Block chain social network opinion leader identification method and device, medium and product

Abstract

The invention discloses a block chain social network opinion leader identification method and device, a medium and a product, and relates to the field of data processing. The method comprises the following steps: constructing a directed network based on user data of a block chain social network; obtaining a network structure characteristic index of the directed network; determining key nodes in the plurality of nodes according to the network structure characteristic indexes; obtaining a contribution behavior characteristic index of the user in the block chain social network; based on an entropy weight method and a standard correlation weight method, determining index weights of the network structure characteristic indexes and the contribution behavior characteristic indexes; determining an influence score of the user according to the network structure feature index, the contribution behavior feature index and the index weight; and determining an opinion leader of the block chain social network from the users according to the influence score. According to the method, the opinion leader in the block chain social network is accurately identified by combining the entropy weight method and the standard correlation weight method and utilizing the network structure characteristic index and the contribution behavior characteristic index.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method, device, medium and product for identifying opinion leaders in a blockchain social network. Background Art

[0002] Blockchain social network is a social network platform built using blockchain technology. It solves the problems of uneven distribution of benefits, resource monopoly and personal privacy leakage in traditional social networks through its decentralized characteristics. In blockchain social networks, each user is both the creator of content and the owner of the network. They can autonomously execute agreed information through smart contracts to ensure that data cannot be tampered with and has high transparency. This network not only improves the quality of information, enhances transparency and authenticity, but also increases user participation, effectively curbs the spread of rumors, and provides the possibility of building a healthier, open and responsive public opinion environment.

[0003] Opinion leaders are important figures in social communication and network interaction in blockchain social networks. With their professional knowledge, skills or influence, they can effectively guide public opinion, accelerate information dissemination, and influence the public's views and behavior patterns. Opinion leaders play a very critical role in guiding public opinion, supervising and participating in the production and dissemination of high-quality content.

[0004] In order to give full play to the important role of opinion leaders in blockchain social networks, how to accurately identify opinion leaders is an urgent problem that needs to be solved. Summary of the invention

[0005] In view of the above technical problems and defects, the purpose of the present invention is to provide a method, device, medium and product for identifying opinion leaders in a blockchain social network, which can accurately identify opinion leaders in a blockchain social network.

[0006] To achieve the above-mentioned purpose, in a first aspect, the present invention provides a method for identifying opinion leaders in a blockchain social network, comprising: constructing a directed network based on user data of the blockchain social network, the directed network comprising a plurality of nodes and a plurality of connection edges, the nodes representing users of the blockchain social network, and the connection edges representing interaction channels between users; obtaining network structure characteristic indicators of the directed network, the network structure characteristic indicators comprising at least one of degree centrality, eigenvector centrality, betweenness centrality and proximity centrality, the degree centrality being used to measure the number of other nodes directly connected to a node in a directed network, the eigenvector centrality being used to reflect the importance of adjacent nodes of a node, the betweenness centrality being used to measure the ability of a node to serve as an information bridge in a directed network, and the proximity centrality being used to measure the ability of a node to serve as an information bridge in a directed network. It is used to reflect the degree of proximity between a node and other nodes, and is measured by the sum of the shortest distances from the node to all other nodes; key nodes are determined among multiple nodes based on network structure characteristic indicators; after the key nodes are determined, the contribution behavior characteristic indicators of users in the blockchain social network are obtained, and the contribution behavior characteristic indicators include at least one of the user's reputation value, number of bounties obtained, originality, number of supporters, number of followers, and activity; based on the entropy weight method and the inter-standard correlation weight method, the indicator weights of the network structure characteristic indicators and the contribution behavior characteristic indicators are determined; based on the network structure characteristic indicators, the contribution behavior characteristic indicators and the indicator weights, the user's influence score is determined; and the opinion leaders of the blockchain social network are determined from the users based on the influence score.

[0007] The present invention adopts the above method to construct a directed network, regards each user in the blockchain social network as a node, and regards the interaction between users as a directed edge, thereby forming a structured network model. Then, the network structure characteristic indicators are calculated, including degree centrality, eigenvector centrality, betweenness centrality and proximity centrality, which measure the importance and influence of nodes in the network from different angles. After that, the user's contribution behavior characteristic indicators are obtained, such as user reputation value, number of bounties, originality, number of supporters and activity, which quantify the user's performance in content creation, community participation and influence. The weights of each indicator are determined by the entropy weight method and the inter-standard correlation weight method, and the multi-dimensional information of network structure and user behavior is comprehensively considered. Finally, the influence score of each user is calculated based on the weighted network structure characteristic indicators and contribution behavior characteristic indicators, and the opinion leaders are identified accordingly. This method not only takes into account the user's position in the network, but also combines the user's behavior and contribution, so as to more comprehensively and accurately identify opinion leaders with actual influence.

[0008] In combination with some embodiments of the first aspect, in some embodiments, the user's influence score is determined based on network structure characteristic indicators, contribution behavior characteristic indicators and indicator weights, including: based on the entropy weight method, the inter-standard correlation weight method and the approximate ideal solution sorting method, using network structure characteristic indicators, contribution behavior characteristic indicators and indicator weights to construct an opinion leader identification model; and obtaining the user's influence score through the opinion leader identification model.

[0009] By adopting the technical solution of the above embodiment, a comprehensive opinion leader identification model is constructed by combining the entropy weight method, the inter-standard correlation weight method and the approximate ideal solution sorting method. This model not only takes into account the network structure characteristics of users, such as degree centrality and betweenness centrality, but also includes the behavioral characteristics of users, such as reputation value and activity. This method converts qualitative indicators into quantitative data by constructing a decision matrix, making the influence score more scientific and objective. In addition, the influence score of the user is determined by calculating the distance between the user and the ideal solution, which further improves the accuracy and reliability of the model. This comprehensive evaluation method helps to more accurately identify opinion leaders who have real influence in blockchain social networks.

[0010] In combination with some embodiments of the first aspect, in some embodiments, the opinion leader identification model is used to: use network structure characteristic indicators and contribution behavior characteristic indicators as indicator elements to construct a decision matrix; normalize the indicator elements to obtain normalized indicator elements; generate a normalized decision matrix based on the normalized indicator elements and indicator weights; determine the element with the largest weighted indicator value in the normalized decision matrix as the positive ideal solution, and determine the element with the smallest weighted indicator value as the negative ideal solution; respectively determine the first Euclidean distance between the normalized indicator element corresponding to the user and the positive ideal solution, and the second Euclidean distance between the normalized indicator element and the negative ideal solution; determine the user's influence score based on the first Euclidean distance and the second Euclidean distance.

[0011] By adopting the technical solution of the above embodiment, by determining the positive and negative ideal solutions, the distance between the user and these ideal states is quantified, thereby evaluating their influence. By calculating the first and second Euclidean distances, the model can distinguish the performance of users and determine their level of influence in the network. The beneficial effect of this method is that it provides a systematic and standardized way to evaluate and compare the influence of different users, making the identification process of opinion leaders more transparent and easy to understand, thereby improving the practicality and operability of the model.

[0012] In combination with some embodiments of the first aspect, in some embodiments, determining the influence score of a user based on the first Euclidean distance and the second Euclidean distance includes: dividing the second Euclidean distance by the sum of the first Euclidean distance and the second Euclidean distance to obtain the proximity of the node to the ideal solution; and determining the influence score of the user based on the proximity.

[0013] The technical solution of the above embodiment is adopted to determine the influence score by calculating the user's proximity to the positive ideal solution and the negative ideal solution. This method not only considers the user's gap with the optimal performance, but also the distance from the worst performance, thereby providing a comprehensive perspective to evaluate the user's influence. By dividing the second Euclidean distance by the sum of the two distances, the resulting ratio can better reflect the user's relative position relative to all other users. This scoring mechanism can more accurately identify users who perform well in multiple key indicators, thereby improving the accuracy and effectiveness of opinion leader identification.

[0014] In combination with some embodiments of the first aspect, in some embodiments, after determining the opinion leaders of the blockchain social network from users according to the influence score, it also includes: obtaining the cumulative number of votes for the content published by the opinion leader on the blockchain social network within a set time period; evaluating the network influence of the opinion leader according to the cumulative number of votes to obtain a first network influence evaluation result; and testing the model validity of the opinion leader identification model according to the first network influence evaluation result and the influence score, and the model validity is used to characterize the accuracy of identifying the opinion leader.

[0015] By adopting the technical solution of the above embodiment, by obtaining the cumulative number of votes of opinion leaders within a set time period, their content popularity and community recognition can be actually evaluated. The introduction of this method makes the output of the model more reliable because it combines the comparison of model prediction and actual performance. In addition, by comparing the cumulative number of votes and influence scores to test the effectiveness of the model, the model parameters can be further optimized to ensure that the model can continuously and accurately identify users with high influence.

[0016] In combination with some embodiments of the first aspect, in some embodiments, after determining the opinion leaders of the blockchain social network from users according to the influence score, it also includes: obtaining the reposting rate of the content published by the opinion leader on the blockchain social network within a set time period; evaluating the network influence of the opinion leader according to the reposting rate to obtain a second network influence evaluation result; and testing the model effectiveness of the opinion leader identification model according to the second network influence evaluation result and the influence score.

[0017] By adopting the technical solution of the above embodiment, the application scope of the opinion leader identification model is expanded, and the repost rate is added as another dimension for evaluating network influence. The repost rate is an important indicator for measuring the dissemination of content. By analyzing the frequency with which the opinion leader's content is reposted, its influence in social networks can be more comprehensively evaluated. The introduction of this method enables the model to capture the user's influence from more angles, thereby improving the accuracy of the identification results. In addition, by combining the repost rate with the influence score to test the model's effectiveness, the model's predictive ability can be further verified and optimized.

[0018] In combination with some embodiments of the first aspect, in some embodiments, based on the entropy weight method and the inter-standard correlation weight method, the indicator weights of the network structure characteristic indicators and the contribution behavior characteristic indicators are determined, including: based on the entropy weight method, determining the information entropy of the normalized indicator elements corresponding to the network structure characteristic indicators and the contribution behavior characteristic indicators; determining the first weight based on the information entropy; determining the standard deviation value of the normalized indicator elements, and the indicator correlation coefficient between the normalized indicator elements; based on the inter-standard correlation weight method, determining the second weight according to the standard deviation value and the indicator correlation coefficient; determining the indicator weight according to the first weight and the second weight.

[0019] The technical solution of the above embodiment is adopted to determine the information entropy of each indicator by the entropy weight method, and then calculate the first weight based on the information entropy. Next, the second weight is determined by calculating the standard deviation value and correlation coefficient of the normalized indicator elements, and using the standard correlation weight method. Finally, the final indicator weight is determined by combining these two weights. The beneficial effect of this comprehensive weight determination method is that it can more scientifically and objectively reflect the importance of each indicator in the evaluation process, thereby improving the accuracy and reliability of opinion leader identification.

[0020] In a second aspect, an embodiment of the present invention provides an electronic device, comprising: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the electronic device to execute the method described in the first aspect or the second aspect, and any possible implementation method of the first aspect or the second aspect.

[0021] In a third aspect, the present invention provides a computer-readable storage medium comprising instructions, which, when executed on the electronic device, causes the electronic device to execute the method described in the first aspect or the second aspect, and any possible implementation of the first aspect or the second aspect.

[0022] In a fourth aspect, the present invention provides a computer program product comprising instructions, which, when the computer program product is run on the electronic device, enables the electronic device to execute the method described in the first aspect or the second aspect, and any possible implementation of the first aspect or the second aspect.

[0023] It is understandable that the electronic device provided in the second aspect, the storage medium provided in the third aspect, and the computer program product provided in the fourth aspect are all used to execute the method provided in the present invention. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding method, which will not be repeated here.

[0024] One or more technical solutions provided by the present invention have at least the following technical effects or advantages: 1. This invention realizes the accurate identification of opinion leaders in blockchain social networks by constructing a directed network model and comprehensively evaluating the network structure characteristics and user contribution behavior characteristics. This method not only considers the number and quality of direct connections of users in the network, but also combines multi-dimensional behavioral characteristics such as user reputation, bounty, originality, etc., and comprehensively evaluates the influence of users through scientifically determined indicator weights, thereby effectively identifying truly influential opinion leaders.

[0025] 2. The present invention adopts the entropy weight method and the inter-criteria correlation weight method to determine the weight of each evaluation index, ensuring the objectivity and scientificity of the evaluation system. By constructing a decision matrix, performing normalization processing, and using the approximate ideal solution sorting method to calculate the user's influence score, this method combines multiple factors, improves the accuracy and reliability of the opinion leader identification model, and makes the identification results more in line with the actual situation.

[0026] 3. The present invention makes the implementation process of the model clear and easy to operate through specific calculation steps and methods, such as determining positive ideal solutions and negative ideal solutions, calculating Euclidean distance, and evaluating cumulative votes and forwarding rates. In addition, the verification of the model's effectiveness further ensures the practicability of the model, enabling it to be widely used in multiple fields such as social network analysis, public opinion monitoring, and market research. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The drawings herein are incorporated into and constitute a part of the specification, showing embodiments consistent with the present invention, and together with the specification, are used to explain the principles of the present invention. Obviously, the drawings described below are only some embodiments of the present invention, and for those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative work. In the drawings: Figure 1 It is a flow chart of a method for identifying opinion leaders in a blockchain social network according to an embodiment of the present invention; Figure 2 This is a flowchart of an application framework of a blockchain social network opinion leader identification method according to an embodiment of the present invention; Figure 3a Schematic diagram of weight distribution of different indicators in an embodiment of the present invention; Figure 3b is a schematic diagram of weight allocation of another different indicator in an embodiment of the present invention, specifically Figure 3a The corresponding grayscale image; Figure 4 This is a comparative analysis diagram of cumulative voting influence in an embodiment of the present invention; Figure 5 This is a comparative analysis diagram of the influence of reposting rate in an embodiment of the present invention; Figure 6 It is a schematic diagram of the architecture of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0028] The terms used in the following embodiments of the present invention are only for the purpose of describing specific embodiments, and are not intended to be limiting of the present invention. As used in the specification of the present invention, the singular expressions "a", "a", "above", "the" and "this" are intended to also include plural expressions, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used in the present invention refers to any or all possible combinations comprising one or more of the listed items.

[0029] In the following, the terms "first" and "second" are used for descriptive purposes only and should not be understood as implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of the present invention, unless otherwise specified, the meaning of "plurality" is two or more.

[0030] It should also be noted that, unless otherwise clearly specified and limited, in the embodiments of the present invention, the terms such as "setting" and "connection" should be understood in a broad sense. For example, "connection" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two components; it can be a wired communication connection or a wireless communication connection. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to the specific circumstances. The embodiments of the present invention are described in detail below.

[0031] The embodiment of the present invention provides a method for identifying opinion leaders in a blockchain social network. Figure 1 As shown, the following steps are included: Step 201: construct a directed network based on user data of the blockchain social network.

[0032] Among them, a directed network is a graph structure. A directed network includes multiple nodes and multiple connecting edges. The nodes represent the users of the blockchain social network, the connecting edges represent the interaction channels between users, and the direction of the edges represents the initiator and receiver of the relationship, thereby capturing the directionality of information flow and influence propagation.

[0033] In order to construct this directed network, it is necessary to first collect user data, including user ID, post content, comments, likes, forwarding, voting and other interactive behaviors. Each user is regarded as a node in the social network, and the interactive behaviors between users are represented as directed edges. Specifically, if user A comments, likes, forwards or votes on user B's post, there will be a directed edge from node A to node B in the network. In this way, complex social interactions can be transformed into a structured directed network graph. This directed network not only reflects the direct connection between users, but also reveals the path of information flow and influence propagation, providing a basis for subsequent analysis and identification of opinion leaders.

[0034] Specifically, the blockchain social network consists of a large number of users, each user can be regarded as a node, and the edge is the channel for interaction between users. In order to simulate the two-way communication between users, a directed network is constructed. ;in, , represents the set of all nodes; Represents the set of edges connecting nodes; , represents the adjacency matrix of G. Obviously, In a directed network, if there is a connection between nodes, then ,otherwise .

[0035] Step 202, obtaining network structure characteristic indicators of the directed network.

[0036] The network structure characteristic index includes at least one of degree centrality, eigenvector centrality, betweenness centrality and closeness centrality.

[0037] Degree centrality is used to measure the number of other nodes that a node is directly connected to in a directed network. As a local attribute ranking indicator, the degree centrality of a node reflects the direct influence of a node on other nodes in the network, which also means that nodes with larger degree centrality have more connections and information channels.

[0038] Furthermore, degree centrality includes in-degree centrality and out-degree centrality. In-degree centrality represents the number of comments, reposts, and votes received by users of the blockchain social media platform from other users. Out-degree centrality represents the number of comments, reposts, and votes received by users of the blockchain social media platform to other users.

[0039] In order to eliminate the influence of network scale on degree centrality, the degree centrality calculation formula of the node is: ; Among them, DC (V i ) represents degree centrality, n represents the number of nodes, i and j are index numbers, and x ij Indicates the connection status of node i to other nodes j.

[0040] Eigenvector centrality is used to reflect the importance of a node's neighboring nodes. The calculation formula for eigenvector centrality is as follows: ; Among them, EC(V i ) represents the eigenvector centrality, λ is the largest eigenvalue of the adjacency matrix X; is the eigenvector corresponding to the largest eigenvalue of the adjacency matrix X. The importance of a node depends on the number of its neighbors and the importance of its neighbors.

[0041] Betweenness centrality is used to measure the ability of a node to serve as an information bridge in a directed network. Betweenness centrality describes how a node controls or influences the relationship between two nodes that are not directly connected to it, reflects the bridging function of the node in the network, and can be used as an important indicator for controlling network information exchange. Betweenness centrality can be measured by the number of times the shortest path between all node pairs passes through the target node.

[0042] The calculation formula of betweenness centrality is as follows: ; Among them, BC(V i ) represents the betweenness centrality, n jk represents the number of shortest paths between node j and node k, n jk (V i ) represents the number of shortest paths between node j and node k through node i.

[0043] Closeness centrality is used to reflect the closeness between a node and other nodes, measured by the sum of the shortest distances from the node to all other nodes. The calculation formula for closeness centrality is as follows: ; Among them, CC(V i ) represents the closeness, d ijRepresents the shortest distance from node i to node j. The higher the closeness centrality value, the closer the relationship between the node and other nodes. Compared with degree centrality, closeness centrality can better reflect the global structure of the network.

[0044] Through these network structure characteristic indicators, we can comprehensively evaluate the position and influence of nodes in the network, providing a scientific basis for identifying opinion leaders.

[0045] Step 203: determining key nodes among multiple nodes according to the network structure characteristic index.

[0046] Specifically, the degree centrality, eigenvector centrality, betweenness centrality, and closeness centrality of each node are first calculated.

[0047] Afterwards, these four indicators are weighted and comprehensively analyzed according to the specific application scenarios and goals. The weights can be determined based on expert knowledge, indicator importance analysis, or through data-driven methods such as principal component analysis (PCA). For example, in a network that focuses on the rapid spread of information, closeness centrality and betweenness centrality may be given higher weights; while in a network that emphasizes node quality and influence, eigenvector centrality is more critical. By weighted summation or constructing a comprehensive indicator score, we can get a comprehensive influence score for each node.

[0048] Finally, all nodes are ranked according to this score, and the nodes with the highest scores are identified as key nodes with high influence, high connectivity and key positions in the network. These nodes are potential opinion leaders and are of great significance in guiding online public opinion, promoting information dissemination and maintaining network stability.

[0049] Step 204, when the key nodes are determined, obtain the user's contribution behavior characteristic indicators in the blockchain social network.

[0050] Among them, the contribution behavior characteristic indicators include at least one of user reputation value, number of bounties received, originality, number of supporters, number of followers and activity.

[0051] As an important way to measure the value that users bring to social networks, the reputation scoring mechanism quantifies the user's contribution behavior. Users gain reputation value through collective consensus by creating high-quality content, attracting user votes, and participating in network governance, which reflects the social network's recognition of the value created by users.

[0052] In blockchain social networks, user reputation becomes an important indicator to measure the recognition of user contributions. As user reputation grows, their influence and status in the social network also increase. Users with high reputation are regarded as being able to publish reliable information sources, which promotes communication and cooperation between users without a trust basis.

[0053] The calculation formula for measuring user reputation value is as follows: ; Among them, RP i is defined as the reputation value (reputation score) of user i, and CP i It represents the energy value obtained by user i for participating in contributing content. These energy values ​​are accumulated through each vote obtained by the user.

[0054] According to the calculation formula of this user reputation value, the user's reputation score will grow rapidly in the initial stage as the number of votes accumulated. However, when the reputation score reaches a certain level, the growth rate becomes slow, which means that in order to maintain the continuous growth of the reputation score, the user needs to accumulate more votes through more frequent content contributions.

[0055] Therefore, continuous content contribution is a key factor in improving user reputation, attracting widespread attention and enhancing influence, emphasizing that in blockchain social networks, the formation of user reputation is a dynamic process based on content engagement, which requires users to continuously participate and contribute high-quality content to achieve reputation growth and enhance personal influence.

[0056] The token-based incentive mechanism encourages users to participate in various activities of blockchain social networks, such as content creation and content promotion, by distributing rewards. The number of bounties obtained is a quantitative indicator of rewards, which is defined as the total amount of rewards obtained by users for contributing high-quality content in blockchain social networks. The distribution of bounties as a recognition of user contributions by social networks not only enhances the status of content creators in the network, but also encourages the willingness of a wide range of user groups to participate, prompting them to actively participate in public opinion discussions and promote the dissemination of high-quality content.

[0057] Content that receives a large number of bounties usually gains higher visibility on social networking platforms, highlighting the contribution of content creators, attracting more users' attention, and thus increasing their influence.

[0058] The calculation formula for the bounty is as follows: ; Among them, R i represents the amount of reward received by user i for a single article, M represents the total amount in the reward pool, and V i Indicates the reward weight of the current article, Represents the sum of the reward weights of all pending articles.

[0059] According to the formula, the total amount of the reward pool M directly determines the amount of reward each article can receive. For a single article, the reward income is based on its relative reward weight among all the articles to be published.

[0060] Therefore, even if the weight of an article is V i If the comprehensive weight of other articles is also high, the actual reward of this article may not be high. This distribution strategy motivates creators to produce high-quality content, because high-quality content often wins a higher weight V i , and then get more rewards.

[0061] Originality is defined as the number of original articles published by users on a certain public opinion topic. Influential users tend to create content in their own words during topic discussions, rather than frequently reposting other people's content. More original content can bring different perspectives to more people and avoid monotony. In addition, the value of original content is much higher than reposted works.

[0062] The calculation formula for originality is as follows: ; Among them, OP i represents the originality of user i, TA i represents the total number of posts published by user i in the topic discussion, OA i Represents the total number of posts published by user i in the topic discussion that have been reposted. Originality reflects the innovation and uniqueness of content in the dissemination of public opinion, which is one of the key factors in measuring user influence. Original content can often trigger wider discussions and trigger deeper thinking, thus standing out in the information-saturated network environment and attracting more attention and dissemination.

[0063] In social networks, users' social attributes and economic aspects influence each other. The number of supporters in blockchain social networks is directly related to cryptocurrency. Users' social activities, such as posting, liking, and commenting, can attract other users to become supporters, which can be converted into token rewards. This tokenized incentive mechanism not only increases the motivation for users to participate, but also makes the number of supporters a key factor in user influence.

[0064] The number of supporters is the number of followers of a user, reflecting the user's popularity and appeal in social networks. Users with a large number of supporters usually have a strong social influence and can effectively spread information and guide public opinion.

[0065] The number of followers is the number of other users followed by the user, reflecting the breadth of the user's interests and the diversity of public opinion information acquisition. Users build their own information flow and social circle by following different people and organizations. This shows the social attributes of users to a certain extent, such as their interests, values, and the social groups they want to communicate and interact with.

[0066] The ratio of the number of supporters to the number of followers (supporters / followers) reflects the quality of the user's interaction with the network and balances the mutual influence between social attributes and economic aspects. Therefore, the ratio of the number of supporters to the number of followers is used as an indicator to evaluate user influence.

[0067] The formula for calculating the ratio of supporters to followers is as follows: ; Among them, FFR i represents the ratio of the number of supporters and followers of user i, followers i Indicates the number of supporters of user i, following i Indicates the number of followers of user i. For users who do not follow others or are not followed by others, set FFR i = 0. A high ratio usually indicates that the user has strong social influence and can attract many supporters without having to follow a large number of others in exchange for attention. This ratio can also reveal the user's social strategy, that is, consciously maintaining a high supporter / following ratio may be an important decision aimed at building a professional image or brand reputation, thereby enhancing one's own influence.

[0068] Activity is defined as the frequency of days in which users participate in or generate activities in blockchain social networks. Such activities may include posting content, commenting, liking, sharing information or participating in discussions. The number of active days reflects the user's continued attention and participation in the platform or a specific topic within a certain period of time. Activity is the ratio of active days to topic discussion days, which represents that users participate in topic interactions more frequently and have formed a certain influence in the network.

[0069] The calculation formula of activity is as follows: ; Among them, AC i Represents the activity of user i, Ad i Represents the total number of days that a topic remains hot, used to describe how long a topic has sparked discussion over a period of time. i It represents the total number of days that user i posts, comments or interacts during the topic discussion period. This indicator can reflect the user i’s interest in a specific topic. A user with long-term activity may have a deeper understanding of a topic.

[0070] The more a user contributes in a blockchain social network, the higher the reward and the higher the value achieved, thus enhancing their influence in the community. By combining the above-mentioned contribution behavior characteristic indicators, we can more comprehensively evaluate the influence of users in blockchain social networks and identify truly influential users. These indicators not only take into account the user's content creation ability, but also the user's community participation and social influence, providing a multi-dimensional evaluation basis for the identification of opinion leaders.

[0071] Step 205, based on the entropy weight method and the inter-criteria correlation weight method, determine the index weights of the network structure characteristic index and the contribution behavior characteristic index.

[0072] Among them, the entropy method is an objective weighting method based on the information entropy of indicators. By calculating the variability (information entropy) of each indicator, indicators with large variability are considered to contain more information and are thus given higher weights.

[0073] Criteria Importance Through Intercriteria Correlation (CRITIC) is a weighting method that takes into account the contrast intensity and conflict between indicators. By calculating the standard deviation and correlation coefficient of each indicator, indicators with large standard deviations and small correlations with other indicators will receive higher weights.

[0074] For the sake of convenience, the network structure characteristic indicators and contribution behavior characteristic indicators are collectively referred to as characteristic indicators or indicators below.

[0075] When determining the indicator weights, the entropy weight method is first used to evaluate the degree of dispersion of each indicator based on the concept of information entropy. Indicators with greater information content are given higher weights to ensure the objectivity of the weights. Then, combined with the CRITI method, the weights are further adjusted by analyzing the correlation and conflict between indicators to ensure the independence and complementarity of the indicators.

[0076] Specifically, the entropy weight method measures the uncertainty of each indicator by calculating its information entropy. The smaller the information entropy, the greater the amount of information provided by the indicator and the higher the weight. The CRITIC rule evaluates the comparative strength and conflict of indicators by calculating the standard deviation of the indicator and the correlation coefficient with other indicators. The larger the standard deviation and the lower the correlation, the higher the weight of the indicator.

[0077] Finally, the final weight of each indicator is determined by combining the results of the entropy weight method and the CRITIC method, ensuring that the weight distribution takes into account both the dispersion of the indicators and the correlation between the indicators, making the weight distribution more scientific and reasonable, and providing a reliable basis for subsequent user influence evaluation.

[0078] Step 206, determining the user's influence score based on the network structure characteristic index, contribution behavior characteristic index and index weight.

[0079] Specifically, firstly, the network structure characteristic index and contribution behavior characteristic index are standardized to ensure that each index is comparable at the same level. Then, the weighted sum of each user's indicators is calculated according to the index weight determined in the previous step to calculate the user's influence score.

[0080] Step 207: Determine opinion leaders of the blockchain social network from the users based on the influence scores.

[0081] Specifically, after calculating the influence scores of all users, opinion leaders can be identified by sorting these scores. Specifically, the influence scores of users are sorted from high to low, and the users with the highest scores are considered to be the most influential opinion leaders. Usually, the top k users with the highest scores can be selected as opinion leaders, where k can be determined based on the specific application scenario and needs. For example, the top 10% of users can be selected as opinion leaders. These opinion leaders have high influence, high connectivity and key positions in the network, can effectively guide the direction of public opinion, promote information dissemination, and play an important role in maintaining network order and improving content quality.

[0082] In this way, truly influential users, namely opinion leaders, can be identified, providing strong support for opinion management and content recommendation in blockchain social networks.

[0083] The method of this embodiment adopts the above series of steps. First, by constructing a directed network, each user in the blockchain social network is regarded as a node, and the interaction between users is regarded as a directed edge, thereby forming a structured network model. Then, the network structure characteristic indicators are calculated, including degree centrality, eigenvector centrality, betweenness centrality and proximity centrality. These indicators measure the importance and influence of nodes in the network from different angles. After that, the user's contribution behavior characteristic indicators are obtained, such as user reputation value, number of bounties, originality, number of supporters and activity. These indicators quantify the user's performance in content creation, community participation and influence. The weight of each indicator is determined by the entropy weight method and the inter-standard correlation weight method, and the multi-dimensional information of network structure and user behavior is comprehensively considered. Finally, the influence score of each user is calculated based on the weighted network structure characteristic indicators and contribution behavior characteristic indicators, and opinion leaders are identified accordingly. This method not only takes into account the user's position in the network, but also combines the user's behavior and contribution, so as to more comprehensively and accurately identify opinion leaders with actual influence. It provides strong support for public opinion management and content recommendation in blockchain social networks, and helps to build a healthier, open and real network ecosystem.

[0084] refer to Figure 2 , showing the application framework process of the blockchain social network opinion leader identification method of this embodiment. First, by analyzing the structure of users in the network, key individuals (i.e., key users) and content contributors are identified, which includes examining the voting relationship, forwarding relationship, and comment relationship of users to construct network centrality indicators (i.e., network structure characteristic indicators), and analyzing users' contribution behaviors such as contribution recognition and active participation to construct contribution content characteristic indicators (i.e., contribution behavior characteristic indicators). Then, these indicator data are broadcast and chained and stored on the blockchain (Block). Then, by comprehensively evaluating the influence of users, including interactive data such as voting, forwarding, and commenting based on blockchain, users are ranked to identify opinion leaders with high influence.

[0085] This process not only utilizes the transparency and immutability of blockchain to ensure the credibility of data, but also more accurately evaluates the influence of users by integrating indicators in multiple dimensions.

[0086] In some embodiments, step 205 may specifically include the following steps: Step 2051, based on the entropy weight method, determine the information entropy of the normalized index elements corresponding to the network structure characteristic index and the contribution behavior characteristic index.

[0087] Among them, the normalized indicator element is the element in the normalized decision matrix, and the normalized decision matrix is ​​obtained by normalizing the decision matrix. The decision matrix is ​​a data structure used to store and organize the values ​​of users in the blockchain social network on multiple evaluation indicators (such as network structure characteristic indicators and contribution behavior characteristic indicators) for further analysis and calculation, so as to identify opinion leaders in the social network.

[0088] The decision matrix is ​​created as follows: The network structure characteristic index and contribution characteristic behavior index calculated above are selected to form a comprehensive evaluation index of user influence, and a decision matrix with n users and m indicators in the blockchain social network is constructed. , the decision matrix is ​​as follows: ; The elements in the decision matrix are indicator elements, representing the network structure characteristic indicators and contribution behavior characteristic indicators.

[0089] In this embodiment, different indicators often have different magnitudes, and direct comparison may result in some indicators having too large or too small an impact on the results. The normalized decision matrix is ​​intended to convert data from different magnitudes to the same scale, so that the data can be fairly compared.

[0090] The calculation formula using normalization is as follows: ; Among them, x ’ ij and x ij denote the normalized value and original value of the jth indicator of the jth user, max x j and min x j They represent the maximum and minimum values ​​of the j-th indicator respectively.

[0091] Step 2052, determining a first weight based on information entropy.

[0092] Specifically, the Entropy method is used to calculate the information entropy or divergence of each indicator. Information entropy is a measure of uncertainty and diversity that captures the amount of information provided by an indicator. The smaller the information entropy, the greater the amount of information provided by the indicator, and therefore a higher weight should be given.

[0093] The information entropy calculation formula is as follows: ; The formula for calculating the first weight is as follows: ; in, is the probability of a specific indicator appearing, defined as , is the jth index value of user i, E j is the information entropy of the jth indicator, It is the first weight calculated by the entropy weight method.

[0094] Step 2053, determining the standard deviation value of the normalized indicator elements and the indicator correlation coefficient between the normalized indicator elements.

[0095] Specifically, the standard deviation value of each normalized indicator element is first calculated. The standard deviation value reflects the degree of dispersion of the indicator. The larger the standard deviation, the wider the range of variation of the indicator, indicating that users have greater differences in performance on the indicator.

[0096] The CRITIC method can be applied to take m indicators and n blockchain social network users as input and normalize the value of each indicator to interval and calculate the standard deviation for each indicator.

[0097] Next, the correlation coefficient between the normalized indicator elements is calculated. The indicator correlation coefficient is used to measure the linear correlation between indicators. The correlation coefficient is used to analyze whether there is information overlap or conflict between different indicators. The lower the correlation coefficient, the stronger the conflict between indicators.

[0098] Step 2054, based on the inter-standard correlation weight method, determine the second weight according to the standard deviation value and the indicator correlation coefficient.

[0099] Using the standard deviation value and the indicator correlation coefficient, the second weight of each indicator can be determined by the CRITIC method. This method takes into account the variability of indicators and the independence between indicators, giving greater weight to indicators with large variability and low correlation with other indicators.

[0100] According to the CRITIC method, the second weight of each indicator is as shown in the formula: ; in, represents the second weight, σ j represents the standard deviation value, r jk represents the index correlation coefficient, j and k represent index numbers.

[0101] Step 2055, determining the indicator weight according to the first weight and the second weight.

[0102] Specifically, this embodiment comprehensively considers the Entropy method and the CRITIC method and defines a new weight allocation method, namely the ECWM (Entropy and CRITIC Weight Method) weight method, to capture the diversity, dispersion and correlation between different features.

[0103] The ECWM weight method is a comprehensive weight determination method that combines the Entropy method and the CRITIC method. It is used to objectively evaluate the weights of each indicator in multi-indicator decision-making. By considering the variability of indicators, the amount of information, and the correlation and comparison strength between indicators, it provides a more scientific and reasonable weight distribution for decision-making.

[0104] The ECWM comprehensive indicator weight calculation formula is as follows: ; Among them, W j is the indicator weight of the jth indicator calculated by the ECWM weight method, represents the first weight, represents the second weight, α and β are constants, and α+β = 1. The sizes of α and β can be defined according to actual needs.

[0105] refer to Figure 3a and 3b ,This figure shows the results of weight allocation for different indicators using Entropy, CRITIC and ECWM weight methods. Figure 3a and 3bIt includes multiple indicators, such as originality, user reputation value, number of bounties, activity, number of supporters / followers, out-degree centrality, in-degree centrality, closeness centrality, betweenness centrality and eigenvector centrality, etc. There are three colors of bar charts under each indicator, representing the weight values ​​obtained by three different weight allocation methods.

[0106] It can be seen that for most indicators, the weight values ​​obtained by the ECWM method are between Entropy and CRITIC, which shows that the ECWM weight method combines the advantages of the two methods and may provide a more balanced and comprehensive weight distribution. In particular, for the two indicators of "user reputation value" and "number of bounties obtained", the weight values ​​obtained by the ECWM method are significantly higher than those of the other two methods, which may mean that user reputation and economic incentives are particularly important factors when identifying opinion leaders.

[0107] In some embodiments, step 206 may specifically include: Step 2061, based on the entropy weight method, the inter-criteria correlation weight method and the approximate ideal solution sorting method, using the network structure characteristic index, the contribution behavior characteristic index and the index weight, construct an opinion leader identification model.

[0108] Among them, TOPSIS is a multi-criteria decision-making method that evaluates and ranks alternatives by calculating the relative proximity of each alternative to the positive ideal solution (optimal solution) and the negative ideal solution (worst solution). The higher the proximity, the better the solution.

[0109] Specifically, firstly, the entropy weight method and inter-criteria correlation weight method are used to determine the index weights of network structure characteristic indexes and user contribution behavior characteristic indexes. These weights reflect the importance of each index in evaluating user influence.

[0110] Next, the approximate ideal solution ranking method is used to calculate the score of each user in the weighted normalized decision matrix to determine its distance from the positive ideal solution (optimal solution) and the negative ideal solution (worst solution). The user's influence score is measured by its closeness to the positive ideal solution. The higher the closeness, the greater the user's influence in the network.

[0111] The opinion leader identification model integrates multi-dimensional information of network structure and user behavior, and can identify opinion leaders more comprehensively and accurately.

[0112] The opinion leader identification model of this embodiment is an analysis tool that comprehensively evaluates the influence of users in the network, obtains user influence scores by analyzing network structure characteristics and user contribution behavior characteristics, and combines weight calculation to identify potential opinion leaders with high influence.

[0113] Step 2062: Obtain the user's influence score through the opinion leader identification model.

[0114] Specifically, after the opinion leader identification model is constructed, the network structure characteristic index and contribution behavior characteristic index of each user are input into the opinion leader identification model, and the Euclidean distance between the user and the ideal optimal solution and the worst solution is calculated to obtain the user's proximity score. The higher the proximity score, the greater the influence of the user, and the higher the influence score.

[0115] In some embodiments, the opinion leader identification model is used to implement the following steps: S301, using the network structure characteristic index and the contribution behavior characteristic index as indicator elements to construct a decision matrix.

[0116] The decision matrix can be created in the manner described above, which will not be elaborated here.

[0117] S302, normalizing the index element to obtain a normalized index element.

[0118] This embodiment normalizes the indicator elements in the decision matrix to eliminate the influence of different indicator dimensions and magnitudes and ensure that the indicators are comparable on the same scale. The normalization process can be referred to the above description and will not be repeated here.

[0119] S303: Generate a normalized decision matrix according to the normalized indicator elements and indicator weights.

[0120] This embodiment can generate a weighted normalized decision matrix based on the normalized indicator elements and the indicator weights calculated by the ECWM weight method. Each element in the normalized decision matrix is ​​the product of the normalized indicator element and the corresponding weight.

[0121] Specifically, the indicator weight calculated by the ECWM weight method is denoted as W j , x ’ ij Represents the value of the normalized indicator element, and the weighted normalized decision matrix is ​​represented by Z ij , Z ij The calculation formula is as follows: .

[0122] S304: Determine the element with the largest weighted index value in the normalized decision matrix as a positive ideal solution, and determine the element with the smallest weighted index value as a negative ideal solution.

[0123] Specifically, each normalized indicator element is multiplied by its corresponding weight to obtain the weighted indicator value. The positive ideal solution is formed by selecting the element with the largest weighted indicator value in each indicator. It represents the best performance level that can be achieved on all indicators and is the ideal state of the decision-making goal. In contrast, the negative ideal solution is determined by the element with the smallest weighted indicator value in each indicator. It reflects the worst performance level and is a state that needs to be avoided in the decision-making process.

[0124] In this way, the difference between each alternative and the ideal and worst-case scenarios can be quantified, and their relative merits can be evaluated. This approach provides a clear and quantitative framework for multi-criteria decision-making, which helps to identify the best option in complex decision-making environments.

[0125] The calculation formulas for the positive ideal solution and the negative ideal solution are as follows: ; ; ; ; Among them, A + It represents the forward ideal solution, which is a vector composed of the maximum weighted normalized values ​​of all indicators, representing the ideal or optimal state under all evaluation indicators.

[0126] A − It represents the negative ideal solution, which is a vector composed of the minimum weighted normalized values ​​of all indicators, representing the worst or lowest standard state under all evaluation indicators.

[0127] Z j + The positive ideal solution of the j-th indicator is the maximum value of the weighted normalized values ​​of all users on the indicator, representing the optimal user performance on a single indicator.

[0128] Z j − It represents the negative ideal value of the j-th indicator, which is the minimum value of the weighted normalized values ​​of all users on this indicator, and represents the worst user performance on a single indicator.

[0129] S305 , respectively determining a first Euclidean distance between the normalized indicator element corresponding to the user and the positive ideal solution, and a second Euclidean distance between the normalized indicator element and the negative ideal solution.

[0130] Specifically, calculate each solution to A + or A − The Euclidean distance (Euclidean distance) of each solution to A + or A −The Euclidean distance represents the deviation between them. The calculation formula is as follows:

[0131]

[0132] in, represents the first Euclidean distance, represents the second Euclidean distance, represents the forward ideal solution, represents the negative ideal solution, represents the normalized decision matrix, i represents the index of the user, j represents the index of the indicator element, and m represents the total number of indicator elements.

[0133] S306: Determine the influence score of the user according to the first Euclidean distance and the second Euclidean distance.

[0134] Among them, a smaller first Euclidean distance means that the user is close to the ideal state in multiple key indicators and has a higher influence; while a larger second Euclidean distance indicates that the user is far away from the worst performance. Combining these two distances, we can quantify the influence of each user relative to all other users, ensuring that users with high scores are closer to the ideal state, thereby more accurately identifying opinion leaders in social networks and achieving objective evaluation of user influence.

[0135] This method not only considers the optimal performance of a single indicator, but also integrates the interactive effects of multiple indicators, providing a comprehensive and detailed means of user influence analysis.

[0136] In some embodiments, the second Euclidean distance may be divided by the sum of the first Euclidean distance and the second Euclidean distance to obtain the proximity between the user (node) and the ideal solution (including the positive ideal solution and the negative ideal solution).

[0137] The smaller the calculated proximity, the closer the user is to the ideal solution, the higher the influence score, and the more likely the user is to be considered an opinion leader.

[0138] This approach allows quantification of each user’s performance relative to both optimal and suboptimal scenarios, identifying those users who excel across multiple key metrics.

[0139] In some embodiments, the proximity can be calculated according to a proximity formula, where the proximity formula includes: ; Among them, S i Indicates proximity; is the first Euclidean distance, which represents the distance between user i and the forward ideal solution; is the second Euclidean distance, which represents the distance between user i and the negative ideal solution.

[0140] Proximity S i The value range of is between 0 and 1. The higher the corresponding influence score is, the closer the performance of user i is to the positive ideal solution and the farther away from the negative ideal solution is, that is, the greater the influence of the user is.

[0141] By calculating the proximity of all users, their influence scores can be obtained, and the influence scores are sorted, and the users with the highest scores are selected as opinion leaders in the blockchain social network. This method can comprehensively consider the performance of users in various indicators and ensure the objectivity and accuracy of the influence score.

[0142] In some embodiments, after step 207, the method further includes: Step 208, obtaining the cumulative number of votes for the content published by the opinion leader on the blockchain social network within a set time period.

[0143] Among them, the cumulative number of votes refers to the total number of all votes (including likes, dislikes, etc.) obtained by the content posted by users in the blockchain social network within a specific time period. It reflects the popularity of user content and the influence of users in the network.

[0144] In blockchain social networks, the number of votes accumulated after content is published is a way to evaluate the quality of content, reflecting the value of the publisher's contribution to the network, thereby enhancing the user's influence and status in the network. This voting process jointly participated by network members not only ensures collective consensus on content quality, but also promotes a form of autonomy based on contribution and recognition. In addition, the number of accumulated votes also represents the authority of the user in the social network, which can be used as another basic fact of user influence to supplement the communication ability based on simulation.

[0145] In order to obtain the cumulative number of votes for the content published by opinion leaders on the blockchain social network within a set time period, it is first necessary to determine the list of opinion leaders, which is usually done through the opinion leader identification model described previously. Next, for each opinion leader on the list, all the posts they published during the time period are retrieved from the database of the blockchain social network. Then, each post is analyzed and the number of votes received by the post is counted, including likes, dislikes, and other forms of votes. Finally, the number of votes for all posts of each opinion leader is accumulated to obtain their cumulative number of votes in the time period. This cumulative number of votes reflects the popularity and community recognition of the content published by the opinion leader, and is an important quantitative indicator for evaluating their network influence.

[0146] Specifically, the total number of votes obtained by analyzing the content posted by the user during the statistical time period can be recorded as V score , to measure the influence of users, the total number of cumulative votes is calculated as follows: ; Among them, u ip represents the pth post published by user i, vote(u ip ) represents the number of votes obtained by user i in post p.

[0147] In blockchain social networks, each vote is not only an affirmation of the quality of the post content, but also directly affects the ranking and dissemination of the content. As more votes are accumulated, the content posted by users will be more likely to be displayed to other users first, thereby achieving wider dissemination. Influential users can accumulate a large number of votes by continuously posting popular posts, enhancing their reputation in the network, making them opinion leaders who can effectively guide the direction of community opinion and discussion focus.

[0148] Step 209, evaluating the network influence of the opinion leader according to the accumulated number of votes to obtain a first network influence evaluation result.

[0149] Specifically, the larger the value of the cumulative number of votes, it usually means that the user's content has received wider attention and recognition, and its network influence is greater.

[0150] By comparing the cumulative votes of different opinion leaders, their network influence can be ranked, or a corresponding cumulative voting influence score can be assigned to obtain the first network influence evaluation result.

[0151] This helps to verify and optimize the accuracy of the opinion leader identification model and ensure that the identified opinion leaders do have high influence and recognition in social networks.

[0152] Step 210: Check the model validity of the opinion leader identification model based on the first network influence evaluation result and the influence score. The model validity is used to characterize the accuracy of identifying opinion leaders.

[0153] Specifically, the first network influence evaluation result, i.e., the cumulative voting influence score, can be compared with the influence score output by the previous opinion leader identification model to verify the model effectiveness of the opinion leader identification model.

[0154] If the opinion leaders identified by the opinion leader identification model do have a high influence in the community, then there should be a positive correlation between the influence score output by the model and the cumulative voting influence score. High model effectiveness means that the model can accurately identify truly influential users, thus providing a reliable basis for public opinion management and content recommendation in blockchain social networks.

[0155] For the opinion leader identification model of this embodiment, the top 105 candidate opinion leaders in a blockchain social network were selected as research objects, aiming to identify and verify potential opinion leaders through the analysis of cumulative voting influence. This analysis process not only involves a quantitative assessment of the user's voting influence, but also includes an in-depth examination of their comprehensive influence in the social network.

[0156] In order to ensure the accuracy and reliability of the analysis results, the above method in this embodiment is combined with a comparative analysis on the same data set, which not only verifies the effectiveness of the proposed model, but also demonstrates the limitations of different identification methods when identifying opinion leaders on social networks.

[0157] This embodiment uses five methods to compare the cumulative influence of votes, such as Figure 4 As shown, the five methods include Proposed, EW-VIKOR, Pagerank, SP and WPM. Among them: Proposed indicates the method proposed in this embodiment.

[0158] EW-VIKOR stands for an extended VIKOR (VlseKriterijum Optimizacija I Kompromis, from Serbian, which means multi-criteria optimization and compromise solution) method, which is a multi-criteria decision analysis method for evaluating and ranking candidate solutions.

[0159] Pagerank is an algorithm developed by Google that measures the importance of web pages. It determines the ranking of web pages by simulating the behavior of random surfers and can also be used for social network analysis.

[0160] SP stands for Steem Power, which refers to the indicator used in the Steemit social network platform to measure the voting weight of users and their impact on revenue distribution.

[0161] WPM (Weighted Product Method) is a multi-attribute decision analysis method that multiplies the value of each evaluation indicator by its weight and accumulates the multiplications to comprehensively evaluate and identify users who perform well on multiple key indicators.

[0162] Figure 4The cumulative number of votes changes. It can be seen that the cumulative voting influence score of the method in this embodiment continues to grow with the increase of the opinion leader candidate set, and shows a significant advantage in the process of expanding the size of the entire candidate set. In the interval with a smaller candidate set, the cumulative influence scores of all methods are relatively close and have the same growth trend. The cumulative scores of each method are not much different, which means that at this stage, all methods can effectively identify the initial influence contributors.

[0163] As the candidate set expands, when the candidate set size is about 20, the cumulative scores begin to show differences. The cumulative voting score curve of the proposed method gradually surpasses other methods, and the score growth rate is significantly higher than EW-VIKOR and Pagerank. This may indicate that the proposed method can more keenly identify influential users as the number of candidate sets increases.

[0164] When the candidate set size increases to the range of 40 to 60, the growth trend of the method in this embodiment maintains a stable linear increase. The growth rate of other methods, especially Pagerank and SP methods, becomes significantly slower. When the size continues to expand to the range of 80 to 100, the cumulative voting score growth of the proposed method still maintains a linear growth trend, while the growth curves of other methods tend to be stable.

[0165] In particular, the SP method has the slowest growth in cumulative scores throughout the process, indicating that in a large-scale candidate set, the SP method has poor performance and is difficult to effectively identify more opinion leaders. Further analysis shows that this may be because users with high SP are more involved in voting for other users and rarely participate in posting.

[0166] Therefore, only a small number of votes were accumulated. It is worth noting that the cumulative score of the EW-VIKOR method was similar to that of the proposed method in the early stage, but as the candidate set increased, its score growth rate began to lag behind, and the final cumulative number of votes was 123553. The method proposed in this paper obtained a total of 136058 votes, accounting for 70% of the total number of votes, which reflects the effectiveness of identifying opinion leaders.

[0167] In some embodiments, after step 207, the method further includes: Step 211, obtaining the reposting rate of the content published by the opinion leader on the blockchain social network within a set time period.

[0168] Among them, the repost rate refers to the frequency with which the content published by opinion leaders is forwarded by other users. This indicator reflects the dissemination power of the content and the influence of opinion leaders on online communities.

[0169] The repost rate is widely considered in existing research to be one of the key indicators for measuring the effectiveness of information dissemination and user influence. The repost rate reflects the degree of user recognition of a certain content or opinion. When a piece of content is widely forwarded, it not only expands the scope of dissemination of the content, but also enhances the original publisher's voice in the social network. In blockchain social networks, this forwarding behavior is particularly important because the network aims to highlight high-quality content, guide public opinion, and reduce the spread of false information. The repost rate can reveal how information flows in social networks and which users play a key role in this process. Content with a high repost rate usually means stronger influence and wider coverage, indicating the importance of the publisher in the network.

[0170] Specifically, the calculation formula for the repost rate is as follows: ; Among them, Rp(u i ) represents the reposting rate of user i, represents the total number of posts of user i, r ij represents the total number of reposts of all posts received by user i, i p represents the pth post of user i, n represents the number of posts of user i; ifu j repostedui represents a conditional judgment, which is used to determine whether user j has forwarded the post of user i.

[0171] Blockchain social networks are designed to highlight high-quality public opinion information, ensuring that users carefully consider when forwarding content. Most users tend to forward posts that have great influence and high-quality content, which not only improves their social network status, but also reduces the spread of rumors and false information, and promotes the flow of healthy and high-quality public opinion information.

[0172] Step 212, evaluating the network influence of the opinion leader according to the forwarding rate, and obtaining a second network influence evaluation result.

[0173] Specifically, the higher the reposting rate, the more popular the opinion leader's content is, the more it can resonate and spread more widely, and therefore the greater its online influence is.

[0174] By comparing the reposting rates of different opinion leaders, we can get a ranking list that reflects the size of their network influence based on their reposting behavior. Or we can assign a corresponding reposting rate score based on the reposting rate to get the second network influence evaluation result, which provides a basis for further analysis and application.

[0175] Step 213: testing the model validity of the opinion leader identification model according to the second network influence evaluation result and the influence score.

[0176] Specifically, the second network influence evaluation result, i.e., the repost rate score, is compared and analyzed with the influence score previously calculated by the opinion leader identification model to test the effectiveness of the model.

[0177] If the opinion leaders identified by the model do perform well in network influence indicators such as repost rate, then there should be a significant positive correlation between the output influence score and the repost rate score. The strength of this correlation can be used to evaluate the accuracy and reliability of the model in identifying truly highly influential opinion leaders, thereby verifying the effectiveness of the model.

[0178] Similarly, the 105 users mentioned in the previous article are used as research objects, and the baseline model and the proposed method are combined for comparative analysis on the same dataset. The reposting rate, defined as the frequency of content being forwarded, reflects the attractiveness of opinion leaders' content and reveals the behavior patterns of social network followers they trigger. In public opinion discussions, an efficient opinion leader identification algorithm can accurately identify those individuals with the potential to maximize information dissemination from the dynamic interactions of social networks.

[0179] Comparative analysis of the influence of the five methods on the repost rate Figure 5 shown.

[0180] exist Figure 5 In the results, all methods show an upward trend as the selected dataset expands. Among them, the proposed method always takes the lead in identifying the top 105 users, and the final cumulative repost rate score is 55, which is significantly better than other methods. As one of the key indicators to measure user influence and content dissemination effect, the repost rate reveals the ability of specific content or users to spread on social networks. It is crucial to analyze the growth trend of this indicator. For example, the EW-VIKOR method and the proposed method have similar growth rates in the early stage, but as the size of the candidate set increases, the proposed method gradually shows stronger growth momentum, which is particularly evident when the candidate set size is 20 to 40. This may be because the proposed method more accurately evaluates the influence of users in this segment, rather than just based on traditional network centrality indicators.

[0181] As the candidate set continues to increase, it is found that various methods have the same advantages and can effectively identify the same potential opinion leaders, until the candidate set reaches 50, showing different differences. When the opinion leader candidate set size is between 60 and 80, the repost rate score of the proposed method is almost more than twice that of other methods. This significant difference reflects the superiority of the proposed method on a large-scale dataset, which can identify those individuals who can truly trigger information dissemination among a wide range of users.

[0182] In addition, it can be observed that as the size of the candidate set increases, the Pagerank algorithm shows a relatively uniform growth trend. However, in terms of both growth rate and final reposting rate, the Pagerank algorithm performs worse than the proposed method. This shows that although the Pagerank algorithm has a certain efficiency in identifying important nodes in the network, more attention needs to be paid to user contribution feature attribute information in blockchain social networks in order to identify more potential opinion leaders.

[0183] When discussing the metric of reposting rate, it is important to understand that it is not a single quantitative metric, but the result of the interaction of multiple social network factors. A user's high reposting rate is not only due to their core position in the network, but may also be affected by factors such as the quality of the content, the strategic choice of posting time, and the reputation established in a specific community. Therefore, the analysis of this metric deepens the understanding of the method of identifying opinion leaders in blockchain social networks.

[0184] In some embodiments, the method of this embodiment can be executed by an electronic device. The following describes the electronic device in the embodiment of the present invention from the perspective of hardware processing. Figure 6 , is a schematic diagram of a physical device structure of an electronic device in an embodiment of the present invention.

[0185] It should be noted that Figure 6 The structure of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0186] like Figure 6 As shown, the electronic device includes a central processing unit (CPU) 401, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 402 or the program loaded from the storage part 408 to the random access memory (RAM) 403, such as executing the method described in the above embodiment. In RAM 403, various programs and data required for system operation are also stored. CPU 401, ROM 402 and RAM 403 are connected to each other through bus 404. Input / output (I / O) interface 405 is also connected to bus 404.

[0187] The following components are connected to the I / O interface 405: an input section 406 including an audio input device, a button switch, etc.; an output section 407 including a liquid crystal display (LCD) and an audio output device, an indicator light, etc.; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to the I / O interface 405 as needed. A removable medium 411, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 410 as needed so that a computer program read therefrom is installed into the storage section 408 as needed.

[0188] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 409, and / or installed from a removable medium 411. When the computer program is executed by the central processing unit (CPU) 401, various functions defined in the present invention are performed.

[0189] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CDROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium that contains or stores a program that may be used by or in combination with an instruction execution system, apparatus, or device.

[0190] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. Each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from that marked in the accompanying drawings.

[0191] Specifically, the electronic device of this embodiment includes a processor and a memory, the memory is coupled to one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and one or more processors call the computer instructions to enable the electronic device to execute the method provided by the above embodiment.

[0192] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiment; or may exist independently without being assembled into the electronic device. The above storage medium carries one or more computer programs, and when the above one or more computer programs are executed by a processor of the electronic device, the electronic device implements the method provided in the above embodiment.

[0193] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention.

[0194] As used in the above embodiments, the term "when..." may be interpreted to mean "if..." or "after..." or "in response to determining..." or "in response to detecting...", depending on the context. Similarly, the phrases "upon determining..." or "if (the stated condition or event) is detected" may be interpreted to mean "if determining..." or "in response to determining..." or "upon detecting (the stated condition or event)" or "in response to detecting (the stated condition or event)", depending on the context.

[0195] Those skilled in the art can understand that to implement all or part of the processes in the above-mentioned embodiments, the processes can be completed by computer programs to instruct related hardware, and the programs can be stored in computer-readable storage media. When the programs are executed, they can include the processes of the above-mentioned method embodiments. The aforementioned storage media include: ROM or random access memory RAM, magnetic disk or optical disk and other media that can store program codes.

Claims

1. A method for identifying opinion leaders in a blockchain social network, characterized in that: include: Constructing a directed network based on user data of the blockchain social network, wherein the directed network includes a plurality of nodes and a plurality of connection edges, wherein the nodes represent users of the blockchain social network, and the connection edges represent interaction channels between the users; Obtaining a network structure characteristic index of the directed network, wherein the network structure characteristic index includes at least one of degree centrality, eigenvector centrality, betweenness centrality and closeness centrality, wherein the degree centrality is used to measure the number of other nodes directly connected to the node in the directed network, the eigenvector centrality is used to reflect the importance of the neighboring nodes of the node, the betweenness centrality is used to measure the ability of the node to serve as an information bridge in the directed network, and the closeness centrality is used to reflect the degree of closeness between the node and other nodes, and is measured by the sum of the shortest distances from the node to all other nodes; Determining a key node among the plurality of nodes according to the network structure characteristic index; When the key nodes are determined, obtaining a contribution behavior characteristic index of the user in the blockchain social network, wherein the contribution behavior characteristic index includes at least one of a user reputation value, a number of bounties obtained, originality, a number of supporters, a number of followers, and activity; Determine the index weights of the network structure characteristic index and the contribution behavior characteristic index based on the entropy weight method and the inter-standard correlation weight method; Determining the influence score of the user according to the network structure characteristic index, the contribution behavior characteristic index and the index weight; An opinion leader of the blockchain social network is determined from among the users according to the influence score.

2. The method according to claim 1, characterized in that Determining the influence score of the user according to the network structure characteristic index, the contribution behavior characteristic index and the index weight includes: Based on the entropy weight method, the inter-criteria correlation weight method and the approximate ideal solution sorting method, the opinion leader identification model is constructed by using the network structure characteristic index, the contribution behavior characteristic index and the index weight; The influence score of the user is obtained through the opinion leader identification model.

3. The method according to claim 2, characterized in that The opinion leader identification model is used to: Using the network structure characteristic index and the contribution behavior characteristic index as indicator elements to construct a decision matrix; Normalizing the index element to obtain a normalized index element; Generating a normalized decision matrix according to the normalized indicator elements and the indicator weights; The element with the largest weighted index value in the normalized decision matrix is ​​determined as a positive ideal solution, and the element with the smallest weighted index value is determined as a negative ideal solution; respectively determining a first Euclidean distance between the normalized indicator element corresponding to the user and the positive ideal solution, and a second Euclidean distance between the normalized indicator element and the negative ideal solution; An influence score of the user is determined according to the first Euclidean distance and the second Euclidean distance.

4. The method according to claim 3, characterized in that The determining the influence score of the user according to the first Euclidean distance and the second Euclidean distance includes: Dividing the second Euclidean distance by the sum of the first Euclidean distance and the second Euclidean distance to obtain the proximity between the node and the ideal solution; An influence score of the user is determined based on the proximity.

5. The method according to claim 2, characterized in that: After determining the opinion leader of the blockchain social network from the users according to the influence score, the method further includes: Obtaining the cumulative number of votes for the content published by the opinion leader on the blockchain social network within a set time period; Evaluate the network influence of the opinion leader according to the accumulated number of votes to obtain a first network influence evaluation result; The model validity of the opinion leader identification model is tested according to the first network influence evaluation result and the influence score, and the model validity is used to characterize the accuracy of identifying opinion leaders.

6. The method according to claim 2, characterized in that After determining the opinion leader of the blockchain social network from the users according to the influence score, the method further includes: Obtaining the reposting rate of the content published by the opinion leader on the blockchain social network within a set time period; Evaluate the network influence of the opinion leader according to the repost rate to obtain a second network influence evaluation result; The model validity of the opinion leader identification model is tested according to the second network influence evaluation result and the influence score.

7. The method according to claim 1, characterized in that The method of determining the index weights of the network structure characteristic index and the contribution behavior characteristic index based on the entropy weight method and the inter-standard correlation weight method includes: Based on the entropy weight method, determining the information entropy of the normalized index elements corresponding to the network structure characteristic index and the contribution behavior characteristic index; Determining a first weight based on the information entropy; Determining a standard deviation value of the normalized indicator elements and an indicator correlation coefficient between the normalized indicator elements; Based on the inter-standard correlation weight method, determining a second weight according to the standard deviation value and the indicator correlation coefficient; The indicator weight is determined according to the first weight and the second weight.

8. An electronic device, characterized in that: including one or more processors and memory; The memory is coupled to the one or more processors, and the memory is used to store computer program codes, wherein the computer program codes include computer instructions, and the one or more processors call the computer instructions to enable the electronic device to execute the method according to any one of claims 1 to 7.

9. A computer-readable storage medium storing computer instructions, characterized in that: When the computer instructions are executed on an electronic device, the electronic device is caused to execute the method as claimed in any one of claims 1 to 7.

10. A computer program product, characterized in that When the computer program product is executed on an electronic device, the electronic device is enabled to execute the method according to any one of claims 1 to 7.

Images