Platform content intelligent recommendation method and system based on natural language processing

Through natural language processing technology, semantic knowledge connections and text reference information are identified in the digital culture article library, and knowledge link nodes are generated, which solves the problems of inaccurate and incomplete search results in the existing technology, and achieves more accurate and personalized Article recommendations improve users' reading experience and knowledge acquisition efficiency.

CN118551031BActive Publication Date: 2024-11-15GUANGZHOU PINGYUN INFORMATION TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410990621.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-07-23
Publication Date
2024-11-15
Estimated Expiration
2044-07-23

AI Technical Summary

Technical Problem

The existing digital culture article search and recommendation methods cannot deeply understand the semantic content and knowledge structure of the article, resulting in inaccurate search results, inaccurate and comprehensive enough, ineffectively organize and display relevant content, and it is difficult to form a systematic knowledge system and reading path.

Method used

Using a natural language processing method, we generate knowledge link nodes through semantic knowledge connections and text citation information in the digital culture article library, organize and recommend related articles, and provide an accurate and personalized reading experience.

Benefits of technology

It improves the accuracy and comprehensiveness of digital cultural article retrieval, enhances users' reading experience and knowledge acquisition efficiency, provides a more systematic and coherent way to obtain information, and improves users' reading satisfaction.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118551031B_ABST
    Figure CN118551031B_ABST
Patent Text Reader

Abstract

The embodiment of the present application provides a platform content intelligent recommendation method and system based on natural language processing. According to the platform text content retrieval requirements, an article retrieval operation is performed in the digital culture article library to generate multiple related digital culture articles, and then not only the semantic knowledge connection between digital culture articles is considered, but also the text reference information is combined to further enrich the depth and breadth of the recommended content. By determining the semantic knowledge connection between digital culture articles and using text reference information, the related digital culture articles can be effectively organized and connected to form a complete knowledge chain. By determining the knowledge chain nodes of digital culture articles, not only the relevance and continuity between articles are considered, but also dynamic adjustment and optimization can be performed according to the user's reading progress and interest preferences, thereby providing a more accurate and personalized recommendation service.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of digital technology, and more specifically, to a method and system for intelligently recommending platform content based on natural language processing. Background Art

[0002] With the rapid development and popularization of digital culture, digital cultural service devices such as smartphones, tablets, and e-book readers have become important tools for people to obtain and read digital cultural articles. However, faced with a massive amount of digital cultural article resources, users often find it difficult to quickly find high-quality content that meets their interests and needs. Therefore, how to effectively retrieve and recommend digital cultural articles has become an urgent problem to be solved.

[0003] Traditional article retrieval and recommendation methods are mainly based on keyword matching, user behavior analysis and other technologies. Although they can meet user needs to a certain extent, they still have some problems. For example, these methods often fail to deeply understand the semantic content and knowledge structure of articles, resulting in inaccurate and incomplete retrieval results; at the same time, they are also unable to effectively organize and display relevant digital cultural articles, making it difficult for users to form a systematic knowledge system and reading path. Summary of the invention

[0004] In order to at least overcome the above-mentioned deficiencies in the prior art, the purpose of this application is to provide a platform content intelligent recommendation method and system based on natural language processing. First, according to the platform text content retrieval requirements of the digital cultural service device, an article retrieval operation is performed in the digital cultural article library to generate multiple related digital cultural articles. Then, by determining the semantic knowledge connection between these articles and the corresponding text reference information, the intrinsic association and extended relationship between the articles are further mined. On this basis, the method uses the concept of knowledge link nodes to effectively organize and connect the relevant digital cultural articles to form a complete knowledge link. Finally, according to the knowledge link nodes, the corresponding digital cultural articles are recommended in the digital cultural service device to provide users with a more accurate and personalized reading experience, which can not only improve the accuracy and comprehensiveness of digital cultural article retrieval, but also help users better understand and master the knowledge content, and improve reading efficiency and experience. At the same time, the method also has a wide range of application prospects and can be applied to various digital cultural service devices and scenarios, such as online education platforms, digital libraries, news information applications, etc.

[0005] In a first aspect, the present application provides a platform content intelligent recommendation method based on natural language processing, which is applied to a digital cultural service system, and the method includes:

[0006] Based on the platform text content retrieval requirements of the digital cultural service device, an article retrieval operation is performed in the digital cultural article library to generate X digital cultural articles, where X is a positive integer greater than 1;

[0007] Determine the semantic knowledge connection between the X digital cultural articles and Y text reference information corresponding to the X digital cultural articles based on a natural language processing algorithm, where one text reference information is used to reference at least one digital cultural article, and Y is a positive integer; the digital cultural articles with semantic knowledge connection for any digital cultural article include: when the corresponding digital cultural article is published, the corresponding derived and extended digital cultural articles;

[0008] Based on the semantic knowledge connections between the X digital cultural articles and the Y text citation information corresponding to the X digital cultural articles, the knowledge link nodes of the X digital cultural articles are determined, wherein the knowledge link node of any digital cultural article is located after the knowledge link nodes of each digital cultural article with which the corresponding digital cultural article has semantic knowledge connections, and the knowledge link nodes of each digital cultural article corresponding to the same text citation information are continuous;

[0009] In the digital culture service device, the X digital culture articles are recommended according to the knowledge link nodes of the X digital culture articles.

[0010] In a second aspect, an embodiment of the present application further provides a digital cultural service system, which includes a processor and a machine-readable storage medium, wherein the machine-readable storage medium stores a computer program, and the computer program is loaded and executed according to the processor to implement the platform content intelligent recommendation method based on natural language processing of the above first aspect.

[0011] Based on any one of the above aspects, the beneficial effects of this application are as follows:

[0012] This application can perform article retrieval operations in the digital culture article library according to the platform text content retrieval requirements, generate multiple related digital culture articles, and meet the user's needs for diversified and personalized content. Secondly, not only the semantic knowledge connection between digital culture articles is considered, but also the text reference information is combined to further enrich the depth and breadth of the recommended content. By determining the semantic knowledge connection between digital culture articles, the intrinsic association and extended relationship between articles can be excavated to provide users with more comprehensive and systematic knowledge services. At the same time, using text reference information, the relevant digital culture articles can be effectively organized and connected to form a complete knowledge chain, which improves the user's reading experience and knowledge acquisition efficiency. Finally, by determining the knowledge chain nodes of digital culture articles, the function of recommending digital culture articles based on knowledge chain nodes in digital culture service equipment is realized. This recommendation method not only takes into account the relevance and continuity between articles, but also can be dynamically adjusted and optimized according to the user's reading progress and interest preferences, thereby providing more accurate and personalized recommendation services. At the same time, it helps to improve the intelligence level and user experience of digital culture service equipment.

[0013] That is, the present application can effectively retrieve the digital culture article library and generate multiple digital culture articles, making full use of digital technology and artificial intelligence, realizing efficient management and accurate retrieval of large-scale data sets, and improving the efficiency and accuracy of article retrieval. By determining the semantic knowledge connection between digital culture articles and the corresponding text reference information, a knowledge link is constructed, which can better understand and mine the intrinsic connection between articles, and can establish the knowledge connection between each article according to the semantic relationship of the article, which is helpful for deep mining and utilization of text information, and enhances the value and use effect of information. Articles are recommended based on the knowledge link nodes of digital culture articles, so that users can quickly obtain other articles related to the article when reading a certain article, which greatly enriches the user's reading experience and knowledge acquisition path. At the same time, it also has good coherence, which makes users more fluent when obtaining information and enhances the user's reading satisfaction. As a result, the efficiency and accuracy of article retrieval are improved, text information is deeply mined and utilized, the user's reading experience is enriched, and the user's reading satisfaction is enhanced. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 A flowchart of a method for intelligently recommending platform content based on natural language processing provided in an embodiment of the present application;

[0015] Figure 2 A block diagram showing the functional structure of a digital cultural service system for implementing the above-mentioned method for intelligent recommendation of platform content based on natural language processing, provided in an embodiment of the present application. DETAILED DESCRIPTION

[0016] See also Figure 1 As shown, the present application provides a platform content intelligent recommendation method based on natural language processing, comprising the following steps.

[0017] Step S110, based on the platform text content retrieval requirements of the digital culture service device, an article retrieval operation is performed in the digital culture article library to generate X digital culture articles, where X is a positive integer greater than 1.

[0018] The digital cultural service device may refer to a hardware device that provides digital cultural content services, such as a computer, tablet computer, smart phone, etc. These digital cultural service devices can access, store and display digital cultural content. For example, if a user accesses a digital library through an application on a smart phone, the smart phone is a digital cultural service device.

[0019] The platform text content retrieval demand may refer to a text content retrieval request made by a user on a digital cultural service platform, which usually contains specific retrieval conditions or parameters. For example, a user may configure a digital cultural article retrieval map for the keyword "ancient Chinese architecture" on a digital library platform to search for articles related to the topic "ancient Chinese architecture".

[0020] The digital cultural article library may refer to a database or warehouse storing a large number of digital cultural articles, which exist in a digital form for easy retrieval, dissemination and preservation. For example, a database in a digital library that specifically stores historical articles is a digital cultural article library.

[0021] Thus, in this embodiment, the server receives a platform text content retrieval demand from a digital culture service device, and illustratively, the platform text content retrieval demand carries a digital culture article retrieval map. The article retrieval map contains specific attributes of the digital culture articles that the user wants to retrieve, such as category attributes (such as history, art), time domain attributes (such as the release time within the past year), user attributes (such as the release user is an expert in a certain field) and status attributes (such as the article status is published).

[0022] Thus, the server can perform a search operation in the digital culture article library according to these attribute parameters in the platform text content retrieval requirements. In detail, by matching the text reference information in the digital culture article library with the attribute parameters in the article retrieval map, the server finds multiple digital culture articles that meet the requirements. These digital culture articles are selected and a digital culture article set is generated. The number of articles in the digital culture set is X, where X is a positive integer greater than 1.

[0023] Step S120, based on a natural language processing algorithm, determines the semantic knowledge connection between the X digital cultural articles and Y text reference information corresponding to the X digital cultural articles, where one text reference information is used to reference at least one digital cultural article, and Y is a positive integer. Any digital cultural article with a semantic knowledge connection includes: when the corresponding digital cultural article is published, the corresponding derived and extended digital cultural article.

[0024] Semantic knowledge connections refer to the semantic associations or connections between digital cultural articles, which may be reflected in the article's subject, content, author, release time, etc. For example, two articles on "Ancient Chinese Architecture" may be written by the same author, or one may be a sequel or comment on the other, which are all semantic knowledge connections.

[0025] For example, when determining the semantic knowledge connection between digital cultural articles and the corresponding text citation information, a natural language processing (NLP) algorithm can be used to implement it. The following is a specific implementation method:

[0026] First, each digital culture article is preprocessed by word segmentation, stop word removal, stem extraction, etc. Then, a word vector model (such as Word2Vec, GloVe, etc.) is used to convert the words in the article into vector representations for semantic calculation.

[0027] Next, you can use pre-trained semantic models (such as BERT, RoBERTa, ERNIE, etc.) to calculate the semantic similarity between articles. These models can capture the contextual information in the text and output a vector representation of the article. By calculating the cosine similarity, Euclidean distance or other similarity indicators between these vectors, the semantic knowledge connection between the articles can be determined.

[0028] In addition, the citation marks in the article (such as footnotes, references, hyperlinks, etc.) can be analyzed to extract the identifiers of other digital cultural articles that are cited. Use regular expressions, natural language understanding tools, or specialized citation parsing algorithms to identify and extract this citation information. Then, combine the calculated semantic similarity with the extracted text citation information to construct a semantic knowledge graph. In this semantic knowledge graph, nodes represent digital cultural articles, and edges represent semantic connections and citation relationships between articles. For example, a graph database (such as Neo4j, OrientDB, etc.) can be used to store and manage this semantic knowledge graph.

[0029] In the semantic knowledge graph, the knowledge link nodes of each digital cultural article are determined based on the semantic connections and citation relationships between articles. Among them, graph traversal algorithms (such as depth-first search, breadth-first search, etc.) can be used to find and determine these knowledge link nodes.

[0030] Suppose there is a collection of three digital cultural articles: Article A, Article B, and Article C. The semantic knowledge connections and textual citation information between them can be determined by following the steps below:

[0031] Preprocess articles A, B, and C to obtain their word vector representations.

[0032] Use the pre-trained semantic model to calculate the semantic similarity between article A and article B, article A and article C, and article B and article C.

[0033] Analyze the reference marks in articles A, B, and C and extract the reference relationship between them. For example, suppose article A cites article B, and article B cites article C.

[0034] Combining semantic similarity and citation relationship, a semantic knowledge graph is constructed. In this graph, articles A, B, and C are nodes, and the edges between them represent semantic connections and citation relationships.

[0035] Use a graph traversal algorithm to determine the knowledge link nodes of each article. In this example, the knowledge link node of article A may be before article B, and the knowledge link node of article B is before article C. At the same time, since article A cites article B, the knowledge link nodes of article A and article B are continuous.

[0036] In this way, the semantic knowledge connections and corresponding text citation information between digital cultural articles can be determined based on natural language processing algorithms, thereby providing strong support for subsequent intelligent recommendations.

[0037] Among them, Y represents the number of text citation information associated with the retrieved X digital cultural articles. One text citation information may cite one or more digital cultural articles. For example, among the 5 retrieved digital cultural articles, 3 different text citation information are used to cite these articles. Here Y is equal to 3. The derived and extended digital cultural articles may refer to other digital cultural articles that are related to or derived from a certain digital cultural article when it is published. These digital cultural articles may be supplements, explanations, comments or extensions of the original articles. For example, an article on "Ancient Chinese Architecture" may trigger a series of derivative articles on specific architectural styles, building materials or architectural history. These derivative articles have semantic knowledge connections with the original articles.

[0038] Therefore, in this embodiment, in the digital culture article collection generated by the server, each digital culture article has its unique text reference information. The server can determine the semantic knowledge connection between these digital culture articles by analyzing the text content of these digital culture articles. For example, a digital culture article may be a sequel, commentary, or expansion of a related topic of another digital culture article.

[0039] At the same time, the server also organizes the Y text citation information corresponding to the X digital cultural articles. Here, Y is a positive integer and may be less than or equal to X, because one text citation information may cite one or more digital cultural articles.

[0040] Step S130, based on the semantic knowledge connections between the X digital cultural articles and the Y text citation information corresponding to the X digital cultural articles, determine the knowledge link nodes of the X digital cultural articles, the knowledge link node of any digital cultural article is located after the knowledge link nodes of each digital cultural article with which the corresponding digital cultural article has semantic knowledge connections, and the knowledge link nodes of each digital cultural article corresponding to the same text citation information are continuous.

[0041] The knowledge link node may refer to the connection point between various knowledge points or information units in the knowledge system. In the context of digital cultural articles, the knowledge link node represents the semantic association and logical order between articles. Each digital cultural article can be regarded as a knowledge link node, and these knowledge link nodes are connected to each other through the semantic knowledge connection between them.

[0042] For example, suppose there are three digital culture articles A, B, and C. Article A introduces a historical event, Article B is an in-depth analysis of the historical event, and Article C discusses the modern impact of the historical event. The knowledge link nodes between these three digital culture articles can be determined as follows: the node of Article A is at the front because it is a basic introduction; the node of Article B is after Article A because it is an in-depth analysis based on the content of Article A; the node of Article C is after Article B because it is a further discussion based on the first two digital culture articles.

[0043] The continuous knowledge link nodes may refer to a series of knowledge points or information units in a knowledge system that are closely connected in a certain logical order to form a continuous knowledge chain. In the context of digital cultural articles, if multiple digital cultural articles are cited by the same text reference information, then the knowledge link nodes of these articles are continuous because they are directly related or continued in content. Continuing with the above articles A, B, and C as an example, if these three digital cultural articles are all cited by the same text reference information, then their knowledge link nodes are continuous. When reading, users can smoothly transition from article A to article B, and then to article C, forming a complete knowledge chain.

[0044] Therefore, in this embodiment, the server can allocate knowledge link nodes to the X digital cultural articles based on the determined semantic knowledge connection and text reference information. The position of the knowledge link node of each digital cultural article is determined based on its relationship with other articles. For example, if a digital cultural article is a sequel to another digital cultural article, then its knowledge link node will be located after the previous digital cultural article.

[0045] In addition, for multiple digital cultural articles cited by the same text citation information, their knowledge link nodes are continuous, which helps users to smoothly jump from one digital cultural article to another related article when reading.

[0046] Step S140: In the digital culture service device, the X digital culture articles are recommended according to the knowledge link nodes of the X digital culture articles.

[0047] Finally, the server recommends these X digital cultural articles to the user of the digital cultural service device according to the determined order of knowledge link nodes. The user can see a list of articles arranged in the order of knowledge links on the device, so that he can read and learn these digital cultural contents in a more organized way. This recommendation method not only improves the user's reading experience, but also helps the user to understand and master the relevant digital cultural knowledge more deeply.

[0048] For example, the server can arrange the articles in the order of knowledge link nodes according to the semantic knowledge connection and logical order between the articles and present them to the user. In this way, users can read and learn these digital cultural contents in a clear and organized order. Taking articles A, B, and C as an example, the server can recommend these three digital cultural articles to the user in sequence according to their knowledge link node order (A→B→C). The list of articles that users see on the digital cultural service device is arranged in this order, so that they can read and learn these articles in a more organized way.

[0049] Based on the above steps, this application can perform article retrieval operations in the digital culture article library according to the platform text content retrieval requirements, generate multiple related digital culture articles, and meet the user's needs for diversified and personalized content. Secondly, not only the semantic knowledge connection between digital culture articles is considered, but also the text reference information is combined to further enrich the depth and breadth of the recommended content. By determining the semantic knowledge connection between digital culture articles, the intrinsic association and extended relationship between articles can be excavated to provide users with more comprehensive and systematic knowledge services. At the same time, using text reference information, the relevant digital culture articles can be effectively organized and connected to form a complete knowledge chain, which improves the user's reading experience and knowledge acquisition efficiency. Finally, by determining the knowledge chain nodes of digital culture articles, the function of recommending digital culture articles based on knowledge chain nodes in digital culture service equipment is realized. This recommendation method not only takes into account the relevance and continuity between articles, but also can be dynamically adjusted and optimized according to the user's reading progress and interest preferences, thereby providing more accurate and personalized recommendation services. At the same time, it helps to improve the intelligence level and user experience of digital culture service equipment.

[0050] That is, the present application can effectively retrieve the digital culture article library and generate multiple digital culture articles, making full use of digital technology and artificial intelligence, realizing efficient management and accurate retrieval of large-scale data sets, and improving the efficiency and accuracy of article retrieval. By determining the semantic knowledge connection between digital culture articles and the corresponding text reference information, a knowledge link is constructed, which can better understand and mine the intrinsic connection between articles, and can establish the knowledge connection between each article according to the semantic relationship of the article, which is helpful for deep mining and utilization of text information, and enhances the value and use effect of information. Articles are recommended based on the knowledge link nodes of digital culture articles, so that users can quickly obtain other articles related to the article when reading a certain article, which greatly enriches the user's reading experience and knowledge acquisition path. At the same time, it also has good coherence, which makes users more fluent when obtaining information and enhances the user's reading satisfaction. As a result, the efficiency and accuracy of article retrieval are improved, text information is deeply mined and utilized, the user's reading experience is enriched, and the user's reading satisfaction is enhanced.

[0051] In a possible implementation, step S130 may include:

[0052] Step S131, based on the semantic knowledge connection between the X digital culture articles, a text label logic diagram set is generated according to the text recognition tags of the X digital culture articles, in which the text recognition tag of any digital culture article is located after the text recognition tags of each digital culture article with which the corresponding digital culture article has a semantic knowledge connection.

[0053] In the present embodiment, the server analyzes the semantic knowledge connection between the X digital culture articles. Each digital culture article has one or more text recognition tags, which describe the theme, content or features of the article. The server can generate a text label logic diagram set based on the text recognition tags of the articles according to the semantic knowledge connection between these articles. In this text label logic diagram set, the text recognition tag of each digital culture article will be placed after the text recognition tags of other articles with which it has semantic knowledge connection, which forms a logical order based on tags, reflecting the theme association and content hierarchy between articles.

[0054] For example, if the topic of the retrieved article is about "ancient Chinese architecture", the server may place the tag about "architectural style" before the tag about "building materials" based on the specific content of the article, because architectural style is usually the primary consideration in building design, and building materials are the means to achieve these styles.

[0055] Step S134, based on the semantic knowledge connection between the X digital cultural articles, a text reference logic graph set is generated according to the Y text reference information corresponding to the X digital cultural articles, and the logic node of any text reference information in the text reference logic graph set is located after the logic node of the text reference information with which the corresponding text reference information has a semantic knowledge connection in the text reference logic graph set. The text reference information with which any text reference information has a semantic knowledge connection includes: the text reference information of the digital cultural article with which the digital cultural article corresponding to the corresponding text reference information has a semantic knowledge connection.

[0056] In this embodiment, in addition to analyzing the semantic knowledge connection between articles, the server also considers the text reference information between articles. These reference information may be direct references to the content of other articles, or indirect references to other articles.

[0057] Specifically, the server can generate a text reference logic graph set based on these reference information. In this text reference logic graph set, each text reference information is placed after other text reference information with semantic knowledge connection, which forms a reference-based logical order, reflecting the content reference and logical dependency relationship between articles.

[0058] For example, if a digital cultural article cites specific data or conclusions about "ancient building materials" in another digital cultural article, the server can place this citation information in the corresponding position to indicate the citation relationship between the two digital cultural articles.

[0059] Step S133, generating a knowledge link based on the text tag logic diagram set and the Y text reference information, the knowledge link including: a reference corresponding link for each text reference information, any reference corresponding link including the text identification tag of each digital cultural article corresponding to the corresponding text reference information, and the logical knowledge association between each text identification tag in any reference corresponding link matches the logical knowledge association of the corresponding text identification tag in the text tag logic diagram set.

[0060] In this embodiment, after generating the text label logic graph set and the text reference logic graph set, the server can generate a knowledge link by combining the text label logic graph set and the text reference logic graph set. The knowledge link is a set of reference corresponding links including each text reference information. Each reference corresponding link includes the text identification label of each digital cultural article corresponding to the corresponding text reference information.

[0061] The logical knowledge associations of these text identification tags in the reference corresponding links match their logical knowledge associations in the text tag logical graph set. This means that the server can ensure that the tag order of the articles and the reference relationships between them are consistent in the knowledge link.

[0062] In detail, in the text label logic graph set, each text recognition label has a logical association with other labels. These logical associations may be based on topic similarity, content hierarchy, or semantic relationship, etc. Therefore, based on these logical associations, the server can establish corresponding label paths. For example, in the topic of "Ancient Chinese Architecture", the path may be from "Architectural History" to "Architectural Style" to "Building Materials". Next, the server will optimize these label paths to ensure that they are both semantically logical and easy for users to understand and navigate. On this basis, the server will analyze Y text citation information and extract feature information such as the number of times each digital cultural article is cited, the specific content of the citation, and the source of the citation. Based on this feature information, the server constructs a citation network in which nodes represent articles and edges represent citation relationships. Then, based on factors such as the frequency, importance, or authority of the citation, the server assigns a weight to each citation edge.

[0063] The server can then combine the established tag path with the constructed citation network. This means that in the knowledge link, not only the topic and content hierarchy of the article are taken into account, but also the citation relationship between them. During the combination process, the server determines the position of each text citation information in the knowledge link. This position is determined based on the position of the article corresponding to the citation information in the tag path and its importance in the citation network. Finally, the server will optimize the generated knowledge link to ensure that it reflects both the semantic connection and citation relationship between the articles, and is also convenient for users to read and learn these articles in a clear and organized order.

[0064] In actual applications, the server may adjust the algorithm for generating knowledge links based on user feedback. For example, if a user finds that the logical associations between certain articles are not close enough or the citation relationships are not accurate enough, the server will adjust these associations and relationships accordingly. As the digital cultural article library continues to be updated and expanded, the server also needs to continuously update and optimize the generated knowledge links to ensure that they are always consistent with the latest article content and user needs.

[0065] In one example, if a digital cultural article quotes content about "ancient building materials" in another digital cultural article, and the labels of these two digital cultural articles are "architectural style" and "building materials" respectively, then when the server generates a knowledge link, it will ensure that the label of "architectural style" comes before the label of "building materials", and this order is consistent with their order in the text label logical graph set.

[0066] Step S134, according to the knowledge link nodes of each text reference information in the text reference logic graph set, positionally arrange the reference corresponding links of each text reference information in the knowledge link to generate node distribution information, wherein the node distribution information represents the knowledge link nodes of the X digital cultural articles.

[0067] Finally, the server can arrange the positions of the referenced links of each text reference information in the knowledge link according to the knowledge link nodes of each text reference information in the text reference logic graph set. This arrangement process will generate a node distribution information, which represents the knowledge link nodes of X digital cultural articles.

[0068] The node distribution information not only reflects the semantic knowledge connection and citation relationship between articles, but also provides users with a clear and organized reading and learning path. Users can read and learn these digital cultural articles in sequence according to this path, so as to better understand and master the relevant digital cultural knowledge.

[0069] In detail, in one example, the server first deeply analyzes the text reference logic graph set, which reflects the logical relationship and knowledge dependency between each text reference information. In the text reference logic graph set, some text reference information may be regarded as key nodes because they are frequently cited or are crucial to understanding the entire knowledge system. The server identifies these key reference information. Next, the server assigns a position weight to each text reference information based on factors such as the importance of the text reference information, the frequency of citation, and their centrality in the logic graph. On this basis, the server applies a sorting algorithm (such as topological sorting, weight sorting, etc.) to sort the reference corresponding links of the text reference information according to the position weight and logical relationship. During the sorting process, the server will specially handle those text reference information with dependencies to ensure that their positions after sorting can reflect such dependencies.

[0070] The server may further optimize the arrangement order to improve the user's reading experience and comprehension efficiency. The optimization may include reducing jumps, merging related topics, etc. The server then maps the sorted text reference information back to the knowledge link to ensure that each text reference information is located in its logically correct position. To facilitate user understanding and navigation, the server may generate visual node distribution information, such as knowledge graphs, mind maps, etc. In addition, the server can also provide an interactive interface that allows users to explore node distribution information according to their interests and needs, such as expanding, collapsing, searching, etc.

[0071] The server may verify whether the generated node distribution information is effective and easy to understand through user testing. The testing may include asking users to complete specific tasks, collecting user feedback, etc. Based on the results of the user testing, the server will adjust and optimize the generation algorithm and arrangement of the node distribution information to improve its usability and user satisfaction.

[0072] Through the above steps, the server can arrange the corresponding links of each text reference information in the knowledge link according to the knowledge link nodes of each text reference information in the text reference logic graph set, and generate clear and organized node distribution information, thereby providing users with a better reading and learning experience.

[0073] In a possible implementation, step S131 may include:

[0074] Step S1311, based on the semantic knowledge connection between the X digital culture articles and the text recognition tags of the X digital culture articles, generate a semantic knowledge connection directory for each digital culture article and a connectivity parameter sequence for each digital culture article, the semantic knowledge connection directory for any digital culture article is used to record: the text recognition tags of each digital culture article whose semantic knowledge is connected to the corresponding digital culture article, the connectivity parameter sequence for any digital culture article is used to record the connectivity parameters of the corresponding digital culture article, the connectivity parameters of any digital culture article represent the article statistics of the digital culture articles with semantic knowledge connection for the corresponding digital culture article.

[0075] In this embodiment, the server first analyzes the semantic knowledge connections between the X digital culture articles, which may be determined by the article's subject, keywords, content similarity, etc. At the same time, the server also considers the text recognition tags of each digital culture article, which are usually provided by the author or editor of the article to describe the main content and features of the article.

[0076] Based on this information, the server generates a semantic knowledge connection directory for each digital culture article, which records the text recognition tags of other articles that have semantic knowledge connections with the current article. In addition, the server also generates a connectivity parameter sequence for each digital culture article, which records the connectivity parameters between the current article and other articles, that is, the number of articles with semantic knowledge connections.

[0077] For example, if there is an article A on "Ancient Chinese Architectural Art", its semantic knowledge connection catalog may include text recognition tags of articles on topics such as "Ancient Architectural Style", "Architectural Materials and Crafts", etc. The connectivity parameter sequence records the number of other articles that have semantic connections with article A.

[0078] Step S1312, generating an initialized text label logic diagram set and an initialized first directory.

[0079] In this embodiment, the server can create an initialized text label logic graph set and an empty first directory. The text label logic graph set will be used to store the logical relationship of the text recognition labels generated according to the semantic knowledge connection and the connectivity parameter.

[0080] Step S1313, when a target text recognition tag is found in the text recognition tags of the X digital cultural articles according to the connectivity parameters recorded in the connectivity parameter sequence of each digital cultural article, the target text recognition tag is added to the first directory, and the target text recognition tag meets the following requirement: the connectivity parameter of the digital cultural article represented is an invalid connectivity parameter.

[0081] In this embodiment, the server parses the connectivity parameters recorded in the connectivity parameter sequence of each digital cultural article. When the connectivity parameters of a text recognition tag are parsed to be invalid connectivity parameters (i.e., the article represented by the tag has no valid semantic knowledge connection with other articles), the text recognition tag is added to the first directory.

[0082] For example, if the connectivity parameter sequence of article B shows that it has no valid semantic connection with article A, then the text recognition tag of article B may be added to the first catalog.

[0083] Step S1314, if the first directory is not an empty directory, the text recognition tag at the first position in the first directory is continuously retrieved, and each time the retrieval operation is performed, the currently retrieved text recognition tag is loaded into the text tag logic diagram set, and the connectivity parameters recorded in the connectivity parameter sequence of each digital cultural article represented by the semantic knowledge connection directory of the current digital cultural article are lowered, and the current digital cultural article refers to the digital cultural article represented by the currently retrieved text recognition tag.

[0084] Step S1315: if the first directory is an empty directory, and the text recognition tags of the X digital culture articles have all been added, it is determined that the text tag logic diagram set is generated.

[0085] In this embodiment, if the first directory is not empty, the server will continuously retrieve text identification tags from the first directory. Each time a text identification tag is retrieved, the server will load it into the text tag logic graph set and update the connectivity parameter sequence of each article represented by the semantic knowledge connection directory of the corresponding article.

[0086] This process is actually to gradually build a logical relationship diagram between text tags based on the semantic knowledge of the article. By continuously retrieving tags from the first directory and updating the logical graph set and connectivity parameters, the server can gradually improve this logical relationship diagram.

[0087] When the first directory is empty and the text recognition tags of all X digital culture articles have been added, the server determines that the text tag logic graph set is generated. This means that the semantic knowledge connections of all articles have been fully considered, and the logical relationships between them have been completely recorded in the text tag logic graph set.

[0088] At this point, the server has completed the process of generating a text label logic graph set based on the semantic knowledge connections and text recognition tags between X digital cultural articles. This text label logic graph set not only reflects the subject associations and content hierarchy between articles, but also provides users with a clear and organized reading and learning path.

[0089] In a possible implementation, step S1311 may include:

[0090] Step S1311-1, generate a semantic knowledge connection directory for each of the X digital cultural articles and a connectivity parameter sequence for each of the digital cultural articles, wherein any generated semantic knowledge connection directory is an empty directory, and any generated connectivity parameter sequence records a connectivity parameter that is an invalid connectivity parameter.

[0091] In this embodiment, the server first generates an empty semantic knowledge connection directory and an initial connection parameter sequence for each of the X digital cultural articles, wherein the connection parameters are set to invalid connection parameters, which means that at the beginning, each digital cultural article is considered to have no semantic knowledge connection with other articles.

[0092] For example, if there are three digital cultural articles A, B and C, the server will generate an empty directory for each digital cultural article (such as A's directory, B's directory, C's directory) and a sequence recording invalid connectivity parameters (such as A's connectivity parameter sequence, B's connectivity parameter sequence, C's connectivity parameter sequence).

[0093] Step S1311 - 2 , wandering through the text recognition tags of the X digital cultural articles, and taking the digital cultural article represented by the currently wandered text recognition tag as the first digital cultural article.

[0094] In this embodiment, the server starts to browse the text recognition tags of the X digital cultural articles. During the browsing process, the server refers to the article represented by the text recognition tag currently being processed as the first digital cultural article.

[0095] For example, the server may first select the text recognition tag of article A as the starting point of the walk, and at this time article A is the first digital culture article.

[0096] Step S1311-3: if, based on the semantic knowledge connection between the X digital cultural articles, it is determined that there is a second digital cultural article among the X digital cultural articles that has a semantic knowledge connection with the first digital cultural article, then the text recognition tag of the first digital cultural article is loaded into the semantic knowledge connection directory of each second digital cultural article, the article statistics of the second digital cultural article are used as the connectivity parameter of the first digital cultural article, loaded into the connectivity parameter sequence of the first digital cultural article, and the text recognition tags of the X digital cultural articles are continued to roam.

[0097] Step S1311-4: if it is determined based on the semantic knowledge connection between the X digital culture articles that there is no second digital culture article with which the first digital culture article has a semantic knowledge connection among the X digital culture articles, then continue to roam the text recognition tags of the X digital culture articles.

[0098] In this embodiment, based on the semantic knowledge connection between the X digital culture articles, the server determines whether there are other articles (referred to as second digital culture articles) that have semantic knowledge connection with the first digital culture article.

[0099] If such a second digital cultural article exists, the server will add the text recognition tag of the first digital cultural article to the semantic knowledge connection directory of the second digital cultural article, and use the article statistics of the second digital cultural article (i.e., the number of articles with semantic connections to the article) as a connection parameter, and update it to the connection parameter sequence of the first digital cultural article. Then, the server continues to walk to the next text recognition tag.

[0100] If there is no second digital cultural article having a semantic knowledge connection with the first digital cultural article, the server directly continues to navigate to the next text recognition tag.

[0101] For example, if the server determines that article B is semantically related to article A, it will add the text recognition tag of article A to the semantic knowledge connection directory of article B and update the connectivity parameter sequence of article A to reflect this connection. The server may then move to the text recognition tag of article C for processing.

[0102] Step S1311-5, after the text recognition tags of the X digital cultural articles are all wandered, it is determined that the semantic knowledge connection directory of each digital cultural article and the connection parameter sequence of each digital cultural article are all generated.

[0103] When the server has walked through the text recognition tags of all X digital culture articles, it determines that the semantic knowledge connection directory and connectivity parameter sequence for each digital culture article have been generated. This means that each digital culture article now has a directory that records the semantic connections with other articles and a connectivity parameter sequence that reflects the number of these connections.

[0104] So far, the server has completed the process of generating the semantic knowledge connection directory and connectivity parameter sequence of each digital cultural article based on the semantic knowledge connection and text recognition tags between X digital cultural articles. These directories and sequences provide basic data for the subsequent generation of text label logic graph sets and knowledge links.

[0105] In a possible implementation, step S134 may include:

[0106] Step S1341, based on the semantic knowledge connections between the X digital cultural articles and the Y text reference information corresponding to the X digital cultural articles, generate a semantic knowledge connection directory for each text reference information and a connectivity parameter sequence for each text reference information, wherein the semantic knowledge connection directory for any text reference information is used to record: each text reference information whose semantic knowledge is connected to the corresponding text reference information, and the connectivity parameter sequence for any text reference information is used to record the connectivity parameters of the corresponding text reference information, and the connectivity parameters of any text reference information represent the article statistics of the text reference information having a semantic knowledge connection with the corresponding text reference information.

[0107] In this embodiment, the server first analyzes the semantic knowledge connection between X digital cultural articles and considers Y text reference information corresponding to these articles. The text reference information may be citations, references, footnotes, etc. in the article, which provide the source or additional information of the article content.

[0108] Based on this information, the server generates a semantic knowledge connection directory and a connectivity parameter sequence for each text reference information. The semantic knowledge connection directory records other text reference information that has semantic connections with the current text reference information. The connectivity parameter sequence records the connectivity parameters between the current text reference information and other text reference information, that is, the number of other text reference information that has semantic connections.

[0109] For example, if there is a text reference information about a book on "ancient architectural art", its semantic knowledge connection directory may include other text reference information about ancient architecture, art history or related topics. The connectivity parameter sequence records the number of other text reference information that have semantic connections with this text reference information.

[0110] Step S1342, generating an initialized text reference logic graph set and an initialized second directory.

[0111] In this embodiment, the server creates an initialized text reference logic graph set and an empty second directory. The text reference logic graph set will be used to store the logical relationship between text reference information generated according to semantic knowledge connections.

[0112] Step S1343, when each time the connectivity parameters recorded in the connectivity parameter sequence of each text reference information are parsed to find that the Y text reference information have target text reference information, the existing target text reference information is added to the second directory, and the target text reference information meets the following requirements: the corresponding connectivity parameters are invalid connectivity parameters.

[0113] In this embodiment, the server performs parsing according to the connectivity parameter sequence of each text reference information. When the connectivity parameter of a text reference information is parsed to be an invalid connectivity parameter (ie, the text reference information has no valid semantic knowledge connection with other text reference information), the text reference information is added to the second directory.

[0114] For example, if there is a text citation that is about an unrelated topic or has no clear semantic connection with other text citations, it may be marked as having broken connectivity parameters and added to the secondary catalog.

[0115] Step S1344, if the second directory is not an empty directory, the text reference information at the first position in the second directory is continuously retrieved, and each time the retrieval operation is performed, the currently retrieved text reference information is loaded into the text reference logic graph set, and the connectivity parameters recorded in the connectivity parameter sequence of each text reference information in the current semantic knowledge contact directory are down-adjusted. The current semantic knowledge contact directory refers to the semantic knowledge contact directory of the currently retrieved text reference information.

[0116] If the second directory is not empty, the server will continuously retrieve text reference information from the first directory. Each time a text reference information is retrieved, the server will load it into the text reference logical graph set and update the connection parameter sequence in the semantic knowledge connection directory of the corresponding text reference information.

[0117] This process is actually to gradually build a text reference logic graph based on the semantic knowledge connection between text reference information. By continuously retrieving text reference information from the second directory and updating the logic graph set and connectivity parameters, the server can gradually improve this logical relationship graph.

[0118] Step S1345: If the second directory is an empty directory, and the Y text reference information have all been added, it is determined that the generation of the text reference logic graph set is completed.

[0119] When the second directory is empty and all Y text reference information have been added, the server determines that the text reference logic graph set is generated. This means that the semantic knowledge connections between all text reference information have been fully considered, and the logical relationships between them have been completely recorded in the text reference logic graph set.

[0120] At this point, the server has completed the process of generating a text citation logic graph collection based on the semantic knowledge connections between X digital cultural articles and Y text citation information. This collection not only reflects the subject associations and content hierarchy between text citation information, but also provides users with a clear and organized reading and learning path, helping them better understand and track citations and references in articles.

[0121] In a possible implementation, step S1341 may include:

[0122] Step S1341-1, for the Y text reference information corresponding to the X digital cultural articles, generate a semantic knowledge connection directory for each text reference information and a connectivity parameter sequence for each text reference information, wherein any generated semantic knowledge connection directory is an empty directory, and any generated connectivity parameter sequence records a connectivity parameter that is an invalid connectivity parameter.

[0123] In this embodiment, the server first generates an empty semantic knowledge connection directory and an initial connection parameter sequence for each of the Y text reference information corresponding to the X digital cultural articles. This means that at the beginning, each text reference information is considered to have no semantic knowledge connection with other text reference information, and its connection parameters are set to invalid connection parameters.

[0124] For example, if there are three articles A, B and C, among which article A has two text reference information a1 and a2, article B has one text reference information b1, and article C has three text reference information c1, c2 and c3, then the server will generate an empty directory (such as a1's directory, a2's directory, b1's directory, c1's directory, c2's directory, c3's directory) and a sequence recording invalid connectivity parameters (such as a1's connectivity parameter sequence, a2's connectivity parameter sequence, b1's connectivity parameter sequence, c1's connectivity parameter sequence, c2's connectivity parameter sequence, c3's connectivity parameter sequence) for these six text reference information (a1, a2, b1, c1, c2, c3).

[0125] Step S1341 - 2 , traverse the Y text reference information, and use the currently traversed text reference information as the first text reference information.

[0126] In this embodiment, the server starts to navigate the Y text reference information. During the navigation process, the server refers to the text reference information currently being processed as the first text reference information.

[0127] For example, the server may first select the text reference information a1 of article A as the starting point of the walk, and a1 is the first text reference information at this time.

[0128] Step S1341-3: if, based on the semantic knowledge connection between the X digital culture articles, it is determined that among the Y text reference information there is second text reference information with which the first text reference information has a semantic knowledge connection, then the first text reference information is loaded into the semantic knowledge connection directory of each second text reference information, the article statistics of the second text reference information are used as the connectivity parameter of the first text reference information, loaded into the connectivity parameter sequence of the first text reference information, and the Y text reference information is continued to be traversed.

[0129] Step S1341-4: if it is determined based on the semantic knowledge connection between the X digital culture articles that there is no second text reference information with which the first text reference information has a semantic knowledge connection among the Y text reference information, then continue to traverse the Y text reference information.

[0130] In this embodiment, based on the semantic knowledge connection between the X digital culture articles, the server determines whether there is other text reference information (referred to as second text reference information) having a semantic knowledge connection with the first text reference information.

[0131] If such a second text reference exists, the server will load the first text reference into the semantic knowledge connection directory of the second text reference, and use the article statistics of the second text reference (i.e., the number of articles that have semantic connections with the text reference) as a connection parameter, and update it to the connection parameter sequence of the first text reference. Then, the server continues to navigate to the next text reference.

[0132] If there is no second text reference information having a semantic knowledge connection with the first text reference information, the server directly continues to navigate to the next text reference information.

[0133] For example, if the server determines that text citation c1 is semantically related to text citation a1 (perhaps because they both cite the same authoritative paper on a topic), it will load a1 into c1's semantic knowledge connection directory and update a1's connectivity parameter sequence to reflect this connection (i.e., add the article statistics of the article in which c1 is located to a1's connectivity parameter sequence). The server may then move to text citation a2 for processing.

[0134] Step S1341-5, after the Y text reference information are all wandered, it is determined that the semantic knowledge connection directory of each text reference information and the connectivity parameter sequence of each text reference information are all generated.

[0135] When the server has walked through all Y text references, it determines that the semantic knowledge connection directory and connectivity parameter sequence for each text reference have been generated. This means that each text reference now has a directory that records the semantic connections with other text references and a connectivity parameter sequence that reflects the number of these connections.

[0136] So far, the server has completed the process of generating a semantic knowledge connection directory for each text reference information and a connectivity parameter sequence for each text reference information based on the semantic knowledge connection between X digital cultural articles and the corresponding Y text reference information. These directories and sequences provide basic data for the subsequent generation of a text reference logic graph set.

[0137] In a possible implementation, the platform text content retrieval requirement carries an article retrieval graph, and step S110 may include:

[0138] Step S111, obtaining the article retrieval map from the platform text content retrieval requirements of the digital culture service device, and based on the text reference information of each digital culture article in the digital culture article library, searching the digital culture article library for the digital culture articles corresponding to each text reference information that conforms to the article retrieval map, and generating a reference digital culture article sequence.

[0139] In this embodiment, the server first obtains an article search graph from the platform text content search requirements of the digital cultural service device. This article search graph is a complex query structure that includes multiple reference attributes of the digital cultural articles that the user wants to retrieve, such as category attributes, time domain attributes, user attributes, and status attributes.

[0140] For example, a user may want to retrieve digital cultural articles in the category of "ancient architectural art". These articles need to be published within a specific time period, published by authors with specific user tags, and in a published state. The server will obtain these specific attribute parameters to form a clear article retrieval graph.

[0141] Step S112, searching the digital culture article library for other digital culture articles that have semantic knowledge association with at least one digital culture article in the reference digital culture article sequence, and generating a sequence of associated digital culture articles.

[0142] The server will then search the digital cultural articles corresponding to the text reference information of each digital cultural article in the digital cultural article library based on the text reference information of each digital cultural article in the digital cultural article library. These retrieved articles constitute the reference digital cultural article sequence.

[0143] Taking "ancient architectural art" as an example, the server will search for digital cultural articles that contain keywords related to "ancient architectural art" in the text reference information, are published within the specified time period, are published by authors with specific user tags, and are in the published state. These articles will be considered as reference articles that meet the user's search requirements.

[0144] Step S113: Integrate the reference digital cultural article sequence and the associated digital cultural article sequence to generate the X digital cultural articles.

[0145] Any text reference information and the article search map include at least one of the following reference attributes: category attribute, time domain attribute, user attribute and status attribute. The attribute parameter of the category attribute is the type of digital culture article, the attribute parameter of the time domain attribute is the publishing node of the digital culture article, the attribute parameter of the user attribute is the user tag of the publishing user of the digital culture article, and the attribute parameter of the status attribute is the publishing status of the digital culture article. When the attribute parameters of each reference attribute in any text reference information respectively match the attribute parameters of the corresponding reference attribute in the article search map, the any text reference information is determined to be in compliance with the article search map.

[0146] After finding the reference digital cultural article sequence, the server will further search the digital cultural article library for other digital cultural articles that have semantic knowledge connections with these reference articles. These retrieved articles constitute the associated digital cultural article sequence.

[0147] For example, if a reference article is about "carving skills in ancient architectural art", the server may retrieve other articles that are related to this article in terms of subject, content or citation, such as "decorative arts in ancient architecture" or "application of carving skills in historical buildings".

[0148] Finally, the server integrates the reference digital cultural article sequence and the associated digital cultural article sequence to generate the final X digital cultural articles. These articles not only meet the user's initial search requirements (i.e., the attribute parameters in the article search graph), but also contain other related articles that are semantically related to these requirements, thus providing users with a more comprehensive and in-depth search result.

[0149] In this process, the server continuously matches the reference attributes of text reference information with the attribute parameters in the article search map, and mines the semantic knowledge connection between digital cultural articles, thus achieving efficient and accurate retrieval of the platform's text content. This not only improves the efficiency of users in obtaining the information they need, but also enhances their ability to understand and appreciate digital cultural content.

[0150] In a possible implementation, any reference attribute in the article search graph has at least one attribute parameter.

[0151] Step S111 may include:

[0152] Step S1111: Generate at least one retrieval guide information based on the attribute parameters of each reference attribute in the article retrieval graph. One retrieval guide information includes one attribute parameter in each reference attribute.

[0153] In this embodiment, first, it is necessary to clarify that each reference attribute in the article search graph has at least one attribute parameter. These attribute parameters are conditions set by the user when performing article search, and are used to screen and locate digital cultural articles that meet the requirements.

[0154] For example, in the article retrieval graph, the attribute parameters of the category attribute may be "history", "art", "technology", etc.; the attribute parameters of the time domain attribute may be a specific year range, such as "2000-2020"; the attribute parameters of the user attribute may be the author's user tag, such as "expert", "scholar", "enthusiast", etc.; the attribute parameters of the status attribute may be the publication status of the article, such as "published", "pending review", "draft", etc.

[0155] Step S1112: Load each generated search guide information into different search instances. One search instance includes one search guide information.

[0156] For example, if a user wants to retrieve digital cultural articles in the "history" category, published in "2000-2020", written by "expert" authors, and with a status of "published", the server will generate a search guide information containing these attribute parameters.

[0157] Then, the server will load each generated search guide information into different search instances. Each search instance contains a search guide information for performing a specific search operation in the digital cultural article library.

[0158] Continuing with the above example, the server might create different retrieval instances for the “history” category, “2000-2020” time domain, “expert” user label, and “published” status, and load the corresponding retrieval guidance information into these instances.

[0159] Step S1113, synchronously running each search instance, searching the digital cultural article library for the digital cultural article corresponding to the text reference information matching the corresponding search guide information.

[0160] The server will run these search instances synchronously to search for digital cultural articles in the digital cultural article library that correspond to the text citation information that matches the corresponding search guide information. This means that the server will simultaneously filter and search in the article library based on multiple attribute parameters.

[0161] In this example, the server will search the article library based on four conditions: "History" category, "2000-2020" time domain, "Expert" user tag, and "Published" status, to find digital cultural articles that meet all these conditions.

[0162] Step S1114, collecting the digital cultural articles retrieved by the various search instances to generate the reference digital cultural article sequence.

[0163] Finally, the server will collect the digital cultural articles retrieved by each search instance and generate a reference digital cultural article sequence. The articles in this sequence all meet the attribute parameters in the article search map set by the user.

[0164] In this example, the server will collect all digital cultural articles that meet the four conditions of "history" category, "2000-2020" time domain, "expert" user label and "published" status, and arrange them into a reference digital cultural article sequence in a certain order (such as publishing time, relevance, etc.). This sequence is one of the results generated by the server based on the user's search requirements.

[0165] In a possible implementation, step S1113 may include: loading each of the retrieval instances into a retrieval running task, synchronously running each of the retrieval instances according to the retrieval running task, and searching the digital cultural articles corresponding to the text reference information matching the corresponding retrieval guide information in the digital cultural article library according to the cloud computing node.

[0166] In this embodiment, after receiving the user's search request, the server generates multiple search guide information according to the article search map, and loads these search guide information into different search instances. Next, the server needs to synchronously run these search instances to search the digital cultural article library for the digital cultural article corresponding to the text reference information matching the corresponding search guide information.

[0167] To achieve synchronous operation, the server will load these search instances into one or more search running tasks. Each search running task is responsible for managing and executing a group of search instances, ensuring that they can perform search operations in the digital cultural article library simultaneously and efficiently.

[0168] For example, the server may allocate the retrieval instances to different retrieval running tasks according to the number and complexity of the retrieval instances, so that each retrieval running task can independently process a part of the retrieval instances, thereby improving the overall retrieval efficiency.

[0169] Once the search instance is loaded into the search run task, the server will run each search instance synchronously according to these tasks. This means that the server will start all search instances at the same time and monitor their running status to ensure that they can complete their respective search operations at the same time.

[0170] In this process, the server may use the parallel processing capabilities of cloud computing nodes to assign different retrieval instances to different nodes for execution. In this way, each node can independently process a part of the retrieval task, thereby further accelerating the overall retrieval process.

[0171] When the search instances start running, they will search for digital cultural articles corresponding to the textual reference information that matches them in the digital cultural article library according to their respective search guide information. These textual reference information may include attributes such as the title, author, release time, and keywords of the article. The server will screen and locate digital cultural articles that meet the requirements according to the degree of match between these attributes and the search guide information.

[0172] For example, if the search guidance information of a search instance contains the "history" category and the "2000-2020" time domain attribute parameters, then the instance will search the digital culture article library for those digital culture articles that belong to the "history" category and were published during the period of "2000-2020". By matching the corresponding attribute parameters in the text reference information, the server can accurately locate these qualified articles.

[0173] Finally, the server will collect the digital cultural articles retrieved by each search instance and generate a reference digital cultural article sequence containing multiple articles that meet the user's needs. This sequence can be displayed to the user as part of the user's search results for viewing and reading.

[0174] Figure 2 A digital cultural service system 100 is provided in an embodiment of the present application, including a processor 1001, a memory 1003 and a program code stored in the memory 1003, and the processor 1001 executes the above program code to implement the steps of a platform content intelligent recommendation method based on natural language processing.

[0175] Figure 2 The digital cultural service system 100 shown includes: a processor 1001 and a memory 1003. The processor 1001 and the memory 1003 are connected, such as through a bus 1002. Optionally, the digital cultural service system 100 may also include a transceiver 1004, which can be used for data interaction between the digital cultural service system 100 and other digital cultural service systems 100, such as data transmission and / or data reception. It should be noted that the transceiver 1004 is not limited to one in actual scheduling, and the structure of the digital cultural service system 100 does not constitute a limitation on the embodiments of the present application.

[0176] Processor 1001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. It may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of this application. Processor 1001 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0177] The bus 1002 may include a path to transmit information between the above components. The bus 1002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus 1002 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 2 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0178] The memory 1003 may be a ROM (Read Only Memory) or other types of static storage devices that can store static information and instructions, a RAM (Random Access Memory) or other types of dynamic storage devices that can store information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, optical disk storage (including compressed optical disk, laser disk, optical disk, digital versatile disk, Blu-ray disk, etc.), magnetic disk storage medium, other magnetic storage devices, or any other medium that can be used to carry or store program code and can be read by a computer, without limitation herein.

[0179] The memory 1003 is used to store program codes for executing the embodiments of the present application, and the execution is controlled by the processor 1001. The processor 1001 is used to execute the program codes stored in the memory 1003 to implement the steps shown in the above method embodiments.

[0180] An embodiment of the present application provides a computer-readable storage medium having program code stored thereon. When the program code is executed by a processor, the steps and corresponding contents of the aforementioned method embodiment can be implemented.

[0181] The above is only an optional implementation method for some implementation scenarios of the present application. It should be pointed out that for ordinary technicians in this technical field, without departing from the technical concept of the scheme of the present application, other similar implementation methods based on the technical ideas of the present application also belong to the protection scope of the embodiments of the present application.

Claims

1. A platform content intelligent recommendation method based on natural language processing, characterized in that: The method comprises: Based on the platform text content retrieval requirements of the digital cultural service device, an article retrieval operation is performed in the digital cultural article library to generate X digital cultural articles, where X is a positive integer greater than 1; Determine the semantic knowledge connection between the X digital cultural articles and Y text reference information corresponding to the X digital cultural articles based on a natural language processing algorithm, where one text reference information is used to reference at least one digital cultural article, and Y is a positive integer; the digital cultural articles with semantic knowledge connection for any digital cultural article include: when the corresponding digital cultural article is published, the corresponding derived and extended digital cultural articles; Based on the semantic knowledge connections between the X digital cultural articles and the Y text citation information corresponding to the X digital cultural articles, the knowledge link nodes of the X digital cultural articles are determined, wherein the knowledge link node of any digital cultural article is located after the knowledge link nodes of each digital cultural article with which the corresponding digital cultural article has semantic knowledge connections, and the knowledge link nodes of each digital cultural article corresponding to the same text citation information are continuous; In the digital culture service device, recommending the X digital culture articles according to the knowledge link nodes of the X digital culture articles; The platform text content retrieval requirement carries an article retrieval graph, and the platform text content retrieval requirement based on the digital cultural service device performs an article retrieval operation in the digital cultural article library to generate X digital cultural articles, including: The article search graph is obtained from the platform text content search demand of the digital culture service device, and based on the text reference information of each digital culture article in the digital culture article library, the digital culture article corresponding to each text reference information that meets the article search graph is searched in the digital culture article library to generate a reference digital culture article sequence; In the digital culture article library, searching for other digital culture articles that have semantic knowledge association with at least one digital culture article in the reference digital culture article sequence, and generating a sequence of associated digital culture articles; Integrating the reference digital cultural article sequence and the associated digital cultural article sequence to generate the X digital cultural articles; Any text reference information and the article search map include at least one of the following reference attributes: category attribute, time domain attribute, user attribute and status attribute, the attribute parameter of the category attribute is the type of digital culture article, the attribute parameter of the time domain attribute is the publishing node of the digital culture article, the attribute parameter of the user attribute is the user tag of the publishing user of the digital culture article, and the attribute parameter of the status attribute is the publishing status of the digital culture article; wherein, when the attribute parameters of each reference attribute in any text reference information respectively match the attribute parameters of the corresponding reference attribute in the article search map, the any text reference information is determined to be in compliance with the article search map; The determining of the knowledge link nodes of the X digital cultural articles based on the semantic knowledge connections between the X digital cultural articles and the Y text citation information corresponding to the X digital cultural articles includes: Based on the semantic knowledge connection between the X digital cultural articles, a text label logic graph set is generated according to the text recognition tags of the X digital cultural articles, in which the text recognition tag of any digital cultural article is located after the text recognition tags of each digital cultural article with which the corresponding digital cultural article has a semantic knowledge connection; Based on the semantic knowledge connection between the X digital cultural articles, a text reference logic graph set is generated according to the Y text reference information corresponding to the X digital cultural articles, wherein the logic node of any text reference information in the text reference logic graph set is located after the logic node of the text reference information with which the corresponding text reference information has a semantic knowledge connection in the text reference logic graph set; the text reference information with which any text reference information has a semantic knowledge connection includes: the text reference information of the digital cultural article with which the digital cultural article corresponding to the corresponding text reference information has a semantic knowledge connection; Generate a knowledge link based on the text tag logic graph set and the Y text reference information, the knowledge link comprising: a reference corresponding link of each text reference information, any reference corresponding link comprising a text identification tag of each digital cultural article corresponding to the corresponding text reference information, and the logical knowledge association between each text identification tag in any reference corresponding link matches the logical knowledge association of the corresponding text identification tag in the text tag logic graph set; According to the knowledge link nodes of each text reference information in the text reference logic graph set, the reference corresponding links of each text reference information in the knowledge link are arranged in position to generate node distribution information, and the node distribution information represents the knowledge link nodes of the X digital cultural articles.

2. The platform content intelligent recommendation method based on natural language processing according to claim 1 is characterized in that: The generating of a text label logic graph set based on the semantic knowledge connection between the X digital culture articles and the text recognition labels of the X digital culture articles includes: Based on the semantic knowledge connection between the X digital culture articles and the text recognition tags of the X digital culture articles, a semantic knowledge connection directory of each digital culture article and a connectivity parameter sequence of each digital culture article are generated, wherein the semantic knowledge connection directory of any digital culture article is used to record: the text recognition tags of each digital culture article whose semantic knowledge is connected to the corresponding digital culture article, and the connectivity parameter sequence of any digital culture article is used to record the connectivity parameters of the corresponding digital culture article, and the connectivity parameters of any digital culture article represent the article statistics of the digital culture articles with semantic knowledge connection to the corresponding digital culture article; Generate an initialized text label logic diagram set and an initialized first directory; When a target text recognition tag is found in the text recognition tags of the X digital cultural articles according to the connectivity parameters recorded in the connectivity parameter sequence of each digital cultural article, the target text recognition tag is added to the first directory, and the target text recognition tag meets the following requirements: the connectivity parameter of the digital cultural article represented is an invalid connectivity parameter; If the first directory is not an empty directory, the text recognition tag at the first position in the first directory is continuously retrieved, and each time the retrieval operation is performed, the currently retrieved text recognition tag is loaded into the text label logic graph set, and the connectivity parameters recorded in the connectivity parameter sequence of each digital cultural article represented by the semantic knowledge connection directory of the current digital cultural article are down-adjusted, and the current digital cultural article refers to the digital cultural article represented by the currently retrieved text recognition tag; If the first directory is an empty directory, and the text recognition tags of the X digital culture articles have all been added, it is determined that the text tag logic diagram set is generated completely.

3. The platform content intelligent recommendation method based on natural language processing according to claim 2 is characterized in that: The generating of a semantic knowledge connection directory of each digital culture article and a connectivity parameter sequence of each digital culture article based on the semantic knowledge connection between the X digital culture articles and the text recognition tags of the X digital culture articles comprises: Generate a semantic knowledge connection directory of each digital cultural article in the X digital cultural articles and a connectivity parameter sequence of each digital cultural article, wherein any generated semantic knowledge connection directory is an empty directory, and connectivity parameters recorded in any generated connectivity parameter sequence are invalid connectivity parameters; Walking through the text recognition tags of the X digital cultural articles, and taking the digital cultural article represented by the currently walked text recognition tag as the first digital cultural article; If it is determined that there is a second digital cultural article with which the first digital cultural article has a semantic knowledge connection among the X digital cultural articles based on the semantic knowledge connection between the X digital cultural articles, then the text recognition tag of the first digital cultural article is loaded into the semantic knowledge connection directory of each second digital cultural article, the article statistics of the second digital cultural article are used as the connectivity parameter of the first digital cultural article, loaded into the connectivity parameter sequence of the first digital cultural article, and the text recognition tags of the X digital cultural articles are continuously roamed; If it is determined based on the semantic knowledge connection between the X digital cultural articles that there is no second digital cultural article with which the first digital cultural article has a semantic knowledge connection among the X digital cultural articles, then continue to walk through the text recognition tags of the X digital cultural articles; After the text recognition tags of the X digital cultural articles are all wandered, it is determined that the semantic knowledge connection directory of each digital cultural article and the connection parameter sequence of each digital cultural article are all generated.

4. The platform content intelligent recommendation method based on natural language processing according to claim 1 is characterized in that: The method of generating a text reference logic graph set based on the semantic knowledge relationship between the X digital cultural articles and according to the Y text reference information corresponding to the X digital cultural articles includes: Based on the semantic knowledge connections between the X digital cultural articles and the Y text citation information corresponding to the X digital cultural articles, a semantic knowledge connection directory for each text citation information and a connectivity parameter sequence for each text citation information are generated, wherein the semantic knowledge connection directory for any text citation information is used to record: the text citation information of the corresponding text citation information that is connected to the semantic knowledge, and the connectivity parameter sequence for any text citation information is used to record the connectivity parameters of the corresponding text citation information, and the connectivity parameters of any text citation information represent the article statistics of the text citation information that has the semantic knowledge connection with the corresponding text citation information; Generate an initialized text reference logic graph set and an initialized second directory; When the Y text reference information is parsed according to the connectivity parameters recorded in the connectivity parameter sequence of each text reference information and the target text reference information exists, the target text reference information is added to the second directory, and the target text reference information meets the following requirements: the corresponding connectivity parameter is an invalid connectivity parameter; If the second directory is not an empty directory, the text reference information at the first position in the second directory is continuously retrieved, and each time the retrieval operation is performed, the currently retrieved text reference information is loaded into the text reference logic graph set, and the connectivity parameters recorded in the connectivity parameter sequence of each text reference information in the current semantic knowledge connection directory are down-adjusted, and the current semantic knowledge connection directory refers to the semantic knowledge connection directory of the currently retrieved text reference information; If the second directory is an empty directory, and the Y text reference information have all been added, it is determined that the text reference logic graph set is generated completely.

5. The method for intelligent platform content recommendation based on natural language processing according to claim 4, characterized in that: The generating of the semantic knowledge connection directory of each text reference information and the connectivity parameter sequence of each text reference information based on the semantic knowledge connection between the X digital cultural articles and the Y text reference information corresponding to the X digital cultural articles includes: For the Y text reference information corresponding to the X digital cultural articles, a semantic knowledge connection directory for each text reference information and a connectivity parameter sequence for each text reference information are generated, wherein any of the generated semantic knowledge connection directories is an empty directory, and the connectivity parameters recorded in any of the generated connectivity parameter sequences are invalid connectivity parameters; Walk through the Y text reference information, and use the currently walked text reference information as the first text reference information; If, based on the semantic knowledge connection between the X digital culture articles, it is determined that there is a second text reference information with which the first text reference information has a semantic knowledge connection among the Y text reference information, the first text reference information is loaded into the semantic knowledge connection directory of each second text reference information, the article statistics of the second text reference information are used as the connectivity parameter of the first text reference information, loaded into the connectivity parameter sequence of the first text reference information, and the Y text reference information is continuously roamed; If it is determined based on the semantic knowledge connection between the X digital culture articles that there is no second text reference information with which the first text reference information has a semantic knowledge connection among the Y text reference information, then continue to walk the Y text reference information; After the Y text reference information are all wandered, it is determined that the semantic knowledge connection directory of each text reference information and the connection parameter sequence of each text reference information are all generated.

6. The method for intelligent recommendation of platform content based on natural language processing according to claim 1, characterized in that: Any one of the reference attributes in the article search graph has at least one attribute parameter; based on the text reference information of each digital cultural article in the digital cultural article library, searching the digital cultural article library for digital cultural articles corresponding to each text reference information that matches the article search graph, and generating a reference digital cultural article sequence, including: Based on the attribute parameters of each reference attribute in the article search map, at least one search guide information is generated; one search guide information includes an attribute parameter of each reference attribute; Loading each generated search guide information into different search instances; one search instance includes one search guide information; Running each search instance synchronously, searching the digital cultural article library for the digital cultural article corresponding to the text reference information matching the corresponding search guide information; The digital cultural articles retrieved by the various search instances are collected to generate the reference digital cultural article sequence.

7. The method for intelligent recommendation of platform content based on natural language processing according to claim 6, characterized in that: The synchronously running of each search instance to search the digital cultural article library for the digital cultural article corresponding to the text reference information matching the corresponding search guide information includes: Loading each of the search instances into the search running task; The various search instances are synchronously run according to the search running task, and the digital cultural articles corresponding to the text reference information matching the corresponding search guide information are searched in the digital cultural article library according to the cloud computing node.

8. A digital cultural service system, characterized in that: The digital cultural service system includes a processor and a machine-readable storage medium, in which machine-executable instructions are stored. The machine-executable instructions are loaded and executed by the processor to implement the platform content intelligent recommendation method based on natural language processing as described in any one of claims 1-7.