Method for improving internal communication efficiency of organization based on social network analysis

Abstract

The invention discloses an organization internal communication efficiency improvement method based on social network analysis. The method comprises the following steps: S1, collecting internal communication data of an organization; s2, preprocessing and standardizing the collected data; s3, constructing a communication network graph, and identifying communication frequency and intensity; s4, identifying communication bottlenecks and key nodes in combination with a social physics model and a graph attention network; s5, introducing a network fluid dynamic model and a swarm intelligence method to optimize a communication strategy; s6, continuously optimizing the network structure based on an adaptive optimization method and a context awareness technology; and S7, generating personalized communication optimization suggestions, and dynamically displaying communication network changes. According to the invention, the communication efficiency and decision-making ability in the organization are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis and mining, and in particular to a method for improving internal communication efficiency of an organization based on social network analysis. Background Art

[0002] First, with the rapid development of information technology and the advancement of globalization, communication within organizations has become increasingly complex, covering multiple dimensions of communication methods such as email, instant messaging, meeting records, and employee behavior data. These communication methods have become the core of daily operations in modern enterprises. However, with the diversification of communication methods and the surge in communication volume, how to effectively manage and optimize these communication channels and improve the communication efficiency within the organization has become an urgent problem to be solved. Existing communication management tools often only provide basic communication records and simple statistical analysis, and lack the ability to deeply analyze and optimize communication efficiency.

[0003] Secondly, traditional communication management methods are usually based on fixed communication channels and hierarchical structures, lack the ability to adjust dynamically, and are difficult to cope with complex organizational structures and changing communication needs. For example, in the traditional hierarchical communication model, information transmission is usually completed through step-by-step reporting between superiors and subordinates. This model may work effectively when the organizational structure is simple and the amount of information is small, but in modern complex large organizations, this model often leads to slow information transmission, serious communication bottlenecks, and an inability to respond quickly to external changes. In addition, traditional communication efficiency evaluation often relies on the subjective judgment of managers, lacks objective quantitative indicators, and cannot provide accurate optimization suggestions.

[0004] Third, existing social network analysis technologies are mainly used for data analysis and marketing promotion of external social networks, and their application within organizations is relatively limited. Even if there are some attempts to apply social network analysis to organizational management, they are usually limited to simple node and edge analysis, and cannot fully utilize the complexity of social networks to optimize communication within organizations. Existing technologies lack in-depth understanding and analysis of complex communication networks within organizations, especially in terms of multi-dimensional data fusion and dynamic optimization. In addition, existing communication optimization tools often only focus on a single communication channel and cannot comprehensively consider the mutual influence between multiple communication methods, resulting in unsatisfactory optimization effects of communication strategies.

[0005] In summary, the existing technologies have the following major defects in dealing with the complex communication needs within modern organizations: communication management tools lack in-depth analysis and dynamic optimization capabilities, making it difficult to improve overall communication efficiency; traditional communication management models lack flexibility and cannot quickly respond to changes in the internal and external environment of the organization; existing social network analysis technologies have not been fully applied to the optimization of internal communication within organizations, and lack multi-dimensional data fusion analysis capabilities. Based on the above problems, how to provide a method for improving internal communication efficiency within organizations based on social network analysis is an urgent problem that technicians in this field need to solve. Summary of the invention

[0006] One purpose of the present invention is to propose a method for improving internal communication efficiency of an organization based on social network analysis. The present invention describes in detail how to effectively collect and preprocess communication data, build and analyze the communication network structure within the organization, and dynamically optimize communication strategies by combining real-time data and context-aware technology, by comprehensively applying social physics models, graph attention networks, network fluid dynamics models, and swarm intelligence optimization methods. This method has the advantages of accurately identifying communication bottlenecks, dynamically adjusting communication paths, and improving overall communication efficiency and decision-making capabilities, especially in complex organizational structures and changing communication environments.

[0007] A method for improving internal communication efficiency of an organization based on social network analysis according to an embodiment of the present invention includes the following steps:

[0008] S1. Collect communication data within the organization, including emails, instant messaging records, meeting minutes, and employee behavior data;

[0009] S2. Preprocess the collected communication data, remove noise data, and perform standardization;

[0010] S3. Construct a communication network diagram within the organization, where nodes represent individuals or departments and edges represent communication frequency and intensity;

[0011] S4. Combining social network analysis technology based on social physics models with graph attention networks, we can identify communication bottlenecks and key nodes of information flow by simulating information dissemination and interaction between individuals, and accurately evaluate the influence and communication efficiency of each node by weight distribution between nodes.

[0012] S5. Introduce the network fluid dynamics model and combine it with the adaptive optimization method based on swarm intelligence to simulate the propagation of information flow in the communication network and dynamically adjust and optimize the communication strategy;

[0013] S6. Use adaptive optimization methods to continuously optimize the communication network structure based on real-time data and simulation results, and combine situational awareness technology to dynamically display changes in the communication network on a visual interface;

[0014] S7. Generate personalized communication optimization suggestions and automated reports, use natural language processing technology based on emotional computing to generate report content, and combine employee feedback and historical data for continuous improvement.

[0015] Optionally, S3 includes the following steps:

[0016] S31, extracting a unique identifier of each individual or department as a node according to the preprocessed comprehensive feature matrix;

[0017] S32, calculating the communication frequency between nodes, and defining a communication frequency matrix, wherein each element represents the number of communications between nodes;

[0018] S33, calculate the communication intensity between nodes, define a communication intensity matrix, where each element represents the communication intensity between nodes, and is calculated based on factors such as communication frequency, message length, and time interval;

[0019] S34. Based on the communication frequency matrix and the communication intensity matrix, construct a preliminary communication network diagram, where nodes represent individuals or departments and edges represent communication frequency and intensity;

[0020] S35. Use graph theory methods to analyze the initially constructed communication network diagram, remove isolated nodes and abnormal edges, and optimize the network structure;

[0021] S36, topologically sorting the optimized communication network graph to generate a directed acyclic graph structure for further analysis and processing;

[0022] S37. Store the finally constructed communication network diagram into the database for subsequent analysis and processing.

[0023] Optionally, S4 includes the following steps:

[0024] S41. Construct a multi-level interaction matrix based on the final communication network diagram and integrate it:

[0025] S411. Decompose the communication behavior within the organization into multiple interaction levels, including direct communication, indirect communication and implicit communication;

[0026] S412. Build an independent interaction matrix for each interaction level in It represents the interaction strength between node i and node j in the k-th layer interaction. The interaction strength is calculated by the following formula:

[0027]

[0028] in, For the communication frequency, is the time interval for information transmission, To ensure timely feedback, is the trust of node i to node j, is the authoritative influence of node i, is the corresponding weight coefficient;

[0029] S413, using the equilibrium model in social physics, the interaction matrices at each level are fused through nonlinear weighting to obtain the comprehensive interaction matrix S ij :

[0030]

[0031] Among them, λ k is the weight coefficient of different interaction levels, reflecting the contribution of each layer’s interaction intensity to the comprehensive interaction;

[0032] S42. Combining coupled dynamics and social physics models to simulate information dissemination:

[0033] S421, based on the comprehensive interaction matrix S ij , a coupled dynamics model is established to describe the nonlinear propagation of information between nodes. The state equation of the model is:

[0034]

[0035] Among them, x i (t) represents the state variable of node i at time t, α and β are system coupling parameters, A ij is the adjacency matrix, tanh(x) represents the hyperbolic tangent function, which is used to simulate the nonlinear effect of information transmission;

[0036] S422, combined with the social physics model, further describe the mutual influence between nodes, and calculate the information transfer probability matrix P in the network through the random walk model ij ;

[0037] S423, using the transition probability matrix P ij Calculate the average time T for information transmission ij :

[0038]

[0039] S43. Applying the graph attention network with dual attention mechanism for deep learning:

[0040] S431, the comprehensive interaction matrix S ij and the average information transmission time matrix T ij As input, a graph attention network with dual attention mechanism is constructed, which combines spatial attention and temporal attention for node feature extraction;

[0041] S432, define the feature update formula of node i as:

[0042] h i ′=σ(∑ j∈N (i)α ij ·δ ij ·W·h j )

[0043] Among them, σ(x) is a nonlinear activation function, α ij is the spatial attention weight, and the calculation formula is:

[0044]

[0045] Among them, ReLU is the rectified linear unit activation function, is the attention weight vector, W is the linear transformation matrix;

[0046] S433, introducing time attention weight δ ij , the weight is combined with the social physics model and calculated based on the time series dependency, and the formula is:

[0047]

[0048] Among them, γ is the time weight coefficient;

[0049] S44, Fusion of coupled dynamics, social physics and graph attention network output for path optimization:

[0050] S441, the node weight α output by the dual attention mechanism graph attention network ij and time weight δ ij Integrate into the coupled dynamics and social physics model and define the global path optimization objective function:

[0051]

[0052] Among them, E represents the edge set in the network;

[0053] S442, by optimizing the objective function L(P), determining the optimal path for information transmission, identifying key nodes and paths in the information flow, and generating corresponding path optimization suggestions;

[0054] S45. Store the optimization results and apply them to policy adjustment:

[0055] S451, storing the optimized node weights and path information in a database for subsequent policy adjustments and communication network dynamic management;

[0056] S452. In management decision-making, based on the optimization results stored in the database, automatically generate improvement plans for the communication network to improve communication efficiency within the organization.

[0057] Optionally, S5 includes the following steps:

[0058] S51, based on the comprehensive interaction matrix S ij and node time attention weight δ ij , initialize the network fluid dynamics model and set the initial conditions, including the initial information density ρ i 90) and initial flow rate v i 90), as model input;

[0059] S52. Construct the control equations of the fluid dynamics model to describe the flow of information in the network:

[0060]

[0061] Among them, ρ i 9t) represents the information density on node i at time t, v i (t) represents the information flow velocity vector at node i, Represents the diffusion and flow of information in the network;

[0062] S53. Define the flow rate v of information flow according to the path optimization result i (t) and viscosity coefficient η:

[0063]

[0064] Among them, A ij is an element of the adjacency matrix, indicating whether there is a direct connection between node i and node j, d ij is the distance between node i and node j, and the viscosity coefficient η determines the resistance to the flow of information in the network;

[0065] S54. Dynamically update information density ρ i (t) and flow velocity v i (t), adjust the viscosity coefficient η according to the real-time situation of information flow:

[0066]

[0067] Among them, γ is the diffusion coefficient, which indicates the diffusion speed of information in the network. Laplacian operator representing information density;

[0068] S55. In each time step, the node weight α output by the graph attention network combined with the dual attention mechanism ij and time weight δ ij , introduces an adaptive optimization method based on swarm intelligence, and dynamically adjusts and optimizes the communication strategy through the particle swarm optimization method PSO:

[0069]

[0070] Where x is the position vector of the particle, T ij is the information transmission time between node i and node j, α ij and β ij are the weight coefficients of node i and node j, respectively, indicating the priority and importance of information transmission;

[0071] S56. Combining the results based on the network fluid dynamics model and the swarm intelligence optimization method, the optimal information flow route of each path is gradually updated. The final path optimization objective function is:

[0072]

[0073] S57. At the end of each iteration, calculate the information flow efficiency in the current state and update the comprehensive interaction matrix S according to the node characteristics and weights. ij and the temporal attention weight δ ij , for the next iterative optimization;

[0074] S58. The finally optimized information flow path, node weight, information density and flow rate data are stored in the database for subsequent analysis and dynamic management.

[0075] Optionally, the S6 comprises the following steps:

[0076] S61, initializing the communication network structure, generating an initial network topology graph G(V,E) based on the optimized information flow path, node weights and flow rate data, where V is a node set and E is an edge set;

[0077] S62, real-time collection of data in the communication network, including the communication frequency, information transmission delay, response time and feedback rate between nodes, and the construction of a real-time data matrix D(t) = [d ij (t)], where d ij (t) represents the real-time communication status between node i and node j at time t. The real-time data matrix is ​​updated by the following formula:

[0078]

[0079] in, are weight coefficients representing communication frequency, transmission delay and feedback rate, respectively, and f ij (t) is the real-time communication frequency between node i and node j, Δt ij (t) is the information transmission delay, r ij (t) is the response time;

[0080] S63. Use adaptive optimization methods to dynamically adjust the communication network structure, with the optimization goal of minimizing network delay and information loss:

[0081]

[0082] Among them, w ij (t) is the dynamic weight between node i and node j, which is determined by the communication frequency, information transmission priority and network load, etc. ij (t) is the information transmission time between node i and node j;

[0083] S64, in each optimization iteration, combining the time attention weight δ in claim 5 ij and node weight α ij , update the network topology G(V,E), the optimization formula is:

[0084]

[0085] Among them, G(V,E)' is the optimized network topology, and the structure with the lowest communication delay and the highest information transmission efficiency is selected;

[0086] S65. Combine situational awareness technology to monitor the environmental variables in the communication network C(t) = [c i (t)], including network load, device performance and user behavior patterns, and calculating the context awareness matrix C ij (t) And further optimize the network structure:

[0087] w ij '(t)=w ij (t)·exp(-λ·C ij (t));

[0088] Where λ is the context-awareness adjustment coefficient, C ij (t) represents the contextual correlation between nodes i and j;

[0089] S66. Real-time calculation and dynamic display of key indicators of the communication network, including average communication delay Information transmission efficiency E9t) and network stability S(t):

[0090]

[0091] S67. Generate personalized optimization suggestions and provide them to management in the form of reports for subsequent strategy adjustments. The suggestions are generated based on key indicators and situational awareness information;

[0092] S68. Store the final optimized network structure, node weights, communication status and situational awareness data in a database for historical analysis, future prediction and long-term optimization.

[0093] Optionally, the S7 includes the following steps:

[0094] S71. Generate personalized communication optimization suggestions based on sentiment computing and adaptive semantic network, specifically including the following steps:

[0095] S711, extracting network feature information from the optimization result in claim 7, including node weight α ij , time weight δ ij 、Communication frequency f ij (t) and communication intensity s ij (t), construct the feature matrix F ij (t):

[0096] F ij (t) = [α ij ,δ ij ,f ij (t),s ij (t)];

[0097] S712. Use the emotional computing model to analyze the emotional state of employees during the communication process and calculate the emotional weight E ij (t):

[0098]

[0099] Among them, β1, β2, β3, and β4 are adjustment parameters used to balance the impact of various factors on the emotional state;

[0100] S713, introduce adaptive semantic network, based on sentiment computing, combine semantic analysis to process communication content, calculate the semantic similarity Sim (C ij (t)):

[0101]

[0102] Among them, TF-IDF(w k ) and TF-IDF(w k ') are the words w that appear in the communication content of nodes i and j respectively k TF-IDF value of

[0103] S714, through the adaptive semantic network, calculate the sentiment-semantic fusion weight Φ between nodes ij (t):

[0104] Φ ij (t) = E ij (t)·Sim(C ij (t));

[0105] S715, based on sentiment-semantic fusion weight Φ ij (t)'s optimization strategy generates personalized communication optimization suggestions:

[0106] ΔΦ ij (t) = γ·(∑ k∈N(i) Φ ik (t)-∑ l∈N(j) Φ lj (t));

[0107] Among them, ΔΦ ij (t) is the optimization adjustment amount of the communication content, and γ is the optimization coefficient, which is used to balance the adjustment of emotion and semantics;

[0108] S72. Generate an automated report based on the adaptive semantic network. The specific steps are as follows:

[0109] S721. Generate a status report of the current communication network based on the analysis results of the adaptive semantic network, the report content including emotional state, semantic consistency, key node analysis and optimization suggestions;

[0110] S722, using historical data and current sentiment-semantic fusion weight Φ ij (t) Dynamically optimize report content:

[0111]

[0112] Among them, λ is the weighting coefficient of historical data, which is used to balance the impact of recent data on optimization suggestions;

[0113] S723. During the report generation process, natural language generation technology is used to generate the final report text:

[0114]

[0115] Among them, T ij (t) is the final report text content, including the current emotional state, semantic consistency analysis and optimization suggestions;

[0116] S73. Combine employee feedback and historical data for continuous improvement:

[0117] S731. Regularly collect employee feedback on communication optimization suggestions and use this feedback data to adjust the affective computing model and adaptive semantic network to ensure that the model can accurately reflect the employee's affective state and semantic needs;

[0118] S732, by updating the parameters of the emotion computing model and the adaptive semantic network to reflect the latest communication status and semantic requirements:

[0119] Θ new =ρ·Θ old +(1-ρ)·Θ feedback ;

[0120] Among them, Θ is the model parameter, ρ is the feedback adjustment coefficient, Θ feedback represents the parameter update based on feedback;

[0121] S74. Store the generated communication optimization suggestions and reports in a database.

[0122] The beneficial effects of the present invention are:

[0123] (1) The present invention proposes a method for improving the efficiency of internal communication in an organization based on social network analysis. By collecting and preprocessing emails, instant messaging records, meeting minutes and employee behavior data, a comprehensive communication feature matrix is ​​constructed. By combining the social physics model with the graph attention network, the communication bottlenecks and key nodes of information flow within the organization are accurately identified, the communication path is optimized, and the communication efficiency is improved, which is particularly effective in complex organizational structures.

[0124] (2) The present invention introduces a network fluid dynamics model and a swarm intelligence adaptive optimization method to simulate the flow of information in a communication network, and dynamically adjusts communication strategies based on real-time data, optimizes information transmission paths, reduces information delays and errors, and ensures the efficiency and accuracy of information transmission in a changing communication environment, thereby improving the response speed and decision-making efficiency within the organization.

[0125] (3) The present invention combines situational awareness technology and adaptive optimization methods to monitor key indicators in the communication network in real time, such as communication frequency, latency and stability, dynamically adjust the network structure, and display network changes through a visual interface, so that management can understand the health status of the communication network in real time and make targeted adjustments, effectively respond to changes in the internal and external environment of the organization, and significantly improve the transparency of communication and decision-making support capabilities.

[0126] (4) The present invention adopts a graph attention network with a dual attention mechanism to perform deep learning and optimization on the internal communication network of the organization. By combining spatial attention with temporal attention, it realizes the precise extraction of node features and the optimization of information flow, ensuring the efficient dissemination of information in the network. By iteratively optimizing path selection, it ultimately improves communication efficiency and the reliability of information transmission, and is suitable for complex multi-level organizational structures. BRIEF DESCRIPTION OF THE DRAWINGS

[0127] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0128] Figure 1 A flowchart of a method for improving internal communication efficiency of an organization based on social network analysis proposed by the present invention;

[0129] Figure 2 A schematic diagram of communication bottleneck identification and optimization based on the combination of social physics model and graph attention network in the present invention;

[0130] Figure 3 A dynamic communication strategy optimization flow chart combining the network fluid dynamics model with the swarm intelligence adaptive optimization method in the present invention. DETAILED DESCRIPTION

[0131] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0132] refer to Figure 1-3 , a method for improving internal communication efficiency of an organization based on social network analysis, comprising the following steps:

[0133] S1. Collect communication data within the organization, including emails, instant messaging records, meeting minutes, and employee behavior data;

[0134] S2. Preprocess the collected communication data, remove noise data, and perform standardization;

[0135] S3. Construct a communication network diagram within the organization, where nodes represent individuals or departments and edges represent communication frequency and intensity;

[0136] S4. Combining social network analysis technology based on social physics models with graph attention networks, we can identify communication bottlenecks and key nodes of information flow by simulating information dissemination and interaction between individuals, and accurately evaluate the influence and communication efficiency of each node by weight distribution between nodes.

[0137] S5. Introduce the network fluid dynamics model and combine it with the adaptive optimization method based on swarm intelligence to simulate the propagation of information flow in the communication network and dynamically adjust and optimize the communication strategy;

[0138] S6. Use adaptive optimization methods to continuously optimize the communication network structure based on real-time data and simulation results, and combine situational awareness technology to dynamically display changes in the communication network on a visual interface;

[0139] S7. Generate personalized communication optimization suggestions and automated reports, use natural language processing technology based on emotional computing to generate report content, and combine employee feedback and historical data for continuous improvement.

[0140] In this implementation, S3 includes the following steps:

[0141] S31, extracting a unique identifier of each individual or department as a node according to the preprocessed comprehensive feature matrix;

[0142] S32, calculating the communication frequency between nodes, and defining a communication frequency matrix, wherein each element represents the number of communications between nodes;

[0143] S33, calculate the communication intensity between nodes, define a communication intensity matrix, where each element represents the communication intensity between nodes, and is calculated based on factors such as communication frequency, message length, and time interval;

[0144] S34. Based on the communication frequency matrix and the communication intensity matrix, construct a preliminary communication network diagram, where nodes represent individuals or departments and edges represent communication frequency and intensity;

[0145] S35. Use graph theory methods to analyze the initially constructed communication network diagram, remove isolated nodes and abnormal edges, and optimize the network structure;

[0146] S36, topologically sorting the optimized communication network graph to generate a directed acyclic graph structure for further analysis and processing;

[0147] S37. Store the finally constructed communication network diagram into the database for subsequent analysis and processing.

[0148] In this implementation, S4 includes the following steps:

[0149] S41. Construct a multi-level interaction matrix based on the final communication network diagram and integrate it:

[0150] S411. Decompose the communication behavior within the organization into multiple interaction levels, including direct communication, indirect communication and implicit communication;

[0151] S412. Build an independent interaction matrix for each interaction level in It represents the interaction strength between node i and node j in the k-th layer interaction. The interaction strength is calculated by the following formula:

[0152]

[0153] in, For the communication frequency, is the time interval for information transmission, To ensure timely feedback, is the trust of node i to node j, is the authoritative influence of node i, is the corresponding weight coefficient;

[0154] S413, using the equilibrium model in social physics, the interaction matrices at each level are fused through nonlinear weighting to obtain the comprehensive interaction matrix S ij :

[0155]

[0156] Among them, λ k is the weight coefficient of different interaction levels, reflecting the contribution of each layer’s interaction intensity to the comprehensive interaction;

[0157] S42. Combining coupled dynamics and social physics models to simulate information dissemination:

[0158] S421, based on the comprehensive interaction matrix S ij , a coupled dynamics model is established to describe the nonlinear propagation of information between nodes. The state equation of the model is:

[0159]

[0160] Among them, x i (t) represents the state variable of node i at time t, α and β are system coupling parameters, A ij is the adjacency matrix, tanh(x) represents the hyperbolic tangent function, which is used to simulate the nonlinear effect of information transmission;

[0161] S422, combined with the social physics model, further describe the mutual influence between nodes, and calculate the information transfer probability matrix P in the network through the random walk model ij ;

[0162] S423, using the transition probability matrix P ij Calculate the average time T for information transmission ij :

[0163]

[0164] S43. Applying the graph attention network with dual attention mechanism for deep learning:

[0165] S431, the comprehensive interaction matrix S ij and the average information transmission time matrix T ij As input, a graph attention network with dual attention mechanism is constructed, which combines spatial attention and temporal attention for node feature extraction;

[0166] S432, define the feature update formula of node i as:

[0167] h i ′=σ(∑ j∈N(i) α ij ·δ ij ·W·h j )

[0168] Among them, σ(x) is a nonlinear activation function, α ij is the spatial attention weight, and the calculation formula is:

[0169]

[0170] Among them, ReLU is the rectified linear unit activation function, is the attention weight vector, W is the linear transformation matrix;

[0171] S433, introducing time attention weight δ ij , the weight is combined with the social physics model and calculated based on the time series dependency, and the formula is:

[0172]

[0173] Among them, γ is the time weight coefficient;

[0174] S44, Fusion of coupled dynamics, social physics and graph attention network output for path optimization:

[0175] S441, the node weight α output by the dual attention mechanism graph attention network ij and time weight δ ij Integrate it into the coupled dynamics and social physics model and define the global path optimization objective function:

[0176]

[0177] Among them, E represents the edge set in the network;

[0178] S442, by optimizing the objective function L(P), determining the optimal path for information transmission, identifying key nodes and paths in the information flow, and generating corresponding path optimization suggestions;

[0179] S45. Store the optimization results and apply them to policy adjustment:

[0180] S451, storing the optimized node weights and path information in a database for subsequent policy adjustments and communication network dynamic management;

[0181] S452. In management decision-making, based on the optimization results stored in the database, automatically generate improvement plans for the communication network to improve communication efficiency within the organization.

[0182] In this implementation, S5 includes the following steps:

[0183] S51, based on the comprehensive interaction matrix S ij and node time attention weight δ ij , initialize the network fluid dynamics model and set the initial conditions, including the initial information density ρ i (0) and the initial flow velocity v i (0), as model input;

[0184] S52. Construct the control equations of the fluid dynamics model to describe the flow of information in the network:

[0185]

[0186] Among them, ρ i (t) represents the information density at node i at time t, v i (t) represents the information flow velocity vector at node i, Represents the diffusion and flow of information in the network;

[0187] S53. Define the flow rate v of information flow according to the path optimization result i (t) and viscosity coefficient η:

[0188]

[0189] Among them, A ij is an element of the adjacency matrix, indicating whether there is a direct connection between node i and node j, d ij is the distance between node i and node j, and the viscosity coefficient η determines the resistance to the flow of information in the network;

[0190] S54. Dynamically update information density ρ i (t) and flow velocity v i (t), adjust the viscosity coefficient η according to the real-time situation of information flow:

[0191]

[0192] Among them, γ is the diffusion coefficient, which indicates the diffusion speed of information in the network. Laplacian operator representing information density;

[0193] S55. In each time step, the node weight α output by the graph attention network combined with the dual attention mechanism ij and time weight δ ij , introduces an adaptive optimization method based on swarm intelligence, and dynamically adjusts and optimizes the communication strategy through the particle swarm optimization method PSO:

[0194]

[0195] Where x is the position vector of the particle, T ij is the information transmission time between node i and node j, α ij and β ij are the weight coefficients of node i and node j, respectively, indicating the priority and importance of information transmission;

[0196] S56. Combining the results based on the network fluid dynamics model and the swarm intelligence optimization method, the optimal information flow route of each path is gradually updated. The final path optimization objective function is:

[0197]

[0198] S57. At the end of each iteration, calculate the information flow efficiency in the current state and update the comprehensive interaction matrix S according to the node characteristics and weights. ij and the temporal attention weight δ ij , for the next iterative optimization;

[0199] S58. The finally optimized information flow path, node weight, information density and flow rate data are stored in the database for subsequent analysis and dynamic management.

[0200] In this implementation, S6 includes the following steps:

[0201] S61, initializing the communication network structure, generating an initial network topology graph G(V,E) based on the optimized information flow path, node weights and flow rate data, where V is a node set and E is an edge set;

[0202] S62, real-time collection of data in the communication network, including the communication frequency, information transmission delay, response time and feedback rate between nodes, and the construction of a real-time data matrix D(t) = [d ij (t)], where dij (t) represents the real-time communication status between node i and node j at time t. The real-time data matrix is ​​updated by the following formula:

[0203]

[0204] in, are weight coefficients representing communication frequency, transmission delay and feedback rate, respectively, and f ij (t) is the real-time communication frequency between node i and node j, Δt ij (t) is the information transmission delay, r ij (t) is the response time;

[0205] S63. Use adaptive optimization methods to dynamically adjust the communication network structure, with the optimization goal of minimizing network delay and information loss:

[0206]

[0207] Among them, w ij (t) is the dynamic weight between node i and node j, which is determined by the communication frequency, information transmission priority and network load, etc. ij (t) is the information transmission time between node i and node j;

[0208] S64, in each optimization iteration, combining the time attention weight δ in claim 5 ij and node weight α ij , update the network topology G(V,E), the optimization formula is:

[0209]

[0210] Among them, G(V,E)' is the optimized network topology, and the structure with the lowest communication delay and the highest information transmission efficiency is selected;

[0211] S65. Combine situational awareness technology to monitor the environmental variables in the communication network C(t) = [c i (t)], including network load, device performance and user behavior patterns, and calculate the context awareness matrix C ij (t) And further optimize the network structure:

[0212] w ij '(t)=w ij (t)exp(-λ·C ij (t));

[0213] Where λ is the context-awareness adjustment coefficient, C ij (t) represents the contextual correlation between nodes i and j;

[0214] S66. Real-time calculation and dynamic display of key indicators of the communication network, including average communication delay Information transmission efficiency E(t) and network stability S(t):

[0215]

[0216] S67. Generate personalized optimization suggestions and provide them to management in the form of reports for subsequent strategy adjustments. The suggestions are generated based on key indicators and situational awareness information;

[0217] S68. Store the final optimized network structure, node weights, communication status and situational awareness data in a database for historical analysis, future prediction and long-term optimization.

[0218] In this implementation, S7 includes the following steps:

[0219] S71. Generate personalized communication optimization suggestions based on sentiment computing and adaptive semantic network, specifically including the following steps:

[0220] S711, extracting network feature information from the optimization result in claim 7, including node weight α ij , time weight δ ij 、Communication frequency f ij (t) and communication intensity s ij (t), construct the feature matrix F ij (t):

[0221] F ij (t) = [α ij ,δ ij ,f ij (t),s ij (t)];

[0222] S712. Use the emotional computing model to analyze the emotional state of employees during the communication process and calculate the emotional weight E ij (t):

[0223]

[0224] Among them, β1, β2, β3, and β4 are adjustment parameters used to balance the impact of various factors on the emotional state;

[0225] S713, introduce adaptive semantic network, based on sentiment computing, combine semantic analysis to process communication content, calculate the semantic similarity Sim (C ij (t)):

[0226]

[0227] Among them, TF-IDF(w k ) and TF-IDF(w k ') are the words w that appear in the communication content of nodes i and j respectively k TF-IDF value of

[0228] S714, through the adaptive semantic network, calculate the sentiment-semantic fusion weight Φ between nodes ij (t):

[0229] Φ ij (t) = E ij (t)·Sim(C ij (t));

[0230] S715, based on sentiment-semantic fusion weight Φ ij (t)'s optimization strategy generates personalized communication optimization suggestions:

[0231] ΔΦ ij (t) = γ·(∑ k∈N ( i )Φ ik (t)-∑ l∈N ( j )Φ lj (t));

[0232] Among them, ΔΦ ij (t) is the optimization adjustment amount of the communication content, and γ is the optimization coefficient, which is used to balance the adjustment of emotion and semantics;

[0233] S72. Generate an automated report based on the adaptive semantic network. The specific steps are as follows:

[0234] S721. Generate a status report of the current communication network based on the analysis results of the adaptive semantic network, the report content including emotional state, semantic consistency, key node analysis and optimization suggestions;

[0235] S722, using historical data and current sentiment-semantic fusion weight Φ ij (t) Dynamically optimize report content:

[0236]

[0237] Among them, λ is the weighting coefficient of historical data, which is used to balance the impact of recent data on optimization suggestions;

[0238] S723. During the report generation process, natural language generation technology is used to generate the final report text:

[0239]

[0240] Among them, T ij (t) is the final report text content, including the current emotional state, semantic consistency analysis and optimization suggestions;

[0241] S73. Combine employee feedback and historical data for continuous improvement:

[0242] S731. Regularly collect employee feedback on communication optimization suggestions and use this feedback data to adjust the affective computing model and adaptive semantic network to ensure that the model can accurately reflect the employee's affective state and semantic needs;

[0243] S732, by updating the parameters of the emotion computing model and the adaptive semantic network to reflect the latest communication status and semantic requirements:

[0244] Θ new =ρ·Θ old +(1-ρ)·Θ feedback ;

[0245] Among them, Θ is the model parameter, ρ is the feedback adjustment coefficient, Θ feedback represents the parameter update based on feedback;

[0246] S74. Store the generated communication optimization suggestions and reports in a database.

[0247] Embodiment 1:

[0248] In a large multinational company, branches around the world maintain daily contact through a variety of communication channels, including email, instant messaging tools, video conferencing, and project management platforms. While these communication methods improve the company's global collaboration capabilities, they also bring complex communication challenges. For example, in a product development project, project team members are distributed in different countries and time zones. Communication is frequent but inefficient, and information transmission is often delayed or lost, resulting in project delays. In order to improve communication efficiency, the project management team decided to introduce the method for improving internal communication efficiency based on social network analysis of the present invention to optimize the communication structure and strategy of the project team and improve overall work efficiency.

[0249] In this product development project, all communication data within the project team was first collected using the method of the present invention, including email records, instant messaging conversations, video conference records, and task updates on the project management platform over the past three months. After preprocessing, these data generated a comprehensive communication feature matrix, which further constructed a communication network diagram of the project team. The network diagram shows the frequency and intensity of communication between team members, and identifies communication bottlenecks and key nodes of information flow through a combination of social physics models and graph attention networks.

[0250] Through analysis, it was found that the information transmission between the project manager and the development team is the key path for the project progress, but due to the time difference, there is a significant delay in the communication between the manager and the team members. In addition, there is an imbalance in the information exchange within the development team. Some core developers are overburdened with communication, while other personnel are not sufficiently involved. In order to solve these problems, the present invention introduces a network fluid dynamics model, which dynamically adjusts the communication strategy and optimizes the communication path by simulating the process of information flow. Specifically, the method recommends adjusting meeting times, increasing the frequency of communication among core members of the team, and introducing more flexible communication tools, such as asynchronously updated documents and task boards.

[0251] To verify the optimization effect, the project management team implemented these adjustments over a two-week period and used the adaptive optimization method and context-aware technology of the present invention to continuously monitor changes in the communication network. After each meeting, the team immediately updated the task status and adjusted the next communication strategy through real-time data analysis. All communication data will be collected and used for further optimization to ensure that the communication strategy is highly consistent with the project progress.

[0252] Concrete data proves our beneficial effects:

[0253] After implementing the method of the present invention, the communication efficiency of the project team has been significantly improved. By comparing the data before and after optimization, the following results were obtained:

[0254] Table 1 Comparison of communication frequency and information transmission accuracy

[0255]

[0256]

[0257] Table 2 Improvement of critical path communication efficiency

[0258] index Before optimization After optimization Improvement Critical path communication delay (average) 8 hours 5.6 hours -30％ Timely rate of project progress tracking 65％ 90％ 25％ Accuracy of information transmission on the critical path 80％ 93％ 13％ Decision response speed in the critical path 10 hours 7 hours -30％

[0259] Table 3 Improvement of overall project efficiency and quality

[0260] index Before optimization After optimization Improvement Project Development Cycle 12 months 11 months -8.30％ Project progress completion (as planned) 75％ 95％ 20％ Number of issues found by code review 100 80 -20％ Team work satisfaction 70％ 85％ 15％ Reduced meeting time 20％ 15％ -25％

[0261] The above three data tables respectively show the comparison of communication frequency and information transmission accuracy, the improvement of critical path communication efficiency, and the improvement of overall project efficiency and quality. These data fully demonstrate the significant effect of the present invention in improving internal communication efficiency, shortening project development cycle and improving project quality.

[0262] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for improving internal communication efficiency of an organization based on social network analysis, characterized in that: The steps include: S1. Collect communication data within the organization, including emails, instant messaging records, meeting minutes, and employee behavior data; S2. Preprocess the collected communication data, remove noise data, and perform standardization; S3. Construct a communication network diagram within the organization, where nodes represent individuals or departments and edges represent communication frequency and intensity; S4. Combining social network analysis technology based on social physics models with graph attention networks, we can identify communication bottlenecks and key nodes of information flow by simulating information dissemination and interaction between individuals, and accurately evaluate the influence and communication efficiency of each node by weight distribution between nodes. S5. Introduce the network fluid dynamics model and combine it with the adaptive optimization method based on swarm intelligence to simulate the propagation of information flow in the communication network and dynamically adjust and optimize the communication strategy; S6. Use adaptive optimization methods to continuously optimize the communication network structure based on real-time data and simulation results, and combine situational awareness technology to dynamically display changes in the communication network on a visual interface; S7. Generate personalized communication optimization suggestions and automated reports, use natural language processing technology based on emotional computing to generate report content, and combine employee feedback and historical data for continuous improvement.

2. The method for improving internal communication efficiency of an organization based on social network analysis according to claim 1, characterized in that: The S3 comprises the following steps: S31, extracting a unique identifier of each individual or department as a node according to the preprocessed comprehensive feature matrix; S32, calculating the communication frequency between nodes, and defining a communication frequency matrix, wherein each element represents the number of communications between nodes; S33, calculate the communication intensity between nodes, define a communication intensity matrix, where each element represents the communication intensity between nodes, and is calculated based on factors such as communication frequency, message length, and time interval; S34. Based on the communication frequency matrix and the communication intensity matrix, construct a preliminary communication network diagram, where nodes represent individuals or departments and edges represent communication frequency and intensity; S35. Use graph theory methods to analyze the initially constructed communication network diagram, remove isolated nodes and abnormal edges, and optimize the network structure; S36, topologically sorting the optimized communication network graph to generate a directed acyclic graph structure for further analysis and processing; S37. Store the finally constructed communication network diagram into the database for subsequent analysis and processing.

3. The method for improving internal communication efficiency of an organization based on social network analysis according to claim 1, characterized in that: The S4 comprises the following steps: S41. Construct a multi-level interaction matrix based on the final communication network diagram and integrate it: S411. Decompose the communication behavior within the organization into multiple interaction levels, including direct communication, indirect communication and implicit communication; S412. Build an independent interaction matrix for each interaction level in It represents the interaction strength between node i and node j in the k-th layer interaction. The interaction strength is calculated by the following formula: in, For the communication frequency, is the time interval for information transmission, To ensure timely feedback, is the trust of node i to node j, is the authoritative influence of node i, is the corresponding weight coefficient; S413, using the equilibrium model in social physics, the interaction matrices at each level are fused through nonlinear weighting to obtain the comprehensive interaction matrix S ij : Among them, λ k is the weight coefficient of different interaction levels, reflecting the contribution of each layer’s interaction intensity to the comprehensive interaction; S42. Combining coupled dynamics and social physics models to simulate information dissemination: S421, based on the comprehensive interaction matrix S ij , a coupled dynamics model is established to describe the nonlinear propagation of information between nodes. The state equation of the model is: Among them, x i ( t ) represents the state variable of node i at time t, α and β are system coupling parameters, A ij is the adjacency matrix, tanh(x) represents the hyperbolic tangent function, which is used to simulate the nonlinear effect of information transmission; S422, combined with the social physics model, further describe the mutual influence between nodes, and calculate the information transfer probability matrix P in the network through the random walk model ij ; S423, using the transition probability matrix P ij Calculate the average time T for information transmission ij : S43. Applying the graph attention network with dual attention mechanism for deep learning: S431, the comprehensive interaction matrix S ij and the average information transmission time matrix T ij As input, a graph attention network with dual attention mechanism is constructed, which combines spatial attention and temporal attention for node feature extraction; S432, define the feature update formula of node i as: h′ i =σ(∑ j∈N(i) a ij ·d ij ·W·h j ) Among them, σ(x) is a nonlinear activation function, α ij is the spatial attention weight, and the calculation formula is: Among them, ReLU is the rectified linear unit activation function, is the attention weight vector, W is the linear transformation matrix; S433, introducing time attention weight δ ij , the weight is combined with the social physics model and calculated based on the time series dependency, and the formula is: Among them, γ is the time weight coefficient; S44, Fusion of coupled dynamics, social physics and graph attention network output for path optimization: S441, the node weight α output by the dual attention mechanism graph attention network ij and time weight δ ij Integrate into the coupled dynamics and social physics model and define the global path optimization objective function: Among them, E represents the edge set in the network; S442, by optimizing the objective function L(P), determining the optimal path for information transmission, identifying key nodes and paths in the information flow, and generating corresponding path optimization suggestions; S45. Store the optimization results and apply them to policy adjustment: S451, storing the optimized node weights and path information in a database for subsequent policy adjustments and communication network dynamic management; S452. In management decision-making, based on the optimization results stored in the database, automatically generate improvement plans for the communication network to improve communication efficiency within the organization.

4. The method for improving internal communication efficiency of an organization based on social network analysis according to claim 1, characterized in that: The S5 comprises the following steps: S51, based on the comprehensive interaction matrix S ij and node time attention weight δ ij , initialize the network fluid dynamics model and set the initial conditions, including the initial information density ρ i (0) and the initial flow velocity v i (0), as model input; S52. Construct the control equations of the fluid dynamics model to describe the flow of information in the network: Among them, ρ i (t) represents the information density at node i at time t, v i (t) represents the information flow velocity vector at node i, Represents the diffusion and flow of information in the network; S53. Define the flow rate v of information flow according to the path optimization result i (t) and viscosity coefficient η: Among them, A ij is an element of the adjacency matrix, indicating whether there is a direct connection between node i and node j, d ij is the distance between node i and node j, and the viscosity coefficient η determines the resistance to the flow of information in the network; S54. Dynamically update information density ρ i ( t) and flow velocity v i ( t), adjust the viscosity coefficient η according to the real-time situation of information flow: Among them, γ is the diffusion coefficient, which indicates the diffusion speed of information in the network. Laplacian operator representing information density; S55. In each time step, the node weight α output by the graph attention network combined with the dual attention mechanism ij and time weight δ ij , introduces an adaptive optimization method based on swarm intelligence, and dynamically adjusts and optimizes the communication strategy through the particle swarm optimization method PSO: Where x is the position vector of the particle, T ij is the information transmission time between node i and node j, α ij and β ij are the weight coefficients of node i and node j, respectively, indicating the priority and importance of information transmission; S56. Combining the results based on the network fluid dynamics model and the swarm intelligence optimization method, the optimal information flow route of each path is gradually updated. The final path optimization objective function is: S57. At the end of each iteration, calculate the information flow efficiency in the current state and update the comprehensive interaction matrix S according to the node characteristics and weights. ij and the temporal attention weight δ ij , for the next iterative optimization; S58. The finally optimized information flow path, node weight, information density and flow rate data are stored in the database for subsequent analysis and dynamic management.

5. The method for improving internal communication efficiency of an organization based on social network analysis according to claim 1, characterized in that: The S6 comprises the following steps: S61, initialize the communication network structure, generate the initial network topology graph G based on the optimized information flow path, node weight and flow rate data ( V, E), where V is the node set and E is the edge set; S62, real-time collection of data in the communication network, including the communication frequency, information transmission delay, response time and feedback rate between nodes, and the construction of a real-time data matrix D(t) = [d ij (t)], where d ij (t) represents the real-time communication status between node i and node j at time t. The real-time data matrix is ​​updated by the following formula: in, are weight coefficients representing communication frequency, transmission delay and feedback rate, respectively, and f ij (t) is the real-time communication frequency between node i and node j, Δt ij (t) is the information transmission delay, r ij (t) is the response time; S63. Use adaptive optimization methods to dynamically adjust the communication network structure, with the optimization goal of minimizing network delay and information loss: Among them, w ij (t) is the dynamic weight between node i and node j, which is determined by the communication frequency, information transmission priority and network load, etc. ij (t) is the information transmission time between node i and node j; S64, in each optimization iteration, combining the time attention weight δ in claim 5 ij and node weight α ij , update the network topology G(V,E), the optimization formula is: Among them, G(V,E)' is the optimized network topology, and the structure with the lowest communication delay and the highest information transmission efficiency is selected; S65. Combine situational awareness technology to monitor the environmental variables in the communication network C(t) = [c i (t)], including network load, device performance and user behavior patterns, and calculate the context awareness matrix C ij (t) And further optimize the network structure: w ij '(t)=w ij (t)·exp(-λ·C ij (t)); Where λ is the context-awareness adjustment coefficient, C ij (t) represents the contextual correlation between nodes i and j; S66. Real-time calculation and dynamic display of key indicators of the communication network, including average communication delay Information transmission efficiency E(t) and network stability S(t): S67. Generate personalized optimization suggestions and provide them to management in the form of reports for subsequent strategy adjustments. The suggestions are generated based on key indicators and situational awareness information; S68. Store the final optimized network structure, node weights, communication status and situational awareness data in a database for historical analysis, future prediction and long-term optimization.

6. The method for improving internal communication efficiency of an organization based on social network analysis according to claim 1, characterized in that: The S7 comprises the following steps: S71. Generate personalized communication optimization suggestions based on sentiment computing and adaptive semantic network, specifically including the following steps: S711, extracting network feature information from the optimization result in claim 7, including node weight α ij , time weight δ ij 、Communication frequency f ij (t) and communication intensity s ij (t), construct the feature matrix F ij (t): F ij (t)=[α ij ,δ ij ,f ij (t),s ij (t)]; S712. Use the emotional computing model to analyze the emotional state of employees during the communication process and calculate the emotional weight E ij (t): Among them, β1, β2, β3, and β4 are adjustment parameters used to balance the impact of various factors on the emotional state; S713, introduce adaptive semantic network, based on sentiment computing, combine semantic analysis to process communication content, calculate the semantic similarity Sim (C ij (t)): Among them, TF-IDF(w k ) and TF-IDF(w k ') are the words w that appear in the communication content of nodes i and j respectively k TF-IDF value of S714, through the adaptive semantic network, calculate the sentiment-semantic fusion weight Φ between nodes ij (t): Φ ij (t)=E ij (t)·Sim(C ij (t)); S715, based on sentiment-semantic fusion weight Φ ij (t)'s optimization strategy generates personalized communication optimization suggestions: DF ij (t)=γ·(∑ k∈N(i) F ik (t)-∑ l∈N(j) F lj (t)); Among them, ΔΦ ij (t) is the optimization adjustment amount of the communication content, and γ is the optimization coefficient, which is used to balance the adjustment of emotion and semantics; S72. Generate an automated report based on the adaptive semantic network. The specific steps are as follows: S721. Generate a status report of the current communication network based on the analysis results of the adaptive semantic network, the report content including emotional state, semantic consistency, key node analysis and optimization suggestions; S722, using historical data and current sentiment-semantic fusion weight Φ ij (t) Dynamically optimize report content: Among them, λ is the weighting coefficient of historical data, which is used to balance the impact of recent data on optimization suggestions; S723. During the report generation process, natural language generation technology is used to generate the final report text: Among them, T ij (t) is the final report text content, including the current emotional state, semantic consistency analysis and optimization suggestions; S73. Combine employee feedback and historical data for continuous improvement: S731. Regularly collect employee feedback on communication optimization suggestions and use this feedback data to adjust the affective computing model and adaptive semantic network to ensure that the model can accurately reflect the employee's affective state and semantic needs; S732, by updating the parameters of the emotion computing model and the adaptive semantic network to reflect the latest communication status and semantic requirements: I new =r·Θ old +(1-r)·Θ feedback ; Among them, Θ is the model parameter, ρ is the feedback adjustment coefficient, Θ feedback represents the parameter update based on feedback; S74. Store the generated communication optimization suggestions and reports in a database.

Images