Method and device for analyzing relevance strength of social network relationship of employees

Abstract

The invention provides a staff social network relation relevance strength analysis method and device, and the method comprises the steps: building a secondary social relation index system, quantifying the direct influence between indexes, and constructing a direct influence matrix; carrying out normalization processing on the matrix by utilizing a DEMATEL method, and sequentially calculating a comprehensive influence matrix, factor weights and the like; constructing a directed network by taking the first-level indexes as clusters and the second-level indexes as nodes based on an ANP method, establishing a weight matrix and a hypermatrix through evaluation, and finally obtaining a mixed weight matrix as a target weight; and finally, obtaining and cleaning human resource structured data of the target object based on a preset data source, and quantifying the social network relationship relevance strength according to the secondary index system and the target weight. According to the technical scheme, the existing data in the human resource management of the specified industry can be effectively integrated, the social relation network between the staff can be accurately identified, the relation strength can be quantified, and a scientific and systematic evaluation model is provided for the human resource management service.

Description

Technical Field

[0001] The present invention relates to the field of data analysis technology, and in particular to a method and device for analyzing the strength of correlation of employee social network relationships. Background Art

[0002] In some industries, companies face unique challenges in their human resource management processes. For example, talent in the railway industry is highly specialized but relatively lacks versatility. This leads to frequent internal talent turnover and limited interaction with the wider community. In practical management practices, specific regulations, such as those regarding avoidance of appointments, require companies to accurately identify the interpersonal networks within their talent pool and quantify the strength of social relationships among employees. However, existing human resource management methods have significant shortcomings in this regard.

[0003] While the industry has accumulated a vast amount of human resources management data, much of this data is scattered across disparate systems, lacking effective integration and in-depth analysis. A mature, systematic assessment method for the key metric of social relationship strength has yet to be established. Traditional methods often offer only a simplistic analysis of social relationships along a single dimension, failing to comprehensively and comprehensively consider the strength of these relationships under the interplay of multiple factors. This limitation makes it difficult for companies to effectively mitigate risks stemming from interpersonal factors, such as conflicts of interest and nepotism, during key stages like recruitment and hiring, thereby impacting the scientific nature and fairness of human resources management. Summary of the Invention

[0004] In view of this, an embodiment of the present invention provides a method and apparatus for analyzing the strength of correlation of employee social network relationships, so as to eliminate or improve one or more defects existing in the prior art, and to solve the problem that the prior art cannot accurately analyze the correlation of employee social relationships.

[0005] One aspect of the present invention provides a method for analyzing the strength of correlation of employee social network relationships, the method comprising the following steps: Establish a secondary social relationship indicator system based on the characteristics of employees in the target field, quantify the direct impact scores between each social relationship indicator configuration, and establish a direct impact matrix; Based on the decision-making experiment and evaluation laboratory method, the direct impact matrix is ​​normalized, and the comprehensive impact matrix, the influence degree, the affected degree, the centrality, the cause degree, the factor weight, the overall impact matrix, the reachable threshold and the reachable matrix of each factor are calculated in turn; Based on the network analysis method, a directed network is established for the reachable matrix with the primary indicators in the social relationship indicator system as clusters and the secondary indicators as nodes. The inter-cluster relationships and inter-node influence relationships are evaluated to establish a cluster weight matrix and an unweighted supermatrix, which are then multiplied to obtain a weighted supermatrix. The limit supermatrix is ​​obtained for the weighted supermatrix to obtain a weight matrix for each social relationship indicator. After adjusting the order of the weight matrix to match the row order of the comprehensive influence matrix, the hybrid weight matrix is ​​normalized to obtain the target weight for each social relationship indicator. The human resources structured data of the two target objects are obtained and cleaned based on a preset data source, and the strength of the social network relationship between the two target objects is quantified based on the secondary social relationship indicator system and the target weight of each social relationship indicator.

[0006] In some embodiments, quantifying the direct influence score between each social relationship indicator configuration includes: The direct impact engineering degree scores between the social relationship indicators are assigned according to the scale of no impact, slight impact, general impact, large impact and serious impact, and the scoring process is calculated using the Pearson correlation coefficient or the Spearman rank correlation coefficient.

[0007] In some embodiments, the secondary social relationship indicator system includes three primary indicators: geographical and cultural relationship, blood and clan relationship, and professional and experience relationship; The geo-cultural relationship includes four secondary indicators: fellow townsmen, neighbors, peers, and peers; The blood and clan relationships include three secondary indicators: direct relatives, collateral relatives and the same ethnic group; The professional experience relationship includes the following five secondary indicators: classmates, same training, same major, same industry and colleagues; The same generation is divided into multiple generation intervals according to the year of birth of the employees for distinction and judgment.

[0008] In some embodiments, the direct impact matrix is ​​normalized based on the decision-making experiment and evaluation laboratory method, and the comprehensive impact matrix, the influence degree, the influenced degree, the centrality, the cause degree, the factor weight, the overall impact matrix, the reachable threshold and the reachable matrix of each factor are calculated in sequence, including: The direct impact matrix A is normalized and the calculation formula is: ; ; in, It indicates the degree of direct influence of factor i on factor j; n indicates the number of factors; The calculation formula of the comprehensive impact matrix is: ; Where I is the identity matrix and k is the matrix such that The smallest integer; The influence calculation formula of the i-th factor is: ; The calculation formula for the influence degree of the i-th factor is: ; The centrality calculation formula of the i-th factor is: ; The calculation formula of the cause degree of the i-th factor is: ; The factor weight calculation formula of the i-th factor is: ; The calculation formula of the overall impact matrix is: H=T+I; Wherein, I is the identity matrix, T is the comprehensive influence matrix; Set the threshold λ to the mean of T and compare the items in H The reachability matrix F is obtained with the size of λ, and the expression is: .

[0009] In some embodiments, the calculation formula of the limit supermatrix is: ; Wherein, W represents the weighted super matrix; The weight matrix is ​​constructed based on the factors in the extreme supermatrix, and the expression is: ; The order of the weight matrix is ​​adjusted to conform to the row order of the comprehensive influence matrix, and the expression is: ; Among them, H represents the overall influence matrix, Y represents the initial mixing weight matrix; The initial mixing weight matrix is ​​used to obtain the mixing weight matrix, which is expressed as follows: ; in, represents the weight value of the i-th social relationship indicator, Represents the value of the corresponding factor of the i-th social relationship indicator in the initial mixed weight matrix.

[0010] In some embodiments, the method further comprises: using a 9-degree partitioning scale to evaluate the inter-cluster relationships and inter-node influence relationships to establish a cluster weight matrix and an unweighted supermatrix.

[0011] In some embodiments, the method further comprises: A log is established for storing the secondary social relationship indicator system, the target weight of each social relationship indicator, and the strength of the social network relationship correlation between the two target objects, and an index is established for backtracking query.

[0012] On the other hand, the present invention also provides a device for analyzing the strength of correlation of employee social network relationships, comprising a processor, a memory, and a computer program / instruction stored in the memory, wherein the processor is used to execute the computer program / instruction, and when the computer program / instruction is executed, the device implements the steps of the above method.

[0013] On the other hand, the present invention further provides a computer-readable storage medium having a computer program / instruction stored thereon, which implements the steps of the above method when executed by a processor.

[0014] On the other hand, the present invention also provides a computer program product, comprising a computer program / instruction, which implements the steps of the above method when executed by a processor.

[0015] The beneficial effects of the present invention are at least: The method and device for analyzing the strength of employee social network relationship correlations, described in the present invention, establishes a secondary social relationship indicator system, quantifies the direct influence between indicators, and constructs a direct influence matrix. The matrix is ​​normalized using the DEMATEL method, and the comprehensive influence matrix and factor weights are sequentially calculated. A directed network is then constructed using the ANP method, with primary indicators as clusters and secondary indicators as nodes. After evaluation, a weight matrix and supermatrix are established, ultimately yielding a hybrid weight matrix as the target weight. Finally, structured human resource data of the target object is acquired and cleaned based on a preset data source, and the strength of social network relationship correlations is quantified based on the secondary indicator system and the target weights. Combined with the document content, this technical solution can effectively integrate existing data from human resource management in a specified industry, accurately identify social relationship networks between employees, and quantify relationship strengths. This addresses the problem of traditional methods that can only perform single-dimensional analysis and cannot comprehensively consider the mutual influence of multiple factors. It reduces the loss of indicator correlation information and calculation errors, providing a scientific and systematic evaluation model for human resource management, helping to effectively prevent risks in key links such as recruitment and employment, and improving the scientific and fair nature of management.

[0016] Additional advantages, objects, and features of the present invention will be set forth in part in the following description and will become apparent to those skilled in the art upon examination of the following or may be learned from practice of the present invention. The objects and other advantages of the present invention may be realized and obtained by the structures particularly pointed out in the description and drawings.

[0017] Those skilled in the art will understand that the purposes and advantages that can be achieved by the present invention are not limited to the above specific descriptions, and the above and other purposes that can be achieved by the present invention will be more clearly understood based on the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings described herein are used to provide a further understanding of the present invention, constitute a part of this application, and do not constitute a limitation of the present invention. In the drawings: Figure 1 Schematic diagram of a flow chart of a method for analyzing the strength of correlation of employees' social network relationships according to an embodiment of the present invention.

[0019] Figure 2 This is a schematic diagram of the relationship between the influence and the degree of influence of various social relationship indicators in the employee social network relationship correlation strength analysis method according to another embodiment of the present invention.

[0020] Figure 3 This is a schematic diagram of the relationship between the centrality and causality of various social relationship indicators in the employee social network relationship correlation strength analysis method according to another embodiment of the present invention.

[0021] Figure 4 This is a schematic diagram of the weights of factors corresponding to various social relationship indicators in the employee social network relationship correlation strength analysis method according to another embodiment of the present invention. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments and the accompanying drawings. Here, the exemplary embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0023] It should also be noted that, in order to avoid obscuring the present invention due to unnecessary details, the accompanying drawings only show structures and / or processing steps closely related to the solutions according to the present invention, while other details that are not closely related to the present invention are omitted.

[0024] It should be emphasized that the term "include / comprises" when used herein refers to the existence of features, elements, steps or components, but does not exclude the existence or addition of one or more other features, elements, steps or components.

[0025] Human resource management in some industries is characterized by a strong focus on specialized talent and limited versatility. This has led to significant internal turnover and limited external social interaction. In implementing specific management regulations, such as those for avoidance of appointments, identifying interpersonal networks within the talent pool and quantifying the strength of social relationships among employees are pressing challenges.

[0026] One aspect of the present invention provides a method for analyzing the strength of correlation of employee social network relationships. Figure 1 As shown, the method includes the following steps S101 to S104: Step S101: Establish a secondary social relationship indicator system based on the characteristics of employees in the target field, quantify the direct impact scores between each social relationship indicator configuration, and establish a direct impact matrix.

[0027] Step S102: normalize the direct impact matrix based on the decision experiment and evaluation laboratory method, and calculate the comprehensive impact matrix, the influence degree, the influenced degree, the centrality, the cause degree, the factor weight, the overall impact matrix, the reachable threshold and the reachable matrix in sequence.

[0028] Step S103: Based on the network analysis method, a directed network is established for the reachable matrix with the first-level indicators in the social relationship indicator system as clusters and the second-level indicators as nodes. The inter-cluster relationship and the inter-node influence relationship are evaluated to establish a cluster weight matrix and an unweighted super matrix, and then multiply them to obtain a weighted super matrix; the limit super matrix of the weighted super matrix is ​​calculated to obtain the weight matrix of each social relationship indicator; after adjusting the order of the weight matrix to match the row order of the comprehensive influence matrix, normalization is performed to obtain a mixed weight matrix as the target weight of each social relationship indicator.

[0029] Step S104: obtaining and cleaning the human resources structured data of the two target objects based on the preset data source, and quantifying the strength of the social network relationship between the two target objects based on the secondary social relationship indicator system and the target weight of each social relationship indicator.

[0030] In step S101, a comprehensive and detailed secondary social relationship indicator system is first established based on the characteristics of employees in the target field. This system must comprehensively consider employees' social relationships across multiple dimensions, including work, life, and education. Specifically, primary indicators can cover key aspects such as geographical and cultural relationships, blood ties, and professional connections, while secondary indicators are further refined into specific factors such as native place, birthplace, relatives, and classmates.

[0031] Taking the railway industry as an example, in some embodiments, the secondary social relationship indicator system includes three primary indicators: geographical and cultural relationships, blood and clan relationships, and professional and experience relationships.

[0032] Geocultural relationships include four secondary indicators: fellow townspeople, neighbors, peers, and same generation. Bloodline and clan ties include three secondary indicators: direct relatives, collateral relatives, and same ethnicity. Professional relationships include five secondary indicators: classmates, same training, same major, same industry, and same colleagues. Among these, the same generation is divided into multiple generations based on the employee's year of birth.

[0033] After establishing the indicator system, the degree of direct influence between each social relationship indicator can be quantified through methods such as expert scoring and questionnaire surveys. For example, relevant experts are organized to score the degree of direct influence between each pair of secondary indicators, and quantification is performed using a standard 4-point scale. In some embodiments, quantifying the degree of direct influence between each social relationship indicator configuration includes assigning a score to the degree of direct influence between the social relationship indicators according to a scale of 0 to 4 points (0 for no influence, 1 for slight influence, 2 for general influence, 3 for significant influence, and 4 for severe influence). The scoring process is calculated using the Pearson correlation coefficient or the Spearman rank correlation coefficient.

[0034] Finally, all the quantified scores are organized into a matrix, the direct impact matrix. The direct impact matrix has a dimension of 12×12, and element a in the matrix represents the degree of direct impact of factor i on factor j.

[0035] By establishing a secondary social relations indicator system, we can comprehensively and systematically cover all aspects of employee social relations, ensuring a comprehensive and detailed analysis. Quantifying the direct impact between indicators and constructing a direct impact matrix provides basic data support for subsequent analysis, making the quantitative analysis of social relations more scientific and accurate, and providing a reliable basis for subsequent calculations and weighting.

[0036] In step S102, the direct influence matrix constructed in step S101 is first normalized using the DEMATEL method. The purpose of normalization is to adjust the elements in the matrix to a relatively comparable range, typically employing the row-column sum maximum method. Specifically, the row and column sums of the direct influence matrix are calculated, the maximum value is found, and each element in the matrix is ​​then divided by this maximum value to obtain the normalized direct influence matrix. Next, using the DEMATEL method's calculation process, the comprehensive influence matrix, the influence degree, the influence degree, the centrality, the causal degree, the factor weight, the overall influence matrix, the reachability threshold, and the reachability matrix are calculated in sequence. By normalizing the direct influence matrix, the differences in dimensions and magnitudes between different indicators are eliminated, making the data more comparable and operational. The subsequent series of calculations systematically reveals the mutual influence and interaction mechanisms between various social relationship indicators. The comprehensive influence matrix reflects the direct and indirect influences between indicators. The influence degree and the influence degree reveal the influence and influence degree of each indicator, while the centrality and causal degree further clarify the importance of the indicator in the network and its driving or passive nature. The determination of factor weights provides a scientific basis for subsequent weight analysis, while the reachable threshold and reachable matrix help to clarify the reachable relationship between indicators, laying the foundation for the subsequent application of network analysis methods.

[0037] In some embodiments, the direct impact matrix is ​​normalized based on the decision-making experiment and the evaluation laboratory method, and the comprehensive impact matrix, the influence degree, the influenced degree, the centrality, the cause degree, the factor weight, the overall impact matrix, the reachable threshold and the reachable matrix of each factor are calculated in sequence, including steps S201 to S09: Step S201: Normalize the direct impact matrix A. The calculation formula is: ; ; in, It represents the degree of direct influence of factor i on factor j; n represents the number of factors.

[0038] Step S202: The calculation formula of the comprehensive impact matrix is: ; Where I is the identity matrix and k is the matrix such that The smallest integer.

[0039] Step S203: The influence calculation formula of the i-th factor is: ; Step S204: The influence degree of the i-th factor is calculated as: ; Step S205: The centrality calculation formula of the i-th factor is: ; Step S206: The calculation formula for the cause degree of the i-th factor is: ; Step S207: The factor weight calculation formula of the i-th factor is: ; Step S208: The calculation formula of the overall impact matrix is: H=T+I; Among them, I is the identity matrix and T is the comprehensive influence matrix.

[0040] Step S209: Set the threshold λ to the mean value of T and compare the items in H The reachability matrix F is obtained by adding the size of λ, and the expression is: .

[0041] In step S103, the reachability matrix obtained in step S102 is first processed using the Analytical Network Process (ANP). A directed network is constructed, using the primary indicators in the social relationship indicator system as clusters (e.g., geo-cultural relationships G, bloodline and clan relationships B, and professional and experience relationships I) and the secondary indicators as nodes (e.g., fellow townsman G1, relative B1, classmate I1, etc.). In this network, the connections between clusters and nodes are determined based on the reachability relationships in the reachability matrix. Next, the inter-cluster relationships and inter-node influence relationships are evaluated to establish a cluster weight matrix and an unweighted supermatrix. The cluster weight matrix is ​​derived from expert assessments of the relative importance of each cluster, while the unweighted supermatrix is ​​an initial judgment matrix derived from expert assessments of inter-node influence relationships, normalized, and subjected to consistency checks. The cluster weight matrix is ​​then multiplied by the unweighted supermatrix to obtain a weighted supermatrix. The weighted supermatrix is ​​then subjected to a limit square to obtain a limit supermatrix, thereby obtaining the weight matrices for each social relationship indicator. Finally, the order of the weight matrix is ​​adjusted to be consistent with the row order of the comprehensive influence matrix in step S102, and then normalized to obtain a mixed weight matrix, that is, the target weight of each social relationship indicator.

[0042] In some embodiments, the calculation formula of the limit supermatrix is: ; Where W represents the weighted super matrix; The weight matrix is ​​constructed based on the factors in the extreme supermatrix, and the expression is: ; The order of the weight matrix is ​​adjusted to match the row order of the comprehensive influence matrix, and the expression is: ; Among them, H represents the overall influence matrix, Y represents the initial mixing weight matrix; The initial mixing weight matrix is ​​obtained by mixing weight matrix, and the expression is: ; in, represents the weight value of the i-th social relationship indicator, Represents the value of the corresponding factor of the i-th social relationship indicator in the initial mixed weight matrix.

[0043] In some embodiments, the method further comprises: using a 9-degree partitioning scale to evaluate the inter-cluster relationships and inter-node influence relationships to establish a cluster weight matrix and an unweighted supermatrix.

[0044] The application of the ANP method fully considers the interdependence and feedback relationships among the various indicators in the social relationship indicator system, making the determination of weights more scientific and reasonable. Constructing a directed network with primary indicators as clusters and secondary indicators as nodes intuitively demonstrates the hierarchical structure and correlations between indicators. The cluster weight matrix FD and the unweighted supermatrix S are constructed to quantify the influence relationships between indicators from the perspectives of inter-cluster and inter-node, respectively, while the weighted supermatrix W integrates information from both levels. The calculation of the extreme supermatrix further reveals the long-term influence relationships between indicators, and the resulting weight matrix reflects the relative importance of each indicator in the entire social relationship network. The resulting hybrid weight matrix Z combines the advantages of the DEMATEL and ANP methods, considering both the direct influence and interdependence between indicators, providing a precise weighting basis for the subsequent quantification of social network relationship strength.

[0045] In step S104, structured human resource data for the two target subjects is first obtained from a pre-defined data source (e.g., railway industry human resource management systems A and B). This data typically includes basic personnel information, key family members, major educational background, and work resumes. After data acquisition, it needs to be cleaned to remove duplicate, erroneous, or incomplete information to ensure data accuracy and usability. For example, missing values ​​can be checked and supplemented, and format inconsistencies can be corrected. Next, based on the secondary social relationship indicator system established in step S101, the cleaned data is analyzed to extract information related to each secondary indicator. Then, combined with the target weights for each social relationship indicator obtained in step S103, the strength of the social network relationship between the two target subjects is quantified. For example, for employees Zhang San and Li Si, their scores on secondary indicators such as fellow townsmen and relatives are determined based on their native place, birthplace, and key family members. The score for each indicator is determined based on the actual situation; if the conditions are met, 10 points are awarded, and if not, 0 points are awarded. For quantitative indicators, scores are calculated based on specific calculation methods. Finally, the score of each indicator is multiplied by the corresponding weight and summed to obtain the total score of the social network relationship strength between the two employees.

[0046] By acquiring and cleaning structured human resources data of target individuals from pre-set data sources, data quality and reliability are ensured. Quantification based on a secondary social relationship indicator system and target weights allows for a comprehensive and integrated assessment of all aspects of the social relationship between two target individuals and their relative importance. This approach not only accurately reflects the strength of the social relationship but also provides a scientific and objective basis for human resources management decisions such as recruitment and employment, effectively mitigating risks associated with interpersonal relationships and enhancing the scientific and impartial nature of management.

[0047] In some embodiments, the method further includes: establishing a log for storing a secondary social relationship indicator system, a target weight of each social relationship indicator, and a strength of a social network relationship correlation between two target objects, and establishing an index for backtracking query.

[0048] On the other hand, the present invention also provides a device for analyzing the strength of correlation of employee social network relationships, comprising a processor, a memory, and a computer program / instruction stored in the memory, wherein the processor is used to execute the computer program / instruction, and when the computer program / instruction is executed, the device implements the steps of the above method.

[0049] The present invention will be described below in conjunction with a specific embodiment: To identify interpersonal networks within talent pools, determine the strength of social relationships among relevant employees, and utilize information technology to provide effective risk prevention in key steps and links of personnel selection and employment, this embodiment provides a method for analyzing the strength of employee social network relationships within the railway industry, utilizing large amounts of management data accumulated in human resource management-related systems.

[0050] First, we distinguish interpersonal social relationships from the perspectives of blood, geography, and industry. Blood relations refer to social relationships formed based on blood or physiological connections. Important blood relations include race, clan, lineage, family, and household. They serve to connect societies and groups, strengthen their cohesion, and ultimately form a solid whole. Geopolitical relations refer to the locational structure of human society, or the relationship between space and geographic location. They can be categorized as closed or open regional relations and contribute to social stability. Industry relations refer to complex social relationships formed based on the extensive division of labor among people. They develop on the basis of blood and geography. The strength of relationships can be reflected by the frequency of interpersonal interactions, emotional intensity, intimacy, reciprocal exchange, homogeneity, and shared interests.

[0051] By sorting out the data structures in two management systems related to railway employee human resources, a statistical analysis was conducted on the data table structures in the structured databases of the two systems involved. From the 384 data tables of four database users, 52 data items that can be directly used or can be analyzed and calculated through existing data were selected according to the classification of blood relationship, geographical relationship and professional relationship. After multiple rounds of screening, three primary indicators and 12 secondary indicators of social relationship indicators were finally formed, as shown in Table 1 below.

[0052] Table 1 Social relationship index table

[0053] To determine whether the people in Table 1 are of the same generation, we first make a generational division. For example, for domestic society, we can divide them into five generations: those born between 1949 and 1959, those born between 1960 and 1974, those born between 1975 and 1990, those born between 1991 and 1999, and those born between 2000 and 2006. We use the "10+X" division method (based on 10 years, with an allowance of 2 to 3 years), and take X as 2 to form the agreed generational division standard.

[0054] For the measurement of qualitative indicators, such as fellow townsmen, neighbors, peers, and peers, those who meet the qualitative judgment standards will get 10 points, and those who do not will get 0 points. The importance of the relevant scores will be further adjusted according to the weights.

[0055] For example, in the "Personnel Basic Information Table" of System A, if Personnel "A" and "B" have the same "Ethnicity," then they are considered to be of the same ethnicity. If they are of the Han ethnicity, 3 points are awarded; if they are of a minority ethnicity, 10 points are awarded.

[0056] In the "Major Educational Experience" table of system A, if the major educational experiences of person "A" and "B" are related, then "A" and "B" can be determined to be classmates. Using the information from the major educational experience table, including the start date of their education, school, department, and class, the relevant classmate relationships are defined as classmates, classmates of the same grade, alumni with time overlap, and alumni without time overlap. These relationships are assigned weights W as shown in Table 2.

[0057] Table 2: Classmate relationship weight table When there is a time overlap, further calculation is required. Let N be the duration of the relevant classmate relationship, expressed in months. The classmate relationship index calculation formula is as follows.

[0058] .

[0059] In the "Main Training Experience" table of system B, if the main training experiences of personnel "A" and "B" are related, then "A" and "B" can be determined to be co-trained. The training course and course type information in the main training experience table are retrieved. Based on the training course type dictionary, five relevant co-training relationships are defined and weighted W, as shown in Table 3 below.

[0060] Table 3: Weight table of peer training relationships Calculate the duration of the relevant peer training relationship as N, in days. The following formula is used to calculate the peer training relationship index. For online topics and online self-study, there is no relationship duration, so the default value is 1 day.

[0061] .

[0062] In the "Resumes" table of system A, if the resumes of personnel "A" and "B" are related, then "A" and "B" can be determined to be colleagues. The unit, department, and start and end date information in the resume table are taken. The relevant colleague relationships are defined into the following six categories and assigned weights W as shown in Table 4.

[0063] Table 4: Colleague relationship weight table Calculate the duration of the relevant colleague relationship as N, in months. The colleague relationship index calculation formula is as follows.

[0064] .

[0065] The DEMATEL-ANP analysis process is as follows: 1. The main process of DEMATEL method to analyze the reachability matrix 1.1 Constructing a direct impact matrix Experts score the degree of direct influence between each factor and construct the direct influence matrix A.

[0066] ; in, It represents the degree of direct influence of factor i on factor j; n represents the number of factors.

[0067] 1.2 Normalized direct impact matrix The row, column and maximum value method is used for normalization. The row, column and maximum value of the direct influence matrix are taken out, and then all elements in the A matrix are divided by the value to calculate the normalized direct influence matrix D.

[0068] 1.3 Calculate the comprehensive impact matrix Calculate the comprehensive influence matrix T.

[0069] ; Where I is the identity matrix and k is the matrix such that The smallest integer.

[0070] 1.4 Calculating the impact Calculate the impact of each factor and the degree of impact .

[0071] ; ; 1.5 Calculating centrality and causality Calculate the centrality of each factor and causal degree .

[0072] ; ; 1.6 Calculating Factor Weights Calculate the weight of each factor .

[0073] ; 1.7 Output Impact Diagram The horizontal axis is , which represents the sum of the degree of influence and the degree of being influenced by the factor, reflecting the importance of the factor; the vertical axis is , which represents the difference between the degree of influence of a factor and the degree of influence, and reflects the influence relationship of the factors.

[0074] 1.8 Calculating the reachability matrix Calculate the overall influence matrix H=T+I, where I is the unit matrix and T is the comprehensive influence matrix. Set the threshold λ to the mean value of T, and compare the various The reachability matrix F is obtained with the size of λ.

[0075] .

[0076] 2. Main process of ANP method analysis of indicator weights 2.1 Building a Directed Network The reachability matrix calculated using the DEMATEL method serves as the input for the correlations and mutual influences between indicators, constructing a directed network. Primary indicators serve as clusters, and secondary indicators serve as nodes. Due to the different definitions of influence relationships in the DEMATEL and ANP methods, the DEMATEL reachability matrix, after row and column transposition, can be used as the input for the correlation and influence relationships between nodes in the ANP method.

[0077] .

[0078] 2.2 Expert evaluation Based on the network relationships formed by the inter-criteria dependence or self-feedback influence, a pairwise comparison matrix was established according to the correlation results evaluated by the reachability matrix, and an expert evaluation was performed using a 9-degree scale of relationship influence.

[0079] First, experts fill in and organize the judgment matrix between clusters and establish the cluster weight matrix FD.

[0080] The experts then fill in and sort out the judgment matrix between the nodes. After obtaining the initial judgment matrix, normalization and consistency testing are performed. After multiple rounds of sorting and feedback, the unweighted supermatrix S is obtained.

[0081] 2.3 Establishing the weighted supermatrix Multiply the cluster weight matrix FD with the unweighted supermatrix S to obtain the weighted supermatrix W.

[0082] 2.4 Calculating the Limit Supermatrix The elements in the weighted supermatrix W calculated by the above steps represent the first-order dominance between elements. In order to calculate the second-order dominance between elements, it is necessary to calculate ; To calculate the cubic dominance between elements, we need to calculate By analogy, the limit weighted supermatrix is ​​obtained. To facilitate the calculation and use of the limit weighted supermatrix, three proven theorems are given.

[0083] Theorem 1, let A be an n-order non-negative matrix, is its modulo maximum eigenvalue, then .

[0084] .

[0085] Theorem 2: Assume that the largest characteristic root 1 of the non-negative column random matrix A is a single root and the moduli of the other characteristic roots are all less than 1, then exists, and All columns of are the same, they are all normalized eigenvectors of A belonging to 1.

[0086] Theorem 3, let A be a non-negative irreducible random matrix, then The necessary and sufficient condition for existence is that A is a prime matrix.

[0087] Therefore, there is is the limit supermatrix. Based on the limit weighted supermatrix, the weight matrix of each indicator can be obtained .

[0088] 3. DEMALTEL-ANP hybrid index weight calculation The DEMALTEL-ANP hybrid indicator weight is calculated by mixing the DEMATEL impact degree and ANP weight.

[0089] 3.1 Constructing the initial mixing weight matrix First, adjust the order of the ANP weight matrix X to match the row order of the DEMALTEL overall influence matrix H. Then the initial hybrid weight matrix Y can be calculated according to the following formula.

[0090] .

[0091] 3.2 Normalized Mixing Weight Matrix Normalizing the initial mixing weight matrix Y can obtain the final mixing weight matrix Z.

[0092] .

[0093] For example, the analysis process of a social relationship strength index system is given as follows: A. Constructing a direct impact matrix A standard 4-point scale (0–4, with 0 = no impact, 1 = slight impact, 2 = moderate impact, 3 = significant impact, and 4 = severe impact) was used to quantify the impact relationships between factors. A survey of 12 experts was conducted. Ultimately, 12 valid questionnaires were collected, and the direct impact matrix A was constructed using arithmetic sum methods, as shown in Table 5.

[0094] Table 6: Direct Impact Matrix A B. Output calculation results using the DEMATEL method After normalizing the direct impact matrix A to form the total impact matrix, the overall impact matrix, influence degree and affected degree, centrality and cause degree, factor weight, reachable threshold and reachable matrix are calculated in turn as shown in Tables 6 and 7, as well as Figure 2 、 Figure 3 、 Figure 4 .

[0095] Table 6: Overall impact matrix Table 7: Influence, influence, centrality, cause and weight of each secondary indicator The mean value of the comprehensive impact matrix is ​​calculated to obtain the reachable matrix threshold λ value of 0.093354. The reachable matrix is ​​calculated using the λ value on the overall impact matrix, as shown in Table 8 below.

[0096] Table 8: Reachability Matrix C. ANP weight analysis of indicator items Construct a directed network and use the reachability matrix F in DEMATEL to construct the directed network of ANP. In this paper, YAANP software is used to perform ANP calculations and comprehensively construct the network structure of social relationship strength evaluation indicators.

[0097] The evaluation table was input to calculate the extreme relationship matrix and weights. The 9-degree scale was used to conduct expert evaluation of the inter-cluster relationship. After sorting, the relevant judgment matrix was obtained. The cluster weight matrix was then calculated as shown in Table 9 below.

[0098] Table 9: Cluster weight matrix Using a 9-degree scale, we conducted expert evaluations of node relationships. This analysis yielded a correlation judgment matrix, which was then normalized and weighted to create a weighted supermatrix. This was then multiplied by its own limit to yield a limit supermatrix. The correlation matrices are shown in Tables 10 and 11.

[0099] Table 10: Weighted Super Matrix Table 11: Extreme Supermatrix Finally, the corresponding weights are calculated based on the data of the extreme supermatrix, as shown in Table 12 below.

[0100] Table 12: Global weights of indicators Taking the global weights of the ANP indicators (Table 12) and the overall influence matrix obtained in the DEMATEL method (Table 6), the mixed weights of the social relationship strength indicators can be obtained using the matrix calculation method described above (Table 13).

[0101] Table 13: Mixed weights of social relationship strength indicators D. Application Examples of the Social Relationship Index System After extracting, desensitizing, cleaning, and organizing data from systems A and B, and structuring it, we generated correlation index data. We then randomly selected a group of individuals within the same department and calculated their social network relationship strength, successfully determining the strength of their social network relationships. The final social relationship results are shown in Table 14.

[0102] Table 14: Examples of application of social relationship strength indicators This paper proposes a weighted analysis method for social network relationship strength indicators based on a combination of DEMATEL and ANP methods, taking into account the mutual influence between indicators, objective limitations of data sources, and differences in indicator metrics. The resulting social network relationship strength assessment model effectively reflects the mutual influence between corresponding indicators, effectively reducing the loss of indicator correlation information and computational errors that would occur in discrete indicator weight assessments. The corresponding assessment model is intended to be integrated into relevant railway human resources management information systems to provide strong support for railway human resources management operations.

[0103] In summary, the method and device for analyzing the strength of employee social network relationship correlations, described in the present invention, establishes a secondary social relationship indicator system, quantifies the direct influence between indicators, and constructs a direct influence matrix. The matrix is ​​normalized using the DEMATEL method, and the comprehensive influence matrix, factor weights, and other factors are calculated sequentially. A directed network is then constructed using the ANP method, with primary indicators as clusters and secondary indicators as nodes. After evaluation, a weight matrix and supermatrix are established, ultimately yielding a hybrid weight matrix as the target weight. Finally, structured human resource data of the target object is acquired and cleaned based on a preset data source, and the strength of social network relationship correlations is quantified based on the secondary indicator system and target weights. Combined with the document content, this technical solution can effectively integrate existing data from human resource management in a specified industry, accurately identify social relationship networks between employees, and quantify relationship strength. This addresses the problem of traditional methods that can only analyze in a single dimension and cannot comprehensively consider the mutual influence of multiple factors. It reduces the loss of indicator correlation information and calculation errors, providing a scientific and systematic evaluation model for human resource management, helping to effectively prevent risks in key links such as recruitment and employment, and improving the scientific and fair nature of management.

[0104] It should be understood that the present invention is not limited to the specific configurations and processes described above and illustrated in the figures. For the sake of brevity, a detailed description of known methods is omitted. In the above embodiments, several specific steps are described and illustrated as examples. However, the method of the present invention is not limited to the specific steps described and illustrated. Those skilled in the art may make various changes, modifications, and additions, or change the order of the steps after understanding the spirit of the present invention.

[0105] In the present invention, features described and / or illustrated for one embodiment may be used in the same or similar manner in one or more other embodiments, and / or combined with or replace features of other embodiments.

[0106] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations to the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A method for analyzing the strength of correlation of employee social network relationships, characterized in that: The method comprises the following steps: Establish a secondary social relationship indicator system based on the characteristics of employees in the target field, quantify the direct impact scores between each social relationship indicator configuration, and establish a direct impact matrix; Based on the decision-making experiment and evaluation laboratory method, the direct impact matrix is ​​normalized, and the comprehensive impact matrix, the influence degree, the affected degree, the centrality, the cause degree, the factor weight, the overall impact matrix, the reachable threshold and the reachable matrix of each factor are calculated in turn; Based on the network analysis method, a directed network is established for the reachable matrix with the primary indicators in the social relationship indicator system as clusters and the secondary indicators as nodes. The inter-cluster relationships and inter-node influence relationships are evaluated to establish a cluster weight matrix and an unweighted supermatrix, which are then multiplied to obtain a weighted supermatrix. The limit supermatrix is ​​obtained for the weighted supermatrix to obtain a weight matrix for each social relationship indicator. After adjusting the order of the weight matrix to match the row order of the comprehensive influence matrix, the hybrid weight matrix is ​​normalized to obtain the target weight for each social relationship indicator. The human resources structured data of the two target objects are obtained and cleaned based on a preset data source, and the strength of the social network relationship between the two target objects is quantified based on the secondary social relationship indicator system and the target weight of each social relationship indicator.

2. The method for analyzing the strength of correlation of employee social network relationships according to claim 1, characterized in that: Quantify the direct impact scores between each social relationship indicator configuration, including: The direct impact engineering degree scores between the social relationship indicators are assigned according to the scale of no impact, slight impact, general impact, large impact and serious impact, and the scoring process is calculated using the Pearson correlation coefficient or the Spearman rank correlation coefficient.

3. The method for analyzing the strength of correlation of employee social network relationships according to claim 1, characterized in that: The secondary social relationship indicator system includes three primary indicators: geographical and cultural relationship, blood and clan relationship, and professional and experience relationship; The geo-cultural relationship includes four secondary indicators: fellow townsmen, neighbors, peers, and peers; The blood and clan relationships include three secondary indicators: direct relatives, collateral relatives and the same ethnic group; The professional experience relationship includes the following five secondary indicators: classmates, same training, same major, same industry and colleagues; The same generation is divided into multiple generation intervals according to the year of birth of the employees for distinction and judgment.

4. The method for analyzing the strength of correlation of employee social network relationships according to claim 1, characterized in that: Based on the decision-making experiment and evaluation laboratory method, the direct impact matrix is ​​normalized, and the comprehensive impact matrix, the influence degree, the affected degree, the centrality, the cause degree, the factor weight, the overall impact matrix, the reachable threshold and the reachable matrix of each factor are calculated in turn, including: The direct impact matrix A is normalized and the calculation formula is: ； ； in, It indicates the degree of direct influence of factor i on factor j; n indicates the number of factors; The calculation formula of the comprehensive impact matrix is: ； Where I is the identity matrix and k is the matrix such that The smallest integer; The influence calculation formula of the i-th factor is: ； The calculation formula for the influence degree of the i-th factor is: ； The centrality calculation formula of the i-th factor is: ； The calculation formula of the cause degree of the i-th factor is: ； The factor weight calculation formula of the i-th factor is: ； The calculation formula of the overall impact matrix is: H=T+I; Wherein, I is the identity matrix, T is the comprehensive influence matrix; Set the threshold λ to the mean of T and compare the items in H The reachability matrix F is obtained with the size of λ, and the expression is: 。 5. The method for analyzing the strength of correlation of employee social network relationships according to claim 1, characterized in that: The calculation formula of the limit supermatrix is: ； Wherein, W represents the weighted super matrix; The weight matrix is ​​constructed based on the factors in the extreme supermatrix, and the expression is: ； The order of the weight matrix is ​​adjusted to conform to the row order of the comprehensive influence matrix, and the expression is: ； Among them, H represents the overall influence matrix, Y represents the initial mixing weight matrix; The initial mixing weight matrix is ​​used to obtain the mixing weight matrix, which is expressed as follows: ； in, represents the weight value of the i-th social relationship indicator, Represents the value of the corresponding factor of the i-th social relationship indicator in the initial mixed weight matrix.

6. The method for analyzing the strength of correlation of employee social network relationships according to claim 1, characterized in that: The method further includes: using a 9-degree partitioning scale to evaluate the inter-cluster relationship and the inter-node influence relationship to establish a cluster weight matrix and an unweighted super matrix.

7. The method for analyzing the strength of correlation of employee social network relationships according to claim 1, characterized in that: The method further comprises: A log is established for storing the secondary social relationship indicator system, the target weight of each social relationship indicator, and the strength of the social network relationship correlation between the two target objects, and an index is established for backtracking query.

8. A device for analyzing the strength of correlation of employee social network relationships, comprising a processor, a memory, and a computer program / instruction stored in the memory, characterized in that: The processor is configured to execute the computer program / instructions. When the computer program / instructions are executed, the device implements the steps of the method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images