Team construction optimization system based on bidirectional semantic matching

Abstract

The invention discloses a team construction optimization system based on bidirectional semantic matching, and belongs to the technical field of bidirectional semantic matching. Through member authorization, natural language expression parameters in the historical team cooperation process are extracted, a member expression style mark set is constructed, and a fluctuation intensity index is calculated in combination with a cooperation score sequence; fusing member character tags and language behavior features to generate a personalized expression style stable set, performing semantic feature extraction on a task text, constructing a multi-dimensional semantic demand set, and judging the matching degree of members and tasks based on weighted semantic matching scores; the initial team is constructed through the preset matching threshold value, the expression behaviors are monitored in real time in the operation process, the team structure and the member composition are dynamically adjusted, high-precision and dynamically-adaptive semantic collaborative matching between the members and the tasks is achieved, the team construction efficiency and the collaborative quality are remarkably improved, and the method has wide application prospects.

Description

Technical Field

[0001] The present invention relates to the technical field of bidirectional semantic matching, and in particular to a team building optimization system based on bidirectional semantic matching. Background Art

[0002] In recent years, with the advancement of natural language processing, personalized modeling, and human-computer interaction technologies, some research has begun to explore supporting automated team building through vector matching between member skill profiles and task skill requirements. These approaches primarily focus on matching and balancing skill dimensions, with limited consideration of the impact of soft factors such as communication style and emotional stability on team collaboration performance. However, in real-world collaboration scenarios, non-structural characteristics such as differences in expression, emotional fluctuations, and language stability among team members significantly impact team efficiency, member synergy, and even the quality of collaboration. These factors remain largely ignored or undervalued in existing semantic matching team building methods.

[0003] Existing technologies struggle to effectively integrate multi-dimensional soft parameters such as members' expression style stability, personality consistency, and task semantic style requirements for comprehensive evaluation. On the one hand, most existing methods employ a one-way matching logic, statically comparing members' skill tags or language features based on task requirements. This approach lacks feedback from the member's perspective on how their style stability and expression consistency impact task suitability. On the other hand, existing systems lack quantitative modeling and structural integration mechanisms for sensitivity to fluctuations in expression style. This can lead to members performing well in the early stages of long-term team operations, but later experiencing emotional instability or expression deviations, ultimately impacting team stability and task completion quality. Furthermore, existing technologies lack sufficient consideration of dynamic evaluation and adjustment mechanisms during team operations, making it difficult to issue timely structural warnings or intervene when team status fluctuates. Summary of the Invention

[0004] The purpose of the present invention is to provide a team building optimization system based on bidirectional semantic matching to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: A team building optimization system based on bidirectional semantic matching, which includes: data acquisition and set building module, stability factor and index calculation module, set building and matching score calculation module and analysis and early warning module; The data acquisition and collection construction module: after authorization by the member, obtains the member's expression parameters and skill keywords in the historical team collaboration process; constructs the member's expression style tag set; and retrieves the member's team collaboration score sequence; The stability factor and index calculation module: obtains the member's fluctuation intensity index based on the team collaboration score sequence; calculates the member's expression style stability factor based on the member's expression style tag set; calculates the member's semantic shift sensitivity index based on the member's expression style stability factor and fluctuation intensity index; The set construction and matching score calculation module: obtains the personality label of the member and constructs a stable set of personalized expression styles; obtains the task description text of the current task and constructs a multidimensional semantic requirement set; calculates the matching score between the stable set of personalized expression styles of the member and the multidimensional semantic requirement set; The analysis and warning module sets the rules for including task team members and builds a task team based on the matching scores. During the team operation process, it determines in real time whether the members meet the rules for including task team members. If not, it issues a warning and reminds relevant staff to intervene and make adjustments.

[0006] Furthermore, the data acquisition and set construction module includes a data acquisition and set construction unit; The data acquisition and set construction unit: after authorization by the members, obtains the natural language communication records and skill keywords of the members in the historical team collaboration process from the historical team collaboration archive system, uses natural language processing technology to extract the expression parameters of the members in the natural language communication records and normalizes them, the expression parameters including the proportion of subjective words and the density of emotional tags; based on the expression parameters, constructs the member's expression style tag set; and retrieves the team collaboration score sequence of the members in the historical team collaboration process from the team evaluation system background.

[0007] Furthermore, the stability factor and index calculation module includes a stability factor calculation unit and an index calculation unit; The stability factor calculation unit: based on the team collaboration score sequence, obtains the team collaboration scores of the members at two adjacent task time points, and calculates the change rate of the team collaboration scores of the members at the two adjacent task time points using the difference method; traverses the task time points to obtain the maximum change rate of the member's team collaboration score, which is recorded as the fluctuation intensity index; and calculates the member's expression style stability factor based on the member's expression style tag set; The index calculation unit calculates the semantic shift sensitivity index of the member based on the expression style stability factor and fluctuation intensity index of the member.

[0008] Furthermore, the set construction and matching score calculation module includes a set construction unit and a matching score calculation unit; The set construction unit: uses MBTI personality test technology to obtain the personality label of the member, integrates the personality label with skill keywords, expression style tag set, expression style stability factor and semantic shift sensitivity index to construct a personalized expression style stability set; obtains the task description text of the current task, extracts subsets of the task description text based on the domain dictionary and word embedding model, and constructs a multi-dimensional semantic requirement set; The matching score calculation unit calculates the matching score between the member's personalized expression style stable set and the multi-dimensional semantic requirement set.

[0009] Furthermore, the analysis and warning module includes an analysis and warning unit; The analysis and early warning unit: sets the task team member inclusion rules, specifically as follows: presets a matching score threshold, and if the matching score between the personalized expression style stable set of the i-th member and the multidimensional semantic requirement set is greater than or equal to the matching score threshold, then determines that the i-th member meets the task team member inclusion rules; obtains all members who meet the task team member inclusion rules, and builds a task team; during the team operation process, obtains the subjective word proportion and emotional tag density of the team members at each task time point in real time, calculates the expression style stability factor of the team members in real time, and updates the personalized expression style stable set of the team members, recalculates the matching score between the personalized expression style stable set of the team members and the multidimensional semantic requirement set, and determines whether the task team member inclusion rules are met; if it is determined that the team member does not meet the task team member inclusion rules, issues a warning and reminds relevant staff to intervene and adjust.

[0010] A team building optimization method based on two-way semantic matching includes the following steps: step S1: after authorization by members, obtaining the expression parameters and skill keywords of members in the historical team collaboration process; constructing the expression style tag set of members; retrieving the team collaboration score sequence of members; step S2: based on the team collaboration score sequence, obtaining the fluctuation intensity index of members; based on the expression style tag set of members, calculating the expression style stability factor of members; based on the expression style stability factor and the fluctuation intensity index of members, calculating the semantic shift sensitivity index of members; step S3: obtaining the personality label of members, constructing a personalized expression style stability set; obtaining the task description text of the current task, constructing a multidimensional semantic requirement set; calculating the matching score between the personalized expression style stability set of members and the multidimensional semantic requirement set; step S4: setting the task team member inclusion rules, combining the matching scores to construct the task team; during the team operation process, judging in real time whether members meet the task team member inclusion rules, if not, issuing an early warning and reminding relevant staff to intervene and adjust.

[0011] As a preferred solution of the team building optimization method based on two-way semantic matching described in the present invention, after authorization by the members, the natural language communication records and skill keywords of the members in the historical team collaboration process are obtained from the historical team collaboration archive system. Using natural language processing technology, the expression parameters of the members in the natural language communication records are extracted and normalized. The expression parameters include the proportion of subjective words (referring to the proportion of words with subjective attitudes (such as "feel", "think", "should", "hate", "surprise", etc.) used by a person in the expression in the overall words) and the density of emotional markers (referring to the density of words carrying emotional polarity (such as "anger", "happy", "worry", "bad", etc.) per unit text length, which is used to measure the degree of emotional exposure in the sentence); based on the expression parameters, the expression style marker set of the i-th member is constructed, which is recorded as ,in, Indicates the proportion of subjective words of the i-th member, represents the emotional label density of the i-th member; The team collaboration score sequence of members in the historical team collaboration process is retrieved from the team evaluation system background, recorded as ,in, represents the team collaboration score of the i-th member at the m-th task time point in the historical team collaboration process, and M represents the total number of task time points of the i-th member in the historical team collaboration process.

[0012] It should be noted that by normalizing the proportion of subjective words and the density of emotional markers and constructing an expression style marker set, it is possible to quantify the modeling of individual language style, rather than the traditional modeling method that relies solely on content; the introduction of a team collaboration scoring sequence provides dynamic behavioral basic data for the subsequent evaluation of members' behavioral stability and emotional changes, making up for the existing system's neglect of "collaborative performance fluctuations", and ultimately forming a mixed file of "semantic style + behavioral feedback" for each member, providing a reliable and systematic data foundation for subsequent accurate and structured person-job matching.

[0013] As a preferred solution of the team building optimization method based on bidirectional semantic matching described in the present invention, based on the team collaboration scoring sequence , obtain the teamwork score of the i-th member at two adjacent task time points, use the difference method to calculate the change rate of the teamwork score of the i-th member at two adjacent task time points; traverse the task time points and obtain the maximum change rate of the teamwork score of the i-th member, which is recorded as the fluctuation intensity index ; Based on the expression style tag set of the i-th member, the expression style stability factor of the i-th member is calculated. The calculation formula is: ,in, and Respectively represent the weight factors of the preset subjective word ratio and emotional tag density, represents the expression style stability factor of the i-th member; Based on the expression style stability factor of the i-th member and the Volatility Strength Index , calculate the semantic shift sensitivity index of the i-th member, the calculation formula is: ,in, and Represents the preset expression style stability factors respectively and the Volatility Strength Index The scaling factor.

[0014] It should be noted that by performing local differential operations on the scoring sequence to quantify the intensity of fluctuations in collaborative performance and introducing a dynamic behavioral stability indicator, this addresses the pain point of traditional models where static features cannot reflect psychological and collaborative fluctuations. Furthermore, by constructing an expression style stability factor and introducing a structural weight of expression style into semantic modeling, this adds stability constraints to semantic similarity determination, effectively reducing the risk of "false matches" caused by fluctuations in language style with emotions. The semantic shift sensitivity index, as a fusion indicator of the two features, can be used to adjust the weight of the member's semantic representativeness during actual task matching, and establish a direct connection between the credibility of the member's language performance and the matching accuracy; this mechanism enables the system to no longer just judge whether the members are "semantically matched", but further identify whether their "semantic stability is sufficient to support the team structure", substantially improving the robustness and predictive ability of team formation.

[0015] As a preferred solution of the team building optimization method based on bidirectional semantic matching described in the present invention, the MBTI personality test technology is used to obtain the personality label of the i-th member, and the personality label is combined with the skill keywords and expression style tag set. , expression style stability factor and semantic shift sensitivity index Fusion is performed to construct a stable set of personalized expression styles, as follows: Based on the expression style stability factor and semantic shift sensitivity index , calculate the control weight factor of the i-th member, the calculation formula is as follows: ; in, represents the control weight factor of the i-th member, Indicates the preset adjustment coefficient; The control weight factor based on the i-th member , construct a stable set of personalized expression styles, recorded as ,in, represents the personality label of the i-th member, Indicates the skill keyword of the i-th member; Obtain the task description text of the current task, perform subset extraction on the task description text based on the domain dictionary and word embedding model, and construct a multidimensional semantic requirement set, as follows: Using natural language processing technology, we extract personality requirement labels, skill requirement keywords, subjective word ratios, and emotional harmony semantics from the task description text, and construct a multidimensional semantic requirement set, which is recorded as ,in, Represents the personality requirement label in the task description text, Indicates the skill requirement keywords in the task description text, Indicates the proportion of subjective words in the task description text, Representing the emotional attunement semantics in the task description text; Calculate the matching score between the stable set of personalized expression styles of the i-th member and the multidimensional semantic requirement set as follows: ; ; ; in, represents the matching score between the stable set of personalized expression style of the i-th member and the multidimensional semantic requirement set, Indicates personality matching, if , then it is recorded as 1, otherwise it is 0. Indicates the matching degree of subjective word proportion, Indicates emotional matching. 、 、 and They respectively represent the influence weights of the preset personality matching, skill keyword matching, subjective word proportion matching, and emotion matching.

[0016] It should be noted that by introducing member personality labels through the MBTI model and integrating them with language style and stability indicators, we can achieve true "psychological-linguistic dual-layer semantic modeling" in team building, solving the problem of traditional matching that only focuses on knowledge structure but ignores personality style. The multidimensional semantic requirement set is no longer limited to skill keywords, but also includes subjective expression tendencies and emotional density, which improves the perception of the task text in the dimension of expression style, making the task itself have "semantic style characteristics"; Introducing the expression style and emotional adaptability matching dimension into the matching function not only improves team language collaboration but also effectively estimates members' adaptability potential in terms of emotional or expression style deviations. By incorporating the style dimension into the semantic matching algorithm, an interpretable, dynamic, and controllable member-task semantic fit mechanism is implemented. This design is significantly different from existing matching models that are only based on content or keywords.

[0017] As a preferred solution of the team building optimization method based on bidirectional semantic matching described in the present invention, the rules for including task team members are set as follows: A matching score threshold is preset. If the matching score between the stable set of personalized expression styles of the i-th member and the multidimensional semantic requirement set is greater than or equal to the matching score threshold, the i-th member is determined to meet the task team member inclusion rule; all members who meet the task team member inclusion rule are obtained, and a task team is constructed; During the team operation process, the subjective word ratio and emotional tag density of team members at each task time point are obtained in real time, the expression style stability factor of team members is calculated in real time, and the personalized expression style stability set of team members is updated. The matching score between the personalized expression style stability set of team members and the multidimensional semantic requirement set is recalculated, and it is determined whether the task team member inclusion rules are met; If it is determined that a team member does not meet the rules for inclusion in the task team, an early warning will be issued and relevant staff will be reminded to intervene and make adjustments.

[0018] It should be noted that the introduction of the matching score threshold as an adjustable inclusion rule controller gives team building explicit structural constraints, avoiding the uncontrollable problem of "full matching without screening"; by collecting expression style parameters in real time and dynamically updating the stable set of expression styles and matching scores, a continuous feedback mechanism for the degree of fit between team member performance and task is implemented, which is predictable and controllable; An early warning mechanism and intervention interface are set up to issue timely signals when it is detected that members deviate from the task expression style requirements or the style stability decreases, significantly improving the team collaboration quality control capabilities; a style stability-driven team maintenance model is proposed to provide a systematic response strategy for individual dynamic fluctuations in complex tasks, with a clear engineering implementation path.

[0019] Compared with the existing technology, the beneficial effects achieved by the present invention are as follows: in the team building optimization system based on bidirectional semantic matching provided by the present invention, the subjective word ratio and emotional tag density in their historical collaboration are extracted through member authorization, and an expression style tag set is constructed. Combined with the team collaboration score sequence, a data archive with language style characteristics and behavioral performance basis is formed; then, the member's fluctuation intensity index is calculated by score difference and integrated with the expression style stability factor to form a semantic shift sensitivity index, thereby characterizing the dynamic stability of the member's language style and enhancing the credibility and anti-interference ability of semantic matching; further, a personalized expression style stability set is constructed by combining personality labels, and subjective expression tendencies and emotional harmony requirements are extracted from task texts to construct a multidimensional semantic requirement set. Style-task bidirectional alignment is achieved through a weighted semantic matching mechanism; finally, a member inclusion rule threshold is set, the expression style changes of team members are monitored in real time, the matching degree is dynamically updated and early warning intervention is implemented, effectively ensuring team language collaboration and expression stability. This method breaks through the limitations of traditional team matching that only relies on skills or static semantic similarity, significantly improves matching accuracy, team stability and dynamic adaptability, and is suitable for complex task team building with high requirements for collaboration style and emotional consistency. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0021] Figure 1 It is a structural diagram of the team building optimization system based on bidirectional semantic matching of the present invention; Figure 2 It is a schematic diagram of the steps of the team building optimization method based on two-way semantic matching of the present invention. DETAILED DESCRIPTION

[0022] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0023] See also Figure 1 ,In this embodiment 1: a team building optimization system based on bidirectional semantic matching is provided, the system including: a data acquisition and set building module, a stability factor and index calculation module, a set building and matching score calculation module and an analysis and early warning module; The data acquisition and collection construction module: after authorization by the member, obtains the member's expression parameters and skill keywords in the historical team collaboration process; constructs the member's expression style tag set; and retrieves the member's team collaboration score sequence; The stability factor and index calculation module: obtains the member's fluctuation intensity index based on the team collaboration score sequence; calculates the member's expression style stability factor based on the member's expression style tag set; calculates the member's semantic shift sensitivity index based on the member's expression style stability factor and fluctuation intensity index; The set construction and matching score calculation module: obtains the personality label of the member and constructs a stable set of personalized expression styles; obtains the task description text of the current task and constructs a multidimensional semantic requirement set; calculates the matching score between the stable set of personalized expression styles of the member and the multidimensional semantic requirement set; The analysis and warning module sets the rules for including task team members and builds a task team based on the matching scores. During the team operation process, it determines in real time whether the members meet the rules for including task team members. If not, it issues a warning and reminds relevant staff to intervene and make adjustments.

[0024] Furthermore, the data acquisition and set construction module includes a data acquisition and set construction unit; The data acquisition and set construction unit: after authorization by the members, obtains the natural language communication records and skill keywords of the members in the historical team collaboration process from the historical team collaboration archive system, uses natural language processing technology to extract the expression parameters of the members in the natural language communication records and normalizes them, the expression parameters including the proportion of subjective words and the density of emotional tags; based on the expression parameters, constructs the member's expression style tag set; and retrieves the team collaboration score sequence of the members in the historical team collaboration process from the team evaluation system background.

[0025] Furthermore, the stability factor and index calculation module includes a stability factor calculation unit and an index calculation unit; The stability factor calculation unit: based on the team collaboration score sequence, obtains the team collaboration scores of the members at two adjacent task time points, and calculates the change rate of the team collaboration scores of the members at the two adjacent task time points using the difference method; traverses the task time points to obtain the maximum change rate of the member's team collaboration score, which is recorded as the fluctuation intensity index; and calculates the member's expression style stability factor based on the member's expression style tag set; The index calculation unit calculates the semantic shift sensitivity index of the member based on the expression style stability factor and fluctuation intensity index of the member.

[0026] Furthermore, the set construction and matching score calculation module includes a set construction unit and a matching score calculation unit; The set construction unit: uses MBTI personality test technology to obtain the personality label of the member, integrates the personality label with skill keywords, expression style tag set, expression style stability factor and semantic shift sensitivity index to construct a personalized expression style stability set; obtains the task description text of the current task, extracts subsets of the task description text based on the domain dictionary and word embedding model, and constructs a multi-dimensional semantic requirement set; The matching score calculation unit calculates the matching score between the member's personalized expression style stable set and the multi-dimensional semantic requirement set.

[0027] Furthermore, the analysis and warning module includes an analysis and warning unit; The analysis and early warning unit: sets the task team member inclusion rules, specifically as follows: presets a matching score threshold, and if the matching score between the personalized expression style stable set of the i-th member and the multidimensional semantic requirement set is greater than or equal to the matching score threshold, then determines that the i-th member meets the task team member inclusion rules; obtains all members who meet the task team member inclusion rules, and builds a task team; during the team operation process, obtains the subjective word proportion and emotional tag density of the team members at each task time point in real time, calculates the expression style stability factor of the team members in real time, and updates the personalized expression style stable set of the team members, recalculates the matching score between the personalized expression style stable set of the team members and the multidimensional semantic requirement set, and determines whether the task team member inclusion rules are met; if it is determined that the team member does not meet the task team member inclusion rules, issues a warning and reminds relevant staff to intervene and adjust.

[0028] See also Figure 2 In the second embodiment, a team building optimization method based on bidirectional semantic matching is provided, which includes the following steps: Step S1: After authorization by the member, obtain the member's expression parameters and skill keywords in the historical team collaboration process; build the member's expression style tag set; and retrieve the member's team collaboration score sequence.

[0029] Specifically, after authorization by the members, the natural language communication records and skill keywords of the members in the historical team collaboration process are obtained from the historical team collaboration archive system. Natural language processing technology is used to extract the expression parameters of the members in the natural language communication records and normalize them. The expression parameters include the proportion of subjective words (referring to the proportion of words with subjective attitudes (such as "feel", "think", "should", "hate", "surprise", etc.) used by a person in the expression in the overall words) and the density of emotional markers (referring to the density of words with emotional polarity (such as "anger", "happy", "worry", "bad", etc.) per unit text length, which is used to measure the degree of emotional exposure in the sentence); based on the expression parameters, the expression style marker set of the i-th member is constructed, which is recorded as ,in, Indicates the proportion of subjective words of the i-th member, represents the emotional label density of the i-th member; The team collaboration score sequence of members in the historical team collaboration process is retrieved from the team evaluation system background, recorded as ,in, represents the team collaboration score of the i-th member at the m-th task time point in the historical team collaboration process, and M represents the total number of task time points of the i-th member in the historical team collaboration process.

[0030] In the present invention, the subjective expression tendency (the proportion of subjective words) ) and emotional exposure (emotional labeling density ) into specific values, solving the problem of language style being difficult to quantify. For example, by counting the proportion of subjective words such as "I think" and "must" in members' historical communications, or the density of emotional words such as "happy" and "anxious", the expression style can be transformed from a vague description into a comparable quantitative indicator.

[0031] Step S2: Based on the team collaboration scoring sequence, obtain the member's fluctuation intensity index; based on the member's expression style tag set, calculate the member's expression style stability factor; based on the member's expression style stability factor and fluctuation intensity index, calculate the member's semantic shift sensitivity index.

[0032] Specifically, based on the team collaboration scoring sequence , obtain the teamwork score of the i-th member at two adjacent task time points, use the difference method to calculate the change rate of the teamwork score of the i-th member at two adjacent task time points; traverse the task time points and obtain the maximum change rate of the teamwork score of the i-th member, which is recorded as the fluctuation intensity index ; Based on the expression style tag set of the i-th member, the expression style stability factor of the i-th member is calculated. The calculation formula is: ,in, and Respectively represent the weight factors of the preset subjective word ratio and emotional tag density, represents the expression style stability factor of the i-th member; Based on the expression style stability factor of the i-th member and the Volatility Strength Index , calculate the semantic shift sensitivity index of the i-th member, the calculation formula is: ,in, and Represents the preset expression style stability factors respectively and the Volatility Strength Index The scaling factor.

[0033] It should be noted that people with stable expression styles tend to maintain consistency in semantic expression; people with high volatility in behavioral data may still experience deviations due to external events or psychological states even if their expression styles are stable. and the Volatility Strength Index , quantifying the sensitivity of members to semantic deviations (such as sudden changes in communication style and decreased efficiency) in long-term collaboration. This design transforms stability from a qualitative judgment (such as "this person is unstable") to a quantitative assessment (such as The higher the value, the higher the risk of deviation), providing an explainable basis for the long-term adaptation of the team.

[0034] Step S3: Obtain the personality label of the member and construct a stable set of personalized expression styles; obtain the task description text of the current task and construct a multidimensional semantic requirement set; calculate the matching score between the stable set of personalized expression styles of the member and the multidimensional semantic requirement set.

[0035] Specifically, the MBTI personality test technology is used to obtain the personality label of the i-th member, and the personality label is combined with the skill keywords and expression style tag set. , expression style stability factor and semantic shift sensitivity index Fusion is performed to construct a stable set of personalized expression styles, as follows: Based on the expression style stability factor and semantic shift sensitivity index , calculate the control weight factor of the i-th member, the calculation formula is as follows: ; in, represents the control weight factor of the i-th member, Indicates the preset adjustment coefficient; The control weight factor based on the i-th member , construct a stable set of personalized expression styles, recorded as ,in, represents the personality label of the i-th member, Indicates the skill keyword of the i-th member; Obtain the task description text of the current task, perform subset extraction on the task description text based on the domain dictionary and word embedding model, and construct a multidimensional semantic requirement set, as follows: Using natural language processing technology, we extract personality requirement labels, skill requirement keywords, subjective word ratios, and emotional harmony semantics from the task description text, and construct a multidimensional semantic requirement set, which is recorded as ,in, Represents the personality requirement label in the task description text, Indicates the skill requirement keywords in the task description text, Indicates the proportion of subjective words in the task description text, Representing the emotional attunement semantics in the task description text; Calculate the matching score between the stable set of personalized expression styles of the i-th member and the multidimensional semantic requirement set as follows: ; ; ; in, represents the matching score between the stable set of personalized expression style of the i-th member and the multidimensional semantic requirement set, Indicates personality matching, if , then it is recorded as 1, otherwise it is 0. Indicates the matching degree of subjective word proportion, Indicates emotional matching. 、 、 and They represent the influence weights of the preset personality matching, skill keyword matching, subjective word ratio matching, and emotion matching, respectively. It represents the skill keyword matching degree, which is the ratio of member skills covering task requirements. For example, if the task requires {programming, communication} and the member has {programming, leadership}, then the matching degree = 1 / 2 = 0.5.

[0036] Step S4: Set the rules for inclusion of task team members, and build a task team based on the matching scores; during the team operation, determine in real time whether the members meet the rules for inclusion of task team members. If not, issue an early warning and remind relevant staff to intervene and make adjustments.

[0037] Specifically, set the rules for including task team members as follows: A matching score threshold is preset. If the matching score between the stable set of personalized expression styles of the i-th member and the multidimensional semantic requirement set is greater than or equal to the matching score threshold, the i-th member is determined to meet the task team member inclusion rule; all members who meet the task team member inclusion rule are obtained, and a task team is constructed; During the team operation process, the subjective word ratio and emotional tag density of team members at each task time point are obtained in real time, the expression style stability factor of team members is calculated in real time, and the personalized expression style stability set of team members is updated. The matching score between the personalized expression style stability set of team members and the multidimensional semantic requirement set is recalculated, and it is determined whether the task team member inclusion rules are met; If it is determined that a team member does not meet the rules for inclusion in the task team, an early warning will be issued and relevant staff will be reminded to intervene and make adjustments.

[0038] Please refer to Table 1-Table 2. In this third embodiment, a team building optimization method based on bidirectional semantic matching is provided. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through simulation experiments.

[0039] For example, a technology company needs to form a "core system development team." Using natural language processing technology, they extract task description text. The personality requirement tag is: ISTJ, the skill requirement keywords are: Python, project management, the subjective word ratio is: low subjective expression (subjective word ratio ≤ 0.2), and the emotional tone semantics is: neutral emotion (emotion tag density ≤ 0.15); There are currently three candidate members (A, B, and C), and the authorized historical data is as follows: Table 1 Historical data ; For member A, the maximum change rate of teamwork score, i.e. the fluctuation intensity index ; For member B, the maximum change rate of teamwork score, i.e. the fluctuation intensity index ; For member C, the maximum change rate of teamwork score, i.e. the fluctuation intensity index ; Assume the proportion of subjective words , the weight factor of sentiment tag density , we calculate that: for member A, ; For member B, ; For member C, ; Hypothesized expression style stability factor The scaling factor , volatility strength indicator The scaling factor , preset adjustment coefficient ; For member A, ; then adjust the weight factor ; For member B, ; then adjust the weight factor ; For member C, ; then adjust the weight factor ; Table 2 Personality, skills, and expression parameters after regulation ; Known, , assuming that the preset personality matching affects the weight , Skill keyword matching affects weight , subjective word proportion matching degree affects weight The influence weight of emotional matching ; For member A, the matching score ; For member B, the matching score ; For member C, the matching score ; Assume that the matching score threshold is 0.7, members A and C meet the conditions and are included in the team; during the team operation, if member C's mood continues ( ), the emotion density after regulation is 0.411, the emotion matching is 0.739, and the total score drops to 0.4+0.1906+0.2×0.739=0.638 (below the threshold), then an early warning is issued and relevant staff are reminded to intervene and adjust.

[0040] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0041] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art will be able to modify the technical solutions described in the aforementioned embodiments or substitute equivalents for some of the technical features. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A team building optimization system based on bidirectional semantic matching, characterized by: The system includes: a data acquisition and set construction module, a stability factor and index calculation module, a set construction and matching score calculation module, and an analysis and early warning module; The data acquisition and collection construction module: after authorization by the member, obtains the member's expression parameters and skill keywords in the historical team collaboration process; constructs the member's expression style tag set; and retrieves the member's team collaboration score sequence; The stability factor and index calculation module: obtains the member's fluctuation intensity index based on the team collaboration score sequence; calculates the member's expression style stability factor based on the member's expression style tag set; calculates the member's semantic shift sensitivity index based on the member's expression style stability factor and fluctuation intensity index; The set construction and matching score calculation module: obtains the personality label of the member and constructs a stable set of personalized expression styles; obtains the task description text of the current task and constructs a multidimensional semantic requirement set; calculates the matching score between the stable set of personalized expression styles of the member and the multidimensional semantic requirement set; The analysis and warning module sets the rules for including task team members and builds a task team based on the matching scores. During the team operation process, it determines in real time whether the members meet the rules for including task team members. If not, it issues a warning and reminds relevant staff to intervene and make adjustments.

2. The team building optimization system based on bidirectional semantic matching according to claim 1 is characterized in that: The data acquisition and set construction module includes a data acquisition and set construction unit; The data acquisition and set construction unit: after authorization by the members, obtains the natural language communication records and skill keywords of the members in the historical team collaboration process from the historical team collaboration archive system, uses natural language processing technology to extract the expression parameters of the members in the natural language communication records and normalizes them, the expression parameters including the proportion of subjective words and the density of emotional tags; based on the expression parameters, constructs the member's expression style tag set; and retrieves the team collaboration score sequence of the members in the historical team collaboration process from the team evaluation system background.

3. The team building optimization system based on bidirectional semantic matching according to claim 2 is characterized in that: The stability factor and index calculation module includes a stability factor calculation unit and an index calculation unit; The stability factor calculation unit: based on the team collaboration score sequence, obtains the team collaboration scores of the members at two adjacent task time points, and calculates the change rate of the team collaboration scores of the members at the two adjacent task time points using the difference method; traverses the task time points to obtain the maximum change rate of the member's team collaboration score, which is recorded as the fluctuation intensity index; and calculates the member's expression style stability factor based on the member's expression style tag set; The index calculation unit calculates the semantic shift sensitivity index of the member based on the expression style stability factor and fluctuation intensity index of the member.

4. The team building optimization system based on bidirectional semantic matching according to claim 3 is characterized in that: The set construction and matching score calculation module includes a set construction unit and a matching score calculation unit; The set construction unit: uses MBTI personality test technology to obtain the personality label of the member, integrates the personality label with skill keywords, expression style tag set, expression style stability factor and semantic shift sensitivity index to construct a personalized expression style stability set; obtains the task description text of the current task, extracts subsets of the task description text based on the domain dictionary and word embedding model, and constructs a multi-dimensional semantic requirement set; The matching score calculation unit calculates the matching score between the member's personalized expression style stable set and the multi-dimensional semantic requirement set.

5. The team building optimization system based on bidirectional semantic matching according to claim 4 is characterized in that: The analysis and early warning module includes an analysis and early warning unit; The analysis and early warning unit sets a task team member inclusion rule, specifically as follows: a matching score threshold is preset, and if the matching score between the stable set of personalized expression styles of the i-th member and the multidimensional semantic requirement set is greater than or equal to the matching score threshold, the i-th member is determined to meet the task team member inclusion rule; all members who meet the task team member inclusion rule are obtained, and a task team is constructed; During the team operation process, the subjective word proportion and emotional tag density of team members at each task time point are obtained in real time, the expression style stability factor of team members is calculated in real time, and the personalized expression style stability set of team members is updated. The matching score between the personalized expression style stability set of team members and the multidimensional semantic demand set is recalculated, and it is judged whether the task team member inclusion rules are met; if it is judged that the team member does not meet the task team member inclusion rules, an early warning will be issued and relevant staff will be reminded to intervene and adjust.

6. A team building optimization method based on bidirectional semantic matching, comprising executing a team building optimization system based on bidirectional semantic matching as claimed in any one of claims 1 to 5, characterized in that: The method comprises the following steps: Step S1: After authorization by the member, obtain the member's expression parameters and skill keywords in the historical team collaboration process; build the member's expression style tag set; and retrieve the member's team collaboration score sequence; Step S2: Based on the team collaboration scoring sequence, obtain the member's fluctuation intensity index; based on the member's expression style tag set, calculate the member's expression style stability factor; based on the member's expression style stability factor and fluctuation intensity index, calculate the member's semantic shift sensitivity index; Step S3: Obtain the member's personality label and construct a stable set of personalized expression styles; obtain the task description text of the current task and construct a multidimensional semantic requirement set; calculate the matching score between the member's stable set of personalized expression styles and the multidimensional semantic requirement set; Step S4: Set the rules for inclusion of task team members, and build a task team based on the matching scores; during the team operation, determine in real time whether the members meet the rules for inclusion of task team members. If not, issue an early warning and remind relevant staff to intervene and make adjustments.

7. The team building optimization method based on bidirectional semantic matching according to claim 6 is characterized in that: The specific implementation process of step S1 includes: After authorization by the members, the natural language communication records and skill keywords of the members in the historical team collaboration process are obtained from the historical team collaboration archive system. Using natural language processing technology, the expression parameters of the members in the natural language communication records are extracted and normalized. The expression parameters include the proportion of subjective words and the density of emotional tags. Based on the expression parameters, the expression style tag set of the i-th member is constructed and recorded as ,in, Indicates the proportion of subjective words of the i-th member, represents the emotional label density of the i-th member; The team collaboration score sequence of members in the historical team collaboration process is retrieved from the team evaluation system background, recorded as ,in, represents the team collaboration score of the i-th member at the m-th task time point in the historical team collaboration process, and M represents the total number of task time points of the i-th member in the historical team collaboration process.

8. The team building optimization method based on bidirectional semantic matching according to claim 7 is characterized in that: The specific implementation process of step S2 includes: Based on teamwork scoring sequence , obtain the teamwork score of the i-th member at two adjacent task time points, use the difference method to calculate the change rate of the teamwork score of the i-th member at two adjacent task time points; traverse the task time points and obtain the maximum change rate of the teamwork score of the i-th member, which is recorded as the fluctuation intensity index ; Based on the expression style tag set of the i-th member, the expression style stability factor of the i-th member is calculated. The calculation formula is: ,in, and Respectively represent the weight factors of the preset subjective word ratio and emotional tag density, represents the expression style stability factor of the i-th member; Based on the expression style stability factor of the i-th member and the Volatility Strength Index , calculate the semantic shift sensitivity index of the i-th member, the calculation formula is: ,in, and Represents the preset expression style stability factors respectively and the Volatility Strength Index The scaling factor.

9. The team building optimization method based on bidirectional semantic matching according to claim 8 is characterized in that: The specific implementation process of step S3 includes: Using MBTI personality test technology, we can obtain the personality label of the i-th member and combine the personality label with the skill keywords and expression style tag set. , expression style stability factor and semantic shift sensitivity index Fusion is performed to construct a stable set of personalized expression styles, as follows: Based on the expression style stability factor and semantic shift sensitivity index , calculate the control weight factor of the i-th member, the calculation formula is as follows: ； in, represents the control weight factor of the i-th member, Indicates the preset adjustment coefficient; The control weight factor based on the i-th member , construct a stable set of personalized expression styles, recorded as ,in, represents the personality label of the i-th member, Indicates the skill keyword of the i-th member; Obtain the task description text of the current task, perform subset extraction on the task description text based on the domain dictionary and word embedding model, and construct a multidimensional semantic requirement set, as follows: Using natural language processing technology, we extract personality requirement labels, skill requirement keywords, subjective word ratios, and emotional harmony semantics from the task description text, and construct a multidimensional semantic requirement set, which is recorded as ,in, Represents the personality requirement label in the task description text, Indicates the skill requirement keywords in the task description text, Indicates the proportion of subjective words in the task description text, Representing the emotional attunement semantics in the task description text; Calculate the matching score between the stable set of personalized expression styles of the i-th member and the multidimensional semantic requirement set as follows: ； ； ； in, represents the matching score between the stable set of personalized expression style of the i-th member and the multidimensional semantic requirement set, Indicates personality matching, if , then it is recorded as 1, otherwise it is 0. Indicates the matching degree of subjective word proportion, Indicates emotional matching. 、 、 and They respectively represent the influence weights of the preset personality matching, skill keyword matching, subjective word proportion matching, and emotion matching.

10. The team building optimization method based on bidirectional semantic matching according to claim 9 is characterized in that: The specific implementation process of step S4 includes: Set the rules for including task team members as follows: A matching score threshold is preset. If the matching score between the stable set of personalized expression styles of the i-th member and the multidimensional semantic requirement set is greater than or equal to the matching score threshold, the i-th member is determined to meet the task team member inclusion rule; all members who meet the task team member inclusion rule are obtained, and a task team is constructed; During the team operation process, the subjective word ratio and emotional tag density of team members at each task time point are obtained in real time, the expression style stability factor of team members is calculated in real time, and the personalized expression style stability set of team members is updated. The matching score between the personalized expression style stability set of team members and the multidimensional semantic requirement set is recalculated, and it is determined whether the task team member inclusion rules are met; If it is determined that a team member does not meet the rules for inclusion in the task team, an early warning will be issued and relevant staff will be reminded to intervene and make adjustments.

Images