Concluding customer social network construction and expansion method and system

Abstract

The invention provides a transaction customer social network construction and expansion method and system, and relates to the technical field of data processing, and the method comprises the steps: obtaining an original data set of a target customer; pre-processing the unstructured data in the original data set; according to the feature vector sets of the structured data and the unstructured data, respectively calculating association indexes between the target customer and other customers; according to the association indexes, constructing a multi-dimensional association matrix of the target customer and other customers; element values in the multi-dimensional incidence matrix are converted into edge weights, according to a preset edge weight threshold value, the relation that the edge weights are lower than the threshold value is eliminated, a graph structure containing nodes and edges is constructed, and a customer social network graph is generated; calculating a centrality index of each client node according to the client social network graph; and sorting according to the centrality indexes, and selecting client nodes with the top N ranks as high-value potential clients. According to the invention, the autonomy and accuracy of social network management are improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method and system for constructing and expanding a social network of transaction customers. Background Art

[0002] With the rapid development of big data and artificial intelligence technologies, enterprises are increasingly relying on social network analysis technology to explore potential customer relationships and provide decision-making basis for marketing strategies. At present, existing technologies collect and analyze customer data and use social network construction algorithms to build relationship networks between customers. These methods usually rely on information such as customer historical behavior, transaction records, and interaction data. Specifically, the system will first collect basic customer data, and generate a correlation matrix between customers through association analysis or machine learning models, and then display the direct and indirect relationships between customers in the form of a graph structure.

[0003] However, existing technologies lack full utilization of unstructured data, especially in scenarios where customer relationship networks are constructed based on text or non-traditional interaction data, such as comment content or non-standard communication data on social media. For example, when companies identify potential customer relationships through customers' social media interactions, such as comments, likes, and reposts, existing technologies usually simply rely on keyword matching or preset rules, ignoring the semantic information implicit in these unstructured data. This approach leads to insufficient mining of complex relationships between customers, and may not accurately identify high-value potential customer nodes, thereby limiting the scalability and practicality of social networks. Summary of the invention

[0004] The purpose of the present invention is to provide a method and system for building and expanding a social network of successful customers, aiming to solve the problems mentioned in the background technology.

[0005] In order to solve the above technical problems, the technical solution of the present invention is as follows:

[0006] In a first aspect, a method for building and expanding a social network of closed customers is provided, the method comprising:

[0007] Obtaining an original data set of target customers, wherein the original data set includes structured data and unstructured data, wherein the unstructured data includes comment content and information platform interaction data;

[0008] Preprocess the unstructured data in the original dataset, including:

[0009] Convert comment content and information platform interaction data into standardized text;

[0010] Apply the text segmentation method to segment the standardized text and generate word sequence data;

[0011] The embedded vectorization model is used to transform word sequence data into high-dimensional vector representation to obtain a set of unstructured data feature vectors;

[0012] According to the feature vector set of structured data and unstructured data, the correlation indexes between the target customer and other customers are calculated respectively. The correlation indexes include transaction frequency correlation and interaction semantic similarity. Among them:

[0013] Calculate transaction frequency correlation based on structured data;

[0014] Calculate the interaction semantic similarity based on the unstructured data feature vector set;

[0015] According to the correlation index, a multi-dimensional correlation matrix between the target customer and other customers is constructed, in which the rows represent the target customers, the columns represent other customers, and the element values ​​of the matrix are the weighted combination values ​​of the correlation index;

[0016] Convert the element values ​​in the multidimensional association matrix into edge weights, and according to the preset edge weight threshold, remove the relationships with edge weights lower than the threshold, build a graph structure containing nodes and edges, and generate a customer social network graph, where nodes represent customers and edges represent the associations between customers;

[0017] According to the customer social network graph, calculate the centrality index of each customer node;

[0018] According to the centrality index sorting, the top N customer nodes are selected as high-value potential customers.

[0019] Preferably, the comment content and information platform interaction data are converted into standardized text, including:

[0020] Parse the comments grammatically, extract the subject, predicate and core semantic relationship in the grammatical structure, and generate the data for this article;

[0021] According to the interaction type of the information platform interaction data, a preset type weight label is assigned to different interaction data, and the preset type weight label is embedded into the text data to generate weighted text data;

[0022] The weighted text data is decomposed into sentences, special symbols and meaningless phrases are removed, and the sentence structure is rearranged according to the semantic level to generate standardized text.

[0023] Preferably, applying a text segmentation method to perform segmentation processing on the standardized text to generate word sequence data includes:

[0024] Analyze the semantic context of standardized text and identify long phrases in the text;

[0025] According to the word frequency statistics, semantic relevance scores, and the dependency strength of long phrases in the context, semantic weight tags are added to long phrases to generate word segmentation semantic data; the calculation formula for semantic weight is: ,in, The first The semantic weight of a long phrase, To include The number of contexts for long phrases, For the The long phrase is in The term frequency in the context - the inverse document frequency value, For the The long phrase is in The grammatical dependency in the context, , It is The depth value of a long phrase in the syntax tree, is the contextual impact factor, , and Respectively represent The long phrase is in context and The embedding vector of the context Dimension value, is the number of dimensions of the embedding vector, For the The long phrase is in The term frequency in the context - the inverse document frequency value, is a constant;

[0026] Generate word frequency distribution data based on word segmentation semantic data, filter out low-frequency words for classification, and generate feature keyword data;

[0027] Preliminary word sequence data is constructed based on the characteristic keyword data, and the word sequence data is generated by optimizing the arrangement according to the semantic weight sorting and context dependency.

[0028] Preferably, an embedded vectorization model is applied to convert word sequence data into a high-dimensional vector representation to obtain a set of unstructured data feature vectors, including:

[0029] Vectorize each word in the word sequence data to generate a word vector matrix;

[0030] The dependency between adjacent words is calculated based on the word vector matrix, and the dependency is dynamically adjusted and the word vector is updated through the weighting function to generate a weighted word vector set;

[0031] Use the attention mechanism to aggregate the weighted word vector set to generate a sentence vector set;

[0032] The vector dimensions of the sentence vector set are unified and normalized to generate a set of unstructured data feature vectors.

[0033] Preferably, calculating the transaction frequency correlation based on structured data includes:

[0034] Extract the structured transaction record data of the target customers based on the structured data to generate transaction behavior data, which includes transaction time, transaction amount and transaction frequency;

[0035] Calculate the transaction interaction frequency between target customers and other customers based on transaction behavior data and generate a transaction frequency matrix;

[0036] According to the transaction frequency matrix, the transaction behaviors in different time periods are weighted to generate a weighted transaction frequency matrix;

[0037] The values ​​in the weighted transaction frequency matrix are normalized, and the correlation values ​​with low transaction frequency are eliminated by presetting the transaction frequency threshold to generate a transaction frequency correlation data set; the calculation formula for transaction frequency correlation is: ,in, For target customers With other customers The transaction frequency correlation of For time period Internal target customers With other customers The number of transactions, For time period Internal target customers With other customers The total transaction amount, and is the weight coefficient, is the adjustment coefficient, is the length of the time period.

[0038] Preferably, the interactive semantic similarity is calculated based on the unstructured data feature vector set, including:

[0039] Obtain unstructured data feature vectors of target customers and unstructured data feature vectors of other customers, and generate a target vector set and a comparison vector set;

[0040] The initial similarity value between the target vector set and the comparison vector set is calculated by the cosine similarity formula to generate an initial semantic similarity matrix;

[0041] Normalizing the values ​​in the initial semantic similarity matrix to generate a normalized semantic similarity matrix;

[0042] According to the similarity values ​​in the normalized semantic similarity matrix, the dynamic weight is calculated in combination with the semantic features of the target customers;

[0043] Applying the dynamic weight to each item in the normalized semantic similarity matrix to generate a dynamically adjusted semantic similarity matrix;

[0044] The interactive semantic similarity data set is generated by grouping calculation based on the semantic association strength of the dynamically adjusted semantic similarity matrix. The calculation formula for interactive semantic similarity is: ,in, For target customers With other customers The interactive semantic similarity of is the total number of long phrases, For the The global semantic weight of a long phrase, For target customers With other customers In the The cosine similarity on long phrases, ,in and Target customers and other customers In the The embedding vector of the long phrase Dimension value, and Target customers and other customers In the The Euclidean norm of the embedding vector on the long phrase, For target customers In the Dynamic weights on long phrases, , For target customers In the The term frequency-inverse document frequency value on the long phrase, For target customers In the The term frequency-inverse document frequency value on a long phrase.

[0045] Preferably, according to the customer social network graph, the centrality index of each customer node is calculated, including:

[0046] According to the customer social network graph, the connection information and edge weight of each node are extracted, the node degree centrality index of each node is calculated, and the node degree centrality index data set is generated;

[0047] Perform path analysis on the customer social network graph, calculate the betweenness centrality index of each node, and generate a node betweenness centrality index dataset;

[0048] According to the node degree centrality index data set and the node betweenness centrality index data set, nodes whose node degree centrality index and betweenness centrality index are both lower than the preset centrality threshold are eliminated to generate an optimized node data set;

[0049] According to the optimized node data set, the closeness between each node and the target stage is calculated to generate a node centrality index data set; the calculation formula of the node centrality index is: ,in, For Node The node centrality index of For Node The degree centrality index of ,in Is a node With Node The edge weights of Is with the node A collection of connected nodes, For Node The betweenness centrality index of ,in It is a slave node To Node The number of shortest paths, Is a node To Node Passing through the node The number of shortest paths, For Node The closeness centrality index of ,in Is a node With Node The shortest path distance between , and is the weight coefficient.

[0050] In a second aspect, a system for building and expanding a social network of closed customers is provided, the system comprising:

[0051] A data acquisition module is used to acquire the original data set of the target customer, wherein the original data set includes structured data and unstructured data, and the unstructured data includes comment content and information platform interaction data;

[0052] The preprocessing module is used to preprocess the unstructured data in the original data set, including:

[0053] Convert comment content and information platform interaction data into standardized text;

[0054] Apply the text segmentation method to segment the standardized text and generate word sequence data;

[0055] The embedded vectorization model is used to transform word sequence data into high-dimensional vector representation to obtain a set of unstructured data feature vectors;

[0056] The correlation index calculation module is used to calculate the correlation index between the target customer and other customers according to the structured data and unstructured data feature vector sets, including:

[0057] Calculate transaction frequency correlation based on structured data;

[0058] Calculate the interaction semantic similarity based on the unstructured data feature vector set;

[0059] The association matrix construction module is used to construct a multi-dimensional association matrix between target customers and other customers based on association indicators, where rows represent target customers, columns represent other customers, and element values ​​of the matrix are weighted combination values ​​of association indicators;

[0060] A social network graph construction module is used to convert the element values ​​in the multidimensional association matrix into edge weights, and according to a preset edge weight threshold, remove the relationships with edge weights lower than the threshold, construct a graph structure containing nodes and edges, and generate a customer social network graph, in which nodes represent customers and edges represent the association relationship between customers;

[0061] A centrality index calculation module is used to calculate the centrality index of each customer node based on the customer social network graph;

[0062] The customer development module is used to select the top N customer nodes as high-value potential customers based on the sorting of centrality indicators.

[0063] The above solution of the present invention includes at least the following beneficial effects:

[0064] First, this method not only uses traditional structured data, such as transaction records and customer historical behaviors, to build associations, but also conducts in-depth semantic mining of unstructured data such as comment content and information platform interaction data. The comment content and interaction data are converted into standardized text, and the subject, predicate and core semantic relationship of the text are extracted by combining grammatical parsing, semantic analysis and other means, which solves the shortcomings of the existing technology that simply relies on keyword matching. Through semantic hierarchical reorganization and weight embedding, the use of implicit semantic information in unstructured data is enhanced, ensuring that the generated customer relationship network has more semantic depth and accuracy.

[0065] Secondly, in terms of data feature expression, this method converts unstructured data into high-dimensional vector representation through steps such as word segmentation and embedded vectorization, and combines it with structured data to generate a multidimensional association matrix. This matrix not only includes the traditional transaction frequency association, but also incorporates the interactive semantic similarity calculated based on unstructured data. Through this multivariate combination of association indicators, the relationship strength between customers can be fully described and reflected in the customer social network graph in the form of edge weights. Compared with the existing technology, this multi-dimensional association matrix construction method breaks through the limitations of a single data source and significantly improves the accuracy and expression ability of the social network graph.

[0066] Thirdly, in the process of constructing the customer social network graph, this method removes low-value edges by presetting edge weight thresholds, thereby ensuring that the generated graph structure is more compact and efficient. At the same time, through the calculation of centrality indicators, this method can not only intuitively display the direct relationship between customers, but also deeply explore the potential relationship between customers. Compared with the customer network generated by simple association analysis in the prior art, the customer social network graph generated by this method is richer in relationship identification and scalability.

[0067] Finally, this method can also quickly screen out the top N high-value potential customers by sorting the centrality of the customer social network graph. This function is not only suitable for potential customer mining in precision marketing, but also provides strong support for other scenarios, such as customer relationship management and customized service recommendations. In particular, with the support of diversified data sources, this method can significantly improve the accuracy of high-value customer node identification, overcoming the problem of insufficient identification caused by keyword matching or simple rules in the existing technology.

[0068] In summary, this method greatly improves the quality of customer social network construction by fully combining structured data and unstructured data and deeply mining semantic information, breaking through the bottleneck of insufficient utilization of unstructured data in existing technologies. At the same time, this method can provide enterprises with more comprehensive and accurate support in customer relationship mining and marketing decision-making through multi-dimensional association indicators and optimized social network graphs, significantly enhancing the practicality and scalability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 It is a flowchart of a method for building and expanding a social network of successful customers provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0070] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0071] like Figure 1 As shown, an embodiment of the present invention provides a method for building and expanding a social network of successful customers, the method comprising:

[0072] S100, obtaining an original data set of target customers, wherein the original data set includes structured data and unstructured data, and the unstructured data includes comment content and information platform interaction data;

[0073] S200, preprocessing the unstructured data in the original data set, including:

[0074] Convert comment content and information platform interaction data into standardized text;

[0075] Apply the text segmentation method to segment the standardized text and generate word sequence data;

[0076] The embedded vectorization model is used to transform word sequence data into high-dimensional vector representation to obtain a set of unstructured data feature vectors;

[0077] S300, calculating the correlation indexes between the target customer and other customers respectively according to the feature vector set of structured data and unstructured data, the correlation indexes including transaction frequency correlation and interaction semantic similarity; wherein:

[0078] Calculate transaction frequency correlation based on structured data;

[0079] Calculate the interaction semantic similarity based on the unstructured data feature vector set;

[0080] S400, constructing a multidimensional association matrix between the target customer and other customers based on the association index, wherein the rows represent the target customers, the columns represent other customers, and the element values ​​of the matrix are weighted combination values ​​of the association index;

[0081] S500, converting element values ​​in the multidimensional association matrix into edge weights, and according to a preset edge weight threshold, eliminating relationships with edge weights lower than the threshold, constructing a graph structure including nodes and edges, and generating a customer social network graph, wherein nodes represent customers and edges represent associations between customers;

[0082] S600, calculating the centrality index of each customer node according to the customer social network graph;

[0083] S700. Select the top N customer nodes as high-value potential customers based on the centrality index ranking.

[0084] In the embodiment of the present invention, through this method, the original data set of the target customer is first comprehensively acquired and processed. The original data set covers two types of structured data and unstructured data. Structured data mainly includes transaction records, which record in detail the customer's transaction time, transaction amount, transaction frequency and other information; unstructured data includes comment content and information platform interaction data, such as likes, comments, and reposts. The diversity of this data source lays the foundation for subsequent comprehensive association calculations.

[0085] In the data preprocessing stage, the consistency of data format was ensured by standardizing the text of unstructured data, and the unstructured text was successfully converted into a feature set represented by a high-dimensional vector through text segmentation and vectorization models. At the same time, the transaction frequency association model and interactive semantic similarity model were used to quantify the various associations between target customers and other customers into unified association indicators. Based on these indicators, a multidimensional association matrix was constructed, in which the rows of the matrix correspond to the target customers, the columns correspond to other customers, and the element values ​​represent the strength of association between customers.

[0086] By converting the multidimensional association matrix into a graph structure, including nodes and edges, the weight of the edge reflects the strength of the association between customers, and at the same time, the low-weight relationship edges are eliminated by presetting the edge weight threshold, and finally a social network graph of the target customers is generated. The centrality calculation is performed on the social network graph to identify the centrality index of each customer node. According to the ranking of the centrality index, the top N customer nodes are selected as high-value potential customers. Through this method, the implicit associations between customers can be accurately mined, and a clearer and more efficient customer relationship network graph can be constructed, which provides strong data support and decision-making basis for subsequent precision marketing.

[0087] More specifically, the original data set of the target customer is obtained, the original data set includes structured data and unstructured data, the unstructured data includes comment content and information platform interaction data, including:

[0088] Structured data refers to data with a clear organizational form, usually in the form of tables or relational databases. The structured data in this method mainly includes customer transaction records and historical behavior information. Transaction records are a key data source, which contains field information such as transaction time, transaction amount, and transaction frequency. By processing structured data, the transaction behavior characteristics between target customers and other customers can be directly extracted. For example, the transaction frequency and amount between a certain customer and the target customer in different time periods can reflect the interaction intensity between the two, and provide data support for the calculation of the subsequent transaction frequency correlation.

[0089] Unstructured data refers to free-form data that has not been strictly organized, including comment content and information platform interaction data.

[0090] Comment content: This part of the data usually comes from the comment records of target customers and other customers on social media or online platforms, such as user evaluations of products or services, discussions with other users, etc. These comments exist in the form of natural language texts and have not been structured. Direct use may lead to information redundancy or misinterpretation. In order to effectively utilize the comment content, this method combines natural language processing technology, such as grammatical parsing and semantic analysis, to extract the semantic relationships contained in the text, such as subject, predicate, and object, to provide support for subsequent semantic similarity calculations.

[0091] Information platform interaction data: This type of data reflects the interaction between target customers and other customers on social platforms, such as likes, comments, and reposts. Different interaction behaviors have different importance, so it is necessary to set corresponding weight labels for each behavior. For example, reposting behavior may reflect a stronger customer association, so a higher weight can be set, while a simple like behavior may reflect a weaker association, so a lower weight can be set. By assigning weight labels to interaction data and embedding them into text data, the utilization value of unstructured data is further improved.

[0092] More specifically, based on the association indicators, a multidimensional association matrix between the target customer and other customers is constructed, where the rows represent the target customers, the columns represent other customers, and the element values ​​of the matrix are weighted combination values ​​of the association indicators, including:

[0093] Definition of matrix rows and columns:

[0094] The rows of the multidimensional association matrix represent target customers, the columns represent other customers, and the elements of the matrix represent the strength of association between the target customer and other customers.

[0095] Calculation of element values:

[0096] The element value of the matrix is ​​the weighted combination value of the associated indicators, and the calculation formula is as follows: ,in, and is the weight coefficient, which is used to balance the influence of the two correlation indicators. For target customers With other customers The transaction interaction frequency, For target customers With other customers The interactive semantic similarity.

[0097] Normalize the matrix elements to ensure that all values ​​of the matrix are in the same range, such as [0, 1].

[0098] In a preferred embodiment of the present invention, the comment content and information platform interaction data are converted into standardized text, including:

[0099] Parse the comments grammatically, extract the subject, predicate and core semantic relationship in the grammatical structure, and generate the data for this article;

[0100] According to the interaction type of the information platform interaction data, a preset type weight label is assigned to different interaction data, and the preset type weight label is embedded into the text data to generate weighted text data;

[0101] The weighted text data is decomposed into sentences, special symbols and meaningless phrases are removed, and the sentence structure is rearranged according to the semantic level to generate standardized text.

[0102] In an embodiment of the present invention, the original comment content and interactive data are comprehensively processed through a series of natural language processing technologies, and the standardized text conversion of unstructured data is successfully achieved. First, the comment content is parsed to extract the subject, predicate and core semantic relationship of each comment. This process effectively filters out redundant information in the comments and strengthens the semantic expression ability of the comment content. Secondly, the interaction type of the information platform interactive data, such as likes, comments, reposts, etc., is combined with the corresponding preset weight tags, and embedded into the comment text to generate weighted text data. This embedding method effectively integrates the user's interactive behavior with semantic information, further enhancing the depth and expression ability of the data.

[0103] Based on the weighted text data, meaningless symbols and phrases are removed through sentence disassembly and reconstruction steps, and the text sentences are rearranged according to the semantic level. This process not only optimizes the semantic logic of the text, but also enhances the standardization and parsability of the text. The standardized text content finally generated has a highly consistent formatting structure, providing a high-quality data foundation for subsequent semantic analysis and feature extraction.

[0104] More specifically, we perform grammatical analysis on the comments, extract the subject, predicate and core semantic relationship in the grammatical structure, and generate the data for this article, including:

[0105] This implementation adopts the syntax parsing algorithm in natural language processing (NLP) technology to analyze the review content sentence by sentence. Commonly used syntax parsing tools include Stanford Parser based on statistical models, dependency tree parsing models, etc. These tools can decompose natural language text into a tree structure containing grammatical relationships.

[0106] Subject extraction: Use the dependency tree structure generated by grammatical parsing to locate the subject part in the grammatical marker, for example, the "nsubj" tag in a sentence, and extract the words or phrases corresponding to the subject.

[0107] Predicate extraction: Find the corresponding predicate part from the dependency tree structure, such as the "ROOT" tag, and extract the predicate verb and its modifiers.

[0108] Core semantic relationship extraction: Extract the dependency relationship between the subject and the predicate, such as the object, adverbial, etc., to ensure that the generated data can fully express the core semantics of the review content.

[0109] The extracted subject, predicate and semantic relationship are reorganized into structured text data according to the preset format. For example, for the comment "user likes this product", the extracted subject is "user", the predicate is "like", and the semantic relationship is "this product". The generated data format is a triple of "subject-predicate-semantic relationship", such as: user-like-this product.

[0110] More specifically, the sentence structure is rearranged according to the semantic level, including:

[0111] Based on the results of grammatical analysis, analyze the primary and secondary levels between the subject, predicate and semantic relationship. For example:

[0112] The part of the sentence that directly describes the core event is the main semantic level;

[0113] The auxiliary information describing time, place and conditions is a secondary semantic level.

[0114] Sort the components of the sentence according to their semantic weights, placing the main semantic components at the core of the sentence and the secondary semantic components in order. For example, the original sentence "Last year, the company's users gave positive comments on the new product" can be rearranged as "Company users positively evaluated the new product (last year)".

[0115] The rearranged text content retains the core semantics of the original sentence and optimizes the logic and readability of the semantic hierarchy.

[0116] In a preferred embodiment of the present invention, a text segmentation method is applied to perform segmentation processing on the standardized text to generate word sequence data, including:

[0117] Analyze the semantic context of standardized text and identify long phrases in the text;

[0118] According to the word frequency statistics, semantic relevance scores, and the dependency strength of long phrases in the context, semantic weight tags are added to long phrases to generate word segmentation semantic data; the calculation formula for semantic weight is: ,in, The first The semantic weight of a long phrase, To include The number of contexts for long phrases, For the The long phrase is in The term frequency in the context - the inverse document frequency value, For the The long phrase is in The grammatical dependency in the context, , It is The depth value of a long phrase in the syntax tree, is the contextual impact factor, , and Respectively represent The long phrase is in context and The embedding vector of the context Dimension value, is the number of dimensions of the embedding vector, For the The long phrase is in The term frequency in the context - the inverse document frequency value, is a constant;

[0119] Generate word frequency distribution data based on word segmentation semantic data, filter out low-frequency words for classification, and generate feature keyword data;

[0120] Preliminary word sequence data is constructed based on the characteristic keyword data, and the word sequence data is generated by optimizing the arrangement according to the semantic weight sorting and context dependency.

[0121] In an embodiment of the present invention, a semantic context analysis is carried out on the standardized text, and the long phrases in the text are accurately identified through intelligent word segmentation technology. On the basis of word segmentation, a semantic weight tag is added to each long phrase according to the word frequency statistics, semantic relevance and the dependency strength of the long phrase in the context. The importance of each long phrase is comprehensively calculated by the semantic weight formula, and the characteristics of each word segmentation unit in the text are further refined. The attached semantic weight not only reflects the relative importance of the word segmentation in the sentence, but also combines the structural depth in the grammatical tree, so that the generated word segmentation result has a higher accuracy.

[0122] The generation of word segmentation semantic data further constructs a word frequency distribution table. Low-frequency words are removed and classified to generate a set of optimized feature keyword data. Feature keywords are sorted by semantic weight and arranged into preliminary word sequence data. On this basis, they are optimized in combination with contextual dependencies. The final generated word sequence data completely retains the semantic information of the original text, while improving data availability and computational efficiency.

[0123] More specifically, the semantic context of the standardized text is analyzed to identify long phrases in the text, including:

[0124] Adopt a phrase recognition algorithm based on a combination of statistics and rules:

[0125] Statistical methods: Use word frequency statistics and mutual information (MI) calculations to identify common word combinations. For example, the ratio of the probability of words A and B appearing together to the probability of them appearing alone is calculated to measure the importance of their combination.

[0126] Rule-based method: According to grammatical rules, such as the noun phrase structure: attributive + noun, directly extract phrases that meet the rules.

[0127] Long phrases are categorized as follows:

[0128] Long phrase: A fixed expression usually consisting of multiple words, such as "product user experience", whose meaning is inseparable in the context.

[0129] Phrase: A simple semantic unit consisting of one or two words, such as "user" and "experience", which can express semantics independently.

[0130] Analyze the contextual dependencies of long phrases in sentences, and further optimize the accuracy of phrase extraction by analyzing their position in sentences and their relationship with other words.

[0131] More specifically, word frequency distribution data is generated based on word segmentation semantic data, and low-frequency words are screened out for classification to generate feature keyword data, including:

[0132] Perform word frequency statistics on the extracted long phrase set, calculate the frequency of each phrase appearing in the text, and generate a word frequency distribution table by sorting the phrases according to the frequency.

[0133] According to the set word frequency threshold, if the word frequency is less than 5, a low-frequency word set is filtered out.

[0134] Categorize phrases in a low-frequency vocabulary set, for example, by clustering them according to semantic categories:

[0135] Category 1: Phrases related to user behavior, such as “like” and “comment”.

[0136] Category 2: Phrases referring to product attributes, such as “performance” and “appearance”.

[0137] The filtered low-frequency vocabulary set is classified to form a new set of feature keyword data, which further optimizes the semantic feature expression.

[0138] More specifically, preliminary word sequence data is constructed based on the characteristic keyword data, and the word sequence data is generated by optimizing the arrangement according to the semantic weight sorting and context dependency, including:

[0139] According to the classification results of the characteristic keyword data, a preliminary word sequence is generated in order of semantic categories. For example, the keywords in category 1 are arranged as "user, interaction, comment", and those in category 2 are arranged as "performance, appearance, design".

[0140] Assign weights to terms using a semantic weight formula.

[0141] The words are sorted according to the calculated semantic weights, with words with higher weights being given priority.

[0142] On the basis of sorting, the word arrangement is optimized in combination with contextual dependencies. For example, if "user" and "experience" frequently appear together, the two are arranged together to form the phrase "user experience".

[0143] In a preferred embodiment of the present invention, an embedded vectorization model is applied to convert word sequence data into a high-dimensional vector representation to obtain a set of unstructured data feature vectors, including:

[0144] Vectorize each word in the word sequence data to generate a word vector matrix;

[0145] The dependency between adjacent words is calculated based on the word vector matrix, and the dependency is dynamically adjusted and the word vector is updated through the weighting function to generate a weighted word vector set;

[0146] Use the attention mechanism to aggregate the weighted word vector set to generate a sentence vector set;

[0147] The vector dimensions of the sentence vector set are unified and normalized to generate a set of unstructured data feature vectors.

[0148] In an embodiment of the present invention, word sequence data is converted into a high-dimensional vector representation through an embedding model. First, the word sequence data is processed word by word using a context-sensitive embedding model to generate a high-dimensional word vector matrix containing all words. On this basis, the word vectors are optimized and adjusted through a dynamic weighting function according to the degree of dependency between words to generate a set of weighted word vectors.

[0149] Furthermore, the attention mechanism is used to aggregate the weighted word vectors and successfully generate a sentence-level vector set. This method not only retains the overall semantic structure of the sentence, but also combines the importance differences between the words within the sentence. Finally, by unifying the dimensions and normalizing the sentence vectors, a complete set of unstructured data feature vectors is generated. This set not only efficiently reduces the dimensionality of the original text, but also provides highly structured basic data for subsequent semantic analysis and relationship calculation.

[0150] More specifically, each word in the word sequence data is vectorized to generate a word vector matrix, including:

[0151] The context-sensitive embedding model is used to dynamically generate word vectors based on the contextual semantics of the vocabulary in the sentence. Commonly used models include:

[0152] Based on bidirectional LSTM, it can capture the dynamic semantics of words in context.

[0153] It uses the Transformer architecture, supports bidirectional context analysis, and is suitable for complex sentences and polysemous word scenarios.

[0154] The word sequence data is used as the model input, and the model extracts the semantic features of the vocabulary layer by layer through multiple deep neural network layers, such as the TransformerEncoder layer.

[0155] For each word, the model dynamically generates a high-dimensional vector based on its contextual information.

[0156] Generate word vectors for all words in the sequence, and arrange these vectors in word order to form a word vector matrix.

[0157] Each row represents a high-dimensional vector of a word, such as 768 dimensions, and each column represents the semantic features of the word in different dimensions.

[0158] More specifically, the dependency between adjacent words is calculated based on the word vector matrix to generate a weighted word vector set, including:

[0159] Use grammar parsing tools, such as Stanford Parser or open source tools based on dependency syntax, to analyze the dependencies between adjacent words in the word embedding matrix.

[0160] The degree of dependency is determined by the semantic distance and syntactic relationship between words. For example, the degree of dependency of subject-predicate relationship may be higher than that of attributive-modification relationship.

[0161] For each pair of adjacent words, a dynamic weight is assigned according to their degree of dependency, and the range of weight values ​​can be defined according to the importance of the syntactic relationship.

[0162] The role of dynamic weights is to adjust the feature expression of words in the vector. For example, for "like" and "product" in "users like products", if the context of the sentence emphasizes the verb "like", it will be given a higher weight.

[0163] According to the assigned weights, the original word vectors are adjusted to generate a new set of weighted word vectors. Each weighted word vector retains the original semantic information and combines the context dependency characteristics, showing higher semantic accuracy.

[0164] More specifically, the attention mechanism is used to aggregate the weighted word vector set to generate a sentence vector set, including:

[0165] The goal of aggregation is to integrate the information in the set of weighted word vectors into a sentence-level vector while retaining the semantic features of important words in the sentence.

[0166] Assign attention weights to each vector in the weighted word vector set. The size of the weight is determined by the importance of the vocabulary. For example, subject-predicate vocabulary is usually more important in the core semantics of a sentence; modifying vocabulary, such as adjectives and adverbs, may be less important.

[0167] The allocation of attention weights is usually calculated by a feedforward neural network, and the specific implementation can adopt the Transformer architecture.

[0168] According to the assigned attention weights, the word vector set is weighted summed and all word vectors are integrated into a sentence vector of fixed length.

[0169] For each sentence data input, a corresponding sentence vector is generated, and all sentence vectors are stored as a set for subsequent processing.

[0170] More specifically, the vector dimensions of the sentence vector set are unified and normalized to generate an unstructured data feature vector set, including:

[0171] The dimensions of different sentence vectors may vary due to the length of the input sentence or context characteristics. To ensure consistency in subsequent processing, all sentence vectors need to be adjusted to a unified dimension, for example, 512 dimensions.

[0172] The operation of unifying dimensions can be done through zero padding or dimensionality reduction, such as PCA dimensionality reduction.

[0173] To make the numerical range of the sentence vector uniform, each component in the vector is mapped to a fixed range, such as [0, 1]. Commonly used normalization methods include: maximum and minimum normalization or L2 normalization.

[0174] The normalized sentence vector set is stored as a feature vector set. Each vector in the feature vector set represents the semantic feature of an unstructured data, providing standardized input for subsequent calculations.

[0175] In a preferred embodiment of the present invention, calculating the transaction frequency correlation based on structured data includes:

[0176] Extract the structured transaction record data of the target customers based on the structured data to generate transaction behavior data, which includes transaction time, transaction amount and transaction frequency;

[0177] Calculate the transaction interaction frequency between target customers and other customers based on transaction behavior data and generate a transaction frequency matrix;

[0178] According to the transaction frequency matrix, the transaction behaviors in different time periods are weighted to generate a weighted transaction frequency matrix;

[0179] The values ​​in the weighted transaction frequency matrix are normalized, and the correlation values ​​with low transaction frequency are eliminated by presetting the transaction frequency threshold to generate a transaction frequency correlation dataset; the calculation formula for transaction frequency correlation is: ,in, For target customers With other customers The transaction frequency correlation of For time period Internal target customers With other customers The number of transactions, For time period Internal target customers With other customers The total transaction amount, and is the weight coefficient, is the adjustment coefficient, is the length of the time period.

[0180] In an embodiment of the present invention, for the target customer's transaction record data, this method successfully quantifies the transaction frequency correlation between the target customer and other customers through a series of formulas and data processing. First, the transaction behavior information, including transaction time, transaction amount and transaction frequency, is extracted from the structured data, and a transaction behavior data set is constructed based on this. Through the transaction frequency formula, the transaction interaction frequency between the target customer and other customers is accurately calculated and quantified into a transaction frequency matrix. The transaction frequency matrix is ​​further combined with the time weight function to give a higher weight to recent transaction behaviors, emphasizing the time sensitivity of dynamic transaction behaviors.

[0181] In addition, the customer relationship data with low transaction frequency was eliminated through normalization processing, generating a high-quality transaction frequency correlation dataset. The final result can accurately reflect the strength of transaction behavior between customers, providing reliable data support for the subsequent construction of correlation matrix and social network graph.

[0182] In a preferred embodiment of the present invention, the interactive semantic similarity is calculated based on the unstructured data feature vector set, including:

[0183] Obtain unstructured data feature vectors of target customers and unstructured data feature vectors of other customers, and generate a target vector set and a comparison vector set;

[0184] The initial similarity value between the target vector set and the comparison vector set is calculated by the cosine similarity formula to generate an initial semantic similarity matrix;

[0185] Normalizing the values ​​in the initial semantic similarity matrix to generate a normalized semantic similarity matrix;

[0186] According to the similarity values ​​in the normalized semantic similarity matrix, the dynamic weight is calculated in combination with the semantic features of the target customers;

[0187] Applying the dynamic weight to each item in the normalized semantic similarity matrix to generate a dynamically adjusted semantic similarity matrix;

[0188] The interactive semantic similarity data set is generated by grouping calculation based on the semantic association strength of the dynamically adjusted semantic similarity matrix. The calculation formula for interactive semantic similarity is: ,in, For target customers With other customers The interactive semantic similarity of is the total number of long phrases, For the The global semantic weight of a long phrase, For target customers With other customers In the The cosine similarity on long phrases, ,in and Target customers and other customers In the The embedding vector of the long phrase Dimension value, and Target customers and other customers In the The Euclidean norm of the embedding vector on the long phrase, For target customers In the Dynamic weights on long phrases, , For target customers In the The term frequency-inverse document frequency value on the long phrase, For target customers In the The term frequency-inverse document frequency value on a long phrase.

[0189] In the embodiment of the present invention, the method successfully calculates the interactive semantic similarity between the target customer and other customers through the unstructured data feature vector set. First, the feature vectors of the target customer and other customers are respectively constructed into a target vector set and a comparison vector set, and the initial semantic similarity value is calculated by the cosine similarity formula. This stage completes the basic semantic similarity quantification.

[0190] On the basis of the initial semantic similarity, we further combined the dynamic semantic features of the target customers to calculate a set of dynamically adjusted weights, which were applied to the normalized semantic similarity matrix to generate a set of dynamically adjusted semantic similarity matrix data. Finally, by grouping the dynamically adjusted matrix, we generated an interactive semantic similarity dataset. This dataset fully reflects the strength of the semantic associations between customers and provides deep support at the semantic level for the construction and optimization of customer social networks.

[0191] In a preferred embodiment of the present invention, the centrality index of each customer node is calculated according to the customer social network graph, including:

[0192] According to the customer social network graph, the connection information and edge weight of each node are extracted, the node degree centrality index of each node is calculated, and the node degree centrality index data set is generated;

[0193] Perform path analysis on the customer social network graph, calculate the betweenness centrality index of each node, and generate a node betweenness centrality index dataset;

[0194] According to the node degree centrality index data set and the node betweenness centrality index data set, nodes whose node degree centrality index and betweenness centrality index are both lower than the preset centrality threshold are eliminated to generate an optimized node data set;

[0195] According to the optimized node data set, the closeness between each node and the target stage is calculated to generate a node centrality index data set; the calculation formula of the node centrality index is: ,in, For Node The node centrality index of For Node The degree centrality index of ,in Is a node With Node The edge weights of Is with the node A collection of connected nodes, For Node The betweenness centrality index of ,in It is a slave node To Node The number of shortest paths, Is a node To Node Passing Node The number of shortest paths, For Node The closeness centrality index of ,in Is a node With Node The shortest path distance between , and is the weight coefficient.

[0196] In the embodiment of the present invention, by constructing the customer social network graph, the method further calculates the centrality index of each customer node, including the degree centrality, betweenness centrality and closeness centrality of the node. The degree centrality of the node directly reflects the connection strength between the node and other nodes, while the betweenness centrality measures the ability of the node to act as a bridge in the customer network. The closeness centrality reflects the importance of the node in the network by calculating the path length between nodes.

[0197] By removing nodes whose degree centrality and betweenness centrality are both below the threshold, an optimized node data set is generated. This optimization not only reduces the interference of redundant nodes on the network structure, but also significantly improves the significance of key nodes in the network. Finally, according to the centrality index ranking of customer nodes, high-value customer nodes that are most closely connected with target customers are successfully screened out, providing reliable support for precision marketing and efficient customer relationship management.

[0198] More specifically, node degree centrality indicators include:

[0199] For unweighted networks, the edge weight is set to 1, indicating the existence of a relationship; for weighted networks, the edge weight value can directly reflect the strength of the association.

[0200] For the target node, count the number of nodes it is directly connected to, for an unweighted network, or the sum of the weights of all its connected edges, for a weighted network.

[0201] For example, in a privileged network, if a customer has multiple high-frequency trading relationships with other customers, then his degree centrality value is high.

[0202] Customers with high degree centrality usually indicate that they have direct connections with more customers in the network and are key interactors in the network.

[0203] Companies can focus on analyzing these customers, for example by giving priority to covering them through marketing activities, thereby achieving higher network penetration.

[0204] The degree centrality index is mainly suitable for measuring active nodes in the network, which helps to find core nodes that can directly influence more customers and is suitable for identifying "active participants" in social networks.

[0205] More specifically, betweenness centrality indicators include:

[0206] Calculate the shortest paths between all pairs of nodes in the network using graph algorithms such as Dijkstra's algorithm or Floyd-Warshall's algorithm. For example, calculate whether the shortest path between nodes s and q passes through node v.

[0207] For the shortest path between each pair of nodes, count whether the target node appears on the path and record the number of shortest paths passing through the node.

[0208] Nodes with high betweenness centrality values ​​are usually irreplaceable bridge nodes in the network, responsible for connecting different customer groups in the network.

[0209] By analyzing these nodes, companies can identify key nodes that connect different customer groups and use their special position in the network to improve overall network connectivity.

[0210] The betweenness centrality index is mainly used to measure the "bridge" role of nodes, and can identify key customers that connect different customer groups in social networks. By optimizing resource allocation to these nodes, the efficiency of information dissemination and the stability of the network structure can be enhanced.

[0211] More specifically, closeness centrality indicators include:

[0212] Use the shortest path algorithm to calculate the path length from the target node to all other nodes. The path length can be expressed as the number of edges, or the sum of edge weights.

[0213] For the target node, count the shortest path lengths from it to all other nodes and calculate their average. The shorter the path from a node to other nodes, the higher the closeness centrality value.

[0214] Customers with high closeness centrality are usually at the center of the network and can quickly reach all other customers in the network.

[0215] Enterprises can prioritize contacting these nodes because they play an important role in information dissemination and customer relationship coordination throughout the network.

[0216] The closeness centrality index is mainly used to measure the "global position" of a node in a network and can identify the core nodes in the network. Enterprises can use customers with high closeness centrality as potential partners or marketing priorities to quickly expand their influence.

[0217] An embodiment of the present invention further provides a system for building and expanding a social network of successful customers, the system comprising:

[0218] A data acquisition module is used to acquire the original data set of the target customer, wherein the original data set includes structured data and unstructured data, and the unstructured data includes comment content and information platform interaction data;

[0219] The preprocessing module is used to preprocess the unstructured data in the original data set, including:

[0220] Convert comment content and information platform interaction data into standardized text;

[0221] Apply the text segmentation method to segment the standardized text and generate word sequence data;

[0222] The embedded vectorization model is used to transform word sequence data into high-dimensional vector representation to obtain a set of unstructured data feature vectors;

[0223] The correlation index calculation module is used to calculate the correlation index between the target customer and other customers according to the structured data and unstructured data feature vector sets, including:

[0224] Calculate transaction frequency correlation based on structured data;

[0225] Calculate the interaction semantic similarity based on the unstructured data feature vector set;

[0226] The association matrix construction module is used to construct a multi-dimensional association matrix between target customers and other customers based on association indicators, where rows represent target customers, columns represent other customers, and element values ​​of the matrix are weighted combination values ​​of association indicators;

[0227] A social network graph construction module is used to convert the element values ​​in the multidimensional association matrix into edge weights, and according to a preset edge weight threshold, remove the relationships with edge weights lower than the threshold, construct a graph structure containing nodes and edges, and generate a customer social network graph, in which nodes represent customers and edges represent the association relationship between customers;

[0228] A centrality index calculation module is used to calculate the centrality index of each customer node based on the customer social network graph;

[0229] The customer development module is used to select the top N customer nodes as high-value potential customers based on the sorting of centrality indicators.

[0230] It should be noted that the system is a system corresponding to the above method, and all implementation methods in the above method embodiment are applicable to this embodiment and can achieve the same technical effect.

[0231] The embodiment of the present invention further provides a computing device, including: a processor, a memory storing a computer program, and when the computer program is executed by the processor, the method described above is executed. All implementations in the above method embodiment are applicable to this embodiment and can achieve the same technical effect.

[0232] The embodiment of the present invention further provides a computer-readable storage medium storing instructions, which, when executed on a computer, enable the computer to execute the above method. All implementations in the above method embodiment are applicable to this embodiment and can achieve the same technical effect.

[0233] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for building and expanding a social network of successful customers, characterized in that: The method comprises: Obtaining an original data set of target customers, wherein the original data set includes structured data and unstructured data, wherein the unstructured data includes comment content and information platform interaction data; Preprocess the unstructured data in the original dataset, including: Convert comment content and information platform interaction data into standardized text; Apply the text segmentation method to segment the standardized text and generate word sequence data; The embedded vectorization model is used to transform word sequence data into high-dimensional vector representation to obtain a set of unstructured data feature vectors; According to the feature vector set of structured data and unstructured data, the correlation indexes between the target customer and other customers are calculated respectively. The correlation indexes include transaction frequency correlation and interaction semantic similarity. Among them: Calculate transaction frequency correlation based on structured data; Calculate the interaction semantic similarity based on the unstructured data feature vector set; According to the correlation index, a multi-dimensional correlation matrix between the target customer and other customers is constructed, in which the rows represent the target customers, the columns represent other customers, and the element values ​​of the matrix are the weighted combination values ​​of the correlation index; Convert the element values ​​in the multidimensional association matrix into edge weights, and according to the preset edge weight threshold, remove the relationships with edge weights lower than the threshold, build a graph structure containing nodes and edges, and generate a customer social network graph, where nodes represent customers and edges represent the associations between customers; According to the customer social network graph, calculate the centrality index of each customer node; According to the centrality index sorting, the top N customer nodes are selected as high-value potential customers.

2. A method for building and expanding a customer social network according to claim 1, characterized in that: Convert comment content and information platform interaction data into standardized text, including: Parse the comments grammatically, extract the subject, predicate and core semantic relationship in the grammatical structure, and generate the data for this article; According to the interaction type of the information platform interaction data, a preset type weight label is assigned to different interaction data, and the preset type weight label is embedded into the data of this paper to generate weighted text data; The weighted text data is decomposed into sentences, special symbols and meaningless phrases are removed, and the sentence structure is rearranged according to the semantic level to generate standardized text.

3. A method for building and expanding a customer social network according to claim 2, characterized in that: Apply the text segmentation method to segment the standardized text and generate word sequence data, including: Analyze the semantic context of standardized text and identify long phrases in the text; According to the word frequency statistics, semantic relevance scores, and the dependency strength of long phrases in the context, semantic weight tags are added to long phrases to generate word segmentation semantic data; the calculation formula for semantic weight is: ,in, The first The semantic weight of a long phrase, To include The number of contexts for long phrases, For the The long phrase is in The term frequency in the context - the inverse document frequency value, For the The long phrase is in The grammatical dependency in the context, , It is The depth value of a long phrase in the syntax tree, is the contextual impact factor, , and Respectively represent The long phrase is in context and The embedding vector of the context Dimension value, is the number of dimensions of the embedding vector, For the The long phrase is in The term frequency in the context - the inverse document frequency value, is a constant; Generate word frequency distribution data based on word segmentation semantic data, filter out low-frequency words for classification, and generate feature keyword data; Preliminary word sequence data is constructed based on the characteristic keyword data, and the word sequence data is generated by optimizing the arrangement according to the semantic weight sorting and context dependency.

4. A method for building and expanding a customer social network according to claim 3, characterized in that: The embedded vectorization model is used to transform word sequence data into high-dimensional vector representation, and a set of unstructured data feature vectors is obtained, including: Vectorize each word in the word sequence data to generate a word vector matrix; The dependency between adjacent words is calculated based on the word vector matrix, and the dependency is dynamically adjusted and the word vector is updated through the weighting function to generate a weighted word vector set; Use the attention mechanism to aggregate the weighted word vector set to generate a sentence vector set; The vector dimensions of the sentence vector set are unified and normalized to generate a set of unstructured data feature vectors.

5. A method for building and expanding a customer social network according to claim 4, characterized in that: Calculate transaction frequency correlation based on structured data, including: Extract the structured transaction record data of the target customers based on the structured data to generate transaction behavior data, which includes transaction time, transaction amount and transaction frequency; Calculate the transaction interaction frequency between target customers and other customers based on transaction behavior data and generate a transaction frequency matrix; According to the transaction frequency matrix, the transaction behaviors in different time periods are weighted to generate a weighted transaction frequency matrix; The values ​​in the weighted transaction frequency matrix are normalized, and the correlation values ​​with low transaction frequency are eliminated by presetting the transaction frequency threshold to generate a transaction frequency correlation data set; the calculation formula for transaction frequency correlation is: ,in, For target customers With other customers The transaction frequency correlation of For time period Internal target customers With other customers The number of transactions, For time period Internal target customers With other customers The total transaction amount, and is the weight coefficient, is the adjustment coefficient, is the length of the time period.

6. A method for building and expanding a customer social network according to claim 5, characterized in that: Calculate the interactive semantic similarity based on the unstructured data feature vector set, including: Obtain unstructured data feature vectors of target customers and unstructured data feature vectors of other customers, and generate a target vector set and a comparison vector set; The initial similarity value between the target vector set and the comparison vector set is calculated by the cosine similarity formula to generate an initial semantic similarity matrix; Normalizing the values ​​in the initial semantic similarity matrix to generate a normalized semantic similarity matrix; According to the similarity values ​​in the normalized semantic similarity matrix, the dynamic weight is calculated in combination with the semantic features of the target customers; Applying the dynamic weight to each item in the normalized semantic similarity matrix to generate a dynamically adjusted semantic similarity matrix; The interactive semantic similarity data set is generated by grouping calculation based on the semantic association strength of the dynamically adjusted semantic similarity matrix. The calculation formula for interactive semantic similarity is: ,in, For target customers With other customers The interactive semantic similarity of is the total number of long phrases, For the The global semantic weight of a long phrase, For target customers With other customers In the The cosine similarity on long phrases, ,in and Target customers and other customers In the The embedding vector of the long phrase Dimension value, and Target customers and other customers In the The Euclidean norm of the embedding vector on the long phrase, For target customers In the Dynamic weights on long phrases, , For target customers In the The term frequency-inverse document frequency value on the long phrase, For target customers In the The term frequency-inverse document frequency value on a long phrase.

7. A method for building and expanding a customer social network according to claim 6, characterized in that: According to the customer social network graph, the centrality index of each customer node is calculated, including: According to the customer social network graph, the connection information and edge weight of each node are extracted, the node degree centrality index of each node is calculated, and the node degree centrality index data set is generated; Perform path analysis on the customer social network graph, calculate the betweenness centrality index of each node, and generate a node betweenness centrality index dataset; According to the node degree centrality index data set and the node betweenness centrality index data set, nodes whose node degree centrality index and betweenness centrality index are both lower than the preset centrality threshold are eliminated to generate an optimized node data set; According to the optimized node data set, the closeness between each node and the target stage is calculated to generate a node centrality index data set; the calculation formula of the node centrality index is: ,in, For Node The node centrality index of For Node The degree centrality index of ,in Is a node With Node The edge weights of Is with the node A collection of connected nodes, For Node The betweenness centrality index of ,in It is a slave node To Node The number of shortest paths, Is a node To Node Passing through the node The number of shortest paths, For Node The closeness centrality index of ,in Is a node With Node The shortest path distance between , and is the weight coefficient.

8. A system for building and expanding a social network of successful customers, characterized in that: Applied to the method according to any one of claims 1 to 7, the system comprises: A data acquisition module is used to acquire the original data set of the target customer, wherein the original data set includes structured data and unstructured data, and the unstructured data includes comment content and information platform interaction data; The preprocessing module is used to preprocess the unstructured data in the original data set, including: Convert comment content and information platform interaction data into standardized text; Apply the text segmentation method to segment the standardized text and generate word sequence data; The embedded vectorization model is used to transform word sequence data into high-dimensional vector representation to obtain a set of unstructured data feature vectors; The correlation index calculation module is used to calculate the correlation index between the target customer and other customers according to the structured data and unstructured data feature vector sets, including: Calculate transaction frequency correlation based on structured data; Calculate the interaction semantic similarity based on the unstructured data feature vector set; The association matrix construction module is used to construct a multi-dimensional association matrix between target customers and other customers based on association indicators, where rows represent target customers, columns represent other customers, and element values ​​of the matrix are weighted combination values ​​of association indicators; A social network graph construction module is used to convert the element values ​​in the multidimensional association matrix into edge weights, and according to a preset edge weight threshold, remove the relationships with edge weights lower than the threshold, construct a graph structure containing nodes and edges, and generate a customer social network graph, in which nodes represent customers and edges represent the association relationship between customers; A centrality index calculation module is used to calculate the centrality index of each customer node based on the customer social network graph; The customer development module is used to select the top N customer nodes as high-value potential customers based on the sorting of centrality indicators.

9. A computing device, characterized in that include: one or more processors; A storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement the method as claimed in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program, which, when executed by a processor, implements the method according to any one of claims 1 to 7.

Images