solutions disclosed herein will be further described below. It should be noted that, unless otherwise specified, the embodiments and features described herein can be combined with each other.

[0018] Numerous specific details are set forth in the following description in order to provide a full understanding of this disclosure, but this disclosure may also be implemented in other ways different from those described herein; obviously, the embodiments in the specification are only some, and not all, of the embodiments of this disclosure.

[0019] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0020] Traditional conference data analysis systems typically rely on predefined statistical reports and fixed query conditions, presenting results through database queries combined with visual charts. These systems require the pre-development of numerous data interfaces and front-end interfaces, with each statistical requirement necessitating independent design, development, and deployment processes, resulting in poor scalability and long development cycles. Furthermore, these systems only provide data visualization capabilities, lacking the ability for in-depth data analysis and decision-making recommendations.

[0021] Therefore, in order to solve the above problems, this application provides a method, apparatus and electronic device for analyzing conference data, as detailed in the following embodiments.

[0022] Figure 1 The example illustrates a flowchart of a method for analyzing meeting data. The executing entity in this example can be a computer system or server running the method, such as... Figure 1 As shown, the method includes: S11. Receive user questions.

[0023] The question information refers to an analysis request for meeting data.

[0024] In some embodiments, the user's question information may be text in the form of natural language text that carries the user's data analysis needs. For example, the user may enter "Statistics on the average usage time of all meeting rooms in Building 1 last week" or "Analyze the changes in Zhang San's attendance rate this month".

[0025] Specifically, user requests from clients are received through corresponding text receiving ports. For example, users can directly input their data analysis needs in natural language text through the client's text input interface, such as "Show statistics on the usage of meeting rooms in Building 1 last month." Users can also issue voice commands through the smart terminal's voice input module, which is then converted into corresponding text commands using speech-to-text technology, i.e., Automatic Speech Recognition (ASR) module, such as "Check how many meetings I have to attend this week." The client then streams the text data to the computer system or server running this method via a communication protocol.

[0026] S12. Perform intent analysis on the user's question information using the target large language model, obtain the intent analysis results, and determine the meeting information to be statistically analyzed based on the intent analysis results.

[0027] The meeting information to be statistically analyzed includes at least one of meeting statistical indicators, meeting statistical dimensions, and meeting statistical parameters.

[0028] Specifically, the received user query information is sent to a pre-built target Large Language Model (LLM); then the LLM performs semantic parsing and understanding of the user query information to identify the user's fundamental purpose, i.e., intent.

[0029] Furthermore, based on the identified intent, LLM further extracts specific meeting information to be statistically analyzed from the text; the meeting information to be statistically analyzed can be understood as a set of structured query elements extracted and output by the target large language model after performing intent analysis and understanding of the user's original question, which is used to accurately guide subsequent data retrieval and analysis operations.

[0030] Among them, meeting statistical indicators refer to the specific data items that need to be measured, such as "number of meeting appointments," "average meeting duration," "attendance rate," and "number of no-shows." Meeting statistical dimensions refer to the angle or grouping basis for analyzing indicators, such as by "meeting room," by "meeting organizer," or by "department." Meeting statistical parameters refer to the limiting conditions or variables required to calculate indicators, the most common being time parameters (such as last week or this month), and also including regional parameters (such as Building 1) and personnel parameters (such as Zhang San).

[0031] S13. Based on the meeting information to be statistically analyzed, determine at least one meeting data statistical model to be invoked.

[0032] In this embodiment, each meeting data statistical model can be understood as a pre-configured, single-function, and clearly defined data processing unit, which is bound to a preset combination of (indicators, dimensions). A preset mapping relationship defines the correspondence between indicators, dimensions, parameters, and the meeting data statistical model.

[0033] Furthermore, by querying the mapping relationship, one or more statistical models that can meet the user's current query needs can be accurately found.

[0034] Specifically, the at least one meeting data statistical model to be invoked includes at least one of the following: a first type of model, a second type of model, and a third type of model, wherein the latter includes the following: The first category of models includes models for counting the number of meetings, the number of participants, the duration of meetings, the utilization rate of meeting rooms, and the satisfaction level of meetings.

[0035] The second category of models includes time-dimensional statistical models, spatial-dimensional statistical models, and personnel-dimensional statistical models.

[0036] The third type of model includes a pre-set time period meeting statistics model, a pre-set meeting room type statistics model, and a pre-set meeting type statistics model.

[0037] It should be noted that the first type of model is divided according to statistical indicators. The specific data items of the indicators to be statistically analyzed, such as "number of meeting appointments", "average meeting duration", "attendance rate", "number of no-shows", etc., can be set up. For example, the first type of model can include: meeting number statistics model, attendee statistics model, meeting duration statistics model, meeting room utilization rate statistics model, meeting satisfaction statistics model, etc.

[0038] The second type of model is divided according to statistical dimensions. Based on the data dimensions required for statistics, such as by "meeting room", "meeting initiator", "department", etc., multiple second-type models can be set up. For example, the second-type model can include: time dimension statistical model (used to perform statistical analysis on meeting data according to time (such as hour, day, week, month, quarter, year), spatial dimension statistical model (used to collect meeting-related data based on spatial attributes such as the geographical location, size, and equipment configuration of the meeting room), and personnel dimension statistical model.

[0039] The third type of model is divided according to statistical parameters. Multiple third-type models can be set according to the parameters to be statistically analyzed. For example, the third-type model may include: a preset time period meeting statistical model (used to focus on meeting data within a specified time interval, and to count the number of meetings, average duration, resource utilization rate, etc. within that time period), a preset meeting room type statistical model (to filter data by the physical attributes of the meeting room, and to count the utilization rate, average meeting duration, conflict rate, etc. of the same type of meeting room), and a preset meeting type statistical model (to collect statistical data based on meeting type labels and compare the efficiency indicators of different types of meetings (such as on-time start rate, target achievement rate)).

[0040] For example, if a user queries "number of bookings for all meeting rooms", the matched meeting data statistical models can include meeting frequency statistical models and spatial dimension statistical models.

[0041] Then, the matched meeting data statistical model is invoked to perform data query and retrieval, ensuring that the data retrieval direction fully matches the user's needs and avoiding the extraction of invalid data. At the same time, when the user's needs involve multi-dimensional and multi-indicator analysis, a multi-model splitting mechanism is used to achieve parallel processing, improving data analysis efficiency while ensuring the independence of the statistical logic of each sub-need.

[0042] S14. Call at least one of the meeting data statistical models to retrieve data, obtain initial meeting data, and perform meeting data analysis on the initial meeting data through the target large language model to generate meeting statistics and meeting analysis results.

[0043] The meeting analysis results include the reasons for meeting anomalies and suggestions for meeting optimization.

[0044] In this step, after determining at least one meeting data statistics model to be invoked, the selected meeting data statistics model can be invoked through the meeting data statistics model's application programming interface (API) to retrieve the raw data (i.e., initial meeting data) that meets the conditions from the meeting database, such as the reservation records and actual meeting records of a meeting room during a preset time period.

[0045] After obtaining the initial data, the target large language model can be called again for data integration and analysis, that is, the initial meeting data along with the user's original request can be submitted to the LLM.

[0046] Subsequently, the target large language model performs in-depth analysis on the initial meeting data, thereby generating meeting analysis results. Specifically, the meeting analysis results may be the summarization and calculation (such as summation, average, percentage, etc.) of the initial meeting data to form structured statistical results. Meanwhile, the meeting analysis results will include reasons for meeting anomalies, which are determined by identifying unusual fluctuations in the data (such as a sudden drop in the usage rate of a certain meeting room) and inferring possible causes (such as equipment malfunction or inconvenient location). For example, "Zhang San's unusually high number of meetings may be because he is a project leader and needs to frequently participate in project coordination meetings." The analysis also includes meeting optimization suggestions, which are actionable recommendations based on the reasons for anomalies or data characteristics. For example, "It is recommended to check the reservation rules for meeting rooms in Building 1 to avoid resource idleness" or "It is recommended to share some meeting tasks with Zhang San to improve his work efficiency." Finally, the meeting statistics and analysis results will be fed back to the user.

[0047] The meeting data analysis method provided in this application is as follows: First, a meeting data analysis request submitted by a user in natural language is received. Then, the user's intent is parsed using a target large language model, and structured query elements are extracted. Next, a pre-set statistical model is matched to these elements, and its interface is called to obtain the data. Finally, the large language model is used to perform in-depth analysis and reasoning on the data to obtain the meeting analysis results. Compared with existing technologies, traditional systems require separate development of interfaces for each analysis request, resulting in fixed interaction methods and a lack of intelligent analysis capabilities. This application uses a large language model for meeting data retrieval and analysis, which not only obtains meeting statistics but also uses a target large language model to analyze meeting data, generating results that include reasons for meeting anomalies and meeting optimization suggestions, providing more valuable references for meeting decision-making.

[0048] As an extension and refinement of the above embodiments, Figure 2 The diagram above illustrates another flowchart of a method for analyzing meeting data, such as... Figure 2 As shown, the method includes: S21. Receive user questions.

[0049] The question information refers to an analysis request for meeting data.

[0050] S22. Semantically analyze the user's question information using a target large language model, and extract key information from the user's question information by combining it with a preset prompt word template, so as to generate the intent analysis result.

[0051] Specifically, semantic parsing of the user's question information can be performed using the target large language model. This can be achieved by utilizing the large amount of language knowledge built into the LLM to perform deep semantic understanding of the complete sentence of the user's question. At the same time, in order to improve the inference accuracy of the subsequent model, preset prompt word templates can be used for guidance and constraints to guide the LLM to extract key information that needs to be focused on from complex natural language.

[0052] For example, the prompt could be: "Please identify the statistical indicators, statistical dimensions, and statistical parameters from the following user query. The statistical indicators should be: frequency, duration, rate, etc.; the statistical dimensions should be: by person, by meeting room, by department, etc.; the statistical parameters should include the time range: last week, this month, and other limiting conditions: Building No. 1, Zhang San."

[0053] Then, based on the instructions of the prompt word template, the LLM analyzes and infers the user's query information, ultimately outputting a structured intent analysis result. This result is typically a structured data JSON object containing explicitly defined fields.

[0054] Furthermore, by combining large models and prompt word templates, vague natural language requirements are transformed into structured intent analysis results, ensuring the accuracy of understanding.

[0055] S23. Based on the intent analysis results, determine the meeting information to be statistically analyzed.

[0056] The preset mapping relationship includes the correspondence between the meeting statistical indicators, the statistical dimensions, and the meeting statistical parameters and the meeting data statistical model.

[0057] In this step, the natural language description in the intent analysis results is converted into three types of structured information that the system can recognize: "meeting statistical indicators, statistical dimensions, and statistical parameters," which serve as the basis for subsequent model invocation.

[0058] The meeting statistics indicators correspond to the "statistical objectives," which are the data types that need to be calculated (such as duration, number of times, number of participants, etc.); the meeting statistics dimensions correspond to the angles used for classification analysis in the "limited dimensions" (such as department, meeting room type, time granularity, etc.); the meeting statistics parameters correspond to specific limiting values ​​(such as time range, meeting room specifications, department name, etc.).

[0059] If there is ambiguous or missing information in the intent analysis results (such as the user not specifying the time range), the system can automatically supplement the default value (such as "the last 30 days") or complete it through preset rules to ensure that the meeting information to be analyzed is complete.

[0060] Ultimately, the intent analysis results are mapped into statistical elements that the system can execute, laying the foundation for subsequent matching of meeting data statistical models and accurate data retrieval.

[0061] S24. Based on at least one of the meeting statistical indicators, statistical dimensions, and meeting statistical parameters in the meeting information to be statistically analyzed, and combined with a preset mapping relationship, match the corresponding meeting data statistical model to determine at least one meeting data statistical model to be invoked.

[0062] Specifically, a preset mapping relationship is generated in advance based on the meeting statistical indicators, statistical dimensions, and meeting statistical parameters in the meeting information to be statistically analyzed, as well as the functional description of the meeting data statistical model required. This preset mapping relationship can be stored. Each mapping relationship clearly defines the correspondence between the meeting data of the statistical model and the query modes it can handle. Then, a matching algorithm (such as exact matching, most specific matching, or priority matching) is used to find one or more meeting data statistical models whose processing capabilities (indicators, dimensions, parameters) can fully cover or best suit the needs of this query.

[0063] For example, the user requirement is to "count the number of meetings in Building 1 last month", and the meeting information to be counted is [metric: number of meetings, dimension: by location, parameter: Building 1, last month].

[0064] Furthermore, by searching the mapping table for the aforementioned meeting information to be analyzed, the system finds models A and B, whose registered capabilities exactly cover the aforementioned meeting information. Therefore, the system matches and determines models A and B as the targets for this call.

[0065] Specifically, the steps for obtaining the preset mapping relationship in this process may include the following steps 1 and 2: Step 1: Pre-register the application programming interfaces (APIs) of the multiple conference data statistical models as callable toolsets.

[0066] In some embodiments, the tool invocation function provided by the target large language model (LLM) (e.g., the tools parameter of the Open AI Assistants API) is used to declare and register the application programming interfaces (APIs) of the existing conference data statistical models with the LLM. That is, for each statistical model, its API invocation specifications (including name, description, parameter list and their format requirements) are encapsulated according to the format required by the LLM to form the tool definition.

[0067] Furthermore, all successfully registered tools are combined into a callable toolset, which LLM can query and select in the subsequent intent recognition stage.

[0068] Step 2: Based on the functional description and parameter specifications corresponding to each application interface, establish the preset mapping relationship between the multiple meeting data statistical models and the meeting statistical indicators, the meeting statistical dimensions, and the meeting statistical parameters.

[0069] Based on the API function description and parameter specifications registered in step 1, extract the corresponding "statistical indicators, statistical dimensions, and statistical parameters". For example, from the API function description "total meeting duration for a specified department within a preset time range", extract the statistical indicator as "total meeting duration"; from the parameters, extract the statistical dimensions as "department dimension" and "time dimension"; and from the parameter value range, extract the statistical parameter as "meeting room type".

[0070] The extracted elements (indicators, dimensions, parameters) are associated with the corresponding meeting data statistical model (or its unique API identifier) ​​to form a mapping relationship, and stored in the mapping table for later use.

[0071] This application embodiment registers the API of the conference data statistical model as a tool that can be called by the large model, and establishes its association with conference statistical indicators, dimensions, and parameters. This enables seamless collaboration between the large model and the statistical model, ensuring accurate matching between user needs and models. It reduces the integration cost of adding new models, shortens the response time to needs, and ensures the consistency and reliability of data analysis results.

[0072] S25. Obtain at least one application programming interface of the meeting data statistical model to perform data retrieval through at least one of the meeting data statistical models to obtain initial meeting data.

[0073] In some embodiments, based on the determined statistical model of the meeting data to be called, the application programming interface (API) information of the corresponding model is extracted; by calling these APIs, the raw data is retrieved from the meeting database according to the parameter requirements in the meeting information to be statistically analyzed (such as time range, department, meeting room type, etc.), and finally the initial meeting data that meets the statistical requirements is obtained.

[0074] For example, in response to the requirement of "meeting data of the Marketing Department every Wednesday afternoon in medium-sized meeting room in August 2025", the system calls the API of the matched "department-time-meeting room type multi-dimensional statistical model" to extract the meeting records under this condition (such as the start time, end time, number of participants, etc. of each meeting) from the database to form an initial meeting data set, which provides a foundation for subsequent data analysis and reasoning.

[0075] S26. Perform meeting data analysis on the initial meeting data using the target large language model to generate meeting statistics and meeting analysis results.

[0076] Specifically, the detailed steps of step S26 (analyzing the initial meeting data using the target large language model to generate meeting statistics and analysis results) may include the following: Step 1: Perform statistical calculations on the initial meeting data using the target large language model to generate the meeting statistics data.

[0077] This step inputs the initial meeting data returned by the data retrieval module (usually raw, fine-grained records from a database) into the target Large Language Model (LLM) and instructs it to act as a statistician. The LLM aggregates, calculates, and organizes this raw data according to common or pre-defined statistical methods, transforming it into higher-level, more readable summary information—i.e., meeting statistics.

[0078] Specifically, statistical calculations include, but are not limited to, calculating sums, averages, maximums, minimums, ratios, frequency distributions, and month-on-month or year-on-year changes. For example, a series of individual meeting records can be used to calculate statistical results such as "Zhang San attended a total of 5 meetings this week" and "The average daily usage rate of the meeting room in Building 1 is 75%".

[0079] This results in structured data that has undergone preliminary processing, typically presented in text, lists, or tables (such as Markdown tables). This data represents the direct answer to the user's requested question.

[0080] Step 2: Perform anomaly detection on the meeting statistics and obtain the anomaly detection results.

[0081] Specifically, LLM analyzes and judges based on context. For example, it can identify situations such as "the utilization rate of a meeting room suddenly drops from an average of 80% to 20%", "an individual's weekly meeting attendance is more than twice the standard deviation of the team average", and "the average meeting duration fluctuates significantly over a certain period of time".

[0082] Furthermore, what is obtained are one or more marked anomalies and their descriptions, indicating "where the problem may exist".

[0083] Step 3: Based on the anomaly detection results, deduce the cause of the anomaly to generate the cause of the meeting anomaly.

[0084] For example, in response to a "sudden drop in meeting room occupancy," an LLM might deduce potential reasons such as "the meeting room equipment is malfunctioning," "the company has a large external event during this period," or "the reservation rules have been modified," providing direction for subsequent decision-making.

[0085] Step 4: Based on the reasons for the meeting anomalies, generate meeting optimization suggestions to obtain meeting statistics and meeting analysis results.

[0086] For example, if the cause of the anomaly is "equipment failure", it is recommended to "notify the maintenance department to carry out repairs"; if the cause is "unreasonable reservation rules", it is recommended to "optimize the meeting room allocation strategy".

[0087] The initial meeting data, meeting statistics, reasons for meeting anomalies, and suggestions for meeting optimization obtained from the above steps are compiled and summarized to form a complete meeting analysis result, which is then returned to the user.

[0088] It should be noted that the meeting data analysis method provided in this application embodiment further includes the following steps: Receive a tool configuration update request, and in response to the tool configuration update request, add the new application interface to the callable toolset, and synchronously update the preset mapping relationship.

[0089] In some embodiments, various meeting needs will continue to emerge as development progresses, and data dimensions will continue to deepen. When new meeting analysis needs arise, the corresponding meeting data statistical model can be updated at any time. Then, the application programming interface of the new model can be added to the callable toolset, and the preset mapping relationships of indicators, dimensions, and parameters can be updated synchronously. This enables the system to quickly identify and call the new model, providing accurate meeting data analysis support for new meeting needs in a timely manner, and ensuring that meeting data analysis capabilities iterate in sync with development.

[0090] Furthermore, through the aforementioned configuration update mechanism, this application enables the dynamic expansion of the meeting data statistical model. New data analysis capabilities can be incorporated without modifying the core system logic, ensuring that the system can quickly respond to new meeting analysis needs and improving the flexibility and scalability of the solution.

[0091] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method.

[0092] The embodiments of this application also provide a device for analyzing conference data. For any omitted virtual device claims, the description should be expanded in the specification, corresponding one-to-one with the method claims.

[0093] The structural schematic diagram of the conference data analysis device provided in this application is as follows: Figure 3 The meeting data analysis device 300 shown includes: a receiving unit 301, an analysis unit 302, a determination unit 303, and a calling unit 304.

[0094] The receiving unit 301 is used to receive user query information; the query information is an analysis request for meeting data. Analysis unit 302 is used to perform intent analysis on the user's question information through a target large language model, obtain intent analysis results, and determine meeting information to be statistically analyzed based on the intent analysis results; the meeting information to be statistically analyzed includes at least one of meeting statistical indicators, meeting statistical dimensions, and meeting statistical parameters; The determining unit 303 is used to determine at least one meeting data statistical model to be invoked based on the meeting information to be statistically analyzed. Calling unit 304 is used to call at least one of the meeting data statistical models to retrieve data, obtain initial meeting data, and perform meeting data analysis on the initial meeting data through a target large language model to generate meeting statistics and meeting analysis results; the meeting analysis results include the reasons for meeting anomalies and meeting optimization suggestions.

[0095] In some feasible examples, the analysis unit 302 is specifically used to perform semantic parsing on the user question information through a target large language model, and extract key information from the user question information in combination with a preset prompt word template to generate the intent analysis result; based on the intent analysis result, determine the meeting information to be statistically analyzed.

[0096] In some feasible examples, the determining unit 303 is specifically used to determine at least one meeting data statistical model to be invoked by matching the corresponding meeting data statistical model based on at least one of the meeting statistical indicators, the statistical dimensions, and the meeting statistical parameters in the meeting information to be statistically analyzed, in conjunction with a preset mapping relationship; the preset mapping relationship includes the correspondence between the meeting statistical indicators, the statistical dimensions, and the meeting statistical parameters and the meeting data statistical model, respectively.

[0097] In some feasible examples, the at least one meeting data statistical model to be invoked includes at least one of a first type of model, a second type of model, and a third type of model. The first type of model includes a meeting frequency statistical model, a participant count statistical model, a meeting duration statistical model, a meeting room utilization rate statistical model, and a meeting satisfaction statistical model. The second type of model includes a time-dimensional statistical model, a spatial-dimensional statistical model, and a personnel-dimensional statistical model. The third type of model includes a preset time period meeting statistical model, a preset meeting room type statistical model, and a preset meeting type statistical model.

[0098] In some implementable examples, the calling unit 304 is specifically used to obtain an application programming interface of at least one of the meeting data statistical models to perform data retrieval through at least one of the meeting data statistical models to obtain initial meeting data.

[0099] In some feasible examples, the calling unit 304 is specifically used to perform statistical calculations on the initial meeting data through the target large language model to generate the meeting statistics data; perform anomaly detection on the meeting statistics data to obtain anomaly detection results; combine the anomaly detection results to infer the cause of the anomaly to generate the meeting anomaly cause; and generate the meeting optimization suggestions based on the meeting anomaly cause to obtain the meeting statistics data and meeting analysis results.

[0100] In some feasible examples, the meeting data analysis device further includes a creation unit, specifically used to pre-register the application programming interfaces (APIs) of multiple meeting data statistical models as a callable toolset; and to establish the preset mapping relationship between the multiple meeting data statistical models and the meeting statistical indicators, the meeting statistical dimensions, and the meeting statistical parameters, based on the functional description and parameter specifications corresponding to each API.

[0101] In some feasible examples, the meeting data analysis device further includes an update unit, specifically configured to receive a tool configuration update request, and in response to the tool configuration update request, add the new application programming interface to the callable toolset, and synchronously update the preset mapping relationship.

[0102] All relevant content of each step involved in the above method embodiments can be referenced from the functional description of the corresponding functional module, and their functions will not be repeated here.

[0103] Of course, the meeting data analysis device provided in this embodiment of the invention includes, but is not limited to, the modules described above. For example, the meeting data analysis device may also include a storage unit 305. The storage unit 305 may be used to store the program code of the meeting data analysis device, and may also be used to store data generated by the meeting data analysis device during operation, such as diagnostic data.

[0104] A schematic diagram of the structure of an electronic device provided in an embodiment of the present invention is shown below. Figure 4 The electronic device shown may include at least one processor 41, a memory 42, a communication interface 43, and a communication bus 44.

[0105] The following is a detailed introduction to the various components of the electronic device: The processor 41 is the control center of the electronic device. It can be a single processor or a collective term for multiple processing elements. For example, the processor 41 can be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention, such as one or more DSPs, or one or more field-programmable gate arrays (FPGAs).

[0106] In a specific implementation, as one embodiment, processor 41 may include one or more CPUs, such as CPU0 and CPU1. Furthermore, as one embodiment, the electronic device may include multiple processors, such as processor 41 and processor 45. Each of these processors may be a single-core processor or a multi-core processor. Here, "processor" may refer to one or more devices, circuits, and / or processing cores used for processing data (e.g., computer program instructions).

[0107] The memory 42 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory 42 may exist independently and be connected to the processor 41 via the communication bus 44. The memory 42 may also be integrated with the processor 41.

[0108] In a specific implementation, memory 42 is used to store data from this invention and the software program for executing this invention. Processor 41 can perform various functions of the air conditioner by running or executing the software program stored in memory 42 and by calling the data stored in memory 42.

[0109] Communication interface 43 uses any transceiver-like device for communicating with other devices or communication networks, such as Radio Access Network (RAN), Wireless Local Area Networks (WLAN), terminals, and the cloud. Communication interface 43 may include an acquisition unit to implement acquisition functions.

[0110] The communication bus 44 can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. This bus can be divided into address bus, data bus, control bus, etc. For ease of representation, it is represented by only one thick line, but this does not indicate that there is only one bus or one type of bus.

[0111] As an example, the receiving unit 301 of the conference data analysis device performs the same function as the communication interface 43, the analysis unit 302 of the conference data analysis device performs the same function as the processor 41, and the storage unit 303 of the conference data analysis device performs the same function as the memory 42.

[0112] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the method in any of the embodiments.

[0113] The above description is merely a specific embodiment of this disclosure, enabling those skilled in the art to understand or implement it. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this disclosure. Therefore, this disclosure is not to be limited to the embodiments described herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method of analyzing conference data, characterized by, Comprise: Receiving user question information; The question information is an analysis request for meeting data; Analyze the intent of the user question information through a target large language model, obtain an intent analysis result, and determine the meeting information to be counted based on the intent analysis result; The meeting information to be counted includes at least one of meeting statistical indicators, meeting statistical dimensions, and meeting statistical parameters; Determine at least one meeting data statistical model to be called based on the meeting information to be counted; Call at least one of the meeting data statistical models to retrieve data to obtain initial meeting data, and analyze the initial meeting data through a target large language model to generate a meeting analysis result containing meeting statistical data; The meeting analysis result includes meeting abnormal reasons and meeting optimization suggestions.

2. The method of claim 1, wherein, The intent analysis result is obtained by analyzing the intent of the user question information through a target large language model, and the meeting information to be counted is determined based on the intent analysis result, comprising: Perform semantic analysis on the user question information through a target large language model, and extract key information from the user question information based on a preset prompt word template to generate the intent analysis result; Determine the meeting information to be counted based on the intent analysis result.

3. The method of claim 1, wherein, The at least one meeting data statistical model to be called is determined based on the meeting information to be counted, comprising: According to at least one of the meeting statistical indicators, the statistical dimensions and the meeting statistical parameters in the meeting information to be counted, combine the preset mapping relationship to match the corresponding meeting data statistical model, to determine at least one of the meeting data statistical models to be called; The preset mapping relationship includes: the corresponding relationship between the meeting statistical indicators, the statistical dimensions and the meeting statistical parameters and the meeting data statistical models.

4. The method of claim 1, wherein, The at least one meeting data statistical model to be called includes at least one of a first type model, a second type model and a third type model, wherein The first type model includes a meeting frequency statistical model, a meeting participant number statistical model, a meeting duration statistical model, a meeting room utilization rate statistical model, and a meeting satisfaction statistical model; The second type model includes a time dimension statistical model, a space dimension statistical model, and a personnel dimension statistical model; The third type model includes a preset time period meeting statistical model, a preset conference room type statistical model, and a preset conference type statistical model.

5. The method of claim 1, wherein, The at least one meeting data statistical model is called to retrieve data to obtain initial meeting data, comprising: Obtain the application programming interface of at least one of the meeting data statistical models to retrieve data through at least one of the meeting data statistical models to obtain initial meeting data.

6. The method of claim 1, wherein, The initial meeting data is analyzed through a target large language model to generate a meeting analysis result containing meeting statistical data, comprising: Perform statistical calculation on the initial meeting data through the target large language model to generate the meeting statistical data; Perform anomaly detection on the meeting statistical data to obtain an anomaly detection result; Inference of abnormal cause is combined with the abnormal detection result to generate a conference abnormal cause; According to the conference abnormal cause, a conference optimization suggestion is generated to obtain the conference statistical data and conference analysis result.

7. The method of claim 3, wherein, The method further comprises: Pre-registering application program interfaces of a plurality of conference data statistical models as a callable tool set; In combination with the function description and parameter specification corresponding to each application program interface, the preset mapping relationship between the plurality of conference data statistical models and the conference statistical indicators, conference statistical dimensions, and conference statistical parameters is established.

8. The method of claim 7, wherein, The method further comprises: Receiving a tool configuration update request, and in response to the tool configuration update request, adding a newly added application program interface to the callable tool set and synchronously updating the preset mapping relationship.

9. An analysis apparatus of conference data, characterized by, Comprise: A receiving unit is configured to receive user query information; The query information is an analysis request for conference data; An analysis unit is configured to perform intent analysis on the user query information through a target large language model, obtain an intent analysis result, and determine to-be-statistical conference information based on the intent analysis result; the to-be-statistical conference information includes at least one of conference statistical indicators, conference statistical dimensions, and conference statistical parameters; A determination unit is configured to determine at least one conference data statistical model to be called based on the to-be-statistical conference information; A calling unit is configured to call at least one conference data statistical model to perform data retrieval to obtain initial conference data, and perform conference data analysis on the initial conference data through a target large language model to generate conference statistical data and conference analysis result; the conference analysis result includes conference abnormal cause and conference optimization suggestion.

10. An electronic device, comprising: The electronic device comprises: A processor; A memory configured to store instructions executable by the processor; The processor is configured to execute the instructions to implement the conference data analysis method of any one of claims 1-8.

Images