Conference room intelligent reservation method and system, and electronic equipment

Abstract

The invention discloses a conference room intelligent reservation method and system, and electronic equipment. The method comprises the following steps: initializing the system, and loading basic data and configuration information; receiving and analyzing demand information input by a user, and identifying a user intention and a demand type to obtain a structured analysis result; generating a candidate conference room list based on the structured analysis result in combination with the conference room information and the time information; performing multi-dimensional scoring on the candidate conference room list to obtain a to-be-selected conference room list which is sequenced according to a total score; performing time conflict screening and optimization on the to-be-selected conference room list to obtain an optimized conference room list; and returning the optimized conference room list to the user side. Text analysis and recommendation weight calculation are carried out in a traditional mode, an intelligent conference room reservation result is given in combination with user intentions, user historical behaviors and conference room occupation conditions, deployment can be carried out in a pure CPU environment, the interface response speed can reach 200 ms, and the reliability reaches 99%.

Description

Technical Field

[0001] This invention relates to the field of intelligent office technology, and more specifically, to a method, system, and electronic device for intelligent meeting room reservation. Background Technology

[0002] Traditional meeting room reservation management systems suffer from cumbersome operations. To address this pain point, the mainstream solution in the industry for adding intelligent reservation functionality relies on the natural language understanding capabilities of large language models to interpret user intent and provide personalized recommendations based on meeting room occupancy. However, this solution has significant drawbacks: firstly, deploying large language models requires substantial hardware and technical costs, a considerable burden for many small and medium-sized enterprises; secondly, its response speed is relatively slow, failing to meet users' immediate needs for reservation efficiency; and more importantly, limited by model training data and contextual understanding capabilities, the reliability of intent recognition cannot reach 100%, potentially leading to misinterpretations of reservation commands and impacting the meeting room user experience.

[0003] Chinese patent document (application number: 202310135202.6, application date: February 17, 2023) discloses a method, apparatus, storage medium, and terminal for scheduling meeting rooms, including: acquiring meeting information of a meeting to be scheduled and meeting room information of available meeting rooms; determining the matching degree between the meeting to be scheduled and the available meeting rooms based on the meeting information and the meeting room information; and determining the target meeting room corresponding to the meeting to be scheduled from the available meeting rooms using a graph algorithm based on the matching degree. This solution only determines the target meeting room corresponding to the meeting to be scheduled from the available meeting rooms using a graph algorithm based on the matching degree between the meeting to be scheduled and the available meeting rooms, and does not solve the problems existing in the large language model solution in the meeting room reservation system: 1. Insufficient response speed; 2. High deployment cost; 3. Insufficient reliability.

[0004] Therefore, the high deployment cost, slow response time, and low reliability of large language model solutions are technical problems that urgently need to be solved in this field. Summary of the Invention

[0005] In view of this, the present invention provides a method for intelligent transformation of meeting reservation that does not rely on large language models, in order to solve the technical problems of high deployment cost, slow response and low reliability of large language model solutions.

[0006] A first aspect of this application provides a method for intelligent meeting room reservation, comprising: Initialize the intelligent meeting room reservation system by loading basic data and configuration information; It receives and parses the user's natural language input request information, and based on multi-level semantic information extraction technology, identifies the user's intent and request type to obtain structured parsing results; Based on the structured parsing results, combined with meeting room information and time information, a candidate meeting room list is generated, which includes a candidate set of meeting rooms; The candidate meeting room list is scored from multiple dimensions to obtain a list of candidate meeting rooms sorted by total score; The list of candidate meeting rooms is filtered and optimized for time conflicts to obtain an optimized list of meeting rooms. The optimized list of meeting rooms is returned to the user.

[0007] Optionally, the initialization of the intelligent meeting room reservation system, including loading basic data and configuration information, includes: Load basic meeting room information from the database, initialize the user preference weight dictionary, start a scheduled task, update the data periodically, and establish a connection with the third-party meeting room system; The method for establishing the user preference weight dictionary includes: Collect reservation records within the user reservation frequency threshold and analyze the meeting room usage frequency; calculate user preference weights according to the weight formula; and normalize the weights. The user preference weights are calculated as follows: ; Where D represents the number of days remaining until the current date, corresponding to the user's reservation threshold.

[0008] Optionally, the process of receiving and parsing the user's natural language input demand information, based on multi-level semantic information extraction technology, to identify the user's intent and demand type, and obtain structured parsing results includes: It receives text input from users, performs text preprocessing and standardization on the text, extracts key information through regular expressions and semantic analysis, identifies user intent and demand type, and outputs structured parsing results. The key information includes office area, floor, room number, appointment time, equipment, and number of personnel.

[0009] Optionally, based on the structured parsing results, combined with meeting room information and time information, a candidate meeting room list is generated; Based on at least one of the office area, floor, and room number in the structured parsing results, candidate meeting room information is filtered. Time information is obtained based on the appointment time. The meeting room information and the time information are combined by performing a Cartesian product to generate the candidate meeting room list. The candidate meeting room list includes a candidate meeting room set, which includes the meeting room information and the time information.

[0010] Optionally, the step of performing multi-dimensional scoring on the candidate meeting room list to obtain a list of candidate meeting rooms sorted by total score includes: Calculate the matching score for each dimension, apply user preference weights, process the bonus for exact matches, sort the results from highest to lowest total score, and generate the list of candidate meeting rooms. The scoring dimensions include: historical preference weight, time matching degree, floor matching degree, personnel capacity matching degree, equipment demand matching degree, and precise matching degree.

[0011] Optionally, the step of filtering and optimizing the candidate meeting room list for time conflicts to obtain an optimized meeting room list includes: Based on the meeting room information in the list of candidate meeting rooms, query the occupancy status of the meeting rooms, eliminate the candidate set of meeting rooms with time conflicts, sort them in descending order of score, and prioritize recommending the candidate set of high-scoring meeting rooms.

[0012] Optionally, the step of filtering and optimizing the candidate meeting room list for time conflicts to obtain an optimized meeting room list further includes: When the specified set of meeting room candidates is unavailable, an alternative set of meeting room candidates is generated from two aspects: time adjustment and / or duration optimization. Usage tips and instructions are added to obtain an optimized list of meeting rooms.

[0013] Optionally, returning the optimized meeting room list to the user includes: The optimized meeting room list is divided into a recommended list and a non-recommended list. Scheme descriptions and prompts are added, the results are formatted, user selections are recorded, and preference data is updated.

[0014] A second aspect of this application provides an intelligent meeting room reservation system, comprising: a natural language understanding module, a solution generation module, an intelligent scoring algorithm module, a solution screening and optimization module, a user preference learning module, and a real-time data synchronization module; The natural language understanding module is coupled to the solution generation module, receives and parses the user's natural language input, identifies the user's intent based on multi-level semantic information extraction technology, distinguishes between precise and fuzzy requirements, obtains structured parsing results, and outputs the structured parsing results to the solution generation module. The scheme generation module is coupled to the natural language understanding module and the real-time data synchronization module respectively. Based on the structured parsing results output by the natural language understanding module and the meeting room information synchronized by the real-time data synchronization module, it generates a candidate meeting room list and outputs it to the intelligent scoring algorithm module. The intelligent scoring algorithm module is coupled to the scheme generation module, the user preference learning module, and the real-time data synchronization module, respectively. It receives the candidate meeting room list output by the scheme generation module, combines the user preference data provided by the user preference learning module and the meeting room information synchronized by the real-time data synchronization module, and obtains a list of candidate meeting rooms sorted by total score based on a multi-dimensional weighted scoring mechanism, which is then transmitted to the scheme screening and optimization module. The scheme screening and optimization module is coupled to the intelligent scoring algorithm module, the user preference learning module, and the real-time data synchronization module, respectively. It receives the candidate meeting room list output by the intelligent scoring algorithm module, combines the user preference data provided by the user preference learning module and the meeting room information synchronized by the real-time data synchronization module, eliminates the candidate set of meeting rooms with time conflicts, optimizes the meeting room list, and outputs the optimized meeting room list. The user preference learning module is coupled to the scheme selection and optimization module and the intelligent scoring algorithm module respectively. The user preference learning module constructs a user preference model based on historical reservation data, which is used to provide user preference data to the scheme selection and optimization module and the intelligent scoring algorithm module. The real-time data synchronization module is coupled to the scheme generation module, the scheme filtering and optimization module, and the intelligent scoring algorithm module, respectively, and is used to provide the scheme generation module, the scheme filtering and optimization module, and the intelligent scoring algorithm module with real-time information of the conference room.

[0015] A third aspect of this application provides an electronic device, comprising: a memory for storing a computer program; A processor is used to execute the computer program to implement the steps of the intelligent meeting room reservation method described above.

[0016] Compared with existing technologies, the intelligent meeting room reservation method, system, and electronic device provided by this invention achieve at least the following beneficial effects: First, this application uses traditional methods for text parsing and recommendation weight calculation, and combines user intent, user history behavior, and meeting room occupancy to provide intelligent meeting room reservation results; Secondly, the system of this application can be deployed in a pure CPU environment, which is suitable for the existing infrastructure of various small and medium-sized enterprises, reducing hardware investment and lowering costs; Third, the system interface response speed of this application can reach 200ms; Fourth, the reliability of user intent recognition in this method can reach 99%, effectively avoiding problems such as misjudgment of reservation instructions.

[0017] Of course, any product implementing this invention does not necessarily need to achieve all of the technical effects described above at the same time.

[0018] Other features and advantages of the invention will become clear from the following detailed description of exemplary embodiments of the invention with reference to the accompanying drawings. Attached Figure Description

[0019] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments of the invention and, together with their description, serve to explain the principles of the invention.

[0020] Figure 1 This is a flowchart of an intelligent meeting room reservation method according to this application; Figure 2 This is a structural block diagram of an intelligent meeting room reservation system according to this application; Figure 3 This is a flowchart of the natural language understanding module in an intelligent meeting room reservation system according to this application; Figure 4 This is a flowchart of the user preference learning module in an intelligent meeting room reservation system according to this application; Figure 5 This is a flowchart of the candidate scheme generation and intelligent scoring and ranking process in an intelligent meeting room reservation method of this application; Figure 6 This is a flowchart of the scheme selection and optimization process in an intelligent meeting room reservation method proposed in this application. Detailed Implementation

[0021] Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments do not limit the scope of the invention.

[0022] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the invention or its application or use.

[0023] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.

[0024] In all the examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0025] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be discussed further in subsequent figures.

[0026] Reference Figures 1-6 , Figure 1 This is a flowchart of an intelligent meeting room reservation method according to this application; Figure 2 This is a structural block diagram of an intelligent meeting room reservation system according to this application; Figure 3 This is a flowchart of the natural language understanding module in an intelligent meeting room reservation system according to this application; Figure 4 This is a flowchart of the user preference learning module in an intelligent meeting room reservation system according to this application; Figure 5 This is a flowchart of the candidate scheme generation and intelligent scoring and ranking process in an intelligent meeting room reservation method of this application; Figure 6 This is a flowchart of the scheme selection and optimization process in an intelligent meeting room reservation method proposed in this application.

[0027] like Figure 1 As shown, this application provides a method for intelligent meeting room reservation, including the following steps (S1-S6): S1. Initialize the intelligent meeting room reservation system, loading basic data and configuration information; S2. Receive and parse the user's natural language input, identify the user's intent and demand type based on multi-level semantic information extraction technology, and obtain the parsing results; S3. Based on the structured parsing results, combined with meeting room information and time information, generate a candidate meeting room list, which includes a candidate set of meeting rooms; S4. Score the candidate meeting room list from multiple dimensions and obtain a list of candidate meeting rooms sorted by total score; S5. Filter and optimize the list of meeting rooms to be selected based on time conflicts to obtain an optimized list of meeting rooms; S6. Return the optimized meeting room list to the user.

[0028] In some optional embodiments provided by the present invention, step S1, which initializes the intelligent meeting room reservation system and loads basic data and configuration information, includes: Load basic meeting room information from the database, initialize the user preference weight dictionary, start a scheduled task, update the data periodically, and establish a connection with the third-party meeting room system; The methods for establishing the user preference weight dictionary include: Collect reservation records within the user's reservation frequency threshold and analyze the meeting room usage frequency; calculate user preference weights according to the weighting formula; normalize the weights; update user preference data through scheduled tasks; The calculation method for user preference weights is as follows: ; Where D represents the number of days remaining until the current date, corresponding to the user's reservation threshold.

[0029] Step S1 is the system initialization step. When the system starts, it loads basic data and configuration information. Specifically, this includes loading basic information about the conference room from the database, initializing the user preference weight dictionary, starting a scheduled task, updating data periodically, and establishing a connection with the third-party conference room system.

[0030] Specifically, the basic information of the meeting room includes the meeting room office area, floor, room, equipment information, capacity, photos, and floor plan.

[0031] Specifically, the methods for constructing the user preference weight dictionary include: Collect the user's past 60 booking records (the 60 most recent booking records are the booking threshold), analyze the frequency of meeting room usage, and calculate preference weights using a time decay function: weight_score = 20 - 10 * log10(past_days), where weight_score is the weight score and past_days is the number of days D since the current date. Normalize the weights to ensure they are within a reasonable range, and update user preference data daily via a scheduled task. Figure 4 As shown, the specific process includes: S101. User historical reservation data: User historical reservation data is updated regularly; S102. Collect and preprocess data, including historical data acquisition, data cleaning, data structuring, and time format standardization. S103, Time decay calculation, including time difference calculation, decay function application, and weight range control; The application of the decay function is to use the weight calculation formula. For example, if the number of days corresponding to the 60 past reservation records of user A for meeting room A is 60 days, then the weight score of meeting room A = 20 - 10 × log10(60) = 2.2. S104. Meeting room preference statistics, including the following steps: Grouping by meeting room: User preference data is grouped according to the meeting room dimension, and all preference weights in the same meeting room are grouped together; Weighted summation: The preference weights within each group are summed to obtain the total preference weight for each meeting room, i.e., the total weight of the meeting room. The total weight of the meeting room = Σ (the weight of each booking for that meeting room). Sorting and filtering: Sort all meeting rooms according to the overall preference weight, in descending order from high to low; Select the top 5 preferred meeting rooms: If the number of meeting rooms is no more than 5, the sorting result is output directly; if the number of meeting rooms is more than 5, the top 5 are selected in descending order. If there is a tie for 5th place, both tied meeting rooms are retained.

[0032] S105. Weight normalization transforms the different original weights into standardized weights that sum to 1, ensuring that all weights are comparable and additivity on the same scale. This includes the following steps: Determine the benchmark weights; Normalization calculation: Normalization is performed according to the formula for normalized weights to obtain normalized weights; Normalized weight = Total weight of meeting rooms ÷ Total weight of the meeting room with the highest weight; Weight dictionary construction: The normalized weights are associated with the corresponding meeting rooms to form a structured weight dictionary; S106. Data storage and updating, specifically including: database storage, caching mechanism, and scheduled updates. The database storage is used to store user appointment records, basic meeting room information, etc. New user appointment records are temporarily stored, written to the cache periodically, and the database is updated. When there are too many user records corresponding to a user in the database, such as more than 60 records, the latest data is written and the oldest record is deleted.

[0033] In some optional embodiments provided by the present invention, step S2, receiving and parsing the user's natural language input demand information, identifying the user's intent and demand type based on multi-level semantic information extraction technology, and obtaining structured parsing results includes: It receives text input from users, performs text preprocessing and standardization on the text, extracts key information through regular expressions and semantic analysis, identifies user intent and demand type, and outputs structured parsing results. Key information includes office area, floor, room number, appointment time, equipment, and number of personnel.

[0034] Step S2 is a user input processing step that receives and analyzes the user's natural language input. Specifically, it includes receiving the text information input by the user, performing text preprocessing and standardization on the text information, extracting key information, and identifying the type of demand of the user's intention based on the key information. The key information includes office area, floor, room number, time, equipment, number of people, etc. The analysis of the text information is carried out by the natural language understanding module 100, which achieves accurate understanding of the user's natural language input based on the multi-level semantic information extraction technology. The specific process is as Figure 3 shown, including: S201. User natural language input. The text information input by the user is one or more demands described in natural language. Specifically, it can include time description (3 pm this afternoon, 10:30 am tomorrow, 2 hours later), location description (16th floor of Yuetan, Tongtai Grand Conference Room, 15th floor of tt, Shunyi 410), duration demand description (2 hours, from 1 pm to 3 pm), personnel capacity demand description (10 people in a meeting, 15 people in a meeting), equipment demand description (need a projector, want a MAXHUB), etc.

[0035] S202. Text preprocessing. Perform standardization processing on the input text, including unified time format, digital conversion, special character processing, etc. Specifically, it includes: receiving the natural language input in S201, performing text standardization processing and character processing, including cleaning the basic content of the text, performing format standardization conversion on the input text: Cleaning the basic content includes: cleaning HTML tags (removing script tags, HTML tags, etc.), special character processing (cleaning meaningless punctuation marks, etc.), text normalization (unifying case, space processing, etc.); Format standardization conversion includes unified time format, unified digital format, location name mapping, equipment term unified, etc.; Unified time format is to convert various time expressions into a standard time format; For example, colons and full stops are unified into "point", morning → morning, noon → afternoon, tomorrow morning → tomorrow morning, ten o'clock → 10 o'clock, two → 2, h → hour, min → minute; Unified digital format is to convert Chinese numbers and English numbers into Arabic numbers; Location name mapping is to convert aliases and abbreviations of locations into standardized names; For example, tt → Tongtai, tongt → Tongtai, sy → Shunyi, shuny → Shunyi, yt → Yuetan, yuet → Yuetan, Block C → Beichen C, Block E → Beichen E, Tongtai Grand Conference Room → Conference Room 1612, Yuetan Live Broadcast → 812, Tongtai Live Broadcast → 902, Round Table → Beichen C1505, Ladder → Beichen C1501, Equipment term unified is to convert synonyms and abbreviations into standard terms, for example, screen mirroring → projector, maxhub → MAXHUB, skype → Skype.

[0036] S203. Keyword extraction: Key information such as office area, floor, room number, time, equipment, and number of personnel is extracted using regular expressions and semantic analysis. Specifically, this includes: equipment information extraction, floor information extraction, number of personnel extraction, office area information extraction, room number extraction, and meeting duration extraction. In detail, user requirements can be categorized into precise requirements and fuzzy requirements. For precisely specified requirements, regular expression pattern matching is used. This involves extracting structured information using predefined regular expression patterns. For example, for an office space, office area information can match expressions like "Shunyi" or "Beichen C"; room number information can match formats like "C1217" or "1612"; floor information can match expressions like "16th floor" or "15th floor"; the number of people can match descriptions like "10 people" or "8 people"; and time can match expressions like "3 o'clock" or "10:30 AM". Time points are converted to the standard format %Y-%m-%d%H:%M:%S; and time periods (durations) are standardized to hours, with duration information matching expressions like "1 hour" or "3 to 5 o'clock". For fuzzy requirements, association analysis is used. For example, regarding office area information, if the request specifies a room number, it will be obtained by matching office areas with that room number; if the request specifies a floor, it will be obtained by matching office areas with that floor; otherwise, it will default to the office area containing the meeting room with the highest weight in the user's preference data. Room number information will not be analyzed if the user does not specify it; floor information will not be analyzed if the user does not specify it; and the number of people will default to 10 if the user does not specify it, but can also be obtained based on the required meeting room capacity. For example, if the request mentions "large meeting room,"... The default number of participants is 15 if the requirements specify a "small meeting room". If the requirements mention a "small meeting room", the default number of participants is 6. For the time information, the system can intelligently adjust to a future valid time based on the relationship between the current time period (0-6:00, 6-12:00, 12-18:00, 18-24:00) and the user-input keywords (morning, afternoon, week x, day of the week x, month x, day x). If the user does not specify a time, the system defaults to the next full 30-minute interval (e.g., if the current time is 14:12, the system defaults to 14:30). For the meeting duration, the system defaults to 1 hour if the user does not specify a time, based on the usual meeting duration.

[0037] S204. Time parsing, specifically including: time point identification, time range processing, fuzzy time parsing, and time verification and adjustment. This step supports multiple time expression methods, including fuzzy time expressions such as "this afternoon," "tomorrow morning," and "next Wednesday." Specifically, the duration (x hours, x half-hours) is extracted using regular expressions; the time points (x hours later, x half-hours later, year, month, day, hour, minute, second, tomorrow, the day after tomorrow, the day after that, week x, weekday x, next week, the week after next, next week x, the week after next x) are extracted using regular expressions; after obtaining the time points, the first time is the start time, the second time (if any) is the end time, and the third and subsequent times are ignored; based on the relationship between the current time period (0-6, 6-12, 12-18, 18-24) and the user's input keywords (morning, afternoon, week x, weekday x, x month, x day), the time is adjusted to a future time that conforms to the user's general understanding. For example: 1) If the hour number is between 1 and 6, check if the user's input contains "early morning". If not, add 12 to the hour number, assuming the user wants a meeting from 1 pm to 6 pm; 2) If the current time is Wednesday and the user's input contains Monday, adjust the time to next Monday.

[0038] S205. Semantic understanding: Identifying user intent and distinguishing between precise and vague requirements.

[0039] S206. The final parsed result is output in a structured format.

[0040] In some optional embodiments provided by the present invention, in step S3, a candidate meeting room list is generated based on the structured parsing results, combined with meeting room information and time information; Based on the structured analysis results, at least one of the following candidate meeting room information (office area, floor, room number) is selected. Time information is obtained based on the appointment time. The meeting room information and time information are combined by performing a Cartesian product to generate a candidate meeting room list. The candidate meeting room list includes a candidate meeting room set, which contains both meeting room information and time information.

[0041] Step S3 is the candidate solution generation step, which generates candidate meeting room solutions based on the analysis results. This includes: selecting relevant meeting rooms based on the office area, generating time plans (considering start time, end time, and meeting duration), combining meeting room and time plans, and applying basic constraints. Candidate solution generation is performed by the solution generation module 200, which generates candidate meeting room solutions based on multi-dimensional constraints. The specific process is as follows: Figure 5 As shown, it includes: S301. Input the parsed structured data. Input the structured parsing result obtained in step S2 into the scheme generation module 200.

[0042] S302. Candidate solution generation, specifically including: Meeting room selection: Select candidate meeting rooms based on criteria such as office area, floor, and room number; Time plan generation: Generate a time plan based on start time, end time, and meeting duration; Solution combination: Combine meeting room information with time plans using the Cartesian product; Constraints include equipment requirements and personnel capacity. Specifically, for office area information, meeting rooms in the office area that are not retained after semantic analysis are not included in the solution generation. There are no constraints on room number information, floor information, time information, and meeting duration information. For personnel number information, the meeting room capacity must be greater than or equal to 60% of the number of people required after semantic analysis.

[0043] Based on the above steps, candidate meeting room solutions are generated.

[0044] In some optional embodiments provided by the present invention, step S4, which involves multi-dimensional scoring of the candidate meeting room list to obtain a list of candidate meeting rooms sorted by total score, includes: Calculate the matching score for each dimension, apply user preference weights, process the bonus for exact matches, sort the results from highest to lowest total score, and generate a list of candidate meeting rooms. The scoring dimensions include: historical preference weight, time matching degree, floor matching degree, personnel capacity matching degree, equipment demand matching degree, and precise matching degree.

[0045] Step S4 is the intelligent scoring and ranking step, which performs multi-dimensional scoring and ranking on the candidate meeting room solutions obtained in Step S3. This includes: calculating the matching score for each dimension, applying user preference weights, processing exact match bonuses, and ranking by total score. Step S4 is performed by the intelligent scoring algorithm module 300, employing a multi-dimensional weighted scoring mechanism to achieve optimal ranking of the solutions. The specific process is as follows: Figure 5 As shown, it includes: S401. Multi-dimensional scoring: Scoring is conducted through several aspects, including historical preference score, time matching score, floor matching score, personnel capacity score, and equipment requirement score. In this embodiment, the historical preference score is 20 points, the time matching score is 40 points, the floor matching score is 10 points, the personnel capacity matching score is 10 points, and the equipment requirement matching score is 20 points. In addition, there is a bonus for precise matching; for precise requirements, i.e., when the user explicitly specifies a certain condition, an extra 50 points are awarded. It should be noted that when calculating the score for the meeting room, the user preference weight is part of the logic for calculating the room number field score. The meeting room score = the user's normalized preference result for the current meeting room × the base score of the room number field.

[0046] Specifically, the fields involved in calculating the score are: office area, room number, floor, equipment, number of people, time, and duration. The standard values ​​for each field used in the score calculation are a combination of precise specification and fuzzy judgment in semantic analysis (e.g., if a meeting room is precisely specified, but the office area and floor are determined based on the meeting room number, then the office area and floor are fuzzy judgment results). When calculating the score, different processing will be applied depending on whether the current field standard value is precisely specified by the user or fuzzy judgment (e.g., the standard value for the floor in this score calculation is the 1st floor, and the meeting room in this solution is also on the 1st floor; if the user precisely specifies "1st floor" in the input text, then the total score of the solution will be increased by 100 points; if the user does not specify, then the total score of the solution will be increased by the base score of the floor multiplied by 1, and the sum of the base scores of all fields will be 100).

[0047] S402. Scheme ranking and optimization: Based on the scoring results of each dimension in step S401, calculate the total score of the relevant meeting rooms, rank the meeting rooms according to their total scores, and mark them as recommended.

[0048] S403, Tag Management: Implement time tags, floor tags, and preference tags for candidate solutions.

[0049] S404. Output results: The output results include the initial recommended solution list, the initial non-recommended solution list, solution scores, and solution tags.

[0050] It should be noted that the difference between recommended and non-recommended solutions here is as follows: a solution that perfectly matches the user's exact needs is a recommended solution (e.g., if the user specifies a meeting room on the 3rd floor at 2 o'clock, then all solutions with meeting rooms on the 2nd floor and a start time of 2 o'clock are recommended solutions). Conversely, solutions that do not perfectly match the user's exact needs are non-recommended solutions.

[0051] In some optional embodiments provided by the present invention, in step S5, the list of meeting rooms to be selected is filtered and optimized for time conflicts to obtain an optimized list of meeting rooms, including: Based on the meeting room information in the list of candidate meeting rooms, check the occupancy status of the meeting rooms, eliminate the candidate set of meeting rooms with time conflicts, sort them in descending order of score, and prioritize recommending the candidate set of high-scoring meeting rooms.

[0052] In some optional embodiments provided by the present invention, step S5, which involves filtering and optimizing the list of meeting rooms to be selected based on time conflicts, to obtain an optimized list of meeting rooms, further includes: When the specified set of meeting room candidates is unavailable, an alternative set of meeting room candidates is generated from two aspects: time adjustment and / or duration optimization. Usage tips and instructions are added to obtain an optimized list of meeting rooms.

[0053] Step S5 is the solution screening and optimization step. It filters the candidate solutions, identifies usable solutions, and performs intelligent optimization, including: checking meeting room occupancy, eliminating time-clashing solutions, generating alternative solutions (time adjustment, duration optimization), and adding usage tips and instructions. Step S5 is performed by the solution screening and optimization module 400. Through multiple rounds of screening and intelligent optimization, it ensures the quality and usability of the recommended solutions. The specific process is as follows: Figure 6 As shown, it includes: S501, Candidate Solution List: This list consists of the alternative solutions obtained in step S4 after multi-dimensional intelligent scoring.

[0054] S502, Conflict Detection and Basic Screening: Query meeting room occupancy status and exclude time-conflicting solutions, including: occupancy status query, conflict detection algorithm, and conflict solution exclusion; specifically, query meeting room occupancy status and judge whether the generated solution time overlaps with the meeting room occupancy based on the conflict detection algorithm. If there is a conflict, it is excluded as a conflicting solution. S503, Scheme sorting and deduplication: Sort by score from high to low, and prioritize recommend high-scoring schemes, including score sorting, room deduplication and scheme deduplication; S504 Intelligent Extension and Optimization: When a specified plan is unavailable, it automatically generates an alternative plan with a similar time, including time boundary detection, duration optimization, and alternative plan generation. Duration optimization automatically adjusts the meeting duration according to user needs, providing multiple options such as half an hour, 45 minutes, and 1 hour. Specifically, the time boundary detection algorithm provides information on the meeting's potential for extension in the generated plan. For example, for a 5:00-6:00 plan, if the meeting room is occupied starting at 6:30, the user will be prompted that "the meeting room can be extended to 6:30 at most." The duration optimization algorithm adds a portion of the duration forward or backward to the generated plan to reduce the waste of meeting room resources. Specifically, duration optimization focuses on the meeting start time. Due to limited meeting room resources, if the user does not specify the exact meeting duration, the meeting duration closest to the standard start time needs to be extended as much as possible to fill the spare time before and after the meeting room start time, thereby improving resource utilization. For example, if the generated plan is from 5:30 to 6:00, but the meeting room is also available from 5:15 to 5:30, the plan time will be adjusted to 5:15 to 6:00 if the user does not specify the meeting duration, trying to get as close as possible to the default meeting duration of 1 hour. On the other hand, time adjustment focuses on the meeting duration. If the user specifies the exact meeting duration, even if the available time slots in the meeting room are fragmented, a plan that meets the specified duration must be added to the plan list for scoring. In this case, the plan will search for available time slots in the meeting room indefinitely until an available time slot that meets the specified meeting duration is found.

[0055] S505. Solution integration and classification: Screen and integrate the solutions, and classify the solutions into recommended solutions and non-recommended solutions. S506. Prompt information generation, including usage prompts, boundary prompts, and scenario prompts. Specifically, it displays tags for individual plans to the user: similar floors, same floors, similar times, same times, and recommended plans. It also informs the user of the maximum extension time for this plan. For example, for a 5:00-6:00 plan, if the meeting room is occupied starting at 6:30, the user will be prompted that "this meeting room can be extended to 6:30 at most." This prompt will appear if the meeting room is occupied within one hour after the plan's end time; otherwise, it will not appear. It also displays a list of plans for this request and an explanation of the response value, such as "No meeting room available for the specified time period was found. The following plans are recommended for you" (i.e., no recommended plans are displayed, only non-recommended plans are shown).

[0056] S507. Output Results: The output results include a solution description, prompts, a list of non-recommended solutions, and a list of recommended solutions.

[0057] In some optional embodiments provided by the present invention, step S6, returning the optimized meeting room list to the user terminal, includes: The optimized meeting room list is divided into a recommended list and a non-recommended list. Solution descriptions and prompts are added, the results are formatted, user selections are recorded, and preference data is updated.

[0058] Step S6 is the result return step, which returns the optimized recommended solutions to the user. This includes: dividing the optimized solutions into a recommended list and a non-recommended list; in this embodiment, the total number of solutions in each list does not exceed nine, and any excess will be removed; adding solution descriptions and prompts; formatting the returned results; recording user selections and updating preference data. After formatting the returned results, the user can select a solution and modify the meeting time (at this time, a new prompt such as "Anytime available for booking is up to x o'clock" will be given based on the selected time). After user confirmation, the solution parameters will be immediately used for meeting room booking, and the booking result will be displayed to the user.

[0059] In one embodiment, an intelligent meeting room reservation system is provided, such as Figure 2 As shown, it includes a natural language understanding module 100, a solution generation module 200, an intelligent scoring algorithm module 300, a solution selection and optimization module 400, a user preference learning module 500, and a real-time data synchronization module 600. The natural language understanding module 100, coupled with the solution generation module 200, receives and parses the user's natural language input. Based on multi-level semantic information extraction technology, it identifies the user's intent, distinguishes between precise and fuzzy requirements, obtains structured parsing results, and outputs the structured parsing results to the solution generation module. The solution generation module 200 is coupled to the natural language understanding module 100 and the real-time data synchronization module 600 respectively. Based on the structured parsing results output by the natural language understanding module 100 and the meeting room information synchronized by the real-time data synchronization module 600, it generates a list of candidate meeting rooms and outputs it to the intelligent scoring algorithm module 300. The intelligent scoring algorithm module 300 is coupled to the solution generation module 200, the user preference learning module 500, and the real-time data synchronization module 600 respectively. It receives the candidate meeting room list output by the solution generation module 200, combines the user preference data provided by the user preference learning module 500 and the meeting room information synchronized by the real-time data synchronization module 600, and obtains a list of candidate meeting rooms sorted by total score based on a multi-dimensional weighted scoring mechanism, which is then transmitted to the solution screening and optimization module 400. The scheme selection and optimization module 400 is coupled to the intelligent scoring algorithm module 300, the user preference learning module 500, and the real-time data synchronization module 600, respectively. It receives the list of candidate meeting rooms output by the intelligent scoring algorithm module 300, combines the user preference data provided by the user preference learning module 500 and the meeting room information synchronized by the real-time data synchronization module 600, eliminates the candidate set of meeting rooms with time conflicts, optimizes the meeting room list, and outputs the optimized meeting room list. The user preference learning module 500 is coupled to the scheme selection and optimization module 400 and the intelligent scoring algorithm module 300 respectively. The user preference learning module 500 builds a user preference model based on historical reservation data, which is used to provide user preference data to the scheme selection and optimization module 400 and the intelligent scoring algorithm module 300. The real-time data synchronization module 600 is coupled to the solution generation module 200, the solution filtering and optimization module 400, and the intelligent scoring algorithm module 300, respectively. It is used to provide the solution generation module 200, the solution filtering and optimization module 400, and the intelligent scoring algorithm module 300 with real-time information about the meeting rooms, such as information about the meeting rooms in each office area and real-time occupancy data of the meeting rooms.

[0060] Specifically, this application establishes a multi-source data synchronization mechanism to ensure the real-time performance and accuracy of system data. In this embodiment, basic meeting room information is synchronized daily at 00:00, meeting room occupancy is queried in real time, a memory cache is established to improve query efficiency, and a retry mechanism is implemented to ensure the reliability of data synchronization.

[0061] In one embodiment, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it performs the following steps: initializing the intelligent meeting room reservation system by loading basic data and configuration information; receiving and parsing the user's natural language input request information, identifying the user's intent and request type based on multi-level text parsing technology, and obtaining structured parsing results; generating candidate meeting room plans based on the structured parsing results, combined with meeting room information and time plans; performing multi-dimensional scoring on the candidate meeting room plans to obtain candidate plans ranked by total score; filtering and optimizing the candidate plans for time conflicts to obtain optimized plans; and returning the optimized plans to the user.

[0062] As can be seen from the above embodiments, the intelligent meeting room reservation method, system, and electronic device provided by the present invention achieve at least the following beneficial effects: This application uses traditional methods for text parsing and recommendation weight calculation, and combines user intent, user history behavior, and meeting room occupancy to provide intelligent meeting room reservation results. It can be deployed in a pure CPU environment, with an interface response speed of up to 200ms and a reliability of 99%.

[0063] While specific embodiments of the invention have been described in detail by way of examples, those skilled in the art should understand that the examples are for illustrative purposes only and not intended to limit the scope of the invention. Those skilled in the art should understand that modifications can be made to the above embodiments without departing from the scope and spirit of the invention. The scope of the invention is defined by the appended claims.

Claims

1. A method for intelligent booking of meeting rooms, characterized in that, include: Initialize the intelligent meeting room reservation system by loading basic data and configuration information; It receives and parses the user's natural language input request information, and based on multi-level semantic information extraction technology, identifies the user's intent and request type to obtain structured parsing results; Based on the structured parsing results, combined with meeting room information and time information, a candidate meeting room list is generated, which includes a candidate set of meeting rooms; The candidate meeting room list is scored from multiple dimensions to obtain a list of candidate meeting rooms sorted by total score; The list of candidate meeting rooms is filtered and optimized for time conflicts to obtain an optimized list of meeting rooms. The optimized list of meeting rooms is returned to the user.

2. The method according to claim 1, characterized in that, The initialization of the intelligent meeting room reservation system, including loading basic data and configuration information, includes: Load basic meeting room information from the database, initialize the user preference weight dictionary, start a scheduled task, update the data periodically, and establish a connection with the third-party meeting room system; The method for establishing the user preference weight dictionary includes: Collect reservation records within the user reservation frequency threshold and analyze the meeting room usage frequency; calculate user preference weights according to the weight formula; and normalize the weights. The user preference weights are calculated as follows: ； Where D represents the number of days remaining until the current date, corresponding to the user's reservation threshold.

3. The method according to claim 1, characterized in that, The process of receiving and parsing the user's natural language input, based on multi-level semantic information extraction technology, identifies the user's intent and demand type, and obtains structured parsing results, includes: It receives text input from users, performs text preprocessing and standardization on the text, extracts key information through regular expressions and semantic analysis, identifies user intent and demand type, and outputs structured parsing results. The key information includes office area, floor, room number, appointment time, equipment, and number of personnel.

4. The method according to claim 1, characterized in that, Based on the structured parsing results, combined with meeting room information and time information, a candidate meeting room list is generated; Based on at least one of the office area, floor, and room number in the structured parsing results, candidate meeting room information is filtered. Time information is obtained based on the appointment time. The meeting room information and the time information are combined by performing a Cartesian product to generate the candidate meeting room list. The candidate meeting room list includes a candidate meeting room set, which includes the meeting room information and the time information.

5. The method according to claim 1, characterized in that, The process of scoring the candidate meeting room list from multiple dimensions to obtain a list of candidate meeting rooms sorted by total score includes: Calculate the matching score for each dimension, apply user preference weights, process the bonus for exact matches, sort the results from highest to lowest total score, and generate the list of candidate meeting rooms. The scoring dimensions include: historical preference weight, time matching degree, floor matching degree, personnel capacity matching degree, equipment demand matching degree, and precise matching degree.

6. The method according to claim 1, characterized in that, The step of filtering and optimizing the list of candidate meeting rooms for time conflicts to obtain an optimized list of meeting rooms includes: Based on the meeting room information in the list of candidate meeting rooms, query the occupancy status of the meeting rooms, eliminate the candidate set of meeting rooms with time conflicts, sort them in descending order of score, and prioritize recommending the candidate set of high-scoring meeting rooms.

7. The method according to claim 1, characterized in that, The step of filtering and optimizing the list of candidate meeting rooms for time conflicts to obtain an optimized list of meeting rooms also includes: When the specified set of meeting room candidates is unavailable, an alternative set of meeting room candidates is generated from two aspects: time adjustment and / or duration optimization. Usage tips and instructions are added to obtain an optimized list of meeting rooms.

8. The method according to claim 1, characterized in that, Returning the optimized meeting room list to the user includes: The optimized meeting room list is divided into a recommended list and a non-recommended list. Scheme descriptions and prompts are added, the results are formatted, user selections are recorded, and preference data is updated.

9. An intelligent meeting room reservation system, characterized in that, include: Natural Language Understanding Module, Solution Generation Module, Intelligent Scoring Algorithm Module, Solution Selection and Optimization Module, User Preference Learning Module, Real-time Data Synchronization Module; The natural language understanding module is coupled to the solution generation module, receives and parses the user's natural language input, identifies the user's intent based on multi-level semantic information extraction technology, distinguishes between precise and fuzzy requirements, obtains structured parsing results, and outputs the structured parsing results to the solution generation module. The scheme generation module is coupled to the natural language understanding module and the real-time data synchronization module respectively. Based on the structured parsing results output by the natural language understanding module and the meeting room information synchronized by the real-time data synchronization module, it generates a candidate meeting room list and outputs it to the intelligent scoring algorithm module. The intelligent scoring algorithm module is coupled to the scheme generation module, the user preference learning module, and the real-time data synchronization module, respectively. It receives the candidate meeting room list output by the scheme generation module, combines the user preference data provided by the user preference learning module and the meeting room information synchronized by the real-time data synchronization module, and obtains a list of candidate meeting rooms sorted by total score based on a multi-dimensional weighted scoring mechanism, which is then transmitted to the scheme screening and optimization module. The scheme screening and optimization module is coupled to the intelligent scoring algorithm module, the user preference learning module, and the real-time data synchronization module, respectively. It receives the candidate meeting room list output by the intelligent scoring algorithm module, combines the user preference data provided by the user preference learning module and the meeting room information synchronized by the real-time data synchronization module, eliminates the candidate set of meeting rooms with time conflicts, optimizes the meeting room list, and outputs the optimized meeting room list. The user preference learning module is coupled to the scheme selection and optimization module and the intelligent scoring algorithm module respectively. The user preference learning module constructs a user preference model based on historical reservation data, which is used to provide user preference data to the scheme selection and optimization module and the intelligent scoring algorithm module. The real-time data synchronization module is coupled to the scheme generation module, the scheme filtering and optimization module, and the intelligent scoring algorithm module, respectively, and is used to provide the scheme generation module, the scheme filtering and optimization module, and the intelligent scoring algorithm module with real-time information of the conference room.

10. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to execute the computer program to implement the steps of the intelligent meeting room reservation method as described in any one of claims 1 to 8.

Images