A method and apparatus for constructing an enterprise personnel relationship network

Abstract

This disclosure provides a method and apparatus for constructing an enterprise personnel relationship network, relating to the field of data processing technology. The method includes: acquiring multi-source data corresponding to relevant personnel within the enterprise; matching the multi-source data based on preset association matching rules to identify direct relationships between relevant personnel; quantifying the relationship strength of direct relationships using a multi-dimensional relationship strength quantification model and calculating relationship strength scores for direct relationships; constructing an enterprise personnel relationship network using a graph structure modeling approach; and mining indirect relationships based on the enterprise personnel relationship network using a graph structure-based association matching algorithm and calculating the relationship strength scores for the mined indirect relationships. This disclosure enables comprehensive and accurate identification and quantification of relationships between personnel, constructing a realistic enterprise personnel relationship network, and thereby achieving efficient management and application of relationship resources.

Description

Technical Field

[0001] This disclosure relates to the field of data processing technology, and in particular to a method and apparatus for constructing an enterprise personnel relationship network. Background Technology

[0002] In the process of enterprise operation, business activities and internal collaboration, relationships are an important intangible asset of the enterprise. Their completeness, accuracy and timeliness directly affect the efficiency of internal collaboration, the efficiency of business cooperation conversion and the level of risk control and compliance management.

[0003] The management and application of relationships in existing technologies have the following technical problems:

[0004] (1) Data sources are limited and the degree of multi-source integration is low: The data sources are single, with enterprise organizational structure data as the core and only a small amount of project data is integrated. It can only identify shallow administrative relationships such as direct superiors and subordinates and the same department, and cannot fully depict the real personnel relationship; it is difficult to meet the application needs of complex scenarios such as business referrals and cross-departmental collaboration.

[0005] (2) Lack of a scientific quantitative system for relationship strength: It can only identify whether two people are related, but cannot distinguish the degree of the relationship. Therefore, it cannot provide a quantitative basis for applications such as relationship path recommendation and relationship matching. Relationship path recommendation and relationship matching can only rely on the number of jumps to sort, which makes the practicality of the relationship path recommendation results extremely low and cannot meet the needs of precise collaboration and business referral for enterprises.

[0006] (3) The relationship network is statically constructed and lacks a dynamic update mechanism: The relationship network is generated by full batch calculation. When the organizational structure is adjusted, project members are changed, or personnel interaction behavior is added, the relationship network cannot synchronize data changes in a timely manner, resulting in the relationship network information being lagging, distorted, and having low practicality, and failing to reflect the dynamic changes of the relationship.

[0007] (4) The association path recommendation function is simple and not optimized according to the actual enterprise scenario: When recommending the network path, only the breadth-first search (BFS) algorithm is used to calculate the association path with the fewest hops. This easily leads to the situation of "the shortest association path but the weakest relationship". As a result, it is easy to recommend weak association paths with no actual business value, which cannot meet the real business introduction, business docking and collaboration needs.

[0008] (5) The resources of personal relationships are scattered in private channels such as personal address books and social software. The departure of personnel can easily lead to the loss of resources of personal relationships, which increases the cost of business development and customer maintenance for enterprises. Summary of the Invention

[0009] To address one or more of the aforementioned technical problems, this disclosure provides a method, apparatus, electronic device, storage medium, and computer program product for constructing an enterprise personnel relationship network.

[0010] This embodiment of the disclosure comprehensively considers both internal personnel and external partners of the enterprise. First, it obtains multi-source data on these personnel's organization, projects, behaviors, business, scientific research, and academic institutions to fully cover the enterprise's business needs. Then, based on preset association matching rules, multi-source data is matched to identify direct relationships between relevant personnel in the enterprise. A multi-dimensional relationship strength quantification model is used to consider the different sensitivities of different relationship types to time factors. Relationships are divided into first-type and second-type relationships. Combining the dynamic weights of each first-type relationship, the relationship strength scores of each first-type relationship in the quantified direct relationships, the association frequency of each first-type relationship in the quantified direct relationships, the time decay factor of the quantified direct relationships, and the relationship strength scores of the second-type relationships, the first-type relationships are weighted and summed collaboratively. The time decay factor is then used to correct the timeliness of the weighted summation result. The second-type relationship, as a static association item, together with the time-corrected result, determines the final relationship strength score. This achieves refined quantification of the relationship strength of direct relationships between different personnel, and the calculated relationship strength score comprehensively and accurately reflects the true closeness of the direct relationships between relevant personnel in the enterprise. Furthermore, a graph-based modeling approach is adopted to establish a network of enterprise personnel relationships with relationship strength scores. Further, a graph-based association matching algorithm is used to comprehensively consider all indirect association paths and all direct associations within each path to calculate the relationship strength score of indirect associations. In calculating the relationship strength score of each indirect association, the relationship strength score for each path is calculated separately and corrected using a path decay coefficient. The total relationship strength score of all paths is then used as the final relationship strength score. This fully considers the influence of the relationship strength of each path and the path length (i.e., the path decay coefficient) on the relationship strength of indirect associations. Simultaneously, the relationship strength score on each path is jointly determined by the relationship strength scores of each segment of direct associations on that path. This fully considers the influence of the relationship strength of each direct association on the path, avoiding the neglect of other direct associations and other effective information. This ensures that the calculated relationship strength score of each indirect association more accurately reflects the closeness of the relationships between personnel. Through the embodiments disclosed herein, it is possible to comprehensively and accurately mine the relationships between relevant personnel within an enterprise and quantify the strength of these relationships, thereby constructing a real and quantifiable network of relationships among enterprise personnel and achieving efficient management and application of enterprise personnel relationship resources.

[0011] According to one aspect of this disclosure, a method for constructing an enterprise personnel relationship network is provided, the method comprising: Acquire multi-source data corresponding to relevant personnel of the enterprise. The multi-source data includes: organizational data, project data, behavioral data, business data, scientific research data, and university data. The relevant personnel of the enterprise include internal personnel and external partners. The multi-source data is matched based on preset association matching rules to identify direct relationships between relevant personnel in the enterprise; among which, direct relationships include the following types: organizational relationship, project relationship, behavioral relationship, business relationship, research relationship, and alumni relationship; A multi-dimensional relationship strength quantification model is used to quantify the relationship strength of each identified direct relationship and calculate the relationship strength score for each direct relationship. Specifically, for any direct relationship, a weighted sum is calculated based on the dynamic weights corresponding to each first-type relationship, the relationship strength scores of each first-type relationship within that direct relationship, and the association frequency of each first-type relationship within that direct relationship. The time decay factor is used to correct the timeliness of the weighted sum, and the sum of the time-corrected result and the relationship strength scores of second-type relationships is taken as the relationship strength score of that direct relationship. The first-type relationships include: organizational relationships, project relationships, behavioral relationships, business relationships, and research relationships; the second-type relationships include: alumni relationships. The dynamic weights corresponding to each first-type relationship are negatively correlated with the interval between the most recent association time of the person corresponding to that direct relationship and the current time. A graph structure modeling approach is used to construct a network of relationships among enterprise personnel, with relevant personnel as nodes and direct relationships as edges. A graph-based association matching algorithm is used to mine indirect relationships between relevant personnel in the enterprise based on the enterprise personnel relationship network, and calculate the relationship strength score of each mined indirect relationship. Specifically, for any indirect relationship, the relationship strength score of each associated path corresponding to the indirect relationship is calculated, and the total score after correcting the relationship strength score of each associated path by the path attenuation coefficient corresponding to each associated path is determined as the relationship strength score of the indirect relationship.

[0012] In one possible implementation, the method further includes: Obtain a user's recommendation request; the recommendation request is used to request a recommendation path from a first person to a second person; wherein, the first person and the second person are any two different people in the enterprise's personnel relationship network; Based on the enterprise personnel relationship network, at least one candidate association path is identified; Based on the relationship strength scores of each direct relationship in each candidate association path, the weights of each direct relationship in each candidate association path are calculated and summed to obtain the weight of each candidate association path; wherein, the weight of any direct relationship is negatively correlated with the relationship strength score of that direct relationship. The candidate association path with the lowest weight is determined as the optimal association path to recommend to the user.

[0013] In one possible implementation, for any direct association, a weighted sum is performed based on the dynamic weights corresponding to each first-type relationship, the relationship strength scores of each first-type relationship in the direct association, and the association frequencies of each first-type relationship in the direct association. The result of the weighted sum is then corrected for timeliness using a time decay factor. The sum of the time-corrected result and the relationship strength scores of the second-type relationships is taken as the relationship strength score of the direct association. This includes: Based on the preset scoring rules corresponding to each first type of relationship and each second type of relationship in the direct relationship, calculate the relationship strength score of each first type of relationship and each second type of relationship in the direct relationship; Obtain the most recent association time of the person corresponding to the direct association, and determine the dynamic weight and time decay factor corresponding to each first type of relationship in the direct association based on the most recent association time; Obtain the association frequency of each first-type relationship in this direct relationship; Based on the dynamic weights corresponding to each first type of relationship in the direct association, the product of the relationship strength score and the association frequency of each first type of relationship in the direct association is weighted and summed to obtain the initial value of the relationship strength of the direct association. Calculate the product of the time decay factor and the initial value of the relationship strength to obtain the timeliness correction value of the relationship strength of the direct correlation; The relationship strength score of the direct association is obtained by summing the timeliness correction value and the relationship strength score of the second type of relationship.

[0014] In one possible implementation, the multi-dimensional relationship strength quantification model is shown in the following equation: +

[0015] in, The score represents the strength of the direct association between person i and person j. This represents the time decay factor corresponding to the direct correlation between person i and person j; Indicates the time of the most recent association between person i and person j; The score represents the strength of the organizational relationship between person i and person j. The score represents the strength of the relationship between personnel i and personnel j. The score represents the strength of the behavioral relationship between person i and person j. This represents the relationship strength score between personnel i and personnel j in terms of their business relationship. The score represents the strength of the research relationship between personnel i and personnel j. The dynamic weight representing the organizational relationship between personnel i and personnel j; The dynamic weight representing the project relationship between personnel i and personnel j; The dynamic weights representing the behavioral relationships between personnel i and personnel j; The dynamic weight representing the business relationship between personnel i and personnel j; The dynamic weight representing the research relationship between personnel i and personnel j; This indicates the frequency of the organizational relationship between person i and person j. This indicates the frequency of project associations between personnel i and personnel j. Indicates the frequency of association of the behavioral relationship between person i and person j; This indicates the frequency of business relationships between personnel i and personnel j. This indicates the frequency of the research relationship between personnel i and personnel j. The score represents the strength of the alumni relationship between person i and person j.

[0016] In one possible implementation, calculating the relationship strength score for each association path corresponding to any indirect association includes: For any association path, the type weight of each direct association in the association path is determined according to the preset weight of each type of relationship in each direct association path; Obtain the relationship strength score of each direct relationship in the associated path; Calculate the product of the type weight of each direct relationship in the association path and the relationship strength score to obtain the relationship strength score of the association path.

[0017] In one possible implementation, the calculation of the relationship strength score for each mined indirect association includes: calculating the relationship strength score for each indirect association using the following formula:

[0018] in, The score represents the strength of the indirect relationship between person i and person k. This represents the set of associated paths originating from person i and ending at person k. This represents the path attenuation coefficient corresponding to path p in the set of associated paths; This indicates the direct association relationship in the associated path p. Type weights; Indicates a direct relationship The strength score of the relationship.

[0019] In one possible implementation, the step of calculating the weights of each direct relationship in each candidate association path based on the relationship strength scores of each direct relationship in each candidate association path and summing them to obtain the weight of each candidate association path includes: Based on the relationship strength scores of each direct relationship in each candidate association path, the weight of each candidate association path is calculated using the following formula:

[0020] in, This represents the weight of the candidate associated path P; The score represents the strength of the direct association between person i and person j. This represents any two individuals in the candidate association path P who have a direct relationship.

[0021] In one possible implementation, the method further includes: The multi-source data is dynamically updated according to the update strategy corresponding to different types of data; wherein the update strategy includes one or more of the following: real-time update strategy, near real-time update strategy, and full update strategy; The enterprise personnel relationship network is adjusted based on the updated multi-source data.

[0022] According to another aspect of this disclosure, an apparatus for building an enterprise personnel relationship network is provided, the apparatus comprising: The acquisition module is used to acquire multi-source data corresponding to relevant personnel of the enterprise. The multi-source data includes: organizational data, project data, behavioral data, business data, scientific research data, and university data; the relevant personnel of the enterprise include internal personnel of the enterprise and external partners of the enterprise. The matching module is used to match the multi-source data based on preset association matching rules to identify direct relationships between relevant personnel of the enterprise; wherein, direct relationships include the following types: organizational relationship, project relationship, behavioral relationship, business relationship, scientific research relationship, and alumni relationship; The quantification module utilizes a multi-dimensional relationship strength quantification model to quantify the relationship strength of each identified direct relationship and calculate a relationship strength score for each direct relationship. Specifically, for any direct relationship, a weighted sum is calculated based on the dynamic weights corresponding to each first-type relationship, the relationship strength scores of each first-type relationship within that direct relationship, and the association frequency of each first-type relationship within that direct relationship. The time decay factor is used to correct the timeliness of the weighted sum, and the sum of the time-corrected result and the relationship strength scores of the second-type relationships is taken as the relationship strength score of that direct relationship. The first-type relationships include: organizational relationships, project relationships, behavioral relationships, business relationships, and research relationships; the second-type relationships include: alumni relationships. The dynamic weights corresponding to each first-type relationship are negatively correlated with the interval between the most recent association time of the person corresponding to that direct relationship and the current time. The module is used to construct a network of relationships among company personnel using a graph structure modeling approach, with relevant personnel as nodes and direct relationships as edges. The mining module is used to employ a graph-based association matching algorithm to mine indirect relationships between relevant personnel within the enterprise based on the enterprise personnel relationship network, and to calculate the relationship strength score for each mined indirect relationship. Specifically, for any indirect relationship, the relationship strength score for each associated path corresponding to that indirect relationship is calculated, and the total score after correcting the relationship strength score of each associated path by using the path decay coefficient corresponding to each associated path is determined as the relationship strength score of that indirect relationship.

[0023] According to another aspect of this disclosure, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the above-described method.

[0024] According to another aspect of this disclosure, a non-volatile computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the steps of the above-described method.

[0025] According to another aspect of this disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of the above-described method.

[0026] The present disclosure has the following beneficial effects: By collecting multi-source data from multiple dimensions, including organizations, projects, behaviors, business operations, research, and universities, and combining precise matching based on preset relationship matching rules with graph-based relationship matching algorithms, this approach can not only accurately identify direct relationships between individuals but also uncover hidden indirect relationships. These relationships encompass both organizational relationships and reflect real-world business collaborations, making relationship discovery more comprehensive and accurate, closely aligned with actual enterprise collaboration scenarios. This allows it to meet the application needs of complex scenarios such as business referrals and cross-departmental collaborations. Furthermore, by acquiring multi-source data from both internal and external collaborators, it covers internal personnel and integrates external relationships, satisfying the diverse needs of cross-institutional and cross-industry collaborations.

[0027] The multi-dimensional relationship strength quantification model transforms abstract direct relationships into quantified relationship strength scores, thereby clarifying the closeness of direct relationships. Specifically, the multi-dimensional relationship strength quantification model combines the dynamic weights corresponding to each type of first-level relationship, the relationship strength scores of each type of first-level relationship in the direct relationships to be quantified, the association frequency of each type of first-level relationship in the direct relationships to be quantified, the time decay factor of the direct relationships to be quantified, and the relationship strength scores of the second-level relationships to achieve refined quantification of the relationship strength of direct relationships between different individuals. Furthermore, considering the varying sensitivities of different relationship types to time factors, relationships are categorized into Type I and Type II relationships. Type I relationships are typically less stable and more susceptible to time-related influences (e.g., the strength of a relationship weakens over time after a collaborative project ends), while Type II relationships are generally more stable and less affected by time (e.g., alumni maintain a high level of trust after graduation). Therefore, different strategies are employed when calculating the relationship strength score for direct connections between Type I and Type II relationships. Type I relationships are weighted and summed collaboratively, with the result adjusted for time decay using a time decay factor. Type II relationships, as static connections, are combined with the time-corrected result to determine the final relationship strength score. This allows the calculated relationship strength score to comprehensively and accurately reflect the true closeness of direct connections between relevant personnel within an enterprise, effectively supporting applications such as precise connection path recommendation and connection matching.

[0028] This algorithm, based on a graph structure, comprehensively considers all indirect association paths and all direct associations within each path to calculate the relationship strength score of indirect associations. Furthermore, in calculating the relationship strength score for each indirect association, the strength score for each association path is calculated separately and corrected using a path attenuation coefficient. The total score of all association paths is then used as the final relationship strength score. This fully considers the influence of the relationship strength of each path and its path length (i.e., the path attenuation coefficient) on the relationship strength of indirect associations. Simultaneously, the relationship strength score on an association path can be jointly determined by the relationship strength scores of each direct association segment along that path, thus fully considering the influence of the relationship strength of each direct association on the association path and avoiding the neglect of other effective information such as considering only some direct associations. The calculated relationship strength score for each indirect association more accurately reflects the closeness of the relationships between individuals.

[0029] By acquiring multi-source data on relevant personnel within the enterprise and constructing an enterprise personnel relationship network using graph structure modeling, personal relationship resources are transformed into standardized graph structure data at the enterprise level. This data can be stored in a unified enterprise graph database, detached from personal private channels, achieving enterprise-level accumulation of relationship resources. In some examples, a dynamic update mechanism ensures that relationship resources continue to accumulate as the enterprise develops, and that relationship resources remain within the enterprise even after personnel leave, preventing the loss of relationship resources. This achieves enterprise-level accumulation of relationship resources, effectively reducing the operational costs of business development and customer maintenance.

[0030] In some examples, when recommending association paths, the weight of candidate association paths is determined by the relationship strength score of each direct relationship and the number of direct relationships in the candidate association path. This comprehensively considers the overall relationship strength of all direct relationships in the candidate association path to achieve optimal association path recommendation, avoiding recommending weak or low-useful association paths to users. This results in accurate association path recommendations that meet the real needs of business referrals, business connections, and collaborations.

[0031] In some examples, a three-level dynamic update mechanism of real-time, near real-time, and full data is used to set differentiated update strategies for different types of data. High-frequency dynamic data is updated in real-time, non-real-time data is updated in near real-time, and low-frequency data is fully calibrated. This can quickly capture data changes such as organizational structure, personnel interaction, and project collaboration, and timely update the enterprise's personnel relationship network. This ensures that the data in the enterprise's personnel relationship network is highly consistent with the actual situation, and greatly improves the accuracy and timeliness of the data.

[0032] Other features and aspects of this disclosure will become clear from the following detailed description of exemplary embodiments with reference to the accompanying drawings. Attached Figure Description

[0033] The accompanying drawings, which are included in and form part of this specification, illustrate exemplary embodiments, features, and aspects of this disclosure together with the specification and serve to explain the principles of this disclosure.

[0034] Figure 1 A flowchart illustrating a method for constructing an enterprise personnel relationship network according to an embodiment of the present disclosure is shown. Figure 2 A flowchart illustrating a method for constructing an enterprise personnel relationship network and recommending associated paths according to an embodiment of the present disclosure is shown. Figure 3 This diagram shows a structural diagram of an enterprise personnel relationship network building apparatus according to an embodiment of the present disclosure; Figure 4 A block diagram of an electronic device 1900 according to an embodiment of the present disclosure is shown. Detailed Implementation

[0035] Various exemplary embodiments, features, and aspects of this disclosure will now be described in detail with reference to the accompanying drawings. The same reference numerals in the drawings denote elements that have the same or similar functions. Although various aspects of the embodiments are shown in the drawings, they are not necessarily drawn to scale unless specifically indicated otherwise.

[0036] As used herein, the terms “comprising,” “including,” “having,” or variations thereof are open-ended and include one or more of the stated features, integrals, elements, steps, components, or functions, but do not exclude the presence or addition of one or more other features, integrals, elements, steps, components, functions, or groups thereof.

[0037] When an element is referred to as “connected,” “coupled,” “responding,” or a variation thereof relative to another element, it may be directly connected, coupled, or responding to another element, or there may be an intermediate element present.

[0038] Although the terms first, second, third, etc., may be used herein to describe various elements / operations, these elements / operations should not be limited by these terms. These terms are only used to distinguish one element / operation from another. Therefore, without departing from the teachings of the inventive concept, a first element / operation in some embodiments may be referred to as a second element / operation in other embodiments.

[0039] The term “exemplary” as used herein means “serving as an example, embodiment, or illustration.” Any embodiment illustrated herein as “exemplary” is not necessarily to be construed as superior to or better than other embodiments.

[0040] Furthermore, to better illustrate this disclosure, numerous specific details are set forth in the following detailed description. Those skilled in the art will understand that this disclosure can be practiced without certain specific details. In some instances, methods, means, components, and circuits well known to those skilled in the art have not been described in detail in order to highlight the main points of this disclosure.

[0041] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, data stored, data displayed, etc.) and signals involved in this application are all authorized by the user or fully authorized by all parties, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant regions.

[0042] Figure 1 This document illustrates a flowchart of a method for constructing an enterprise personnel relationship network according to an embodiment of this disclosure. Exemplarily, this method can be executed by electronic devices such as terminal devices and servers, or a portion of such electronic devices (e.g., a processor). The terminal device can be a desktop terminal or a mobile terminal, such as a laptop, tablet, desktop computer, smartphone, smart speaker, smartwatch, smart TV, in-vehicle terminal, or various other types of electronic devices. The server can be a standalone physical server, a server cluster consisting of multiple physical servers, or a distributed system. The terminal devices, servers, and other electronic devices are all deployed within the enterprise, and the algorithms and models involved in this application are implemented by these locally deployed electronic devices within the enterprise. Figure 1 As shown, the method may include the following steps: Step 101: Obtain multi-source data corresponding to relevant personnel of the enterprise. The multi-source data includes: organizational data, project data, behavioral data, business data, scientific research data, and university data. The relevant personnel of the enterprise include internal personnel of the enterprise and external cooperative personnel of the enterprise.

[0043] External partners refer to individuals who have collaborative relationships with the company. For example, in the case of a pharmaceutical company, external partners could include: hospital experts and doctors, medical university professors and administrators, medical research institute researchers, pharmacy sales staff, and so on. By acquiring multi-source data on both internal and external partners, the company can cover both its internal and external networks, thereby building a network of these relationships and providing support for its business development.

[0044] For example, multi-source data can include data from six dimensions: organizational data, project data, behavioral data, business data, scientific research data, and institutional data. This can more comprehensively cover the dimensions of relationship mining among relevant personnel in an enterprise, making the relationship mining more comprehensive.

[0045] As an example, organizational data can include: basic personnel information and organizational structure information; basic personnel information can include: employee ID, name, department, position, location, etc.; organizational structure information can include: hierarchical relationships, department affiliation, historical organizational changes, and association information with departments within the same hospital, etc. Project data can include: project participant information, project division of labor, number of shared tasks, project start and end dates, etc. Behavioral data can include: number of academic conferences, number of meals shared, number of phone calls, number of business trips shared, number of registered academic visits, research information, etc. Business data can include: customer affiliation information, number of times permissions were shared, number of times clients were jointly contacted, etc. Research data can include: research achievements such as patents, articles, and copyrights. University data includes: graduating institution and major.

[0046] For example, multi-source data corresponding to relevant personnel in an enterprise can be automatically and seamlessly collected through internal enterprise systems and public databases.

[0047] In one possible implementation, raw multi-source data corresponding to relevant personnel within the enterprise can be collected. This raw multi-source data can then be preprocessed to obtain the corresponding multi-source data for each person. By preprocessing the collected raw multi-source data, invalid information can be removed, and the data format can be standardized, providing high-quality data support for subsequent multi-source data matching. This effectively avoids multi-source data matching errors caused by inconsistent data formats.

[0048] For example, preprocessing can include data cleaning, format standardization, and adding association identifiers. Data cleaning can include removing invalid data such as null values, incorrect formats, and duplicate data, as well as correcting abnormal data such as incorrect department names and personnel codes. Format standardization can include unifying data in different formats (e.g., standardizing personnel codes to a fixed length, department names to follow company naming conventions, and time formats to YYYY-MM-DD HH:MM:SS). Adding association identifiers can include adding unique association identifiers (e.g., employee ID, project ID, customer ID) to each type of data to ensure cross-matching between different data sources. The employee ID, project ID, and customer ID identifiers are set internally by the company according to its own needs. Thus, by adopting a three-level preprocessing workflow of "cleaning-standardization-association," data format standardization unifies the data format, and unique association identifiers are added to each type of data, thereby providing high-quality, matchable structured data for subsequent identification of direct relationships between relevant personnel within the company and for mining indirect relationships among relevant personnel.

[0049] Step 102: Match the multi-source data based on preset association relationship matching rules to identify the direct association relationships between relevant personnel of the enterprise.

[0050] Direct relationships include the following types: organizational relationships, project relationships, behavioral relationships, business relationships, research relationships, and alumni relationships.

[0051] For example, the types of direct relationships correspond one-to-one with the data dimensions of multi-source data. Based on the six data dimensions of multi-source data—organizational data, project data, behavioral data, business data, scientific research data, and institutional data—direct relationships can be categorized into six major types: organizational relationships, project relationships, behavioral relationships, business relationships, scientific research relationships, and alumni relationships.

[0052] Direct association refers to a relationship between two individuals that exists directly without the need for intermediaries; it can also be called a first-degree association (or first-degree network). For any two individuals with a direct association, they may have only one type of direct association, or multiple types of direct associations simultaneously. For example, the direct association between two individuals could be an organizational relationship, a project relationship, or a behavioral relationship; furthermore, it could be both an organizational and project relationship, or a combination of these relationships.

[0053] In one possible implementation, each type of relationship can be further divided into one or more subtypes based on the data characteristics of the data dimension corresponding to each type of relationship, and then the standardized classification of direct relationships can be achieved based on the divided subtypes.

[0054] As an example, organizational relationships can include subtypes such as: direct superior-subordinate relationships, relationships within the same department, relationships between colleagues within the same department at the same level, relationships between different levels within the same department, relationships within the same hospital department, and historical relationships within the same department. Project relationships can include subtypes such as joint project participation, collaborative task sharing, joint project leadership, upstream and downstream project coordination, and historical joint project participation. Behavioral relationships can include subtypes such as meeting collaboration, visit and communication, dining interaction, telephone communication, and joint business trips. Business relationships can include subtypes such as shared client liaison, shared permissions, and historical shared business liaison. Research relationships can include subtypes such as joint publication of articles, joint patent applications, and joint copyright applications.

[0055] It should be noted that the above sub-type classification of each type of relationship is only an example, and can be flexibly classified according to the actual situation or needs of the enterprise's business, operation, etc., without limitation.

[0056] In this step, multi-source data corresponding to relevant personnel of the enterprise are matched based on preset association matching rules, automatically identifying potential associations, i.e., direct associations, between different personnel within the enterprise. For example, the multi-source data obtained through the above preprocessing can be matched to achieve cross-matching of data from different data sources.

[0057] In one possible implementation, the preset relationship matching rules can include relationship matching rules corresponding to each type of relationship, meaning that the relationship matching rules correspond one-to-one with the data dimensions. Matching the multi-source data based on the preset relationship matching rules to identify direct relationships between relevant personnel within the enterprise can include: matching multi-source data based on the relationship matching rules corresponding to each type of relationship, thereby ensuring the accuracy of the matching results through refined relationship matching rules.

[0058] As an example, the pre-defined matching rules for organizational relationships could include: if two individuals have the same department code, they are in the same department; if one individual's superior code equals the other's employee ID, they are in a direct superior-subordinate relationship, and so on. Pre-defined matching rules for project relationships could include: if two individuals have the same project ID and their participation times overlap, they are in a joint project participation relationship, and so on. Pre-defined matching rules for behavioral relationships could include: if two individuals attend meetings together at least once within 30 days, they have a meeting collaboration relationship; if two individuals visit and communicate with each other at least twice within 30 days, they have a visit and communication relationship; if two individuals travel together on business trips at least once within 30 days, they have a joint business trip relationship, and so on. Pre-defined matching rules for business relationships could include: if two individuals have the same client ID and are both the client's contact person, they have a joint client contact relationship, and so on. Pre-defined matching rules for research relationships could include: if two individuals publish the same articles, they have a joint article publication relationship, and so on. The pre-defined association matching rules for alumni relationships may include: if two people graduated from the same institution, then the two people are alumni.

[0059] It should be noted that the relationship matching rules for each type of relationship mentioned above are only examples and can be flexibly configured according to the actual situation or needs of the enterprise's business and operations, without any limitations.

[0060] One possible implementation involves using a fast hash-based matching algorithm to match multi-source data based on predefined association matching rules. This combines precise matching based on predefined association matching rules with fast hash-based matching, enabling accurate identification of direct relationships between relevant personnel within an enterprise and efficient matching of large-scale data.

[0061] For example, data (such as "author name + affiliation" or "employee ID + project ID") can be hashed, and a unique hash value can be obtained by taking the modulo, which can be used to quickly compare and match relationships.

[0062] As an example, taking MD5 hashing as an example, it is shown in the following formula:

[0063] in, Representing data The hash value; This indicates the data to be encoded (e.g., “Zhang San + Chinese Academy of Sciences”, “5206456 + XM2024001”). Indicates data Perform an MD5 hash operation to obtain a 128-bit hash value; This indicates the size of the hash table, used to control the range of hash values ​​and avoid hash collisions.

[0064] Step 103: Using a multi-dimensional relationship strength quantification model, the relationship strength of each identified direct relationship is quantified, and a relationship strength score is calculated for each direct relationship. Specifically, for any direct relationship, a weighted sum is performed based on the dynamic weights corresponding to each first-type relationship, the relationship strength scores of each first-type relationship within that direct relationship, and the association frequency of each first-type relationship within that direct relationship. The time decay factor is then used to correct the timeliness of the weighted sum. The sum of the time-corrected result and the relationship strength scores of the second-type relationships is taken as the relationship strength score of that direct relationship. The first-type relationships include: organizational relationships, project relationships, behavioral relationships, business relationships, and research relationships; the second-type relationships include: alumni relationships.

[0065] In this step, a multi-dimensional relationship strength quantification model is employed, integrating six dimensions—organization, project, behavior, business, research, and institution—to quantify the relationship strength of each direct connection identified in step 102, thereby transforming abstract relationships into quantifiable indicators. Each direct connection thus has a relationship strength score, providing data support for subsequent enterprise personnel relationship network construction, indirect relationship mining, and connection path recommendation. Furthermore, when quantifying relationship strength, the multi-dimensional relationship strength quantification model combines the dynamic weights corresponding to each first-type relationship, the relationship strength scores of each first-type relationship in the direct connections to be quantified, the association frequency of each first-type relationship in the direct connections to be quantified, the time decay factor of the direct connections to be quantified, and the relationship strength scores of the second-type relationships, achieving refined quantification of the relationship strength of direct connections between different individuals. Furthermore, considering the varying sensitivities of different relationship types to time factors, relationships are categorized into Type I and Type II relationships. Type I relationships are typically less stable and more susceptible to time-related influences (e.g., the strength of a relationship weakens over time after a collaborative project ends), while Type II relationships are generally more stable and less affected by time (e.g., alumni maintain a high level of trust after graduation). Therefore, different strategies are employed when calculating the relationship strength score for direct connections between Type I and Type II relationships. Type I relationships are weighted and summed collaboratively, with the result adjusted for time decay using a time decay factor. Type II relationships, as static connections, are combined with the adjusted result to determine the final relationship strength score. This allows the calculated relationship strength score for direct connections to comprehensively and accurately reflect the true closeness of direct relationships among relevant personnel within the company. This provides precise data support for subsequent efforts to establish accurate corporate personnel relationship networks, mine indirect relationships, and recommend connection paths.

[0066] The relationship strength score indicates the closeness of the relationship between individuals, and is positively correlated with the closeness of the relationship. For example, for a given direct relationship, a higher relationship strength score indicates a closer relationship between the two individuals; conversely, a lower score indicates a weaker relationship. As an example, the relationship strength score can range from 0 to 100, where 0 indicates no direct relationship between the two individuals; and if the calculated relationship strength score exceeds 100, then 100 is taken as the final relationship strength score.

[0067] In one possible implementation, for any direct association, a weighted sum is performed based on the dynamic weights corresponding to each first type of relationship, the relationship strength scores of each first type of relationship in the direct association, and the association frequency of each first type of relationship in the direct association. The timeliness of the weighted sum is then corrected by a time decay factor, and the sum of the timeliness-corrected result and the relationship strength scores of the second type of relationship is taken as the relationship strength score of the direct association. This includes: calculating the relationship strength scores of each first type of relationship and the second type of relationship in the direct association based on preset scoring rules corresponding to each first type of relationship and the second type of relationship in the direct association. Obtain the most recent association time of the person corresponding to the direct relationship, and determine the dynamic weight and time decay factor corresponding to each first type of relationship in the direct relationship based on the most recent association time; obtain the association frequency of each first type of relationship in the direct relationship; based on the dynamic weight corresponding to each first type of relationship in the direct relationship, perform a weighted summation of the product of the relationship strength score and the association frequency of each first type of relationship in the direct relationship to obtain the initial value of the relationship strength of the direct relationship; calculate the product of the time decay factor and the initial value of the relationship strength to obtain the timeliness correction value of the relationship strength of the direct relationship; sum the timeliness correction value and the relationship strength score of the second type of relationship to obtain the relationship strength score of the direct relationship.

[0068] For example, preset scoring rules can be set for each type of first-level and second-level relationship based on factors such as the subtype of the relationship, the frequency of personnel interaction, project importance, business importance, and the number of research achievements. Then, for any direct relationship, the relationship strength score of each type of first-level and second-level relationship within that direct relationship can be calculated based on the preset scoring rules. In this way, the preset scoring rules for each type of first-level and second-level relationship, which consider factors such as the subtype of the relationship, the frequency of personnel interaction, project importance, business importance, and the number of research achievements, are used as scoring standards to score the relationship strength of the corresponding types of relationships, thereby ensuring that the scoring results (i.e., the relationship strength scores) are more closely aligned with the actual collaborative scenarios of the enterprise.

[0069] For example, if the direct relationship between two individuals includes some of the first-type relationships such as organizational relationship, project relationship, behavioral relationship, business relationship, and research relationship, then the relationship strength score of these first-type relationships in the direct relationship is not zero, while the score of other types of relationships is zero. For instance, if two individuals have an organizational relationship and a project relationship, then the relationship strength scores of the first-type relationships in the direct relationship between these two individuals are as follows: organizational relationship has a positive score, project relationship has a positive score, behavioral relationship has a zero score, business relationship has a zero score, and research relationship has a zero score. For example, if the direct relationship between two individuals includes a second-type relationship, then the relationship strength score of the second-type relationship in the direct relationship is not zero (e.g., a positive score); otherwise, the relationship strength score of the second-type relationship in the direct relationship is zero.

[0070] As an example, the preset scoring rules for organizational relationships can be set according to different subtypes of organizational relationships. For example, they can include: a relationship strength score of 80-100 points for direct superior-subordinate relationships (e.g., 90 points); a relationship strength score of 60-80 points for same-department, same-level relationships (e.g., 70 points); a relationship strength score of 40-60 points for same-department, different-level relationships (e.g., 50 points); a relationship strength score of 20-40 points for historical same-department relationships (e.g., 30 points), and so on; the preset scoring rules for project relationships... The scoring rules can be set according to different subtypes of project relationships and project importance. For example, they can include: a relationship strength score of 80-100 points (e.g., 90 points) for jointly leading projects; a relationship strength score of 60-80 points (e.g., 75 points) for jointly participating in core projects; a relationship strength score of 40-60 points (e.g., 55 points) for jointly participating in ordinary projects; a relationship strength score of 20-40 points (e.g., 30 points) for historically jointly participating projects; and preset scoring rules for behavioral relationships. The relationship strength score can be set based on the frequency of interaction between individuals. For example, it can include: high-frequency interaction (such as attending meetings together, traveling on business together, dining together, visiting and communicating together, and making phone calls together at least 10 times within 30 days) with a score in the range of 80-100 (e.g., 90 points); medium-frequency interaction (such as attending meetings together, traveling on business together, dining together, visiting and communicating together, and making phone calls together at least 3-9 times within 30 days) with a score in the range of 50-80 (e.g., 70 points); and low-frequency interaction (such as attending meetings together, traveling on business together, dining together, visiting and communicating together, and making phone calls together at least 1 time within 30 days) with a score in the range of 100 (e.g., 90 points). The relationship strength score for ~2 times is a score in the range of 10 to 50 (e.g., 30 points), etc.; the preset scoring rules for business relationships can be set according to different sub-types of business relationships and business importance. For example, it can include: the relationship strength score for joint contact with core customers is a score in the range of 80 to 100 (e.g., 90 points), the relationship strength score for joint contact with ordinary customers is a score in the range of 40 to 80 (e.g., 50 points), the relationship strength score for joint contact with historical business is a score in the range of 20 to 40 (e.g., 30 points), etc.The preset scoring rules for research relationships can be set based on the number of research achievements. For example, they could include: a large number of joint research achievements (e.g., more than 5 joint patent applications, published articles, etc.) would correspond to a relationship strength score of 80-100 points (e.g., 90 points); a medium number of joint research achievements (e.g., 3-5 joint patent applications, published articles, etc.) would correspond to a relationship strength score of 40-80 points (e.g., 60 points); and a small number of joint research achievements (e.g., less than 3 joint patent applications, published articles, etc.) would correspond to a relationship strength score of 20-40 points (e.g., 30 points), and so on. Similarly, the preset scoring rules for alumni relationships could be: the existence of an alumni relationship would correspond to a relationship strength score of 5-15 points (e.g., 5 points).

[0071] It should be noted that the preset scoring rules for each of the first and second types of relationships mentioned above are only examples and can be set according to the actual situation or needs of the enterprise's business and operations.

[0072] The most recent association time can also be called the most recent interaction time. For example, for any direct association, the most recent association time of the people corresponding to that direct association represents the time when the two people with that direct association most recently became connected. If the direct association is only a first-type relationship, for example, if the two people only have a project relationship, then the most recent association time for these two people is the time they most recently participated in a project together. If the direct association is multiple first-type relationships, for example, if the two people simultaneously have a project relationship (e.g., a joint project participation relationship), a behavioral relationship (e.g., a visit / communication relationship), and a research relationship (e.g., a co-published article relationship), then the most recent association time for these two people is the shortest time among the most recent joint project participation time, the most recent visit / communication time, and the most recent co-published article time. For example, the most recent association time of the people corresponding to that direct association can be obtained from the data of the dimension corresponding to that direct association.

[0073] For example, determining the dynamic weights of each first-type relationship in a direct relationship based on the most recent association time can include: for any direct relationship, the dynamic weights of each first-type relationship in that direct relationship are negatively correlated with the interval between the most recent association time and the current time for the person corresponding to that direct relationship; that is, the longer the interval between the most recent association time and the current time, the smaller the dynamic weight; the shorter the interval between the most recent association time and the current time, the larger the dynamic weight. The specific function or formula for determining the dynamic weights can be configured according to requirements, ensuring that the dynamic weights are negatively correlated with the interval between the most recent association time and the current time for the person corresponding to that direct relationship, and is not limited in this respect. As an example, in addition to the most recent association time for the person corresponding to the direct relationship, the importance of each first-type relationship can be further considered. The dynamic weights of each first-type relationship in the direct relationship can be jointly determined by the interval between the most recent association time and the current time and the importance of each first-type relationship, so that the dynamic weights can further reflect the importance of different first-type relationships (e.g., the weight of a research relationship can be higher than the weight of a behavioral relationship).

[0074] For example, the time decay factor of a direct association can be determined based on the most recent association time using the following formula:

[0075] in, This represents the time decay factor corresponding to the direct correlation between person i and person j; Indicates the time of the most recent association between person i and person j; Indicates the time interval between the most recently associated time and the current time; This represents the attenuation coefficient, used to weaken direct correlations with long intervals. Its specific value can be set according to requirements; for example, it can be a positive number.

[0076] In this way, considering the timeliness of direct relationships, the time decay factor of direct relationships is determined based on the most recent association time of the corresponding personnel. This allows for a reduction in the quantitative score of long-standing direct relationships and an increase in the quantitative score of recent direct relationships when calculating the relationship strength score, thereby improving the accuracy of quantification.

[0077] For example, for any direct association, the association frequency of any first-type relationship within that direct association represents the number of times that first-type relationship occurs between the two individuals with that direct association. If the direct association is only one type of first-type relationship, for example, if the two individuals only have a project relationship (such as a joint project participation relationship), then the number of projects the two individuals jointly participate in can be used as the association frequency of the project relationship within the direct association between the two individuals. If the direct association is multiple types of first-type relationships, for example, if the two individuals simultaneously have a project relationship (such as a joint project participation relationship), a behavioral relationship (such as a visit and communication relationship), and a research relationship (such as a co-authored article relationship), then the number of projects the two individuals jointly participate in can be used as the association frequency of the project relationship within the direct association between the two individuals, the number of visits and communications between the two individuals can be used as the association frequency of the behavioral relationship within the direct association between the two individuals, and the number of articles co-authored by the two individuals can be used as the association frequency of the research relationship within the direct association between the two individuals. For example, the association frequency of each first-type relationship within the direct association can be obtained from the data of the corresponding dimension of each first-type relationship within the direct association.

[0078] This approach considers the impact of association frequency on the strength of relationships; a higher association frequency indicates a stronger relationship. Therefore, it obtains the association frequency of each first-type relationship within direct associations. This allows for the reinforcement of high-frequency first-type relationships when calculating the relationship strength score of direct associations. Consequently, the quantification results better reflect actual collaboration scenarios, further improving quantification accuracy.

[0079] For example, based on the dynamic weights corresponding to each first-type relationship in the direct association, a weighted sum is performed on the product of the relationship strength score and the association frequency of each first-type relationship in the direct association to obtain the initial value of the relationship strength of the direct association. This can include: for any first-type relationship in any direct association, calculating the product of the relationship strength score of the first-type relationship and the association frequency of the first-type relationship, and calculating the product of the product and the dynamic weight corresponding to the first-type relationship to obtain the weighted score of the first-type relationship; then, traversing all first-type relationships in the direct association and repeating the above operation to obtain the weighted scores of all first-type relationships; and summing the weighted scores of all first-type relationships to obtain the initial value of the relationship strength of the direct association.

[0080] In one possible implementation, the multi-dimensional relationship strength quantification model is shown in the following equation: +

[0081] in, The score represents the strength of the direct relationship between person i and person j. This represents the time decay factor corresponding to the direct correlation between person i and person j; This represents the time of the most recent association between person i and person j. The score represents the strength of the organizational relationship between person i and person j. The score represents the strength of the relationship between personnel i and personnel j. The score represents the strength of the behavioral relationship between person i and person j. This represents the relationship strength score between personnel i and personnel j in terms of their business relationship. The score represents the strength of the research relationship between personnel i and personnel j. The dynamic weight representing the organizational relationship between personnel i and personnel j; The dynamic weight representing the project relationship between personnel i and personnel j; The dynamic weights representing the behavioral relationships between personnel i and personnel j; The dynamic weight representing the business relationship between personnel i and personnel j; This represents the dynamic weight corresponding to the research relationship between personnel i and personnel j. This indicates the frequency of the organizational relationship between person i and person j. This indicates the frequency of project associations between personnel i and personnel j. Indicates the frequency of association of the behavioral relationship between person i and person j; This indicates the frequency of business relationships between personnel i and personnel j. This indicates the frequency of the research relationship between personnel i and personnel j. The score represents the strength of the alumni relationship between person i and person j.

[0082] In this way, by combining the dynamic weights, relationship strength scores, and association frequencies corresponding to organizational relationships, project relationships, behavioral relationships, business relationships, and research relationships (i.e., the first type of relationship), the time decay factor of direct relationships, and the relationship strength score of alumni relationships (i.e., the second type of relationship), this formula quantifies the relationship strength of the direct relationship between person i and person j, and calculates the relationship strength score of the direct relationship between person i and person j. It can accurately reflect the true closeness of the relationship between person i and person j.

[0083] Step 104: Using a graph structure modeling approach, construct a network of relationships among enterprise personnel, with relevant personnel as nodes and direct relationships as edges.

[0084] Since the relationship strength of direct connections has been quantified above, each side in the enterprise personnel relationship network constructed in this step has the basic information of relationship strength score.

[0085] For example, each side in an enterprise personnel relationship network may also include core information such as the type of relationship contained in the direct relationship, the source of the relationship, and the time of the most recent association, so as to comprehensively characterize the features of the direct relationship.

[0086] For example, each node in an enterprise personnel relationship network may also include basic information such as a unique personnel identifier (e.g., employee ID), name, position, and department, thereby ensuring that the node can be uniquely identified.

[0087] For example, an enterprise personnel relationship network can support functions such as visualization and hierarchical expansion. As an example, in response to a user's instruction to view the enterprise personnel relationship network, the network can be visualized on the user's terminal; different identifiers can also be used to display internal personnel and external partners in the enterprise personnel relationship network to facilitate user differentiation.

[0088] In this step, a graph structure modeling approach is used to construct a network of relationships among enterprise personnel, thereby transforming direct relationships into graph structure data; for example, a graph database can be used to store the relationships between nodes and edges.

[0089] Step 105: Employ a graph-based association matching algorithm to mine indirect relationships between relevant personnel within the enterprise based on the enterprise personnel relationship network, and calculate the relationship strength score for each mined indirect relationship. Specifically, for any indirect relationship, calculate the relationship strength score for each associated path corresponding to that indirect relationship, and determine the total score after correcting the relationship strength score of each associated path using the path decay coefficient corresponding to each associated path as the relationship strength score of that indirect relationship.

[0090] In this step, a graph-based association matching algorithm comprehensively considers all indirect association paths and all direct associations within each path to calculate the relationship strength score of indirect associations. This gives each indirect association a relationship strength score, providing data support for applications such as association path recommendation and indirect association querying. Furthermore, in calculating the relationship strength score of each indirect association, the relationship strength score is calculated separately for each association path and corrected using a path decay coefficient. The total relationship strength score of all association paths is then used as the final relationship strength score. This fully considers the influence of the relationship strength of each association path and the path length (i.e., the path decay coefficient) on the relationship strength of indirect associations. Simultaneously, the relationship strength score on an association path can be jointly determined by the relationship strength scores of each direct association segment on that path, thus fully considering the influence of the relationship strength of each direct association on the association path and avoiding the neglect of other effective information such as considering only some direct associations (e.g., only considering the direct association with the lowest relationship strength score). The calculated relationship strength score of each indirect association can more accurately reflect the closeness of the relationships between people.

[0091] Indirect connections refer to relationships between two individuals that require an intermediary and are not directly related. Indirect connections can also be called multi-degree connections (or multi-level networks or multi-hop indirect connections). For example, if two individuals are connected through one intermediary, their indirect connection can be called a second-degree connection (or second-level network). Similarly, if two individuals are connected through two intermediaries, their indirect connection can be called a third-degree connection (or third-level network). For a given indirect connection, a higher relationship strength score indicates a closer relationship between the two individuals; conversely, a lower score indicates a weaker relationship.

[0092] As an example, mining indirect relationships between relevant personnel within a company can include second-degree and third-degree relationships. Since fourth-degree and higher relationships typically indicate weaker levels of trust and closeness, mining second- and third-degree relationships effectively saves costs.

[0093] In one possible implementation, calculating the relationship strength score for each association path corresponding to any indirect association includes: for any association path, determining the type weight of each direct association in the association path based on the preset weights of each type of relationship in each direct association in the association path; obtaining the relationship strength score of each direct association in the association path; and calculating the product of the type weight and the relationship strength score of each direct association in the association path to obtain the relationship strength score of the association path. In this way, the relationship strength score on the association path is jointly determined by the relationship strength scores of each direct association segment on the association path, thus fully considering the impact of the relationship strength of each direct association on the association path. Furthermore, by combining the type weights of each direct association, the calculated relationship strength score of the association path can more accurately reflect the relationship strength.

[0094] Here, an association path refers to a path that starts from any person among the relevant personnel of the enterprise and ends at any other person among the relevant personnel of the enterprise through the enterprise's personnel relationship network. For any indirect association relationship, the indirect association relationship can correspond to one or more association paths (i.e., a set of association paths), and each association path has a person as the starting point, a person as the ending point, at least one intermediate person (i.e., at least three nodes in the enterprise's personnel relationship network), and a direct association relationship between these people (i.e., an edge in the enterprise's personnel relationship network). The person as the starting point and the person as the ending point are two people who have the indirect association relationship.

[0095] The preset weights for each type of relationship (including Type 1 and Type 2 relationships) can be set according to needs and are not limited thereto. For example, they can be set according to the importance of different types of relationships. As an example, the preset weights for organizational relationships, research relationships, business relationships, project relationships, alumni relationships, and behavioral relationships can decrease sequentially. For instance, for any direct relationship in any association path, if the direct relationship is only of one type, the preset weight of that type of relationship can be used as the type weight of the direct relationship; if the direct relationship is of multiple types, the preset weight with the highest preset weight among the multiple types of relationships can be used as the type weight of the direct relationship.

[0096] For example, calculating the product of the type weight and the relationship strength score of each direct relationship in the association path to obtain the relationship strength score of the association path may include: for any direct relationship in any association path, calculating the product of the type weight and the relationship strength score of the direct relationship as the score corresponding to the direct relationship; thus, traversing all direct relationships in the association path to obtain the scores of all direct relationships in the association path; and then calculating the product of the scores of all direct relationships in the association path, which is the relationship strength score of the association path.

[0097] In one possible implementation, the total score, after correcting the relationship strength score of each associated path using the path attenuation coefficient corresponding to each associated path, is determined as the relationship strength score of the indirect association. This can include: for any indirect association, correcting the relationship strength score of each associated path using the path attenuation coefficient corresponding to each associated path to obtain the corrected relationship strength score of each associated path; then summing the corrected relationship strength scores of all associated paths corresponding to the indirect association, and the total score obtained is the relationship strength score of the indirect association. In this way, in the process of calculating the relationship strength score of each indirect association, the relationship strength score of each associated path is calculated separately, corrected using the path attenuation coefficient, and the total score of the relationship strength scores of all associated paths is used as the final relationship strength score. This fully considers the influence of the relationship strength of each associated path and the path length (i.e., the path attenuation coefficient) on the relationship strength of the indirect association, thus making the calculated relationship strength score of the indirect association more accurate.

[0098] The path decay coefficient quantifies how much the influence of a relationship diminishes as the number of intermediaries increases. For any given path, the more intermediaries in the path, the smaller the path decay coefficient; conversely, the fewer intermediaries in the path, the larger the path decay coefficient. For example, the path decay coefficient can be determined based on the number of intermediaries, and its specific value can be set according to requirements. For instance, when there is one intermediary, the path decay coefficient is 0.9; when there are two intermediaries, the path decay coefficient is 0.8; when there are three intermediaries, the path decay coefficient is 0.7, and so on.

[0099] For example, for any indirect relationship, the relationship strength score of each associated path is corrected by the path attenuation coefficient corresponding to each associated path to obtain the corrected relationship strength score of each associated path. This may include: for any associated path, determining the path attenuation coefficient corresponding to the associated path based on the number of intermediate persons in the associated path; calculating the product of the path attenuation coefficient corresponding to the associated path and the relationship strength score of the associated path, and the score obtained by the product is the corrected relationship strength score of the associated path.

[0100] In one possible implementation, the calculation of the relationship strength score for each mined indirect association includes: calculating the relationship strength score for each indirect association using the following formula:

[0101] in, The score represents the strength of the indirect relationship between person i and person k. This represents the set of associated paths originating from person i and ending at person k. Represents the set of associated paths The path attenuation coefficient corresponding to the associated path p in the middle; This indicates the direct association relationship in the associated path p. Type weights; Indicates a direct relationship The strength score of the relationship.

[0102] In this way, by combining the path attenuation coefficient, relationship strength score, and type weight, this formula quantifies the relationship strength of the indirect association between person i and person k, and calculates the relationship strength score of the indirect association between person i and person k. It can accurately reflect the true closeness of the relationship between person i and person k.

[0103] In one possible implementation, the constructed enterprise personnel relationship network can support hierarchical display of first-degree and multi-degree relationships. As an example, in response to a user's command to view first-degree relationships in the enterprise personnel relationship network, the edges and nodes corresponding to the first-degree relationships are visualized on the user's terminal; in response to a user's command to view multi-degree relationships in the enterprise personnel relationship network, the edges and nodes corresponding to the multi-degree relationships are visualized on the user's terminal, thereby satisfying the user's need to view first-degree or multi-degree relationships.

[0104] In this embodiment, multi-source data corresponding to relevant personnel of an enterprise is obtained. The multi-source data includes: organizational data, project data, behavioral data, business data, research data, and alumni data. The relevant personnel of the enterprise include internal personnel and external collaborators. The multi-source data is matched based on preset association matching rules to identify direct relationships between relevant personnel of the enterprise. Direct relationships include the following types: organizational relationships, project relationships, behavioral relationships, business relationships, research relationships, and alumni relationships. A multi-dimensional relationship strength quantification model is used to quantify the relationship strength of each identified direct relationship, calculating a relationship strength score for each direct relationship. For any direct relationship, a weighted sum is calculated based on the dynamic weights corresponding to each first-type relationship, the relationship strength scores of each first-type relationship within the direct relationship, and the association frequency of each first-type relationship within the direct relationship, and then adjusted using a time decay factor. The weighted summation result is adjusted for timeliness, and the sum of the adjusted result and the relationship strength score of the second type of relationship is taken as the relationship strength score of the direct association. The first type of relationship includes: organizational relationships, project relationships, behavioral relationships, business relationships, and research relationships; the second type of relationship includes: alumni relationships. A graph structure modeling approach is used, with relevant personnel as nodes and direct relationships as edges, to construct a network of enterprise personnel relationships. A graph-based association matching algorithm is used to mine indirect relationships between relevant personnel in the enterprise based on the network, and the relationship strength score of each mined indirect relationship is calculated. Specifically, for any indirect relationship, the relationship strength score of each associated path is calculated, and the total score after adjusting the relationship strength score of each associated path using the path decay coefficient is determined as the relationship strength score of that indirect relationship. Thus, by acquiring multi-source data on these personnel's organizations, projects, behaviors, businesses, research, and institutions, the enterprise's business needs are comprehensively covered.Then, based on preset association matching rules, multi-source data is matched to identify direct relationships between relevant personnel in the enterprise. A multi-dimensional relationship strength quantification model is used to consider the different sensitivities of different relationship types to time factors. Relationships are divided into first-type and second-type relationships. Combining the dynamic weights of each first-type relationship, the relationship strength scores of each first-type relationship in the quantified direct relationships, the association frequency of each first-type relationship in the quantified direct relationships, the time decay factor of the quantified direct relationships, and the relationship strength scores of the second-type relationships, the first-type relationships are weighted and summed collaboratively. The time decay factor is then used to correct the timeliness of the weighted summation result. The second-type relationship, as a static association item, together with the time-corrected result, determines the final relationship strength score. This achieves refined quantification of the relationship strength of direct relationships between different personnel, and the calculated relationship strength score comprehensively and accurately reflects the true closeness of the direct relationships between relevant personnel in the enterprise. Furthermore, a graph-based modeling approach is adopted to establish a network of enterprise personnel relationships with relationship strength scores. Further, a graph-based association matching algorithm is used to comprehensively consider all indirect association paths and all direct associations within each path to calculate the relationship strength score of indirect associations. In calculating the relationship strength score of each indirect association, the relationship strength score for each path is calculated separately and corrected using a path decay coefficient. The total relationship strength score of all paths is then used as the final relationship strength score. This fully considers the influence of the relationship strength of each path and the path length (i.e., the path decay coefficient) on the relationship strength of indirect associations. Simultaneously, the relationship strength score on each path is jointly determined by the relationship strength scores of each segment of direct associations on that path. This fully considers the influence of the relationship strength of each direct association on the path, avoiding the neglect of other direct associations and other effective information. This ensures that the calculated relationship strength score of each indirect association more accurately reflects the closeness of the relationships between personnel. This enables a comprehensive and accurate understanding of the relationships between relevant personnel within an enterprise and quantifies the strength of these relationships. It allows for the construction of a real and quantifiable network of relationships among enterprise personnel, thereby achieving efficient management and application of enterprise personnel relationship resources.

[0105] In one possible implementation, the method further includes: dynamically updating the multi-source data according to update strategies corresponding to different types of data; wherein the update strategy includes one or more of the following: real-time update strategy, near real-time update strategy, and full update strategy; and adjusting the enterprise personnel relationship network according to the updated multi-source data.

[0106] Among them, the real-time update strategy refers to an update method driven by data change events, which synchronizes the changed content at the moment or within a very short time when the source data is added, deleted, or modified; the near real-time update strategy refers to an update method driven by timed scheduling or batch events, which actively pulls or receives incremental data at fixed time intervals (or when a preset batch threshold is reached), and completes the synchronization in a relatively short time; the full update strategy refers to an update method driven by fixed periods or manual instructions, which completely extracts, compares, and overwrites the source data at a predetermined time point.

[0107] Update strategies for different types of data can be pre-configured according to requirements. For example, high-frequency dynamic data can be configured with a real-time update strategy, non-real-time data can be configured with a near-real-time update strategy, and data with low change frequency can be configured with a full update strategy. As an example, organizational data can adopt a near-real-time update strategy, project data can adopt a near-real-time update strategy, behavioral data can adopt a real-time update strategy or a near-real-time update strategy; business data can adopt a near-real-time update strategy, scientific research data can adopt a full update strategy, and institutional data can adopt a full update strategy.

[0108] For example, data of the corresponding type can be updated based on the update strategy corresponding to each type of data, thereby realizing the dynamic updating of multi-source data corresponding to relevant personnel of the enterprise.

[0109] For example, adjusting the enterprise personnel relationship network based on the updated multi-source data may include: adjusting the nodes and edges in the enterprise personnel relationship network that involve the updated data based on the updated multi-source data corresponding to the relevant personnel of the enterprise; or, performing the above-mentioned adjustments based on the updated multi-source data corresponding to the relevant personnel of the enterprise. Figure 1 The operations in steps 102-105 will build the latest corporate personnel relationship network.

[0110] As an example, when data changes due to employee departures, the relevant data of departing employees can be retained when adjusting the company's personnel relationship network. This ensures that personnel relationship resources continue to accumulate as the company develops, and that these resources remain within the company after an employee leaves, thus preventing the loss of personnel relationship resources.

[0111] In this way, differentiated update strategies are set for different types of data, and a three-level dynamic update mechanism of real-time + near real-time + full update is adopted to update multi-source data. This enables the rapid capture of data changes such as organizational structure, personnel interaction, and project collaboration, and timely synchronization and updating of the enterprise's personnel relationship network to ensure that the data in the enterprise's personnel relationship network is highly consistent with the actual situation.

[0112] Furthermore, based on the enterprise personnel relationship network constructed above, efficient management and multi-scenario application of related relationship resources can be achieved, meeting the needs of enterprise relationship connection in actual operation.

[0113] The following provides examples illustrating possible application scenarios.

[0114] I. Relationship Query Users can enter an employee's ID or name to quickly query all personnel who are associated with that person (which can be direct or indirect relationships), as well as information such as the type of each relationship, relationship strength score, and the time of the most recent association, thus achieving a visual display of the relationships.

[0115] II. Path Recommendation Users can input the identification information of two people, and the system will automatically calculate and recommend the optimal connection path between the two people, identify the intermediary referrer, and solve the problem of finding people across departments and business lines.

[0116] III. Relationship Matching Based on users' personalized needs (such as "connecting with a hospital's oncology department" or "solving a technical problem"), the system automatically matches nodes with relevant resources, capabilities, or personnel relationships to achieve precise resource matching.

[0117] IV. Risk Control Verification: By combining a company's personnel relationship network, risk control verification can be carried out on behaviors such as personnel visits and customer interactions, identifying potential conflicts of interest and sensitive relationships, thereby improving the company's risk control and compliance level.

[0118] Thus, this embodiment of the disclosure adopts a full-process approach: multi-source processing acquisition – preprocessing – multi-source data matching – quantification of the strength of direct relationships – construction of enterprise personnel relationship networks – quantification of the strength of indirect relationships – dynamic updates – application. By collecting multi-source data from organizations, projects, behaviors, businesses, scientific research, and universities, and after preprocessing, matching, and strength quantification, an enterprise personnel relationship network is constructed with personnel as nodes and direct relationships with relationship strength scores as edges. Combined with real-time, near real-time, and full-data update mechanisms, the accuracy of the enterprise personnel relationship network is ensured, and application services such as relationship query and relationship path recommendation can be provided. This enables intelligent mining of personnel relationships, construction of enterprise personnel relationship networks, and multi-scenario applications.

[0119] This disclosure also provides a method for constructing an enterprise personnel relationship network and recommending associated paths.

[0120] Figure 2This diagram illustrates a flowchart of a method for constructing an enterprise personnel relationship network and recommending associated paths according to an embodiment of the present disclosure. Figure 2 As shown, it includes the following steps: Step 401: Obtain multi-source data corresponding to relevant personnel of the enterprise. The multi-source data includes: organizational data, project data, behavioral data, business data, scientific research data, and university data.

[0121] Step 402: Match the multi-source data based on the preset association relationship matching rules to identify the direct association relationships between relevant personnel of the enterprise.

[0122] Step 403: Using a multi-dimensional relationship strength quantification model, quantify the relationship strength of each identified direct relationship and calculate the relationship strength score for each direct relationship.

[0123] Step 404: Using a graph structure modeling approach, construct a network of relationships among enterprise personnel, with relevant personnel as nodes and direct relationships as edges.

[0124] Step 405: Using a graph-based association matching algorithm, indirect relationships between relevant personnel in the enterprise are mined from the enterprise personnel relationship network, and the relationship strength score of each mined indirect relationship is calculated.

[0125] Among them, steps 401-405 are the same as those mentioned above. Figure 1 Steps 101-105 are the same and will not be repeated here.

[0126] Step 406: Obtain the user's recommendation request; the recommendation request is used to request a recommendation path from the first person to the second person; wherein, the first person and the second person are any two different people in the enterprise's personnel relationship network.

[0127] For example, a user can input a recommendation request through a user terminal. The recommendation request may include identification information of two people (such as employee ID, name, etc.) and indicate the person who is the starting point (i.e., the first person) and the person who is the ending point (i.e., the second person) in the association path.

[0128] Step 407: Based on the enterprise personnel relationship network, determine at least one candidate association path.

[0129] For example, starting with the first person and ending with the second person, the direct and indirect relationships between these two nodes in the enterprise's personnel relationship network can be traversed to obtain all candidate relationship paths.

[0130] For example, if the candidate association path is not empty, it means that there is an association between the first person and the second person, and the following steps are continued; if the candidate association path is empty, it means that there is no association between the first person and the second person, and the following steps are stopped, and the user is informed that there is no association between the two people.

[0131] Step 408: Based on the relationship strength score of each direct relationship in each candidate association path, calculate the weight of each direct relationship in each candidate association path and sum them up to obtain the weight of each candidate association path; wherein, the weight of any direct relationship is negatively correlated with the relationship strength score of that direct relationship.

[0132] Since the weight of any direct relationship is negatively correlated with its relationship strength score (i.e., the higher the relationship strength score, the smaller the corresponding weight), for any candidate relationship path, considering both the relationship strength and the number of direct relationships, the weights of each direct relationship in the candidate relationship path are summed. The calculated weight of the candidate relationship path is determined by both the relationship strength score of each direct relationship and the number of direct relationships in the candidate relationship path. This comprehensively reflects the overall relationship strength of all direct relationships in the candidate relationship path. The smaller the calculated weight of the candidate relationship path, the higher the overall relationship strength of all direct relationships in the candidate relationship path; conversely, the larger the weight of the candidate relationship path, the lower the overall relationship strength of all direct relationships in the candidate relationship path.

[0133] In one possible implementation, the step of calculating the weights of each direct relationship in each candidate association path based on the relationship strength scores of each direct relationship in each candidate association path and summing them to obtain the weight of each candidate association path includes: calculating the weight of each candidate association path based on the relationship strength scores of each direct relationship in each candidate association path using the following formula:

[0134] in, This represents the weight of the candidate associated path P; The score represents the strength of the direct association between person i and person j. This represents any two individuals in a candidate association path P who have a direct relationship. For example, if the candidate association path P is: Person A - Person B - Person C - Person D, then... This includes (Personnel A, Personnel B), (Personnel B, Personnel C), and (Personnel C, Personnel D).

[0135] For example, The value range is 0-100.

[0136] In this way, the weight of the candidate association path is calculated based on the relationship strength score of each direct relationship in the candidate association path. The weight of the candidate association path is jointly determined by the relationship strength score of each direct relationship and the number of direct relationships in the candidate association path, thereby realizing the comprehensive consideration of the relationship strength of all direct relationships in the candidate association path to calculate the weight of the candidate association path.

[0137] Step 409: Determine the candidate association path with the smallest weight as the optimal association path to recommend to the user.

[0138] Since the smaller the weight of the calculated candidate association path, the higher the overall relationship strength of all direct relationships in the candidate association path, this step selects the candidate association path with the smallest weight as the optimal association path from the first person to the second person.

[0139] As an example, Dijkstra's Algorithm can be used to find the candidate association path with the smallest weight, which is the optimal association path recommended to the user; where the weight of the candidate association path can be calculated by the formula in step 408 above.

[0140] In this embodiment, a user's recommendation request is obtained; based on the enterprise personnel relationship network, at least one candidate association path is determined; based on the relationship strength score of each direct relationship in each candidate association path, the weight of each direct relationship in each candidate association path is calculated and summed to obtain the weight of each candidate association path; wherein, the weight of any direct relationship is negatively correlated with the relationship strength score of that direct relationship; the candidate association path with the smallest weight is determined as the optimal association path recommended to the user. Thus, the weight of a candidate association path is jointly determined by the relationship strength score of each direct relationship and the number of direct relationships in the candidate association path, thereby comprehensively considering the overall relationship strength of all direct relationships in the candidate association path, achieving optimal association path recommendation, and avoiding recommending weakly related or low-utility association paths to the user.

[0141] The following example uses a pharmaceutical company to illustrate how to construct a personnel relationship network for that company using the method described in this disclosure.

[0142] Step S1: Collect multi-source data corresponding to relevant personnel within the enterprise and perform preprocessing. (Corresponding to...) Figure 1 (Step 101) The multi-source data may include: Organizational data: employee ID (e.g., employee ID 5206456), name, department (e.g., "Oncology Innovative Products Division"), position (e.g., "Marketing Specialist"), superior-subordinate relationship (e.g., employee ID 5206456's superior employee ID is 5206450), historical organizational change records; Project data: project ID (e.g., XM2024001), project name (e.g., "Oncology Innovative Products Promotion Project"), employee IDs of participants (e.g., employee IDs 5206456, 5206450, 5206451), project division of labor, start and end dates (e.g., 2024-01-01 to 2024-12-31); Behavioral data: number of academic conferences (e.g., employees IDs 5206456, 5206450, and 5206452 all attended an academic conference on 2024-05-01). Number of meetings (1), number of academic visits registered (e.g., employee ID 5206456 and employee ID 5206450 have had 5 registered academic visits in the past 30 days), number of phone calls (e.g., employee ID 5206456 and employee ID 5206451 have had 3 phone calls in the past 30 days); Business data: Customer ID (e.g., KH2024001), Customer name (e.g., "Oncology Department of a certain hospital"), Contact person's employee ID (employee ID 5206456, employee ID 5206452); Research data: Article ID (e.g., WZ2026050), Article title (e.g., "Research on Tumor Treatment Based on xxxx"), Author ID (e.g., KH2024001, employee ID 5206450); University data: University (e.g., YX2026111), Graduated employees (e.g., KH2024001, employee ID 5206450).

[0143] S2: Multi-source data matching and identification of direct correlations (corresponding to...) Figure 1 (Step 102) The pre-defined matching rules for organizational relationships include: if two individuals work in the same department, they are considered to be in the same department; if one individual's superior employee ID equals the other's employee ID, they are considered to be in a direct superior-subordinate relationship. Pre-defined matching rules for project relationships include: if two individuals participate in the same project ID, they are considered to be in a joint project participation relationship. Pre-defined matching rules for behavioral relationships include: if two individuals attend meetings together at least once within 30 days, they are considered to have a meeting collaboration relationship; if two individuals visit and communicate with each other at least twice within 30 days, they are considered to have a visit and communication relationship; if two individuals talk on the phone at least three times within 30 days, they are considered to have a phone communication relationship. Pre-defined matching rules for business relationships include: if two individuals have the same customer ID and are both the contact person for that customer, they are considered to have a joint customer contact relationship. Pre-defined matching rules for research relationships include: if two individuals publish the same article, they are considered to have a joint article publication relationship. The pre-defined association matching rules for alumni relationships may include: if two people graduated from the same institution, then the two people are alumni.

[0144] Regarding organizational relationships: Employee ID 5206456's superior employee ID is 5206450, therefore, employee ID 5206456 and employee ID 5206450 have a direct superior-subordinate relationship. Employee IDs 5206456, 5206450, and 5206451 all belong to the Oncology Division, therefore, each employee ID is within the same department. Regarding project relationships: Employee IDs 5206456, 5206450, and 5206451 all participate in project ID XM2024001, therefore, each employee ID is involved in a shared project. Regarding behavioral relationships: Employee ID 5206456 and Employee ID 5206450 have visited each other 5 times in the past 30 days, therefore their relationship is one of visiting communication. Employee ID 5206456 and Employee ID 5206451 have spoken on the phone 3 times in the past 30 days, therefore their relationship is one of speaking on the phone. Employee IDs 5206456, 5206450, and 5206452 all participated in a meeting together, therefore each of them has a meeting collaboration relationship. Regarding business relationships: Employee IDs 5206456 and 5206452 both have the client ID KH2024001, therefore their relationship is one of shared client contact. Regarding research relationships: Employee ID 5206450 and client KH2024001 co-authored article WZ2026050, therefore, employee ID 5206450 and client KH2024001 have a co-authored article relationship. Regarding alumni relationships: Both employee ID 5206450 and client KH2024001 graduated from institution YX2026111, therefore, employee ID 5206450 and client KH2024001 have an alumni relationship.

[0145] There are direct relationships between employee IDs 5206456 and 5206450 (direct superior-subordinate relationship, same department relationship, joint project participation relationship, visit and communication relationship, meeting collaboration relationship); between employee IDs 5206456 and 5206451 (same department relationship, joint project participation relationship, telephone communication relationship); between employee IDs 5206456 and 5206452 (meeting collaboration relationship, joint client liaison relationship); between employee IDs 5206450 and 5206451 (same department relationship, joint project participation relationship); between employee IDs 5206450 and 5206452 (meeting collaboration relationship); and between employee ID 5206450 and client KH2024001 (jointly published articles relationship, alumni relationship).

[0146] S3: Quantification of the strength of direct relationships (corresponding to...) Figure 1 (Step 103) The direct association between employee ID 5206456 and employee ID 5206450 scores 77 points; the direct association between employee ID 5206456 and employee ID 5206451 scores 72 points; the direct association between employee ID 5206456 and employee ID 5206452 scores 65 points; the direct association between employee ID 5206450 and employee ID 5206451 scores 70 points; the direct association between employee ID 5206450 and employee ID 5206452 scores 45 points; and the direct association between employee ID 5206450 and customer KH2024001 scores 73 points.

[0147] S4: Building an Enterprise Personnel Relationship Network (corresponding to...) Figure 1 (Step 104) The edge between node-Employee ID5206456 and node-Employee ID5206450 has the following information: (Direct superior-subordinate relationship + same department relationship + joint project participation relationship + visit communication relationship + meeting collaboration relationship, 77 points, most recent association time 2024-05-01, relationship source: organizational data + project data + behavioral data); The edge between node-Employee ID5206456 and node-Employee ID5206451 has the following information: (Same department relationship + joint project participation relationship + phone call relationship, 72 points, most recent association time 2024-05-02, relationship source: organizational data + project data + behavioral data); The edge between node-Employee ID5206456 and node-Employee ID5206452 has the following information: (Meeting collaboration relationship + joint client liaison relationship, 65 points, most recent association time...). May 1, 2024 (Relationship Source: Behavioral Data + Business Data); Edge between Node - Employee ID 5206450 and Node - Employee ID 5206451, edge information is (same department relationship + joint project participation relationship, 70 points, most recent association time 2025-01-01, relationship source: organizational data + project data); Edge between Node - Employee ID 5206450 and Node - Employee ID 5206452, edge information is (meeting collaboration relationship, 45 points, most recent association time 2024-05-01, relationship source: behavioral data); Edge between Node - Employee ID 5206450 and Node - Customer KH2024001, edge information is (jointly published article relationship + alumni relationship, 73 points, most recent association time 2025-01-15, relationship source: scientific research data, university data).

[0148] S5: Indirect Relationship Mining (corresponding to...) Figure 1 (Step 105) There is an indirect relationship between Node-Employee ID 5206456 and Node-Customer KH2024001, with a relationship strength score of 48; there is an indirect relationship between Node-Employee ID 5206451 and Node-Employee ID 5206452, with a relationship strength score of 35; there is an indirect relationship between Node-Employee ID 5206452 and Node-Customer KH2024001, with a relationship strength score of 27.

[0149] Compared with the prior art, the above embodiments of this disclosure have at least the following beneficial effects: (1) By breaking through the limitations of existing technical data dimensions, it acquires multi-dimensional, multi-source data from organizations, projects, behaviors, businesses, scientific research, and universities, thereby reflecting the real personnel relationships in actual business collaborations within enterprises and ensuring the integrity and accuracy of these relationships. This allows it to meet the application needs of complex scenarios such as business referrals and cross-departmental collaborations. Furthermore, by acquiring multi-source data from both internal and external personnel, it covers internal personnel and integrates external relationships, thus meeting the diversified needs of cross-institutional and cross-industry collaborations within enterprises.

[0150] (2) A two-level collaborative strategy of basic matching and advanced matching is adopted to achieve accurate and comprehensive mining of association relationships. In basic matching, precise matching based on preset association relationship matching rules and fast matching based on hash are combined to achieve accurate identification of direct association relationships and efficient matching of large-scale data. In advanced matching, graph-based association matching algorithms are used to mine indirect association relationships. All association paths of indirect association relationships and all direct association relationships of each association path are considered to calculate the relationship strength score of indirect association relationships. The calculated relationship strength score of each indirect association relationship can more accurately reflect the closeness of the relationship between people. Thus, through the collaborative matching of association relationships by multiple algorithms, a complete matching process of "precise matching + efficient mining + deep reasoning" is formed, which takes into account both the accuracy of matching results and mining efficiency.

[0151] (3) Overcoming the limitations of existing technologies that lack a scientific system for quantifying relationship strength. Based on actual enterprise collaboration scenarios, six core dimensions are selected: organization, project, behavior, business, scientific research, and university. Through a multi-dimensional relationship strength quantification model, abstract direct relationships are transformed into quantified relationship strength scores, thereby clarifying the closeness of direct relationships. Among them, the multi-dimensional relationship strength quantification model combines the dynamic weights corresponding to each type of first-class relationship, the relationship strength scores of each type of first-class relationship in the direct relationships to be quantified, the association frequency of each type of first-class relationship in the direct relationships to be quantified, the time decay factor of the direct relationships to be quantified, and the relationship strength scores of the second-class relationships to achieve refined quantification of the relationship strength of direct relationships between different personnel. Furthermore, considering the varying sensitivities of different relationship types to time factors, relationships are categorized into Type I and Type II relationships. Type I relationships are typically less stable and more susceptible to time-related influences, while Type II relationships are generally more stable and less affected by time. Therefore, when calculating the relationship strength score for direct connections, different strategies are employed for Type I and Type II relationships. All Type I relationships are weighted and summed collaboratively, with the result adjusted for time decay using a time decay factor. Type II relationships, as static connections, together with the time-corrected result, determine the final relationship strength score. In this way, the calculated relationship strength score for direct connections comprehensively and accurately reflects the true closeness of direct relationships among relevant personnel within an enterprise; providing core quantitative evidence for connection path recommendation and connection matching, and meeting the needs of enterprises for precise collaboration and business referrals.

[0152] (4) Breaking through the limitations of the existing technology storage and modeling methods, the graph structure modeling method is used to construct the enterprise personnel relationship network. With personnel as nodes and the relationship with relationship strength score as edges, the relationship is transformed into a visual graph structure data, which supports the hierarchical display and fast query of indirect relationship. The graph database is used to store the relationship between nodes and edges, which greatly improves the efficiency of relationship query and relationship path calculation, thus adapting to the application scenario of large-scale enterprise personnel.

[0153] (5) Breaking through the limitation of existing technology in recommending related paths by only pursuing the shortest number of hops, when recommending related paths, the weight of candidate related paths is determined by the relationship strength score of each direct relationship and the number of direct relationships in the candidate related paths. Thus, the overall relationship strength of all direct relationships in the candidate related paths is comprehensively considered to achieve optimal related path recommendation, avoiding recommending weak and low-use related paths to users; achieving accurate optimal related path recommendation, solving the core pain point of "difficulty in finding people and slow connection" for enterprises across departments and business lines.

[0154] (6) By acquiring multi-source data corresponding to relevant personnel of the enterprise and constructing the enterprise personnel relationship network by adopting graph structure modeling, the personal relationship resources scattered in the private domain of individuals are transformed into enterprise-level standardized graph structure data, which can be stored in the unified graph database of the enterprise, away from the private domain channels, and realize the enterprise-level accumulation of relationship resources; and through the dynamic update mechanism, the relationship resources are continuously accumulated and iterated, ensuring that the relationship resources become intangible assets that can be reused by the enterprise for a long time, and reducing the cost of loss of relationship resources caused by personnel turnover.

[0155] Based on the same inventive concept in the above method embodiments, this disclosure also provides an enterprise personnel relationship network construction device, which can be used to execute the technical solutions described in the above method embodiments.

[0156] Figure 3 This diagram illustrates a structural diagram of an enterprise personnel relationship network building apparatus according to an embodiment of the present disclosure, such as... Figure 3As shown, the device may include: an acquisition module 501, used to acquire multi-source data corresponding to relevant personnel of an enterprise, the multi-source data including: organizational data, project data, behavioral data, business data, scientific research data, and university data; the relevant personnel of the enterprise include internal personnel and external collaborators; a matching module 502, used to match the multi-source data based on preset association matching rules to identify direct relationships between relevant personnel of the enterprise; wherein, direct relationships include the following types: organizational relationships, project relationships, behavioral relationships, business relationships, scientific research relationships, and alumni relationships; a quantification module 503, used to quantify the relationship strength of each identified direct relationship using a multi-dimensional relationship strength quantification model, and calculate the relationship strength score of each direct relationship; wherein, for any direct relationship, a weighted sum is performed based on the dynamic weights corresponding to each first type of relationship, the relationship strength scores of each first type of relationship in the direct relationship, and the association frequency of each first type of relationship in the direct relationship, and the timeliness correction of the weighted sum result is performed through a time decay factor. The sum of the timeliness-corrected result and the relationship strength score of the second type of relationship is taken as the relationship strength score of the direct relationship. The first type of relationship includes: organizational relationship, project relationship, behavioral relationship, business relationship, and research relationship; the second type of relationship includes: alumni relationship. The dynamic weight of each first type of relationship is negatively correlated with the interval between the most recent association time and the current time for the person corresponding to the direct relationship. A construction module 504 is used to construct an enterprise personnel relationship network using a graph structure modeling approach, with enterprise-related personnel as nodes and direct relationships as edges. A mining module 505 is used to use a graph-based association matching algorithm to mine indirect relationships between enterprise-related personnel based on the enterprise personnel relationship network, and calculate the relationship strength score of each mined indirect relationship. Specifically, for any indirect relationship, the relationship strength score of each association path corresponding to that indirect relationship is calculated, and the total score after correcting the relationship strength score of each association path using the path decay coefficient corresponding to each association path is determined as the relationship strength score of that indirect relationship.

[0157] In this embodiment, firstly, multi-source data on the organization, projects, behaviors, business, research, and institutions of these personnel are acquired to comprehensively cover the enterprise's business needs. Then, based on preset relationship matching rules, the multi-source data is matched to identify direct relationships between relevant personnel within the enterprise. A multi-dimensional relationship strength quantification model is used, considering the different sensitivities of different relationship types to time factors. Each type of relationship is divided into first-type relationships and second-type relationships. Combining the dynamic weights corresponding to each first-type relationship, the relationship strength scores of each first-type relationship in the quantified direct relationships, the association frequency of each first-type relationship in the quantified direct relationships, the time decay factor of the quantified direct relationships, and the relationship strength scores of the second-type relationships, the first-type relationships are collaboratively weighted and summed. The time decay factor is then used to correct the timeliness of the weighted summation result. The second-type relationship, as a static association item, together with the time-corrected result, determines the final relationship strength score. This achieves refined quantification of the relationship strength of direct relationships between different personnel, and the calculated relationship strength score comprehensively and accurately reflects the true closeness of the direct relationships between relevant personnel within the enterprise. Furthermore, a graph-based modeling approach is adopted to establish a network of enterprise personnel relationships with relationship strength scores. Further, a graph-based association matching algorithm is used to comprehensively consider all indirect association paths and all direct associations within each path to calculate the relationship strength score of indirect associations. In calculating the relationship strength score of each indirect association, the relationship strength score for each path is calculated separately and corrected using a path decay coefficient. The total relationship strength score of all paths is then used as the final relationship strength score. This fully considers the influence of the relationship strength of each path and the path length (i.e., the path decay coefficient) on the relationship strength of indirect associations. Simultaneously, the relationship strength score on each path is jointly determined by the relationship strength scores of each segment of direct associations on that path. This fully considers the influence of the relationship strength of each direct association on the path, avoiding the neglect of other direct associations and other effective information. This ensures that the calculated relationship strength score of each indirect association more accurately reflects the closeness of the relationships between personnel. This enables a comprehensive and accurate understanding of the relationships between relevant personnel within an enterprise and quantifies the strength of these relationships. It allows for the construction of a real and quantifiable network of relationships among enterprise personnel, thereby achieving efficient management and application of enterprise personnel relationship resources.

[0158] In one possible implementation, the apparatus further includes a recommendation module, configured to: obtain a user's recommendation request; the recommendation request is for requesting a recommendation path from a first person to a second person; wherein the first person and the second person are any two different persons in the enterprise personnel relationship network; determine at least one candidate association path based on the enterprise personnel relationship network; calculate the weight of each direct association in each candidate association path based on the relationship strength score of each direct association in each candidate association path and sum them up to obtain the weight of each candidate association path; wherein the weight of any direct association is negatively correlated with the relationship strength score of that direct association; and determine the candidate association path with the smallest weight as the optimal association path to recommend to the user.

[0159] In one possible implementation, the quantization module 503 is further configured to: calculate the relationship strength score of each first-type relationship and second-type relationship in the direct association based on the preset scoring rules corresponding to each first-type relationship and second-type relationship in the direct association; obtain the most recent association time of the person corresponding to the direct association, and determine the dynamic weight and the time decay factor corresponding to each first-type relationship in the direct association based on the most recent association time; obtain the association frequency of each first-type relationship in the direct association; perform a weighted summation of the product of the relationship strength score and the association frequency of each first-type relationship in the direct association based on the dynamic weight corresponding to each first-type relationship in the direct association, to obtain the initial value of the relationship strength of the direct association; calculate the product of the time decay factor and the initial value of the relationship strength to obtain the timeliness correction value of the relationship strength of the direct association; and sum the timeliness correction value and the relationship strength score of the second-type relationship to obtain the relationship strength score of the direct association.

[0160] In one possible implementation, the multi-dimensional relationship strength quantification model is shown in the following equation: +

[0161] in, The score represents the strength of the direct association between person i and person j. This represents the time decay factor corresponding to the direct correlation between person i and person j; Indicates the time of the most recent association between person i and person j; The score represents the strength of the organizational relationship between person i and person j. The score represents the strength of the relationship between personnel i and personnel j. The score represents the strength of the behavioral relationship between person i and person j. This represents the relationship strength score between personnel i and personnel j in terms of their business relationship. The score represents the strength of the research relationship between personnel i and personnel j. The dynamic weight representing the organizational relationship between personnel i and personnel j; The dynamic weight representing the project relationship between personnel i and personnel j; The dynamic weights representing the behavioral relationships between personnel i and personnel j; The dynamic weight representing the business relationship between personnel i and personnel j; The dynamic weight representing the research relationship between personnel i and personnel j; This indicates the frequency of the organizational relationship between person i and person j. This indicates the frequency of project associations between personnel i and personnel j. Indicates the frequency of association of the behavioral relationship between person i and person j; This indicates the frequency of business relationships between personnel i and personnel j. This indicates the frequency of the research relationship between personnel i and personnel j. The score represents the strength of the alumni relationship between person i and person j.

[0162] In one possible implementation, the mining module 505 is further configured to: for any association path, determine the type weight of each direct association in the association path according to the preset weight of each type of relationship in each direct association in the association path; obtain the relationship strength score of each direct association in the association path; and calculate the product of the type weight and the relationship strength score of each direct association in the association path to obtain the relationship strength score of the association path.

[0163] In one possible implementation, the mining module 505 is further configured to: calculate the relationship strength score for each indirect association using the following formula:

[0164] in, The score represents the strength of the indirect relationship between person i and person k. This represents the set of associated paths originating from person i and ending at person k. This represents the path attenuation coefficient corresponding to path p in the set of associated paths; This indicates the direct association relationship in the associated path p. Type weights; Indicates a direct relationship The strength score of the relationship.

[0165] In one possible implementation, the recommendation module is further configured to: Based on the relationship strength scores of each direct relationship in each candidate association path, the weight of each candidate association path is calculated using the following formula:

[0166] in, This represents the weight of the candidate associated path P; The score represents the strength of the direct association between person i and person j. This represents any two individuals in the candidate association path P who have a direct relationship.

[0167] In one possible implementation, the device further includes an update module, configured to: dynamically update the multi-source data according to update strategies corresponding to different types of data; wherein the update strategy includes one or more of the following: real-time update strategy, near real-time update strategy, and full update strategy; and adjust the enterprise personnel relationship network according to the updated multi-source data.

[0168] In some embodiments, the functions or modules of the apparatus provided in this disclosure can be used to perform the methods described in the above method embodiments. The specific implementation can be referred to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0169] This disclosure also provides an electronic device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the above method.

[0170] This disclosure also provides a non-volatile computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0171] This disclosure also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method.

[0172] Figure 4 A block diagram of an electronic device 1900 according to an embodiment of the present disclosure is shown. For example, the electronic device 1900 may be provided as a server or a terminal device. (Refer to...) Figure 4The electronic device 1900 includes a processing component 1922, which further includes one or more processors, and memory resources represented by memory 1932 for storing instructions, such as application programs, that can be executed by the processing component 1922. The application programs stored in memory 1932 may include one or more modules, each corresponding to a set of instructions. Furthermore, the processing component 1922 is configured to execute instructions to perform the methods described above.

[0173] Electronic device 1900 may also include a power supply component 1926 configured to perform power management of electronic device 1900, a wired or wireless network interface 1950 configured to connect electronic device 1900 to a network, and an input / output interface 1958 (I / O interface). Electronic device 1900 can operate on an operating system, such as Windows Server, stored in memory 1932. TM Mac OS X TM Unix TM Linux TM FreeBSD TM Or similar.

[0174] In an exemplary embodiment, a non-volatile computer-readable storage medium is also provided, such as a memory 1932 including computer program instructions that can be executed by a processing component 1922 of an electronic device 1900 to perform the above-described method.

[0175] Computer-readable storage media can be tangible devices capable of holding and storing programs / instructions used by instruction execution devices. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination of the foregoing. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0176] The computer program (or computer-readable program instructions) described herein can be downloaded from a computer-readable storage medium to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage medium in the respective computing / processing device.

[0177] The computer program (or computer program instructions) used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing state information from the computer-readable program instructions to implement various aspects of this disclosure.

[0178] Various aspects of this disclosure are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0179] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processor of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner; thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0180] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0181] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0182] The various embodiments of this disclosure have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or technical improvements to the embodiments in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A method for constructing an enterprise personnel relationship network, characterized in that, The method includes: Acquire multi-source data corresponding to relevant personnel of the enterprise. The multi-source data includes: organizational data, project data, behavioral data, business data, scientific research data, and university data. The relevant personnel of the enterprise include internal personnel and external partners. The multi-source data is matched based on preset association matching rules to identify direct relationships between relevant personnel in the enterprise; among which, direct relationships include the following types: organizational relationship, project relationship, behavioral relationship, business relationship, research relationship, and alumni relationship; A multi-dimensional relationship strength quantification model is used to quantify the relationship strength of each identified direct relationship and calculate the relationship strength score for each direct relationship. Specifically, for any direct relationship, a weighted sum is calculated based on the dynamic weights corresponding to each first-type relationship, the relationship strength scores of each first-type relationship within that direct relationship, and the association frequency of each first-type relationship within that direct relationship. The time decay factor is used to correct the timeliness of the weighted sum, and the sum of the time-corrected result and the relationship strength scores of second-type relationships is taken as the relationship strength score of that direct relationship. The first-type relationships include: organizational relationships, project relationships, behavioral relationships, business relationships, and research relationships; the second-type relationships include: alumni relationships. The dynamic weights corresponding to each first-type relationship are negatively correlated with the interval between the most recent association time of the person corresponding to that direct relationship and the current time. A graph structure modeling approach is used to construct a network of relationships among enterprise personnel, with relevant personnel as nodes and direct relationships as edges. A graph-based association matching algorithm is used to mine indirect relationships between relevant personnel in the enterprise based on the enterprise personnel relationship network, and calculate the relationship strength score of each mined indirect relationship. Specifically, for any indirect relationship, the relationship strength score of each associated path corresponding to the indirect relationship is calculated, and the total score after correcting the relationship strength score of each associated path by the path attenuation coefficient corresponding to each associated path is determined as the relationship strength score of the indirect relationship.

2. The method according to claim 1, characterized in that, The method further includes: Obtain a user's recommendation request; the recommendation request is used to request a recommendation path from a first person to a second person; wherein, the first person and the second person are any two different people in the enterprise's personnel relationship network; Based on the enterprise personnel relationship network, at least one candidate association path is identified; Based on the relationship strength scores of each direct relationship in each candidate association path, the weights of each direct relationship in each candidate association path are calculated and summed to obtain the weight of each candidate association path; wherein, the weight of any direct relationship is negatively correlated with the relationship strength score of that direct relationship. The candidate association path with the lowest weight is determined as the optimal association path to recommend to the user.

3. The method according to claim 1, characterized in that, For any direct association, a weighted sum is performed based on the dynamic weights corresponding to each first-type association, the relationship strength scores of each first-type association in the direct association, and the association frequencies of each first-type association in the direct association. The result of the weighted sum is then corrected for timeliness using a time decay factor. The sum of the time-corrected result and the relationship strength scores of the second-type associations is taken as the relationship strength score of the direct association, including: Based on the preset scoring rules corresponding to each first type of relationship and each second type of relationship in the direct relationship, calculate the relationship strength score of each first type of relationship and each second type of relationship in the direct relationship; Obtain the most recent association time of the person corresponding to the direct association, and determine the dynamic weight and time decay factor corresponding to each first type of relationship in the direct association based on the most recent association time; Obtain the association frequency of each first-type relationship in this direct relationship; Based on the dynamic weights corresponding to each first type of relationship in the direct association, the product of the relationship strength score and the association frequency of each first type of relationship in the direct association is weighted and summed to obtain the initial value of the relationship strength of the direct association. Calculate the product of the time decay factor and the initial value of the relationship strength to obtain the timeliness correction value of the relationship strength of the direct correlation; The relationship strength score of the direct association is obtained by summing the timeliness correction value and the relationship strength score of the second type of relationship.

4. The method according to claim 3, characterized in that, The multi-dimensional relationship strength quantification model is shown in the following equation: + in, The score represents the strength of the direct association between person i and person j. This represents the time decay factor corresponding to the direct correlation between person i and person j; Indicates the time of the most recent association between person i and person j; The score represents the strength of the organizational relationship between person i and person j. The score represents the strength of the relationship between personnel i and personnel j. The score represents the strength of the behavioral relationship between person i and person j. This represents the relationship strength score between personnel i and personnel j in terms of their business relationship. The score represents the strength of the research relationship between personnel i and personnel j. The dynamic weight representing the organizational relationship between personnel i and personnel j; The dynamic weight representing the project relationship between personnel i and personnel j; The dynamic weights representing the behavioral relationships between personnel i and personnel j; The dynamic weight representing the business relationship between personnel i and personnel j; The dynamic weight representing the research relationship between personnel i and personnel j; This indicates the frequency of the organizational relationship between person i and person j. This indicates the frequency of project associations between personnel i and personnel j. Indicates the frequency of association of the behavioral relationship between person i and person j; This indicates the frequency of business relationships between personnel i and personnel j. This indicates the frequency of the research relationship between personnel i and personnel j. The score represents the strength of the alumni relationship between person i and person j.

5. The method according to claim 1, characterized in that, The step of calculating the relationship strength score for each association path corresponding to any indirect association includes: For any association path, the type weight of each direct association in the association path is determined according to the preset weight of each type of relationship in each direct association path; Obtain the relationship strength score of each direct relationship in the associated path; Calculate the product of the type weight of each direct relationship in the association path and the relationship strength score to obtain the relationship strength score of the association path.

6. The method according to claim 5, characterized in that, The calculation of the relationship strength score for each indirect relationship discovered includes: calculating the relationship strength score for each indirect relationship using the following formula: in, The score represents the strength of the indirect relationship between person i and person k. This represents the set of associated paths originating from person i and ending at person k. This represents the path attenuation coefficient corresponding to path p in the set of associated paths; This indicates the direct association relationship in the associated path p. Type weights; Indicates a direct relationship The strength score of the relationship.

7. The method according to claim 2, characterized in that, The weight of each candidate association path is calculated based on the relationship strength score of each direct relationship in each candidate association path, and the weights of each direct relationship in each candidate association path are summed to obtain the weight of each candidate association path, including: Based on the relationship strength scores of each direct relationship in each candidate association path, the weight of each candidate association path is calculated using the following formula: in, This represents the weight of the candidate associated path P; The score represents the strength of the direct association between person i and person j. This represents any two individuals in the candidate association path P who have a direct relationship.

8. The method according to any one of claims 1-7, characterized in that, The method further includes: The multi-source data is dynamically updated according to the update strategy corresponding to different types of data; wherein the update strategy includes one or more of the following: real-time update strategy, near real-time update strategy, and full update strategy; The enterprise personnel relationship network is adjusted based on the updated multi-source data.

9. A device for building an enterprise personnel relationship network, characterized in that, The device includes: The acquisition module is used to acquire multi-source data corresponding to relevant personnel of the enterprise. The multi-source data includes: organizational data, project data, behavioral data, business data, scientific research data, and university data; the relevant personnel of the enterprise include internal personnel of the enterprise and external partners of the enterprise. The matching module is used to match the multi-source data based on preset association matching rules to identify direct relationships between relevant personnel of the enterprise; wherein, direct relationships include the following types: organizational relationship, project relationship, behavioral relationship, business relationship, scientific research relationship, and alumni relationship; The quantification module utilizes a multi-dimensional relationship strength quantification model to quantify the relationship strength of each identified direct relationship and calculate a relationship strength score for each direct relationship. Specifically, for any direct relationship, a weighted sum is calculated based on the dynamic weights corresponding to each first-type relationship, the relationship strength scores of each first-type relationship within that direct relationship, and the association frequency of each first-type relationship within that direct relationship. The time decay factor is used to correct the timeliness of the weighted sum, and the sum of the time-corrected result and the relationship strength scores of the second-type relationships is taken as the relationship strength score of that direct relationship. The first-type relationships include: organizational relationships, project relationships, behavioral relationships, business relationships, and research relationships; the second-type relationships include: alumni relationships. The dynamic weights corresponding to each first-type relationship are negatively correlated with the interval between the most recent association time of the person corresponding to that direct relationship and the current time. The module is used to construct a network of relationships among company personnel using a graph structure modeling approach, with relevant personnel as nodes and direct relationships as edges. The mining module is used to employ a graph-based association matching algorithm to mine indirect relationships between relevant personnel within the enterprise based on the enterprise personnel relationship network, and to calculate the relationship strength score for each mined indirect relationship. Specifically, for any indirect relationship, the relationship strength score for each associated path corresponding to that indirect relationship is calculated, and the total score after correcting the relationship strength score of each associated path by using the path decay coefficient corresponding to each associated path is determined as the relationship strength score of that indirect relationship.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 8.

Images