A Method for Constructing a Multi-Layer Scientific Research Collaboration Network Based on Professional Papers and Patent Documents

Abstract

This invention discloses a method for constructing a multi-layered research collaboration network based on professional papers and patent documents, belonging to the field of knowledge base construction technology. It addresses the problem of inaccurate acquisition of relevance data between professional papers and patent documents. This invention improves the accuracy of collaboration identification through key data verification and multiple matching strategies. Specifically, to address the issue of name duplication, institutional matching is introduced as a supplement, effectively reducing misjudgments. It constructs five different types of collaboration data, each reflecting different aspects of research collaboration from different perspectives. Through network overlap and integration, it not only preserves the collaboration relationships within each professional paper and patent document but also reveals the direct and indirect collaboration relationships between them.

Description

Technical Field

[0001] This invention relates to the field of knowledge base construction technology, specifically a method for constructing a multi-layered scientific research collaboration network based on professional papers and patent documents. Background Technology

[0002] Research collaborations involving professional papers and patent documents mainly refer to the cooperation and connections formed between different researchers or research teams in the field of scientific research through the publication and application of professional papers and patent documents.

[0003] Patent application CN110533253B discloses a method for predicting research collaboration relationships based on heterogeneous information networks. It primarily uses a graph autoencoder model to obtain the vector representation of nodes, thus solving the problem of excessively high dimensionality caused by sparse vectors in research collaboration networks. By fusing two similarity calculation methods, it avoids the computational error caused by using only one similarity method, thereby improving the accuracy of author collaboration relationship prediction. While the aforementioned patent solves the problem of research collaboration prediction, the following issues still exist in practical application:

[0004] 1. After acquiring the raw data from professional papers and patent documents, no further data processing was performed, resulting in poor data quality and making accurate analysis impossible later.

[0005] 2. The lack of a more detailed analysis of the correlation between professional papers and patent documents made it impossible to confirm the cooperation relationship.

[0006] 3. The relationship between professional papers and patent documents was not effectively assessed, resulting in an inability to clearly understand the correlation between the two. Summary of the Invention

[0007] The purpose of this invention is to provide a method for constructing a multi-layered research collaboration network based on professional papers and patent documents. Through key data verification and multiple matching strategies, the accuracy of collaboration identification is improved. Specifically, to address the issue of name duplication, institutional matching is introduced as a supplement, effectively reducing false positives. Five different types of collaboration data are constructed, each reflecting different aspects of research collaboration from different perspectives. Through network overlap and integration, not only are the collaboration relationships within each professional paper and patent document preserved, but also the direct and indirect collaboration relationships between them are revealed. This multi-layered collaboration revealing method helps to comprehensively understand the complexity and diversity of research collaboration and can solve problems in existing technologies.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] A method for constructing a multi-layered research collaboration network based on professional papers and patent documents includes the following steps:

[0010] S1: Scientific research collaboration data acquisition: Acquire professional papers and patent documents from different data acquisition centers, confirm the basic information of professional papers and patent documents, and mark professional papers and patent documents whose basic information has been confirmed as scientific research data to be processed.

[0011] S2: Scientific research collaboration data processing: Preprocess the scientific research data to be processed, and obtain the collaboration data to be extracted after the data preprocessing is completed;

[0012] S3: Confirmation of Cooperation Relationship: Confirm the key data of the cooperation data to be extracted, and extract the relationship based on the confirmed key data. After the relationship extraction is completed, the target cooperation relationship data is obtained.

[0013] S4: Construction of Cooperation Network: The target cooperation data is used to construct networks for professional papers and patent documents respectively. After the networks for professional papers and patent documents are constructed, the networks are overlapped. After the networks are merged, the overlapping network of professional papers and patent documents is obtained and labeled as the target multi-layer scientific research data.

[0014] S5: Network Construction and Relationship Analysis: Perform relationship network analysis on the target multi-layer scientific research data. After the relationship network analysis, obtain the cooperation characteristics data of professional papers and patent documents, and display the cooperation characteristics data on mobile terminals.

[0015] Preferably, in step S1, professional papers and patent documents are acquired from different data acquisition centers, and the basic information of the professional papers and patent documents is confirmed, including:

[0016] Professional papers are retrieved from journal databases, and patent documents are retrieved from patent databases;

[0017] Furthermore, the basic information of the acquired professional papers and patent documents was confirmed.

[0018] The basic information for professional papers includes the paper title, field of study, content, author's name, publishing institution, and publication date.

[0019] Basic information in patent documents includes patent title, patent field, patent office, patent content, applicant and inventor names, application date and publication date;

[0020] The acquired professional papers and patent documents, along with their basic information, are matched and then uniquely coded.

[0021] The professional papers and patent documents with unique codes, along with their basic information, are labeled as research data to be processed.

[0022] Preferably, the data preprocessing for the research data to be processed in S2 includes:

[0023] The scientific research data to be processed is sequentially cleaned, integrated, and transformed.

[0024] Data cleaning includes removing duplicate values, missing values, outliers, and noise from the research data to be processed.

[0025] After data cleaning, data integration is performed, and the dataset is formed by integrating the cleaned scientific research data into a single dataset.

[0026] After data integration, data conversion is performed to standardize the format of the integrated scientific research data to be processed.

[0027] After the data conversion is completed, the preprocessed scientific research data to be processed is obtained and marked as collaborative data to be extracted.

[0028] Preferably, in S3, key data confirmation is performed on the cooperation data to be extracted, and relationship extraction is performed based on the confirmed key data, including:

[0029] The key data in the collaborative data to be extracted was confirmed, including the basic information of professional papers and patent documents.

[0030] Relationship extraction is performed after key data is confirmed;

[0031] First, obtain the names of the authors and applicants or inventors of professional papers and patent documents from the basic information.

[0032] The string matching method is used to match the names of authors and applicants or inventors in professional papers and patent documents;

[0033] Once a name match is found, the first cooperation relationship data is obtained.

[0034] When names are duplicated, the institution is matched based on the publishing institution of the professional papers and patent documents and the patent institution. If the patent institution and the publishing institution have a cooperative relationship, the institution is successfully matched and the second cooperative relationship data is obtained.

[0035] Preferably, in S3, key data confirmation of the cooperation data to be extracted, and relationship extraction based on the confirmed key data, further includes:

[0036] Extract the content of professional papers and patent documents from basic information;

[0037] Data mining techniques are used to correlate the content of academic papers and patents. If there is a correlation between academic papers and patent documents, the correlation mining will reveal the common authors, common fields, and common institutions of the academic papers and patent documents.

[0038] The professional papers and patent documents obtained through correlation mining will be used as third-party collaboration data.

[0039] Based on the third-party cooperation data, the citation relationships between the content of professional papers and patent documents are searched.

[0040] When there are mutual citations in the references and cited documents in the content of academic papers and patents, the academic papers and patent documents are regarded as the fourth cooperation relationship data.

[0041] Preferably, in S3, key data confirmation of the cooperation data to be extracted, and relationship extraction based on the confirmed key data, further includes:

[0042] Obtain the publication date and application date of professional papers and patent documents from the basic information;

[0043] Cross-correlation analysis was used to analyze the cross-correlation coefficient between publication date and application date, and the correlation data and time delay data between professional papers and patent documents were obtained after the analysis.

[0044] The correlation data is used to determine whether there is a relationship between professional papers and patent documents. When there is a relationship between professional papers and patent documents, the time sequence between professional papers and patent documents is determined based on the time delay data.

[0045] Professional papers and patent documents completed in chronological order are used as the fifth collaborative relationship data.

[0046] The data for the first, second, third, fourth, and fifth cooperative relationships are uniformly labeled as the target cooperative relationship data.

[0047] Preferably, in S4, the target cooperation relationship data is used to construct networks for professional papers and patent documents respectively. After the networks for professional papers and patent documents are constructed, the networks are merged, including:

[0048] Based on the target cooperation relationship data, nodes are generated for professional papers and patent documents respectively;

[0049] This involves generating paper nodes from professional papers.

[0050] After the paper node is generated, patent sub-nodes are established based on the cooperation relationship data with patent documents in the target cooperation relationship data. After the patent sub-nodes are established, the cooperation relationship edges between the paper node and several patent sub-nodes are obtained.

[0051] By extending the cooperative relationships between paper nodes and several patent child nodes to edges, and using the Pajek tool in graph theory, a cooperative network for professional papers is constructed.

[0052] Once the collaborative network for professional papers is constructed, a collaborative relationship network for professional papers is obtained.

[0053] Preferably, in step S4, the network construction of professional papers and patent documents for the target cooperation relationship data is performed separately, and the networks are merged after the professional papers and patent documents are completed. This also includes:

[0054] Generate patent nodes from patent documents;

[0055] After the patent node is generated, paper sub-nodes are established based on the cooperation relationship data with professional papers in the target cooperation relationship data. After the paper sub-nodes are established, the cooperation relationship edges between the patent node and several paper sub-nodes are obtained.

[0056] By extending the cooperative relationships between patent nodes and several paper sub-nodes to edges, the Pajek tool in graph theory is used to construct a cooperative network of patent documents.

[0057] Once the collaborative network of patent documents is constructed, a collaborative relationship network of patent documents is obtained.

[0058] Preferably, in step S4, the network construction of professional papers and patent documents for the target cooperation relationship data is performed separately, and the networks are merged after the professional papers and patent documents are completed. This also includes:

[0059] Overlap the patent document collaboration network and the professional paper collaboration network;

[0060] When patent document collaboration networks and professional paper collaboration networks overlap, overlapping nodes are generated when there are co-authors and public institutions.

[0061] If there is a direct collaboration between the patent document collaboration network and the professional paper collaboration network, add corresponding outgoing edges to the overlapping nodes to connect the nodes with the collaboration relationship.

[0062] If there is an indirect cooperative relationship between the patent document cooperative relationship network and the professional paper cooperative relationship network, the co-citation analysis method is used to add potential directed edges in the overlapping nodes;

[0063] After the network nodes of the patent literature collaboration network and the professional paper collaboration network overlap and are generated, the network is integrated to obtain the target multi-layer scientific research data.

[0064] Preferably, in S5, a relationship network analysis is performed on the target multi-layer scientific research data. After the relationship network analysis, data on the collaborative characteristics of professional papers and patent documents are obtained, including:

[0065] Perform topology analysis on the network nodes in the target multi-layer scientific research data;

[0066] Among them, topology analysis involves using node degree analysis to confirm the number of each network node and the number of directed edges in the target multi-layer scientific research data.

[0067] Then, clustering coefficient analysis was used to confirm the tight clusters between each network node in the target multi-layer scientific research data;

[0068] Finally, path length analysis was used to determine the average shortest path length between any two nodes in each network of the target multi-layer scientific research data.

[0069] Based on the number of each network node and its edges, the number of tight clusters between each network node, and the average shortest path length between any two nodes, we obtain data on the collaborative characteristics between academic papers and patent documents.

[0070] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0071] 1. This invention provides a method for constructing a multi-layered research collaboration network based on professional papers and patent documents. Through key data verification and multiple matching strategies, it improves the accuracy of collaboration identification. In particular, to address the issue of name duplication, it introduces institutional matching as a supplement, effectively reducing misjudgments. It constructs five different types of collaboration data, each reflecting different aspects of research collaboration from different perspectives.

[0072] 2. This invention provides a method for constructing a multi-layered research collaboration network based on professional papers and patent documents. Through network overlap and integration, it not only preserves the collaboration relationships within each professional paper and patent document but also reveals the direct and indirect collaboration relationships between them. This multi-layered collaboration revealing method helps to comprehensively understand the complexity and diversity of research collaboration. Utilizing Pajek tools in graph theory for network construction and visualization makes the presentation of research collaboration relationships more intuitive and clear.

[0073] 3. This invention provides a method for constructing a multi-layered scientific research collaboration network based on professional papers and patent documents. Through node degree analysis, clustering coefficient analysis, and path length analysis, it comprehensively and deeply analyzes the topological structure of the scientific research collaboration network. This comprehensive analysis reveals the complex relationships between nodes in the network and the overall structural characteristics of the network. By comparing the collaboration characteristics data of different networks or different time periods, it can reveal the changes and evolutionary patterns of scientific research collaboration trends. Attached Figure Description

[0074] Figure 1 This is a schematic diagram illustrating the collaborative steps between the professional papers and patent documents of this invention.

[0075] Figure 2 This is a schematic diagram illustrating the collaborative process between the professional papers and patent documents of this invention. Detailed Implementation

[0076] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0077] To address the problem in existing technologies where raw data from academic papers and patent documents is acquired but not further processed, resulting in poor data quality and hindering accurate analysis later, please refer to [the relevant documentation / reference]. Figure 1 and Figure 2 This embodiment provides the following technical solution:

[0078] A method for constructing a multi-layered research collaboration network based on professional papers and patent documents includes the following steps:

[0079] S1: Scientific research collaboration data acquisition: Acquire professional papers and patent documents from different data acquisition centers, confirm the basic information of professional papers and patent documents, and mark professional papers and patent documents whose basic information has been confirmed as scientific research data to be processed.

[0080] Among them, obtaining professional papers and patent documents and their detailed information provides a rich data source for building a multi-layered network of scientific research collaborations;

[0081] S2: Scientific research collaboration data processing: Preprocess the scientific research data to be processed, and obtain the collaboration data to be extracted after the data preprocessing is completed;

[0082] Among them, the characteristics of professional papers and patent documents were fully considered in the data preprocessing process, so that the constructed multi-layer scientific research cooperation network can more comprehensively reflect the actual situation of scientific research cooperation.

[0083] S3: Confirmation of Cooperation Relationship: Confirm the key data of the cooperation data to be extracted, and extract the relationship based on the confirmed key data. After the relationship extraction is completed, the target cooperation relationship data is obtained.

[0084] Among these measures, the accuracy of identifying cooperative relationships was improved through key data verification and multiple matching strategies.

[0085] S4: Construction of Cooperation Network: The target cooperation data is used to construct networks for professional papers and patent documents respectively. After the networks for professional papers and patent documents are constructed, the networks are overlapped. After the networks are merged, the overlapping network of professional papers and patent documents is obtained and labeled as the target multi-layer scientific research data.

[0086] In particular, through network overlap and integration, not only were the internal collaborative relationships of professional papers and patent documents preserved, but also the direct and indirect collaborative relationships between them were revealed;

[0087] S5: Network Construction and Relationship Analysis: Perform relationship network analysis on the target multi-layer scientific research data. After the relationship network analysis, obtain the cooperation characteristics data of professional papers and patent documents, and display the cooperation characteristics data on mobile terminals.

[0088] By comparing the characteristics of collaboration across different networks or time periods, we can reveal the changes and evolution of research collaboration trends. By analyzing these characteristics, researchers can understand their position and influence within the network.

[0089] For S1, professional papers and patent documents were acquired from different data acquisition centers, and the basic information of the professional papers and patent documents was confirmed, including:

[0090] Professional papers are retrieved from journal databases, and patent documents are retrieved from patent databases;

[0091] Furthermore, the basic information of the acquired professional papers and patent documents was confirmed.

[0092] The basic information for professional papers includes the paper title, field of study, content, author's name, publishing institution, and publication date.

[0093] Basic information in patent documents includes patent title, patent field, patent office, patent content, applicant and inventor names, application date and publication date;

[0094] The acquired professional papers and patent documents, along with their basic information, are matched and then uniquely coded.

[0095] The professional papers and patent documents with unique codes, along with their basic information, are labeled as research data to be processed.

[0096] Specifically, acquiring professional papers and patent documents from different data acquisition centers (such as journal databases and patent databases) ensures the comprehensiveness and breadth of the data. This multi-channel data collection approach helps reduce data omissions and biases, and matching the basic information of professional papers and patent documents with unique codes facilitates data standardization and unification. This approach provides convenience for subsequent data analysis and relationship network construction, avoiding analytical difficulties caused by inconsistent data formats. Obtaining professional papers and patent documents and their detailed information provides a rich data source for building multi-layered research collaboration networks. In-depth analysis of this data can reveal research collaboration relationships at different levels.

[0097] For S2, data preprocessing of the research data to be processed includes:

[0098] The scientific research data to be processed is sequentially cleaned, integrated, and transformed.

[0099] Data cleaning includes removing duplicate values, missing values, outliers, and noise from the research data to be processed.

[0100] After data cleaning, data integration is performed, and the dataset is formed by integrating the cleaned scientific research data into a single dataset.

[0101] After data integration, data conversion is performed to standardize the format of the integrated scientific research data to be processed.

[0102] After the data conversion is completed, the preprocessed scientific research data to be processed is obtained and marked as collaborative data to be extracted.

[0103] Specifically, data cleaning, integration, and transformation avoid analytical biases caused by duplicate data, ensuring that nodes and edges in the network are unique, thereby improving the accuracy and reliability of the network. Appropriate padding or deletion operations reduce the impact of incomplete data on the analysis results, ensuring network integrity. Identifying and correcting outliers or noisy data reduces their interference with the network structure, improving the accuracy of network analysis. Integrating cleaned data from different sources into a single dataset simplifies data complexity, making subsequent data analysis and network construction more efficient and convenient. Standardizing data formats ensures that all data meets the requirements of subsequent analysis algorithms, improving algorithm compatibility and operational efficiency. Because the data preprocessing fully considers the characteristics of professional papers and patent documents, the constructed multi-layered research collaboration network can more comprehensively reflect the actual situation of research collaboration.

[0104] To address the issue in existing technologies where the correlation between academic papers and patent documents is not analyzed in greater detail, thus hindering the confirmation of collaborative relationships, please refer to [link to relevant documentation]. Figure 1 and Figure 2 This embodiment provides the following technical solution:

[0105] In S3, key data for the cooperation data to be extracted is confirmed, and relationships are extracted based on the confirmed key data, including:

[0106] The key data in the collaborative data to be extracted was confirmed, including the basic information of professional papers and patent documents.

[0107] Relationship extraction is performed after key data is confirmed;

[0108] First, obtain the names of the authors and applicants or inventors of professional papers and patent documents from the basic information.

[0109] The string matching method is used to match the names of authors and applicants or inventors in professional papers and patent documents;

[0110] Once a name match is found, the first cooperation relationship data is obtained.

[0111] When names are duplicated, the institution is matched based on the publishing institution of the professional papers and patent documents and the patent institution. If the patent institution and the publishing institution have a cooperative relationship, the institution is successfully matched and the second cooperative relationship data is obtained.

[0112] Extract the content of professional papers and patent documents from basic information;

[0113] Data mining techniques are used to correlate the content of academic papers and patents. If there is a correlation between academic papers and patent documents, the correlation mining will reveal the common authors, common fields, and common institutions of the academic papers and patent documents.

[0114] The professional papers and patent documents obtained through correlation mining will be used as third-party collaboration data.

[0115] Based on the third-party cooperation data, the citation relationships between the content of professional papers and patent documents are searched.

[0116] When there are mutual citations in the references and cited documents in the content of academic papers and patents, the academic papers and patent documents are regarded as the fourth cooperation relationship data.

[0117] Obtain the publication date and application date of professional papers and patent documents from the basic information;

[0118] Cross-correlation analysis was used to analyze the cross-correlation coefficient between publication date and application date, and the correlation data and time delay data between professional papers and patent documents were obtained after the analysis.

[0119] The correlation data is used to determine whether there is a relationship between professional papers and patent documents. When there is a relationship between professional papers and patent documents, the time sequence between professional papers and patent documents is determined based on the time delay data.

[0120] Professional papers and patent documents completed in chronological order are used as the fifth collaborative relationship data.

[0121] The data for the first, second, third, fourth, and fifth cooperative relationships are uniformly labeled as the target cooperative relationship data.

[0122] Specifically, using string matching to match the author names and applicant or inventor names of professional papers with those of patent documents is an efficient and relatively simple method. This method can quickly identify potential collaborations, providing a foundation for subsequent in-depth analysis. When name matching encounters difficulties or uncertainties, institutional matching can be used to supplement and verify the collaboration, providing more evidence for confirming the collaboration. Data mining methods are used to correlate and analyze the content of papers and patents, identifying common authors, common fields, and common institutions between professional papers and patent documents. This method goes beyond superficial collaboration information (such as co-authorship) and delves into the content level, revealing deeper research collaborations. It further searches for references and cited documents in the content of papers and patents to identify mutual citation relationships, and uses professional papers and patent documents with citation relationships as fourth-level collaboration data. This expands the boundaries of collaborative relationships to include indirect, knowledge-flow-based collaborations. Beyond traditional author and institutional information, it introduces paper and patent content as data dimensions, enriching and comprehensively building the data foundation for the network. Because the network construction process considers content-level connections and citations, subsequent analysis can be more in-depth and detailed. For example, it allows for analysis of knowledge flows between different fields, evaluation of the efficiency and quality of research collaborations, and analysis of the cross-correlation coefficient between the publication dates of professional papers and the application dates of patent documents, revealing their temporal correlation and potential influence. This analytical method enhances the ability to identify research collaborations, uncovering those that might be overlooked by traditional methods. The calculation of time delay data allows us to determine the chronological order of professional papers and patent documents, which helps in understanding the temporal characteristics of research activities. In research collaborations, sometimes a study may be published as a paper first, followed by a patent application; or conversely, the proposal of a patented technology may inspire subsequent research papers. This time-dimensional analysis provides an important perspective for understanding the dynamic process of scientific research collaboration. By combining relevance data and time-delay data, it is possible to more accurately determine whether there is a genuine connection between academic papers and patent documents. This avoids misjudgments that may arise from matching only author or institutional information, thus improving the accuracy of collaboration relationships. In summary, the accuracy of collaboration relationship identification has been improved through key data confirmation and multiple matching strategies. In particular, for the issue of name duplication, institutional matching was introduced as a supplement, effectively reducing misjudgments. First to fifth different types of collaboration relationship data were constructed, each reflecting different aspects of scientific research collaboration from different perspectives.

[0123] To address the problem in existing technologies where the correlation between academic papers and patent documents is not effectively determined, thus hindering a clear understanding of their relationship, please refer to [link to relevant documentation]. Figure 1 and Figure 2 This embodiment provides the following technical solution:

[0124] In S4, target cooperation relationship data is used to construct networks for professional papers and patent documents separately. After the networks for professional papers and patent documents are constructed, they are merged, including:

[0125] Based on the target cooperation relationship data, nodes are generated for professional papers and patent documents respectively;

[0126] This involves generating paper nodes from professional papers.

[0127] After the paper node is generated, patent sub-nodes are established based on the cooperation relationship data with patent documents in the target cooperation relationship data. After the patent sub-nodes are established, the cooperation relationship edges between the paper node and several patent sub-nodes are obtained.

[0128] By extending the cooperative relationships between paper nodes and several patent child nodes to edges, and using the Pajek tool in graph theory, a cooperative network for professional papers is constructed.

[0129] Once the collaborative network for professional papers is constructed, a collaborative relationship network for professional papers is obtained.

[0130] Generate patent nodes from patent documents;

[0131] After the patent node is generated, paper sub-nodes are established based on the cooperation relationship data with professional papers in the target cooperation relationship data. After the paper sub-nodes are established, the cooperation relationship edges between the patent node and several paper sub-nodes are obtained.

[0132] By extending the cooperative relationships between patent nodes and several paper sub-nodes to edges, the Pajek tool in graph theory is used to construct a cooperative network of patent documents.

[0133] Once the collaborative network of patent documents is constructed, a collaborative relationship network of patent documents is obtained.

[0134] Overlap the patent document collaboration network and the professional paper collaboration network;

[0135] When patent document collaboration networks and professional paper collaboration networks overlap, overlapping nodes are generated when there are co-authors and public institutions.

[0136] If there is a direct collaboration between the patent document collaboration network and the professional paper collaboration network, add corresponding outgoing edges to the overlapping nodes to connect the nodes with the collaboration relationship.

[0137] If there is an indirect cooperative relationship between the patent document cooperative relationship network and the professional paper cooperative relationship network, the co-citation analysis method is used to add potential directed edges in the overlapping nodes;

[0138] After the network nodes of the patent literature collaboration network and the professional paper collaboration network overlap and are generated, the network is integrated to obtain the target multi-layer scientific research data.

[0139] Specifically, node generation and collaboration relationship construction for both academic papers and patent documents allow for a more accurate reflection of the collaboration within these two types of documents. This refined network construction method helps to deeply explore the collaboration characteristics of each, providing a solid foundation for subsequent network overlap and integration. During network overlap, overlapping nodes are generated by identifying co-authors and public institutions, accurately reflecting the direct connections between academic papers and patent documents. Simultaneously, the use of co-citation analysis to add potential directed edges in indirect collaboration relationships further enhances the effect of network overlap, making the integrated network more comprehensive and accurate. Through network overlap and integration, not only are the collaboration relationships within academic papers and patent documents preserved, but also the direct and indirect collaboration relationships between them are revealed. This multi-layered collaboration relationship revealing method helps to comprehensively understand the complexity and diversity of scientific research collaboration. Utilizing Pajek tools in graph theory for network construction and visualization makes the presentation of scientific research collaboration relationships more intuitive and clear.

[0140] In S5, a network analysis of the target multi-layered scientific research data is performed. This analysis yields data on the collaborative characteristics of professional papers and patent documents, including:

[0141] Perform topology analysis on the network nodes in the target multi-layer scientific research data;

[0142] Among them, topology analysis involves using node degree analysis to confirm the number of each network node and the number of directed edges in the target multi-layer scientific research data.

[0143] Then, clustering coefficient analysis was used to confirm the tight clusters between each network node in the target multi-layer scientific research data;

[0144] Finally, path length analysis was used to determine the average shortest path length between any two nodes in each network of the target multi-layer scientific research data.

[0145] Based on the number of each network node and its edges, the number of tight clusters between each network node, and the average shortest path length between any two nodes, we obtain data on the collaborative characteristics between academic papers and patent documents.

[0146] Specifically, a comprehensive and in-depth analysis of the topology of the scientific research collaboration network was conducted using node degree analysis, clustering coefficient analysis, and path length analysis. This comprehensive analysis reveals the complex relationships between the nodes (i.e., academic papers and patent documents) in the network, as well as the overall structural characteristics of the network. Node degree analysis reveals the activity level of each network node (i.e., academic papers or patent documents), that is, its influence and participation in the network. Clustering coefficient analysis reveals the tight clusters between nodes, that is, which nodes are more inclined to form close cooperative relationships. Path length analysis provides the "distance" information between any two nodes in the network, which helps to understand the efficiency of information dissemination and the diversity of cooperation paths, presenting cooperation characteristics data in a quantitative form, such as node degree, clustering coefficient, and average shortest path length. This quantitative data is not only intuitive and easy to understand, but also facilitates subsequent data processing and analysis. By comparing the cooperation characteristics data of different networks or different time periods, the changes and evolution patterns of scientific research collaboration trends can be revealed. By analyzing the cooperation characteristics data, researchers can understand their position and influence in the network, as well as the cooperation potential and obstacles with other nodes.

[0147] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0148] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention.

Claims

1. A method for constructing a multi-layered scientific research collaboration network based on professional papers and patent documents, characterized in that, Includes the following steps: S1: Scientific research collaboration data acquisition: Acquire professional papers and patent documents from different data acquisition centers, confirm the basic information of professional papers and patent documents, and mark professional papers and patent documents whose basic information has been confirmed as scientific research data to be processed. S2: Scientific research collaboration data processing: Preprocess the scientific research data to be processed, and obtain the collaboration data to be extracted after the data preprocessing is completed; S3: Confirmation of Cooperation Relationship: Confirm the key data of the cooperation data to be extracted, and extract the relationship based on the confirmed key data. After the relationship extraction is completed, the target cooperation relationship data is obtained. S4: Construction of Cooperation Network: The target cooperation data is used to construct networks for professional papers and patent documents respectively. After the networks for professional papers and patent documents are constructed, the networks are overlapped. After the networks are merged, the overlapping network of professional papers and patent documents is obtained and labeled as the target multi-layer scientific research data. S5: Network Construction and Relationship Analysis: Perform relationship network analysis on the target multi-layer scientific research data. After the relationship network analysis, obtain the cooperation characteristics data of professional papers and patent documents, and display the cooperation characteristics data on mobile terminals. In S4, target cooperation relationship data is used to construct networks for professional papers and patent documents separately. After the networks for professional papers and patent documents are constructed, they are merged, including: Based on the target cooperation relationship data, nodes are generated for professional papers and patent documents respectively; This involves generating paper nodes from professional papers. After the paper node is generated, patent sub-nodes are established based on the cooperation relationship data with patent documents in the target cooperation relationship data. After the patent sub-nodes are established, the cooperation relationship edges between the paper node and several patent sub-nodes are obtained. By extending the cooperative relationships between paper nodes and several patent child nodes to edges, and using the Pajek tool in graph theory, a cooperative network for professional papers is constructed. Once the collaborative network for professional papers is constructed, a collaborative relationship network for professional papers is obtained. Generate patent nodes from patent documents; After the patent node is generated, paper sub-nodes are established based on the cooperation relationship data with professional papers in the target cooperation relationship data. After the paper sub-nodes are established, the cooperation relationship edges between the patent node and several paper sub-nodes are obtained. By extending the cooperative relationships between patent nodes and several paper sub-nodes to edges, the Pajek tool in graph theory is used to construct a cooperative network of patent documents. Once the collaborative network of patent documents is constructed, a collaborative relationship network of patent documents is obtained; Overlap the patent document collaboration network and the professional paper collaboration network; When patent document collaboration networks and professional paper collaboration networks overlap, overlapping nodes are generated when there are co-authors and public institutions. If there is a direct collaboration between the patent document collaboration network and the professional paper collaboration network, add corresponding outgoing edges to the overlapping nodes to connect the nodes with the collaboration relationship. If there is an indirect cooperative relationship between the patent document cooperative relationship network and the professional paper cooperative relationship network, the co-citation analysis method is used to add potential directed edges in the overlapping nodes; After the network nodes of the patent literature collaboration network and the professional paper collaboration network overlap and are generated, the network is integrated to obtain the target multi-layer scientific research data.

2. The method for constructing a multi-layered scientific research collaboration network based on professional papers and patent documents according to claim 1, characterized in that, For S1, professional papers and patent documents were acquired from different data acquisition centers, and the basic information of the professional papers and patent documents was confirmed, including: Professional papers are retrieved from journal databases, and patent documents are retrieved from patent databases; Furthermore, the basic information of the acquired professional papers and patent documents was confirmed; among them, the basic information of the professional papers included the paper title, field of study, content of the paper, author's name, publishing institution, and publication date. Basic information in patent documents includes patent title, patent field, patent office, patent content, applicant and inventor names, application date and publication date; The acquired professional papers and patent documents, along with their basic information, are matched and then uniquely coded. The professional papers and patent documents with unique codes, along with their basic information, are labeled as research data to be processed.

3. The method for constructing a multi-layered scientific research collaboration network based on professional papers and patent documents according to claim 2, characterized in that, For S2, data preprocessing of the research data to be processed includes: The scientific research data to be processed is sequentially cleaned, integrated, and transformed. Data cleaning includes removing duplicate values, missing values, outliers, and noise from the research data to be processed. After data cleaning, data integration is performed, and the dataset is formed by integrating the cleaned scientific research data into a single dataset. After data integration, data conversion is performed to standardize the format of the integrated scientific research data to be processed. After the data conversion is completed, the preprocessed scientific research data to be processed is obtained and marked as collaborative data to be extracted.

4. The method for constructing a multi-layered scientific research collaboration network based on professional papers and patent documents according to claim 3, characterized in that, In S3, key data for the cooperation data to be extracted is confirmed, and relationships are extracted based on the confirmed key data, including: The key data in the collaborative data to be extracted was confirmed, including the basic information of professional papers and patent documents. Relationship extraction is performed after key data is confirmed; First, obtain the names of the authors and applicants or inventors of professional papers and patent documents from the basic information. The string matching method is used to match the names of authors and applicants or inventors in professional papers and patent documents; Once a name match is found, the first cooperation relationship data is obtained. When names are duplicated, the institution is matched based on the publishing institution of the professional papers and patent documents and the patent institution. If the patent institution and the publishing institution have a cooperative relationship, the institution is successfully matched and the second cooperative relationship data is obtained.

5. The method for constructing a multi-layered scientific research collaboration network based on professional papers and patent documents according to claim 4, characterized in that, In S3, key data confirmation for the cooperation data to be extracted, and relationship extraction based on the confirmed key data, also includes: Extract the content of professional papers and patent documents from basic information; Data mining techniques are used to correlate the content of academic papers and patents. If there is a correlation between academic papers and patent documents, the correlation mining will reveal the common authors, common fields, and common institutions of the academic papers and patent documents. The professional papers and patent documents obtained through correlation mining will be used as third-party collaboration data. Based on the third-party cooperation data, the citation relationships between the content of professional papers and patent documents are searched. When there are mutual citations in the references and cited documents in the content of academic papers and patents, the academic papers and patent documents are regarded as the fourth cooperation relationship data.

6. The method for constructing a multi-layered scientific research collaboration network based on professional papers and patent documents according to claim 5, characterized in that, In S3, key data confirmation for the cooperation data to be extracted, and relationship extraction based on the confirmed key data, also includes: Obtain the publication date and application date of professional papers and patent documents from the basic information; Cross-correlation analysis was used to analyze the cross-correlation coefficient between publication date and application date, and the correlation data and time delay data between professional papers and patent documents were obtained after the analysis. The correlation data is used to determine whether there is a relationship between professional papers and patent documents. When there is a relationship between professional papers and patent documents, the time sequence between professional papers and patent documents is determined based on the time delay data. Professional papers and patent documents completed in chronological order are used as the fifth collaborative relationship data. The data for the first, second, third, fourth, and fifth cooperative relationships are uniformly labeled as the target cooperative relationship data.

7. The method for constructing a multi-layered scientific research collaboration network based on professional papers and patent documents according to claim 6, characterized in that, In S5, a network analysis of the target multi-layered scientific research data is performed. This analysis yields data on the collaborative characteristics of professional papers and patent documents, including: Perform topology analysis on the network nodes in the target multi-layer scientific research data; Among them, topology analysis involves using node degree analysis to confirm the number of each network node and the number of directed edges in the target multi-layer scientific research data. Then, clustering coefficient analysis was used to confirm the tight clusters between each network node in the target multi-layer scientific research data; Finally, path length analysis was used to determine the average shortest path length between any two nodes in each network of the target multi-layer scientific research data. Based on the number of each network node and its edges, the number of tight clusters between each network node, and the average shortest path length between any two nodes, we obtain data on the collaborative characteristics between academic papers and patent documents.

Images