Semantic big data analysis method and business intelligence system

Abstract

The present invention belongs to the field of artificial intelligence technology, and in particular relates to a semantic big data analysis method and a business intelligence system. The method first obtains historical task completion status and undertaking team and individual attribute data, and uses an integrated model to build a dynamic task undertaking portrait pool; secondly, the published tasks are analyzed for attributes, and are decomposed into subtasks using a rule splitting model; based on the dynamic task undertaking portrait pool and subtask attributes, the matching scores between teams and tasks, members and subtasks, and team task coverage are calculated through a double matching model; according to preset matching rules, tasks are intelligently assigned to teams or individuals, and the interface configuration between subtasks is optimized; finally, task progress is monitored in real time, the completion status is scored using an evaluation model, and the feedback is used to dynamically adjust or update the task undertaking portrait pool, thereby realizing intelligent management of task allocation and completion status evaluation, and improving task allocation efficiency and completion quality.

Description

Technical Field

[0001] The present invention belongs to the field of artificial intelligence technology, and in particular relates to a semantic big data analysis method and a business intelligence system. Background Art

[0002] With the acceleration of digital transformation, more and more companies are beginning to rely on flexible labor markets to meet their changing human resource needs; especially in the IT field, a large number of freelancers and contractors have become an indispensable part of the company; in this context, business intelligence (BI) technology is widely used to optimize human resource management and task allocation processes. At present, there are platforms such as "Freelancers" on the market, which provide a bridge for companies and freelancers through the "platform + technology + service" model. Such platforms can support various types of human resource needs including software development, IT consulting, project management, testing, database and big data analysis, Internet operation and promotion, etc.; however, these platforms are unable to split the released tasks, which results in a low degree of matching of corresponding teams or talents;

[0003] For example, the patent with authorization announcement number CN115577983B discloses a blockchain-based enterprise task matching method, server and storage medium. The method includes: obtaining the first standard professional information of the target task; for each recommended node, obtaining the second standard professional information of the recommended node personnel; calculating the spatial distance between the first standard professional information and the second standard professional information, and determining the task matching degree between the personnel and the target task based on the spatial distance; selecting personnel whose task matching degree is within a preset matching degree range to form a target task team to complete the target task.

[0004] For example, the patent with publication number CN115269771A discloses a semantic-based big data analysis system, which includes a data collection unit, a data identification unit, a data analysis unit and a data visualization unit. The data collection unit is used to store and update big data in real time, the data identification unit is used to identify and preliminarily filter the information required by the user, the data analysis unit is used to integrate, classify and correlate the big data information, form analysis results, and provide real-time data required for analysis, and the data visualization unit is used to present the data analysis results in graphics and voice that can be recognized by the user.

[0005] The above existing technologies have the following problems: the existing platforms lack sufficient intelligent means for task splitting, which may lead to unreasonable splitting results, resulting in low matching degree of corresponding talents or teams. For this reason, the present invention provides a semantic big data analysis method and a business intelligence system. Summary of the Invention

[0006] In response to the shortcomings of the existing technology, the present invention proposes a semantic big data analysis method and a business intelligence system. The method first obtains historical task completion status and the attribute data of the undertaking team and individuals, and uses an integrated model to build a dynamic task undertaking portrait pool; secondly, the published tasks are analyzed for attributes and decomposed into subtasks using a rule splitting model; based on the dynamic task undertaking portrait pool and subtask attributes, the matching scores of teams and tasks, members and subtasks, and team task coverage are calculated through a double matching model; according to preset matching rules, tasks are intelligently assigned to teams or individuals, and the interface configuration between subtasks is optimized; finally, task progress is monitored in real time, the completion status is scored using an evaluation model, and the feedback is used to dynamically adjust or update the task undertaking portrait pool, thereby realizing intelligent management of task allocation and completion status evaluation, and improving task allocation efficiency and completion quality.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] Based on the semantic big data analysis method, the steps include:

[0009] S1. Obtain the completion status data of historically released tasks and the corresponding attribute distribution status data of the undertaking teams and individuals. Based on the acquired data, a dynamic task undertaking profile pool is obtained through the configured integration model.

[0010] S2. Obtain the attribute data of the published task, split the published task using the configured rule splitting model, and obtain the list of subtasks after splitting and the corresponding attribute data;

[0011] S3. Based on the dynamic task acceptance profile pool, the list of split subtasks, and the corresponding attribute data, the configured dual matching model is used to obtain the matching score between the team and the task, the matching score between team members and the corresponding subtasks, and the team's coverage of the task;

[0012] S4. Set matching rules. Match the corresponding teams according to the set matching rules and the obtained matching scores and coverage. If all teams do not meet the matching rules, match the split subtasks with individual individuals, and configure the upstream and downstream subtask interface attributes for the subtasks that have been matched with individuals. At the same time, continue to match the remaining subtasks with teams until all subtasks are assigned.

[0013] S5. Obtain the matching team or individual task completion status data in real time, and evaluate the team or individual task completion status through the configured evaluation model to obtain the completion status evaluation score, and feed the obtained evaluation score back to the dynamic task acceptance portrait pool to adjust or delete the team or individual attribute status in the pool.

[0014] Specifically, the dynamic task acceptance portrait pool includes a team task acceptance portrait sub-pool and a personal task acceptance portrait sub-pool; the steps for constructing the team task acceptance portrait sub-pool include:

[0015] S101. Obtain information on the professional skill types and proficiency of team members, the number of complete or partial tasks undertaken per historical unit time, the maximum task load, the number of remaining tasks at the current moment, the types of tasks undertaken, the time percentage of task completion, the task undertaking fee, the initial delivery success rate of different types of tasks, the delivery success rate of different types of tasks, and textual information on satisfaction with delivered tasks;

[0016] S102: Input the acquired text information into the configured entity extraction algorithm to sequentially obtain team member skill attribute keyword information and task attribute keyword information, and automatically annotate the acquired team member skill attribute keyword information and task attribute keyword information with score tags using the trained automatic annotation model;

[0017] S103: Input the labeled team member skill attribute keyword information into the evaluation algorithm 1 to obtain the team's overall skill proficiency evaluation score and the corresponding skill proficiency evaluation score of each member;

[0018] S104: Input the annotated task attribute keyword information into the evaluation algorithm 2 to obtain the team's task acceptance evaluation score and the team's task satisfaction score;

[0019] S105. Obtain a comprehensive team assessment score based on the obtained team overall skill proficiency score, team task capability assessment score, and team task satisfaction score;

[0020] S106. Input the obtained keyword information and corresponding evaluation scores of different teams into the graph database to construct a team task acceptance portrait sub-pool. Use the team name as the first-level node in the team task acceptance portrait sub-pool. Use the obtained comprehensive evaluation score of each team as the sorting label of the first-level node to sort the teams in descending order.

[0021] S107: Use the corresponding team member's professional skill type and proficiency keyword information as a secondary node, and use the obtained team's overall skill proficiency score as a secondary node sorting label to sort the team in descending order of skill level;

[0022] S108. The keyword information of the attribute of the historical tasks undertaken by the corresponding team is used as the third-level node, and the obtained evaluation score of the team's ability to undertake tasks is used as the first-level sorting label of the third-level node, and the task satisfaction score is used as the second-level sorting label.

[0023] Specifically, the steps for building a sub-pool for personal task acceptance portraits include:

[0024] S109: Obtain the text information of S101 corresponding to the non-team individual, and obtain the skill proficiency assessment score and personal task satisfaction score of the non-team individual through the process of S102-S104;

[0025] S110: Input the acquired non-team individual skill proficiency into the graph database to construct a sub-pool of individual task acceptance portraits. Use the corresponding individual name keywords as first-level nodes in the sub-pool of individual task acceptance portraits, and use the acquired individual skill proficiency assessment scores as first-level ranking labels.

[0026] S111. Map the occupational type and proficiency attribute information of the team members in the team task acceptance profile sub-pool to the first-level nodes in the individual task acceptance profile sub-pool, and use the obtained skill proficiency assessment scores of the individual members as the corresponding first-level ranking labels;

[0027] S112. Arrange the personal information in the first-level nodes of the personal task acceptance profile sub-pool in descending order based on the obtained first-level ranking labels of personal skill proficiency. Simultaneously, use the task attribute information of non-team individuals as second-level nodes, and use the corresponding personal task satisfaction scores as ranking labels within the second-level nodes.

[0028] S113. Map the team task satisfaction score to the team member's corresponding individual task satisfaction score as a ranking label in the second-level node of the team member's individual task acceptance profile sub-pool;

[0029] S114. Sort the task attribute information of team members and non-team members according to the obtained sorting labels in the secondary nodes, and obtain a constructed sub-pool of individual task acceptance portraits.

[0030] Specifically, the steps for building a dynamic task acceptance profile pool include:

[0031] S115. Based on the mapping relationship between team members and personal task acceptance portrait sub-pools in S113, the personal task acceptance portrait sub-pool and the team task acceptance portrait sub-pool are connected to obtain a constructed dynamic task acceptance portrait pool;

[0032] S116. Set a team comprehensive evaluation score threshold and an individual task satisfaction score threshold, and embed the team comprehensive evaluation score threshold into the team task acceptance profile sub-pool, and embed the individual task satisfaction score threshold into the individual task acceptance profile sub-pool. When the corresponding team comprehensive evaluation score is lower than the team comprehensive evaluation score threshold, the corresponding team will be deleted from the team task acceptance profile sub-pool;

[0033] S117. If the personal task satisfaction score of a non-team member is lower than the personal task satisfaction score threshold, the corresponding individual will be deleted from the personal task acceptance portrait sub-pool. If the personal task satisfaction score of an individual in the team is lower than the personal task satisfaction score threshold, an alarm will be issued to the team as a whole.

[0034] Specifically, the steps of obtaining the split subtask list in S2 include:

[0035] S201. Obtain overall information of the released task, and extract the task cycle information and overall required skill type information corresponding to the released task through the configured entity extraction algorithm;

[0036] S202: Set a period split threshold. If the obtained task period is greater than the period split threshold, input the overall information of the released task into the comprehensive fuzzy evaluation algorithm to evaluate the urgency score of each subtask in the overall task;

[0037] S203, dividing the overall task into an early stage subtask set, a mid-term subtask set, and a late stage subtask set based on the obtained urgency scores of the subtasks at each stage;

[0038] S204: For the obtained early subtask set, mid-term subtask set, and late subtask set, using the configured skill recognition model, obtain the occupational skill type required for the early subtask set, the occupational skill type required for the mid-term subtask set, and the occupational skill type required for the late subtask set;

[0039] S205. Functionally split the early subtask set, the mid-term subtask set and the late subtask set according to the required professional skill types obtained to obtain the early functional subtask set, the mid-term functional subtask set and the late functional subtask set, and use the obtained early functional subtask set, the mid-term functional subtask set and the late functional subtask set to construct a subtask list.

[0040] Specifically, the steps of obtaining the split subtask list in S2 include:

[0041] S206: If the acquired task period is less than or equal to the period splitting threshold, then directly repeat S204-S205 to functionally split the overall task to obtain a functional subtask set and a subtask list;

[0042] S207. If, through the configured skill recognition model, it is determined that there is only one type of professional skill required for the overall task, the task will not be split, and the corresponding task will be marked for individual matching.

[0043] Specifically, the steps of task matching in S3 and S4 include:

[0044] S301. Match the subtask list after the corresponding task is split with the team task acceptance profile subpool. Using matching algorithm 1, obtain a matching score 1 between each team as a whole and the overall task. Simultaneously, using matching algorithm 2, obtain a matching score 2 between each team member and the required professional skill type in the subtask list.

[0045] S302: Calculate the coverage of each team for the required professional skills in the subtask list based on the intersection of the professional skill types required in the subtask list and the professional skill types corresponding to the team members, and the professional skill types required in the subtask list;

[0046] S303: Set a match score 1 threshold, a match score 2 threshold, and a coverage threshold. If there is more than one team whose match score 1, match score 2, and coverage are all greater than the corresponding thresholds, match the task to the team with the highest initial task delivery success rate based on the initial task delivery success rate of the corresponding teams.

[0047] S304. If the initial task delivery success rate of all teams that meet the conditions in S303 is 0, the task will be matched to the team with the lowest task undertaking fee;

[0048] S305: If all teams do not meet the coverage threshold, then compare the matching score 1 and matching score 2 with the corresponding thresholds. If there are more than one team that meets both the matching score 1 threshold and the matching score 2 threshold, match the functional subtask set covered by the team to the team with the largest coverage.

[0049] S306: If there is more than one team with the largest coverage, S303 and S304 will be repeated to match teams and tasks.

[0050] Specifically, the task matching steps in S3 and S4 also include:

[0051] S307. If there is more than one uncovered functional subtask after the matching process in S306, the remaining functional subtask is matched with the unmatched team that meets the conditions in S305. If there is a team whose matching score 1, matching score 2, and coverage are all greater than the corresponding threshold, the remaining functional subtask is assigned to the team that meets the conditions. At the same time, the interface attributes of the functional subtask assigned in S306 that is adjacent to the remaining functional subtask are configured to the remaining functional subtask assigned in S307.

[0052] S308. If no team satisfies the matching score 1, matching score 2, and coverage greater than the corresponding thresholds, the remaining functional subtasks are matched with individuals in the personal task undertaking profile subpool. If there are more than one individual with a matching score greater than the matching score 2 threshold for the same remaining functional subtask, the corresponding remaining functional subtask is matched to the individual with the highest personal task satisfaction score until all remaining functional subtasks are matched.

[0053] S309: If the number of individuals with the highest individual task satisfaction scores is greater than 1, the task is matched to the individual with the highest corresponding skill proficiency assessment score, and the interface attributes of the functional subtasks assigned in S306 adjacent to the remaining functional subtasks are configured into the remaining functional subtasks assigned in S307;

[0054] S310: If the obtained task period is less than or equal to the period split threshold, repeat S206 and S301-S303. If there is a team that meets the conditions of S303 and the number of teams is greater than 1, match the task to the team with the smallest proportion of task completion time.

[0055] S311. If there is no team that meets the conditions of S303, repeat the process of S305-S309 to assign tasks;

[0056] S312. When there is only one type of professional skill required for the overall task, repeat S308-S309 to match individuals with tasks from the personal task acceptance portrait sub-pool, and match the task to the individual who meets the conditions in S308-S309.

[0057] A business intelligence system based on semantic big data analysis, including: a portrait pool building module, a task splitting module, and a matching module;

[0058] The portrait pool construction module includes a data processing unit, a portrait unit, and a portrait pool unit;

[0059] The data processing unit is used to obtain the completion status data of historical release tasks and the corresponding attribute distribution status data of the undertaking teams and individuals, and perform standardized preprocessing on the obtained data;

[0060] The portrait unit constructs hierarchical team and individual portraits based on pre-processed team and individual attribute distribution data using a configured natural language algorithm. The portrait pool unit uses the constructed hierarchical team and individual portraits to construct a dynamic task acceptance portrait pool through a graph database.

[0061] The task splitting module is used to obtain the attribute data of the published task, split the published task through the configured rule splitting model, and obtain the differential subtask list and corresponding attribute data;

[0062] The matching module includes a team matching unit, a matching judgment unit, and a matching optimization unit;

[0063] The team matching unit uses the configured dual matching model based on the dynamic task acceptance profile pool, the list of split subtasks, and the corresponding attribute data to obtain the matching score between the team and the task, the matching score between team members and the corresponding subtasks, and the team's coverage of the task;

[0064] The matching judgment unit matches the corresponding team based on the obtained matching score and coverage by setting matching rules. If all teams do not meet the matching rules, the split subtasks are matched with a single person;

[0065] The matching optimization unit is used to configure the upstream and downstream subtask interface attributes for the subtasks that have been matched with individuals, and to continue matching the remaining subtasks with the team until all subtasks are assigned.

[0066] A computer-readable storage medium stores computer instructions, which, when executed, execute the semantic big data analysis method.

[0067] Compared with the prior art, the present invention has the following beneficial effects:

[0068] In response to the deficiencies of the existing technology, the present invention, through the configured rule splitting model, can scientifically and rationally split the task into multiple subtasks according to the complexity of the task and the required skill type, and generate a detailed subtask list and corresponding attribute data, so as to ensure that each subtask maintains both independence and overall consistency of the task; secondly, in the matching stage, the dynamic task acceptance portrait pool and the dual matching model are used to accurately evaluate the matching scores between the team and the task, the matching scores between the team members and the subtasks, and the team's coverage of the task, thereby greatly improving the accuracy and efficiency of the matching, effectively solving the deficiencies of the existing platform in task splitting and talent matching, ensuring the rationality and efficiency of task allocation, and thus significantly improving the overall execution quality of the project and the management level of the enterprise. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 This is a flow chart of the semantic big data analysis method of the present invention;

[0070] Figure 2 This is a module diagram of the business intelligence system based on semantic big data analysis of the present invention. DETAILED DESCRIPTION

[0071] Example 1

[0072] See also Figure 1 , an embodiment provided by the present invention: based on a semantic big data analysis method, the steps include:

[0073] S1. Obtain the completion status data of historically released tasks and the corresponding attribute distribution status data of the undertaking teams and individuals. Based on the acquired data, a dynamic task undertaking profile pool is obtained through the configured integration model.

[0074] Furthermore, in this embodiment, the dynamic task undertaking portrait pool includes a team task undertaking portrait sub-pool and an individual task undertaking portrait sub-pool;

[0075] Furthermore, in this embodiment, the steps of constructing a team task acceptance portrait sub-pool include:

[0076] S101. Obtain information on the professional skill types and proficiency of team members, the number of complete or partial tasks undertaken per historical unit time, the maximum task load, the number of remaining tasks at the current moment, the types of tasks undertaken, the time percentage of task completion, the task undertaking fee, the initial delivery success rate of different types of tasks, the delivery success rate of different types of tasks, and textual information on satisfaction with delivered tasks;

[0077] Furthermore, in this embodiment, professional skill proficiency is measured by the number of years that the corresponding team members have engaged in the current professional skills. For example, in the IT field, professional skill types include front-end, back-end, data analysis, operation and maintenance, testing, UG, art, product, etc. The professional skills corresponding to different fields are set by those skilled in the art.

[0078] Furthermore, in this embodiment, the time ratio for task completion is the actual length of time the team completes the task after accepting it and the length of the task completion cycle specified in the task release; the type of task accepted in this embodiment is the industry type corresponding to the task as a whole, such as the IT industry, the painting industry, etc.; the initial delivery success rate in this embodiment is the success rate of the task being passed on the first delivery to Party A after completion; the task delivery success rate in this embodiment is the success rate of tasks delivered after multiple revisions, excluding tasks that were successfully delivered on the first try;

[0079] S102: Input the acquired text information into the configured entity extraction algorithm to sequentially obtain team member skill attribute keyword information and task attribute keyword information, and automatically annotate the acquired team member skill attribute keyword information and task attribute keyword information with score tags using the trained automatic annotation model;

[0080] In this embodiment, the team member skill attribute keyword information, such as member A, engaged in back-end, ten years of experience; task attribute keyword information, such as task cycle length, industry field, quotation, etc.;

[0081] S103: Input the labeled team member skill attribute keyword information into the evaluation algorithm 1 to obtain the team's overall skill proficiency evaluation score and the corresponding skill proficiency evaluation score of each member;

[0082] Furthermore, the evaluation algorithm in this embodiment is a comprehensive fuzzy evaluation algorithm;

[0083] Furthermore, in this embodiment, the skill proficiency assessment score of a single member is marked by the number of years each team member has been engaged in the corresponding skill type; for example, if member A has been engaged in back-end work for 5 years, member A will be given a score of 5 points;

[0084] Furthermore, the overall skill proficiency assessment score of the team is the sum of the product of the skill proficiency assessment score of each team member and the probability of successfully completing the corresponding skill subtask in the task they undertake in one go. The probability of successfully completing the corresponding skill subtask in a task in one go is the ratio of the number of corresponding skill subtasks successfully completed in one go in all the tasks of each team member in the past to the total number.

[0085] S104: Input the annotated task attribute keyword information into the evaluation algorithm 2 to obtain the team's task acceptance evaluation score and the team's task satisfaction score;

[0086] Furthermore, in this embodiment, the steps for obtaining the evaluation score of the task that the team can undertake include:

[0087] Based on the number of complete or partial tasks undertaken in the historical unit time, the maximum task load, the number of remaining tasks at the current moment, and the type of tasks undertaken, a comprehensive fuzzy algorithm evaluation factor and evaluation index are constructed;

[0088] Input the constructed evaluation factors and evaluation indicators into the comprehensive fuzzy algorithm to obtain the evaluation score of the team's ability to undertake tasks;

[0089] Similarly, the team task satisfaction score is obtained through a comprehensive fuzzy algorithm using the time ratio of task completion, task undertaking costs, the initial delivery success rate of different types of tasks, the delivery success rate of different types of tasks, and the textual information of satisfaction with the delivered tasks.

[0090] S105. Obtain a comprehensive team assessment score based on the obtained team overall skill proficiency score, team task capability assessment score, and team task satisfaction score;

[0091] S106. Input the obtained keyword information and corresponding evaluation scores of different teams into the graph database to construct a team task acceptance portrait sub-pool. Use the team name as the first-level node in the team task acceptance portrait sub-pool. Use the obtained comprehensive evaluation score of each team as the sorting label of the first-level node to sort the teams in descending order.

[0092] S107: Use the corresponding team member's professional skill type and proficiency keyword information as a secondary node, and use the obtained team's overall skill proficiency score as a secondary node sorting label to sort the team in descending order of skill level;

[0093] S108. The keyword information of the attribute of the historical tasks undertaken by the corresponding team is used as the third-level node, and the obtained evaluation score of the team's ability to undertake tasks is used as the first-level sorting label of the third-level node, and the task satisfaction score is used as the second-level sorting label.

[0094] During the construction of the team task acceptance profile sub-pool, the process comprehensively evaluates the professional skill types, proficiency, and other key indicators of team members (such as the proportion of task completion time, task acceptance costs, delivery success rate, etc.), and uses the configured entity extraction algorithm and automatic labeling model to ensure the high accuracy of the team's overall skill proficiency score and the corresponding skill proficiency assessment scores of individual members; secondly, the task attributes are evaluated through evaluation algorithm 2 to obtain the team's task acceptance assessment score and task satisfaction score, further optimizing the rationality of task allocation; finally, by inputting these assessment scores into the graph database to construct the team task acceptance profile sub-pool and sorting them according to the comprehensive assessment scores, dynamic monitoring and optimization of the team's task acceptance capability is achieved; this process not only improves the scientificity and rationality of task allocation, but also enhances the transparency and controllability of team management.

[0095] Furthermore, in this embodiment, the steps of constructing the sub-pool of personal task acceptance portraits include:

[0096] S109: Obtain the text information of S101 corresponding to the non-team individual, and obtain the skill proficiency assessment score and personal task satisfaction score of the non-team individual through the process of S102-S104;

[0097] S110: Input the acquired non-team individual skill proficiency into the graph database to construct a sub-pool of individual task acceptance portraits. Use the corresponding individual name keywords as first-level nodes in the sub-pool of individual task acceptance portraits, and use the acquired individual skill proficiency assessment scores as first-level ranking labels.

[0098] S111. Map the occupational type and proficiency attribute information of the team members in the team task acceptance profile sub-pool to the first-level nodes in the individual task acceptance profile sub-pool, and use the obtained skill proficiency assessment scores of the individual members as the corresponding first-level ranking labels;

[0099] S112. Arrange the personal information in the first-level nodes of the personal task acceptance profile sub-pool in descending order based on the obtained first-level ranking labels of personal skill proficiency. Simultaneously, use the task attribute information of non-team individuals as second-level nodes, and use the corresponding personal task satisfaction scores as ranking labels within the second-level nodes.

[0100] S113. Map the team task satisfaction score to the team member's corresponding individual task satisfaction score as a ranking label in the second-level node of the team member's individual task acceptance profile sub-pool;

[0101] S114. Sort the task attribute information of team members and non-team members according to the obtained sorting labels in the secondary nodes, and obtain a constructed sub-pool of individual task acceptance portraits.

[0102] Furthermore, in this embodiment, the steps of constructing a dynamic task acceptance portrait pool include:

[0103] S115. Based on the mapping relationship between team members and personal task acceptance portrait sub-pools in S113, the personal task acceptance portrait sub-pool and the team task acceptance portrait sub-pool are connected to obtain a constructed dynamic task acceptance portrait pool;

[0104] S116. Set a team comprehensive evaluation score threshold and an individual task satisfaction score threshold, and embed the team comprehensive evaluation score threshold into the team task acceptance profile sub-pool, and embed the individual task satisfaction score threshold into the individual task acceptance profile sub-pool. When the corresponding team comprehensive evaluation score is lower than the team comprehensive evaluation score threshold, the corresponding team will be deleted from the team task acceptance profile sub-pool;

[0105] S117. If the personal task satisfaction score of a non-team member is lower than the personal task satisfaction score threshold, the corresponding individual will be deleted from the personal task acceptance portrait sub-pool. If the personal task satisfaction score of an individual in the team is lower than the personal task satisfaction score threshold, an alarm will be issued to the team as a whole.

[0106] The dynamic task acceptance portrait pool constructed by this process ensures the objectivity and accuracy of the evaluation results by comprehensively considering multi-dimensional information such as the professional skills proficiency, task completion history, etc. of team members and non-team individuals; secondly, by setting score thresholds to screen teams and individuals, entities with weak task acceptance capabilities are effectively eliminated, ensuring the scientificity and rationality of task allocation; thirdly, the team task acceptance portrait sub-pool is connected with the individual task acceptance portrait sub-pool to achieve seamless connection between teams and individuals, facilitating the rapid matching of the most suitable team or individual according to project requirements; finally, this process not only optimizes resource allocation, but also improves the level of refinement of team management, providing a strong guarantee for the smooth completion of tasks, thereby achieving a dual improvement in the efficiency and quality of task allocation.

[0107] S2. Obtain the attribute data of the published task, split the published task using the configured rule splitting model, and obtain the list of subtasks after splitting and the corresponding attribute data;

[0108] Furthermore, in this embodiment, the step of obtaining the split subtask list includes:

[0109] S201. Obtain overall information of the released task, and extract the task cycle information and overall required skill type information corresponding to the released task through the configured entity extraction algorithm;

[0110] Furthermore, in this embodiment, the entity extraction algorithm includes a Bert model pre-trained in Chinese and English;

[0111] S202: Set a period split threshold. If the obtained task period is greater than the period split threshold, input the overall information of the released task into the comprehensive fuzzy evaluation algorithm to evaluate the urgency score of each subtask in the overall task;

[0112] Furthermore, in this embodiment, each stage of the subtask is a stage division of the overall task based on Party A's progress requirements for a certain part of the overall task;

[0113] Furthermore, in this embodiment, the period splitting threshold is determined according to the completion time length of the task setting. For example, tasks with a period of more than one month are split, and tasks within one month are not split.

[0114] S203. Divide the overall task into an early subtask set, a mid-term subtask set, and a late subtask set based on the obtained urgency scores of the subtasks at each stage. Further, the urgency scores of the subtasks at each stage are divided according to the deadlines corresponding to each sub-function in the overall task. Sub-functions with shorter deadlines are placed in the early subtask set. For example, if a task has an overall cycle of two months and includes six functional subtasks, but two of the functional subtasks have a cycle length of only 15 days, one subtask has a cycle length of one month, and the other three have a cycle length of two months, then the 15-day subtasks are placed in the early subtask set, the one-month subtasks are placed in the mid-term subtask set, and the two-month subtasks are placed in the late subtask set.

[0115] S204: For the obtained early subtask set, mid-term subtask set, and late subtask set, using the configured skill recognition model, obtain the occupational skill type required for the early subtask set, the occupational skill type required for the mid-term subtask set, and the occupational skill type required for the late subtask set;

[0116] Furthermore, in this embodiment, the skill recognition model is constructed by the pre-trained RoBERTa model to identify the types of skills required for each subtask. For example, to build a data platform, the required skill types include: front-end, back-end, database construction, etc.

[0117] S205. Functionally split the early subtask set, the mid-term subtask set and the late subtask set according to the required professional skill types obtained to obtain the early functional subtask set, the mid-term functional subtask set and the late functional subtask set, and use the obtained early functional subtask set, the mid-term functional subtask set and the late functional subtask set to construct a subtask list.

[0118] Furthermore, in this embodiment, each functional subtask in the corresponding early functional subtask set, mid-term functional subtask set, and late functional subtask set in the subtask list is numbered in the order of logical implementation;

[0119] S206: If the acquired task period is less than or equal to the period splitting threshold, then directly repeat S204-S205 to functionally split the overall task to obtain a functional subtask set and a subtask list;

[0120] S207. If, through the configured skill recognition model, it is determined that there is only one type of professional skill required for the overall task, the task will not be split, and the corresponding task will be marked for individual matching.

[0121] This process extracts task cycle information and required skill type information through the configured entity extraction algorithm, ensuring the accuracy of basic task information; secondly, a cycle splitting threshold is set to split tasks with a task cycle of more than one month, and the urgency score of the subtasks in each stage is evaluated through an evaluation algorithm, dividing the tasks into early, mid-term and late subtask sets; this process determines the urgency of each subtask by considering its deadline, ensuring the rationality of task arrangement; thirdly, the professional skill type required for each subtask is identified through the configured skill recognition model, further refining the task splitting; finally, the subtask set is functionally split according to the required skill type to construct a subtask list; for tasks with a task cycle less than or equal to the cycle splitting threshold, functional splitting is directly performed, while for tasks with a single required skill type, no splitting is performed and individual matching labeling is directly performed. This process not only improves the accuracy of task splitting, but also ensures the rationality and efficiency of task allocation.

[0122] S3. Based on the dynamic task acceptance profile pool, the list of split subtasks, and the corresponding attribute data, the configured dual matching model is used to obtain the matching score between the team and the task, the matching score between team members and the corresponding subtasks, and the team's coverage of the task;

[0123] S4. Set matching rules. Match the corresponding teams according to the set matching rules and the obtained matching scores and coverage. If all teams do not meet the matching rules, match the split subtasks with individual individuals, and configure the upstream and downstream subtask interface attributes for the subtasks that have been matched with individuals. At the same time, continue to match the remaining subtasks with teams until all subtasks are assigned.

[0124] Furthermore, the steps of performing task matching in this embodiment include:

[0125] S301. Match the subtask list after the corresponding task is split with the team task acceptance profile subpool. Using matching algorithm 1, obtain a matching score 1 between each team as a whole and the overall task. Simultaneously, using matching algorithm 2, obtain a matching score 2 between each team member and the required professional skill type in the subtask list.

[0126] S302: Calculate the coverage of each team for the required professional skills in the subtask list based on the intersection of the professional skill types required in the subtask list and the professional skill types corresponding to the team members, and the professional skill types required in the subtask list;

[0127] S303: Set a match score 1 threshold, a match score 2 threshold, and a coverage threshold. If there is more than one team whose match score 1, match score 2, and coverage are all greater than the corresponding thresholds, match the task to the team with the highest initial task delivery success rate based on the initial task delivery success rate of the corresponding teams.

[0128] S304. If the initial task delivery success rate of all teams that meet the conditions in S303 is 0, the task will be matched to the team with the lowest task undertaking fee;

[0129] S305: If all teams do not meet the coverage threshold, then compare the matching score 1 and matching score 2 with the corresponding thresholds. If there are more than one team that meets both the matching score 1 threshold and the matching score 2 threshold, match the functional subtask set covered by the team to the team with the largest coverage.

[0130] S306: If there is more than one team with the largest coverage, S303 and S304 will be repeated to match teams and tasks.

[0131] S307. If there is more than one uncovered functional subtask after the matching process in S306, the remaining functional subtask is matched with the unmatched team that meets the conditions in S305. If there is a team whose matching score 1, matching score 2, and coverage are all greater than the corresponding threshold, the remaining functional subtask is assigned to the team that meets the conditions. At the same time, the interface attributes of the functional subtask assigned in S306 that is adjacent to the remaining functional subtask are configured to the remaining functional subtask assigned in S307.

[0132] Furthermore, in this embodiment, the interface attributes of the functional subtasks assigned by the S306 process are configured to the remaining functional subtasks assigned by S307. The interface attributes here are the early functional subtask set, the mid-term functional subtask set and the late functional subtask set obtained after the overall task is split, and each functional subtask is numbered in the order of logical implementation; here, the functional subtasks assigned to the corresponding team of S306 and the corresponding team of S307 and with adjacent numbers are configured with relevant information according to the information of the other functional subtask, so that even if they are assigned to two teams, they can be fully integrated in the later stage; for example, functional subtask 1 in the early functional subtask set is assigned to the corresponding team of S306, and functional subtask 2 is assigned to the corresponding team of S307. Before the assignment, the task information corresponding to functional subtask 1 is made into an information interface configured to functional subtask 2, and the task information corresponding to functional subtask 2 is made into an information interface configured to functional subtask 1, so that even if they are assigned to two different teams, when the final functional subtasks are integrated, they can be matched and integrated without distinction, and there will be no failure in the splicing of functional subtasks.

[0133] S308. If no team satisfies the matching score 1, matching score 2, and coverage greater than the corresponding thresholds, the remaining functional subtasks are matched with individuals in the personal task undertaking profile subpool. If there are more than one individual with a matching score greater than the matching score 2 threshold for the same remaining functional subtask, the corresponding remaining functional subtask is matched to the individual with the highest personal task satisfaction score until all remaining functional subtasks are matched.

[0134] S309: If the number of individuals with the highest individual task satisfaction scores is greater than 1, the task is matched to the individual with the highest corresponding skill proficiency assessment score, and the interface attributes of the functional subtasks assigned in S306 adjacent to the remaining functional subtasks are configured into the remaining functional subtasks assigned in S307;

[0135] S310: If the obtained task period is less than or equal to the period split threshold, repeat S206 and S301-S303. If there is a team that meets the conditions of S303 and the number of teams is greater than 1, match the task to the team with the smallest proportion of task completion time.

[0136] Furthermore, in step S310, the functional subtask set obtained by functional splitting that meets the conditions of S206 is allocated;

[0137] S311. If there is no team that meets the conditions of S303, repeat the process of S305-S309 to assign tasks;

[0138] S312. When there is only one type of professional skill required for the overall task, repeat S308-S309 to match individuals with tasks from the personal task acceptance portrait sub-pool, and match the task to the individual who meets the conditions in S308-S309.

[0139] This process ensures the quality and efficiency of task execution by breaking tasks into multiple subtasks and matching the specific professional skills required by the subtasks with the professional skills of the team or individuals. Secondly, through a dual matching model and coverage calculation, it not only considers the overall capabilities of the team and the matching of individual members' professional skills, but also takes into account the team's comprehensive coverage of the skills required for subtasks, thereby improving the reliability and success rate of task execution. Furthermore, this process establishes clear matching rules and thresholds, dynamically adjusting the matching strategy based on multiple dimensions such as the initial task delivery success rate, task acceptance costs, and skill coverage, thereby ensuring more reasonable task allocation. When a team cannot fully cover a subtask, the process also provides a mechanism to match remaining subtasks with individuals, ensuring that all tasks are handled promptly and effectively. Finally, by considering the task cycle, the time management of task allocation is further optimized, ensuring that resources are properly allocated within a tight timeframe.

[0140] S5. Obtain the matching team or individual task completion status data in real time, and evaluate the team or individual task completion status through the configured evaluation model to obtain the completion status evaluation score, and feed the obtained evaluation score back to the dynamic task acceptance portrait pool to adjust or delete the team or individual attribute status in the pool.

[0141] Example 2

[0142] See also Figure 2 , another embodiment provided by the present invention: a business intelligence system based on semantic big data analysis, comprising: a portrait pool construction module, a task splitting module, a matching module and a feedback module;

[0143] The portrait pool construction module is used to build a dynamic task-accepting portrait pool; the portrait pool construction module includes a data processing unit, a portrait unit, and a portrait pool unit;

[0144] The data processing unit is used to obtain the completion status data of historical release tasks and the corresponding attribute distribution status data of the undertaking teams and individuals, and perform standardized preprocessing on the obtained data;

[0145] The portrait unit constructs hierarchical team and individual portraits based on pre-processed team and individual attribute distribution data using a configured natural language algorithm. The portrait pool unit uses the constructed hierarchical team and individual portraits to construct a dynamic task acceptance portrait pool through a graph database.

[0146] The task splitting module is used to obtain the attribute data of the published task, split the published task through the configured rule splitting model, and obtain the differential subtask list and corresponding attribute data;

[0147] The matching module is used to match the published tasks with teams or individuals in the dynamic task acceptance profile pool; the matching module includes a team matching unit, a matching judgment unit, and a matching optimization unit;

[0148] The team matching unit uses the configured dual matching model based on the dynamic task acceptance profile pool, the list of split subtasks, and the corresponding attribute data to obtain the matching score between the team and the task, the matching score between team members and the corresponding subtasks, and the team's coverage of the task;

[0149] The matching judgment unit matches the corresponding team based on the obtained matching score and coverage by setting matching rules. If all teams do not meet the matching rules, the split subtasks are matched with a single person;

[0150] The matching optimization unit is used to configure the upstream and downstream subtask interface attributes for the subtasks that have been matched with individuals, and to continue matching the remaining subtasks with teams until all subtasks have been assigned;

[0151] The feedback module is used to match the team or individual task completion status for evaluation, and adjust or delete the team or individual attribute status in the dynamic task acceptance portrait pool based on the evaluation results.

[0152] Example 3

[0153] A computer-readable storage medium stores computer instructions, which, when executed, execute a semantic big data analysis method.

[0154] An electronic device includes a memory and a processor, wherein the memory stores a computer program, and the processor implements a semantic big data analysis method when executing the computer program.

[0155] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present invention, ordinary technicians in this field can also change, modify, replace and modify the above-mentioned embodiments without departing from the scope of protection of the purpose of the present invention and the claims, and all of these are protected by the present invention.

[0156] If the technical solution disclosed herein involves personal information, the product using the technical solution disclosed herein has clearly informed the individual of the personal information processing rules and obtained the individual's voluntary consent before processing the personal information. If the technical solution disclosed herein involves sensitive personal information, the product using the technical solution disclosed herein has obtained the individual's separate consent before processing the sensitive personal information and at the same time meets the "explicit consent" requirement. For example, on personal information collection devices such as cameras, a clear and prominent sign is set to inform that the individual has entered the personal information collection scope and that personal information will be collected. If the individual voluntarily enters the collection scope, it is deemed that they agree to the collection of their personal information; or on the personal information processing device, when the personal information processing rules are notified by obvious signs / information, the individual's authorization is obtained through pop-up information or by asking the individual to upload their personal information. The personal information processing rules may include information such as the personal information processor, the purpose of personal information processing, the processing method, and the type of personal information processed.

Claims

1. Based on the semantic big data analysis method, it is characterized by the following steps: include: S1. Obtain the completion status data of historically released tasks and the corresponding attribute distribution status data of the undertaking teams and individuals. Based on the acquired data, a dynamic task undertaking profile pool is obtained through the configured integration model. S2. Obtain the attribute data of the published task, split the published task using the configured rule splitting model, and obtain the list of subtasks after splitting and the corresponding attribute data; S3. Based on the dynamic task acceptance profile pool, the list of split subtasks, and the corresponding attribute data, the configured dual matching model is used to obtain the matching score between the team and the task, the matching score between team members and the corresponding subtasks, and the team's coverage of the task; S4. Set matching rules. Match the corresponding teams according to the set matching rules and the obtained matching scores and coverage. If all teams do not meet the matching rules, match the split subtasks with individual individuals, and configure the upstream and downstream subtask interface attributes for the subtasks that have been matched with individuals. At the same time, continue to match the remaining subtasks with teams until all subtasks are assigned. S5. Real-time acquisition of matching team or individual task completion status data, and evaluation of the team or individual task completion status using the configured evaluation model to obtain a completion status evaluation score. The obtained evaluation score is then fed back to the dynamic task acceptance profile pool to adjust or delete the team or individual attribute status in the pool. The dynamic task acceptance portrait pool includes a team task acceptance portrait sub-pool and a personal task acceptance portrait sub-pool; The steps of constructing the team task portrait sub-pool include: S101. Obtain information on the professional skill types and proficiency of team members, the number of complete or partial tasks undertaken per historical unit time, the maximum task load, the number of remaining tasks at the current moment, the types of tasks undertaken, the time percentage of task completion, the task undertaking fee, the initial delivery success rate of different types of tasks, the delivery success rate of different types of tasks, and textual information on satisfaction with delivered tasks; S102: Input the acquired text information into the configured entity extraction algorithm to sequentially obtain team member skill attribute keyword information and task attribute keyword information, and automatically annotate the acquired team member skill attribute keyword information and task attribute keyword information with score tags using the trained automatic annotation model; S103: Input the labeled team member skill attribute keyword information into the evaluation algorithm 1 to obtain the team's overall skill proficiency evaluation score and the corresponding skill proficiency evaluation score of each member; S104: Input the annotated task attribute keyword information into the evaluation algorithm 2 to obtain the team's task acceptance evaluation score and the team's task satisfaction score; S105. Obtain a comprehensive team assessment score based on the obtained team overall skill proficiency score, team task capability assessment score, and team task satisfaction score; S106. Input the obtained keyword information and corresponding evaluation scores of different teams into the graph database to construct a team task acceptance portrait sub-pool. Use the team name as the first-level node in the team task acceptance portrait sub-pool. Use the obtained comprehensive evaluation score of each team as the sorting label of the first-level node to sort the teams in descending order. S107: Use the corresponding team member's professional skill type and proficiency keyword information as a secondary node, and use the obtained team's overall skill proficiency score as a secondary node sorting label to sort the team in descending order of skill level; S108: Use the keyword information of the attribute of the historical tasks undertaken by the corresponding team as a third-level node, use the obtained evaluation score of the team's ability to undertake tasks as the first-level ranking label of the third-level node, and use the task satisfaction score as the second-level ranking label; The steps of constructing the personal task acceptance portrait sub-pool include: S109: Obtain the text information of the non-team individual in S101, and obtain the skill proficiency assessment score and personal task satisfaction score of the non-team individual through the processes S102-S104; S110: Input the acquired non-team individual skill proficiency into the graph database to construct a sub-pool of individual task acceptance portraits. Use the corresponding individual name keywords as first-level nodes in the sub-pool of individual task acceptance portraits, and use the acquired individual skill proficiency assessment scores as first-level ranking labels. S111. Map the occupational type and proficiency attribute information of the team members in the team task acceptance portrait sub-pool to the first-level nodes in the individual task acceptance portrait sub-pool, and use the obtained skill proficiency assessment scores of the individual members as the corresponding first-level ranking labels; S112. Arrange the personal information in the first-level nodes of the personal task acceptance profile sub-pool in descending order based on the obtained first-level ranking labels of personal skill proficiency. Simultaneously, use the task attribute information of non-team individuals as second-level nodes, and use the corresponding personal task satisfaction scores as ranking labels within the second-level nodes. S113. Map the team task satisfaction score to the team member's corresponding individual task satisfaction score as a ranking label in the second-level node of the team member's individual task acceptance profile sub-pool; S114. Sort the task attribute information of team members and non-team members according to the obtained sorting labels in the secondary nodes to obtain a constructed sub-pool of individual task acceptance profiles; The step of obtaining the split subtask list in S2 includes: S201: Set a period splitting threshold, obtain overall information of the released task, and extract the task period information and overall required skill type information corresponding to the released task through the configured entity extraction algorithm; The steps of performing task matching in S3 and S4 include: S301. If the obtained task cycle is greater than the cycle split threshold, the subtask list after the corresponding task split is matched with the team task undertaking profile subpool. Using matching algorithm 1, a matching score 1 is obtained between each team as a whole and the overall task. At the same time, using matching algorithm 2, a matching score 2 is obtained between each team member and the required professional skill type in the subtask list. S302: Calculate the coverage of each team for the required professional skills in the subtask list based on the intersection of the professional skill types required in the subtask list and the professional skill types corresponding to the team members, and the professional skill types required in the subtask list; S303: Set a match score 1 threshold, a match score 2 threshold, and a coverage threshold. If there is more than one team whose match score 1, match score 2, and coverage are all greater than the corresponding thresholds, match the task to the team with the highest initial task delivery success rate based on the initial task delivery success rate of the corresponding teams. S304. If the initial task delivery success rate of all teams that meet the conditions in S303 is 0, the task will be matched to the team with the lowest task undertaking fee; S305: If all teams do not meet the coverage threshold, then compare the matching score 1 and matching score 2 with the corresponding thresholds. If there are more than one team that meets both the matching score 1 threshold and the matching score 2 threshold, match the functional subtask set covered by the team to the team with the largest coverage. S306: If there is more than one team with the largest coverage, repeat S303 and S304 to match teams and tasks.

2. The semantic big data analysis method according to claim 1, characterized in that: The steps of constructing the dynamic task acceptance portrait pool include: S115. Based on the mapping relationship between team members and personal task acceptance portrait sub-pools in S113, the personal task acceptance portrait sub-pool and the team task acceptance portrait sub-pool are connected to obtain a constructed dynamic task acceptance portrait pool; S116. Set a team comprehensive evaluation score threshold and an individual task satisfaction score threshold, and embed the team comprehensive evaluation score threshold into the team task acceptance profile sub-pool, and embed the individual task satisfaction score threshold into the individual task acceptance profile sub-pool. When the corresponding team comprehensive evaluation score is lower than the team comprehensive evaluation score threshold, the corresponding team will be deleted from the team task acceptance profile sub-pool; S117. If the personal task satisfaction score of a non-team member is lower than the personal task satisfaction score threshold, the corresponding individual will be deleted from the personal task acceptance portrait sub-pool. If the personal task satisfaction score of an individual in the team is lower than the personal task satisfaction score threshold, an alarm will be issued to the team as a whole.

3. The semantic big data analysis method according to claim 2, characterized in that: The step of obtaining the split subtask list in S2 further includes: S202: If the obtained task period is greater than the period splitting threshold, the overall information of the released task is input into the comprehensive fuzzy evaluation algorithm to evaluate the urgency score of each subtask in the overall task; S203, dividing the overall task into an early stage subtask set, a mid-term subtask set, and a late stage subtask set based on the obtained urgency scores of the subtasks at each stage; S204: For the obtained early subtask set, mid-term subtask set, and late subtask set, using the configured skill recognition model, obtain the occupational skill type required for the early subtask set, the occupational skill type required for the mid-term subtask set, and the occupational skill type required for the late subtask set; S205. Functionally split the early subtask set, the mid-term subtask set and the late subtask set according to the required professional skill types obtained to obtain the early functional subtask set, the mid-term functional subtask set and the late functional subtask set, and use the obtained early functional subtask set, the mid-term functional subtask set and the late functional subtask set to construct a subtask list.

4. The semantic big data analysis method according to claim 3, characterized in that: The step of obtaining the split subtask list in S2 further includes: S206: If the acquired task period is less than or equal to the period splitting threshold, then directly repeat S204-S205 to functionally split the overall task to obtain a functional subtask set and a subtask list; S207. If, through the configured skill recognition model, it is determined that there is only one type of professional skill required for the overall task, the task will not be split, and the corresponding task will be marked for individual matching.

5. The semantic big data analysis method according to claim 4, characterized in that: The steps of performing task matching in S3 and S4 further include: S307. If there is more than one uncovered functional subtask after the matching process S306, the remaining functional subtask is matched with the unmatched team that meets the conditions of S305. If there is a team whose matching score 1, matching score 2, and coverage are all greater than the corresponding threshold, the remaining functional subtask is assigned to the team that meets the conditions. At the same time, the interface attributes of the functional subtask assigned in the S306 process adjacent to the remaining functional subtask are configured to the remaining functional subtask assigned in S307. S308. If no team satisfies the matching score 1, matching score 2, and coverage greater than the corresponding thresholds, the remaining functional subtasks are matched with individuals in the personal task undertaking profile subpool. If there are more than one individual with a matching score greater than the matching score 2 threshold for the same remaining functional subtask, the corresponding remaining functional subtask is matched to the individual with the highest personal task satisfaction score until all remaining functional subtasks are matched. S309: If the number of individuals with the highest individual task satisfaction scores is greater than 1, the task is matched to the individual with the highest corresponding skill proficiency assessment score, and the interface attributes of the functional subtasks assigned in S306 adjacent to the remaining functional subtasks are configured into the remaining functional subtasks assigned in S307; S310: If the obtained task period is less than or equal to the period split threshold, repeat S206 and S301-S303. If there is a team that meets the conditions of S303 and the number of teams is greater than 1, match the task to the team with the smallest proportion of task completion time. S311. If there is no team that meets the conditions of S303, repeat the process of S305-S309 to assign tasks; S312. When there is only one type of professional skill required for the overall task, repeat S308-S309 to match individuals with tasks from the personal task acceptance portrait sub-pool, and match the task to the individual who meets the conditions in S308-S309.

6. A business intelligence system based on semantic big data analysis, which is used to implement the semantic big data analysis method according to any one of claims 1 to 5, characterized in that: include: Portrait pool construction module, task splitting module and matching module; The portrait pool construction module includes a data processing unit, a portrait unit and a portrait pool unit; The data processing unit is used to obtain the completion status data of historically released tasks and the corresponding attribute distribution status data of the undertaking teams and individuals, and perform standardized preprocessing on the obtained data; The portrait unit constructs hierarchical team portraits and individual portraits based on the pre-processed team and individual attribute distribution status data through a configured natural language algorithm; the portrait pool unit uses the constructed hierarchical team portraits and individual portraits to construct a dynamic task acceptance portrait pool through a graph database; The task splitting module is used to obtain the attribute data of the published task, split the published task according to the configured rule splitting model, and obtain the differential subtask list and corresponding attribute data; The matching module includes a team matching unit, a matching judgment unit and a matching optimization unit; The team matching unit obtains the matching score between the team and the task, the matching score between the team members and the corresponding subtasks, and the team's coverage of the task through the configured dual matching model based on the dynamic task acceptance portrait pool, the split subtask list, and the corresponding attribute data; The matching judgment unit matches the corresponding teams by setting matching rules according to the obtained matching scores and coverage. If all teams do not meet the matching rules, the split subtasks are matched with a single person; The matching optimization unit is used to configure the upstream and downstream subtask interface attributes for the subtasks that have been matched with individuals, and continue to match the remaining subtasks with teams until all subtasks are assigned.

7. A computer-readable storage medium, characterized in that Computer instructions are stored thereon, and when the computer instructions are executed, the semantic big data analysis method described in any one of claims 1 to 5 is executed.

Images