A Social Network Relationship Prediction Method Based on Monte Carlo Simulation Network Augmentation Techniques

By using Monte Carlo simulation network augmentation technology, a social network graph model is constructed and Monte Carlo augmentation sampling of a multilayer perceptron model is performed. This solves the uncertainty problem caused by noise and error in social relationship prediction and achieves higher accuracy and robustness.

CN120807192BActive Publication Date: 2026-11-17ANHUI UNIVERSITY OF TRADITIONAL CHINESE MEDICINE
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202510973560.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2025-07-15
Publication Date
2026-11-17
Estimated Expiration
2045-07-15

AI Technical Summary

Technical Problem

Existing social relationship prediction methods are affected by factors such as noise pollution, sampling bias, and measurement error, which lead to uncertainty in network topology, affecting the stability and generalization ability of the model and reducing its accuracy.

Method used

The Monte Carlo simulation network augmentation technique is adopted. By constructing a social network graph model, structural and attribute features are extracted for initial probability modeling. Then, Monte Carlo augmentation sampling is performed using a multilayer perceptron model to generate multiple augmentation networks. The prediction results are fused to improve accuracy and robustness.

Benefits of technology

It effectively mitigates the topological uncertainty caused by noise interference and measurement errors, significantly improves the accuracy and robustness of social relationship prediction, and enhances the stability and adaptability of the model.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120807192B_ABST
    Figure CN120807192B_ABST
Patent Text Reader

Abstract

This invention relates to a social network relationship prediction method based on Monte Carlo simulation network augmentation technology. Compared with existing technologies, it solves the defects of decreased prediction accuracy and insufficient robustness caused by directly modeling based on incomplete or noisy network structures. The invention includes the following steps: constructing a social network graph model; extracting structural and attribute features between each pair of users; performing initial probability modeling; performing Monte Carlo augmentation sampling based on the initial connection probabilities; and obtaining the social network relationship prediction results. This invention effectively mitigates the topological uncertainty caused by noise interference and measurement errors in real-world networks through Monte Carlo augmentation strategies, significantly improving the accuracy and robustness of social relationship prediction.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of big data analytics, specifically to a social network relationship prediction method based on Monte Carlo simulation network enhancement technology. Background Technology

[0002] With the rapid development of internet technology and social platforms, social networks have become an important carrier of human behavior and information dissemination. Social relationship prediction, as a key issue in network science and data mining, has received widespread attention. Its goal is to predict potential relationships between users based on existing network connection information, and it has significant application value in many practical scenarios such as community segmentation, personalized recommendations, fraud detection, and propagation path analysis.

[0003] Currently, most mainstream social relationship prediction methods are based on network topology information, utilizing similarity metrics between nodes (such as the number of common neighbors, Adamic-Adar index, resource allocation indicators, etc.) or constructing prediction models through machine learning models (such as logistic regression, support vector machines, graph embedding methods). These methods do indeed show good performance on static, fully observable networks, and have achieved breakthroughs, especially under ideal conditions where data is complete and errors are small.

[0004] However, while previous studies have achieved significant results in certain specific contexts, their core preconditions are often based on the assumption of perfectly accurate network structures. In reality, real-world social networks are generally affected by various factors such as noise pollution, sampling bias, measurement errors, node masking, and information gaps, resulting in a high degree of uncertainty and incompleteness in the obtained network topology. These unavoidable disturbances can lead to biases or even errors in network connection information. Against this backdrop, traditional relationship prediction methods, if directly modeled based on network data with inherent errors, often amplify the uncertainty in the initial data, thereby affecting the model's stability and generalization ability. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies that directly model based on incomplete or noisy network structures, leading to decreased prediction accuracy and insufficient robustness. This invention provides a social network relationship prediction method based on Monte Carlo simulation network enhancement technology to solve the above problems.

[0006] To achieve the above objectives, the technical solution of the present invention is as follows:

[0007] A social network relationship prediction method based on Monte Carlo simulation network enhancement technology includes the following steps:

[0008] 11) Construct a social network graph model;

[0009] 12) Extract the structural and attribute features between each pair of users;

[0010] 13) Perform initial probability modeling;

[0011] 14) Monte Carlo augmented sampling based on initial connectivity probability;

[0012] 15) Obtaining the prediction results of social network relationships: Repeat the feature extraction and training process in the augmented network, average the prediction results of multiple augmented networks, obtain the edge probability estimate, and obtain the prediction results of social network relationships.

[0013] The construction of the social network graph model includes the following steps:

[0014] Extracting social networks ,

[0015] Represents the set of users on a social network. Indicates the first One user, Indicates the number of users;

[0016] Represents the set of relationships between any two users. For the first individual users With the individual users Is there a relationship between them?

[0017] if Indicates the first individual users With the individual users They have a relationship and are neighbors; if Indicates the first individual users With the individual users It doesn't matter.

[0018] The extraction of structural and attribute features between each pair of users includes the following steps:

[0019] 31) For any two users and Extracting feature vectors ;

[0020] 32) The extracted features include:

[0021] Number of common neighbors ,

[0022] , (1)

[0023] in, Indicates user Neighbors User The neighbors;

[0024] Resource allocation indicators :

[0025] , (2)

[0026] in, User The neighbors;

[0027] Adamic-Adar indicator ,

[0028] ; (3)

[0029] Node degree similarity ,

[0030] ; (4)

[0031] Age similarity ,

[0032] ,

[0033] in, and Representing users respectively and Age;

[0034] Interest similarity ,

[0035] (6)

[0036] in, and Representing users respectively and A collection of interests and hobbies;

[0037] Geographical distance similarity ,

[0038] , (7)

[0039] in, and Representing users respectively and Geographical location;

[0040] 33) User and eigenvectors Defined as:

[0041] (8)

[0042] The initial probability modeling involves using a multilayer perceptron model to perform initial probability modeling on whether a relationship exists between users, which includes the following steps:

[0043] Using a multilayer perceptron model, with feature vectors As input, As output For the first individual users With the individual users Is there a relationship between them?

[0044] The training formula is as follows:

[0045] , (9)

[0046] in, , , , For learnable parameters, It is the Sigmoid activation function.

[0047] , ,

[0048] For users and The initial connection probability.

[0049] The Monte Carlo enhancement sampling based on the initial connectivity probability includes the following steps:

[0050] 51) Utilizing users and Initial connection probability Monte Carlo sampling-based generative enhancement network ,

[0051] in, For the enhanced social network, A collection of users on a social network. To enhance the edge set of the social network;

[0052] ,

[0053] , (10)

[0054] In formula (10) For the generated random numbers, To enhance the network individual users With the individual users Is there a relationship between them?

[0055] 52) Repeat step 51) to generate An enhanced network ,

[0056] ,in Indicates the number of augmentation networks.

[0057] For the first A set of relationships between two users in an enhanced network. For the first The first enhanced network individual users With the individual users Is there a relationship between them?

[0058] if Indicates the first In the enhanced network, the first individual users With the individual users They are related and neighbors;

[0059] if Indicates the first In the enhanced network, the first individual users With the individual users It doesn't matter.

[0060] Obtaining the social network relationship prediction results includes the following steps:

[0061] 61) For each augmentation network Following the steps of extracting structural and attribute features between each pair of users, the corresponding features are extracted again. ,

[0062] use As input, As output For feature vectors, For the first The first enhanced network individual users With the individual users Is there a relationship between them?

[0063] Train the following multilayer perceptron model:

[0064] (11)

[0065] in, This represents vector concatenation. , , , For learning parameters, For users and In the first Connection probability in an enhanced network Use the Sigmoid activation function;

[0066] 62) Fusion of multi-network prediction results:

[0067] Predicting the first using formula (5) individual users With the individual users Probability of connecting edges :

[0068] (5)

[0069] in, To increase the number of networks.

[0070] A computer-readable storage medium storing a computer program that, when executed by a processor, enables a social network relationship prediction method based on Monte Carlo simulation network enhancement technology.

[0071] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, which, when executed by the processor, enables a social network relationship prediction method based on Monte Carlo simulation network enhancement technology.

[0072] Beneficial effects

[0073] The social network relationship prediction method based on Monte Carlo simulation network enhancement technology of the present invention effectively alleviates the topological uncertainty caused by noise interference and measurement errors in real networks through Monte Carlo enhancement strategy compared with the prior art, significantly improving the accuracy and robustness of social relationship prediction, and has strong practical application value.

[0074] In the process of relationship prediction modeling, this invention not only considers the local structural similarity between nodes (such as the number of common neighbors, resource allocation index, Adamic-Adar index, etc.), but also introduces individual-level attribute features (such as node degree, age, interests, geographical location degree, etc.), realizing the organic integration of structural information and attribute information, thereby improving the model's ability to express different types of network relationships.

[0075] This invention estimates the connection probability between user pairs using a multilayer perceptron model and performs Monte Carlo sampling based on this probability to generate multiple augmentation networks. The feature extraction and training prediction processes are repeated in these augmentation networks, and multiple prediction results are fused, effectively reducing prediction bias caused by observation errors, incomplete information, or structural noise in the original network, thus enhancing the robustness and stability of the model.

[0076] The method of this invention has a clear structure, is simple to implement, and possesses good scalability and practicality. It can be adapted to various types of social network data and can also be integrated into existing recommendation systems and community identification system frameworks, showing broad application prospects. In terms of accuracy, robustness, and practicality, it outperforms existing traditional prediction methods and can more effectively address the modeling challenges brought about by the uncertainty of data in real-world social networks. Attached Figure Description

[0077] Figure 1 This is a sequence diagram of the method of the present invention;

[0078] Figure 2 This is a flowchart of the algorithm of the present invention. Detailed Implementation

[0079] To provide a better understanding of the structural features and effects achieved by the present invention, a detailed description is provided below, accompanied by preferred embodiments and accompanying drawings:

[0080] like Figure 1 and Figure 2 As shown, the social network relationship prediction method based on Monte Carlo simulation network enhancement technology described in this invention achieves a paradigm shift from "deterministic graphs" to "probabilistic graph atlases" by introducing structural probabilistic modeling and simulation enhancement mechanisms, effectively mitigating model biases caused by measurement errors and information gaps. This design approach breaks the limitations of the traditional single-graph prediction framework and solves the problem of how to robustly model the connection prediction process when network structure is uncertain, demonstrating strong theoretical innovation and methodological versatility.

[0081] Includes the following steps:

[0082] The first step is to construct a social network graph model.

[0083] Extracting social networks ,

[0084] Represents the set of users on a social network. Indicates the first One user, Indicates the number of users;

[0085] Represents the set of relationships between any two users. For the first individual users With the individual users Is there a relationship between them?

[0086] if Indicates the first individual users With the individual users They have a relationship and are neighbors; if Indicates the first individual users With the individual users It doesn't matter.

[0087] The second step is to extract the structural and attribute features between each pair of users, thereby achieving the organic integration of structural and attribute information and improving the model's ability to express different types of network relationships.

[0088] (1) For any two users and Extracting feature vectors .

[0089] (2) The extracted features include:

[0090] Number of common neighbors ,

[0091] , (1)

[0092] in, Indicates user Neighbors User The neighbors;

[0093] Resource allocation indicators :

[0094] , (2)

[0095] in, User The neighbors;

[0096] Adamic-Adar indicator ,

[0097] ; (3)

[0098] Node degree similarity ,

[0099] ; (4)

[0100] Age similarity ,

[0101] ,

[0102] in, and Representing users respectively and Age;

[0103] Interest similarity ,

[0104] (6)

[0105] in, and Representing users respectively and A collection of interests and hobbies;

[0106] Geographical distance similarity ,

[0107] , (7)

[0108] in, and Representing users respectively and Its geographical location.

[0109] (3) User and eigenvectors Defined as:

[0110] (8)

[0111] The third step is to perform initial probabilistic modeling, which enables probabilistic modeling when there is uncertainty in the network structure, effectively mitigating model bias caused by measurement errors, missing information, and other issues.

[0112] The initial probability modeling of whether relationships exist between users is performed using a multilayer perceptron model, which includes the following steps:

[0113] Using a multilayer perceptron model, with feature vectors As input, As output For the first individual users With the individual users Is there a relationship between them?

[0114] The training formula is as follows:

[0115] , (9)

[0116] in, , , , For learnable parameters, It is the Sigmoid activation function.

[0117] , ,

[0118] For users and The initial connection probability.

[0119] The fourth step involves Monte Carlo augmentation sampling based on the initial connection probability, which breaks the limitations of the traditional single-graph prediction framework and uses multiple sampled graphs to more realistically approximate the real social network.

[0120] (1) Utilizing users and Initial connection probability Monte Carlo sampling-based generative enhancement network ,

[0121] in, For the enhanced social network, A collection of users on a social network. To enhance the edge set of the social network;

[0122] ,

[0123] , (10)

[0124] In formula (10) For the generated random numbers, To enhance the network individual users With the individual users Is there a relationship between them?

[0125] (2) Repeat step (1) to generate An enhanced network ,

[0126] ,in Indicates the number of augmentation networks.

[0127] For the first A set of relationships between two users in an enhanced network. For the first The first enhanced network individual users With the individual users Is there a relationship between them?

[0128] if Indicates the first In the enhanced network, the first individual users With the individual users They are related and neighbors;

[0129] if Indicates the first In the enhanced network, the first individual users With the individual users It doesn't matter.

[0130] The fifth step is to obtain the social network relationship prediction results: repeat the feature extraction and training process in the augmented network, average the prediction results of multiple augmented networks, obtain the edge probability estimate, and obtain the social network relationship prediction results.

[0131] (1) For each augmentation network Following the steps of extracting structural and attribute features between each pair of users, the corresponding features are extracted again. ,

[0132] use As input, As output For feature vectors, For the first The first enhanced network individual users With the individual users Is there a relationship between them?

[0133] Train the following multilayer perceptron model:

[0134] (11)

[0135] in, This represents vector concatenation. , , , For learning parameters, For users and In the first Connection probability in an enhanced network This is the Sigmoid activation function.

[0136] (2) Fusion of multi-network prediction results:

[0137] Predicting the first using formula (5) individual users With the individual users Probability of connecting edges :

[0138] (5)

[0139] in, To increase the number of networks.

[0140] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A social network relationship prediction method based on Monte Carlo simulation network enhancement technology, characterized in that, Includes the following steps: 11) Construct a social network graph model; 12) Extract the structural and attribute features between each pair of users; The extraction of structural and attribute features between each pair of users includes the following steps: 121) For any two users and Extracting feature vectors ; 122) The extracted features include: Number of common neighbors , , (1) in, Indicates user Neighbors User The neighbors; Resource allocation indicators : , (2) in, User The neighbors; Adamic-Adar indicator , ; (3) Node degree similarity , ; (4) Age similarity , , in, and Representing users respectively and Age; Interest similarity , , (6) in, and Representing users respectively and A collection of interests and hobbies; Geographical distance similarity , , (7) in, and Representing users respectively and Geographical location; 123) User and eigenvectors Defined as: ; 13) Perform initial probability modeling; 14) Monte Carlo augmented sampling based on initial connectivity probability; The Monte Carlo enhancement sampling based on the initial connectivity probability includes the following steps: 141) Utilizing users and Initial connection probability Monte Carlo sampling-based generative enhancement network , in, For the enhanced social network, A collection of users on a social network. To enhance the edge set of the social network; , , (10) In formula (10) For the generated random numbers, To enhance the network individual users With the individual users Is there a relationship between them? 142) Repeat step 141) to generate An enhanced network , , ,in Indicates the number of augmentation networks. For the first A set of relationships between two users in an enhanced network. For the first The first enhanced network individual users With the individual users Is there a relationship between them? if Indicates the first In the enhanced network, the first individual users With the individual users They are related and neighbors; if Indicates the first In the enhanced network, the first individual users With the individual users It doesn't matter; 15) Obtaining the prediction results of social network relationships: Repeat the feature extraction and training process in the augmented network, average the prediction results of multiple augmented networks, obtain the edge probability estimate, and obtain the prediction results of social network relationships. Obtaining the social network relationship prediction results includes the following steps: 151) For each augmentation network Following the steps of extracting structural and attribute features between each pair of users, the corresponding features are extracted again. , use As input, As output For feature vectors, For the first The first enhanced network individual users With the individual users Is there a relationship between them? Train the following multilayer perceptron model: ,(11) in, This represents vector concatenation. , , , For learning parameters, For users and In the first Connection probability in an enhanced network Use the Sigmoid activation function; 152) Fusion of multi-network prediction results: Predicting the first using formula (5) individual users With the individual users Probability of connecting edges : ,(5) in, To increase the number of networks.

2. The social network relationship prediction method based on Monte Carlo simulation network enhancement technology according to claim 1, characterized in that, The construction of the social network graph model includes the following steps: Extracting social networks , Represents the set of users on a social network. Indicates the first One user, Indicates the number of users; Represents the set of relationships between any two users. For the first individual users With the individual users Is there a relationship between them? if Indicates the first individual users With the individual users They have a relationship and are neighbors; if Indicates the first individual users With the individual users It doesn't matter.

3. The social network relationship prediction method based on Monte Carlo simulation network enhancement technology according to claim 1, characterized in that, The initial probability modeling involves using a multilayer perceptron model to perform initial probability modeling on whether a relationship exists between users, which includes the following steps: Using a multilayer perceptron model, with feature vectors As input, As output For the first individual users With the individual users Is there a relationship between them? The training formula is as follows: , (9) in, , , , For learnable parameters, It is the Sigmoid activation function. , , For users and The initial connection probability.

4. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, enables the social network relationship prediction method based on Monte Carlo simulation network enhancement technology as described in any one of claims 1-3.

5. A computer device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it can implement the social network relationship prediction method based on Monte Carlo simulation network enhancement technology as described in any one of claims 1-3.