Scientific and technological innovation research theme auxiliary decision-making method based on relation network matrix

Abstract

The invention relates to a scientific and technological innovation research theme auxiliary decision-making method based on a relation network matrix, and the method comprises the steps: obtaining a scientific and technological literature data set, carrying out the preprocessing, and constructing a keyword library and a technical keyword library; through statistical analysis of keywords, a keyword relation matrix and a technical keyword relation matrix are obtained, user interested keywords are obtained, a user interested keyword co-occurrence network and a technical keyword co-occurrence network are constructed, and the keyword co-occurrence network and the technical keyword co-occurrence network are displayed to a user through a visualization technology. And the user is helped to make scientific and technological innovation research theme auxiliary decisions. According to the method, the co-occurrence network between the keywords is constructed through the co-occurrence frequency of the keywords, the interrelation and the knowledge structure between the scientific and technical literatures are revealed, and researchers are helped to understand the complex relation between the scientific and technical literatures, so that the structure and the development trend of a certain research field are better understood; and researchers are helped to make scientific and technological innovation research theme auxiliary decision-making.

Description

Technical Field

[0001] The present invention belongs to the field of computer technology, and in particular relates to a method for assisting decision-making of scientific and technological innovation research topics based on a relationship network matrix. Background Art

[0002] In the current market, technologies that assist in the selection of research topics are developing rapidly and have become an important support tool for researchers' daily research. These technologies mainly use advanced algorithms such as natural language processing (NLP), machine learning (ML), and data mining to help researchers identify research hotspots, trends, and potential research directions from a large amount of scientific literature and data. For example, natural language processing technology can analyze the content of academic articles, extract key terms and topics, and help researchers quickly understand the research dynamics in a certain field. Through topic modeling, researchers can identify areas that have not been fully explored, so as to better select research topics. In addition, machine learning algorithms can predict possible future research trends by analyzing historical research data, helping researchers make more informed topic selection decisions.

[0003] Many scientific research management platforms (such as ResearchGate, Mendeley, and Academia.edu) have integrated the function of assisting topic selection and providing literature recommendation and analysis services. These platforms provide personalized recommendations by tracking users' interests and reading habits, allowing researchers to quickly find relevant literature and topics. In addition, new AI-driven tools (such as Semantic Scholar, CiteSeerX, and Connected Papers) can provide customized topic analysis and literature recommendations based on researcher input or existing literature, further improving the efficiency and accuracy of topic selection. The emergence of these tools helps researchers organize literature and screen topics more efficiently, reducing the time for literature research.

[0004] Although the technology of assisting research topic selection has made significant progress, there are still some shortcomings in practical applications. For example, the current technology can only simply give recommended topics, rather than helping users to make their own choices, and cannot better help users understand the structure and development dynamics of a research field. Summary of the invention

[0005] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and to provide a method for assisting decision-making of scientific and technological innovation research topics based on a relationship network matrix.

[0006] The purpose of the present invention can be achieved by the following technical solutions:

[0007] On the one hand, the present invention provides a method for assisting decision-making of scientific and technological innovation research topics based on a relationship network matrix, comprising the following steps:

[0008] Obtain scientific literature datasets;

[0009] Preprocess the scientific literature dataset;

[0010] Use the TextRank algorithm to extract keywords from the abstracts of papers and articles, and together with the keywords of the papers and articles, form a keyword table for scientific literature. Use expert experience to identify the corresponding technical keywords of each paper and article from the keyword table.

[0011] Merge all keyword tables and remove duplicates to obtain a keyword library;

[0012] Merge all technical keywords and remove duplicates to obtain a technical keyword library;

[0013] Traverse each technical keyword in the technical keyword library, count the number of times each technical keyword appears in each article, and obtain a keyword relationship matrix according to the number of times each technical keyword in the technical keyword library appears in each article. The relationship matrix is ​​a two-dimensional matrix, and the numbers in the matrix are the number of times a certain keyword appears in a certain paper;

[0014] Traverse each keyword in the keyword library, count the number of times each keyword appears in each article, and obtain the technical keyword relationship matrix based on the number of times each technical keyword in the technical keyword library appears in each article;

[0015] Get the keywords of interest entered by the user;

[0016] Construct a keyword co-occurrence network based on the keywords of interest and the keyword relationship matrix;

[0017] Construct a technical keyword co-occurrence network based on the relationship matrix between the keywords of interest and the technical keywords;

[0018] The keyword co-occurrence network and the technology keyword co-occurrence network are displayed to users through visualization technology, assisting users in selecting research topics for scientific and technological innovation.

[0019] Furthermore, the preprocessing includes: cleaning data, removing duplicates, and correcting spelling errors.

[0020] Furthermore, the method of extracting keywords from the abstract of the paper article using the TextRank algorithm specifically includes:

[0021] Preprocess the article abstracts, including removing stop words, punctuation marks, and performing word segmentation;

[0022] Build a vocabulary, count the frequency of each word, and assign a unique ID to each word;

[0023] Build a word similarity graph;

[0024] Use the TextRank algorithm to iteratively calculate on this graph to determine the importance score of each word;

[0025] Sort the keywords according to the calculated scores, and select the top N words with the highest scores as keywords.

[0026] Furthermore, the nodes of the similarity graph are words, and the edges represent the similarities between words.

[0027] Furthermore, the similarity is calculated by the co-occurrence relationship of words, and the specific process includes:

[0028] Define a window size, slide word by word in the text, record the number of times each pair of words co-occurs in the window, and obtain a co-occurrence matrix. The rows and columns of the matrix correspond to the words in the text, and each element represents the co-occurrence frequency of the corresponding word pair. Use the co-occurrence matrix to calculate the Jaccard similarity between words and obtain the similarity matrix between words.

[0029] Furthermore, the importance score of each word is determined, and the specific process includes: initializing the score of each word, updating the score of the current word according to the scores of adjacent words and the connection weights, and after multiple iterations, the score tends to be stable and finally converges.

[0030] Furthermore, the keyword relationship matrix is:

[0031]

[0032] Among them, n is the number of keywords in the keyword library, m is the number of articles in scientific literature; a mn Indicates the number of times keyword n ​​appears in article m;

[0033] The technical keyword relationship matrix is:

[0034]

[0035] Among them, l is the number of technical keywords in the technical keyword library, b ml Indicates the number of times the technical keyword l appears in article m.

[0036] Furthermore, the construction of a keyword co-occurrence network based on the keywords of interest and the keyword relationship matrix specifically includes the following steps:

[0037] According to the keyword of interest, traverse the keyword library and find the number k corresponding to the keyword of interest;

[0038] Traverse the keyword relationship matrix, count all the occurrences of the keyword of interest, and calculate the weight value of the keyword of interest in each row according to the number of occurrences of the keyword of interest in this column. The calculation formula is:

[0039]

[0040] Among them, w i is the weight value of the keyword of interest in the i-th row;

[0041] According to the weight value, calculate the co-occurrence value of the remaining keywords and the keyword of interest. The calculation formula is:

[0042]

[0043] Among them, 1≤j≤n,j≠k,g jk represents the co-occurrence value of keyword j and the keyword of interest;

[0044] Sort the calculated keyword co-occurrence values, and select the keywords with the largest co-occurrence values ​​before the first preset value;

[0045] Based on the selected keywords and their co-occurrence values ​​with the keywords of interest, a keyword co-occurrence network of the keywords of interest is constructed.

[0046] Furthermore, in the keyword co-occurrence network, the network nodes are keywords, and the network center node is the keyword input by the user. The thickness of the edges of the network is different, which is determined by the co-occurrence value between the nodes. The larger the co-occurrence value, the thicker the edge. The node sizes of the network are different, which is determined by the co-occurrence value between the nodes. The larger the co-occurrence value, the larger the two nodes.

[0047] On the other hand, the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the method for assisting decision-making of scientific and technological innovation research topics based on the relationship network matrix as described in any one of the above items.

[0048] Compared with the prior art, the present invention has the following advantages:

[0049] (1) The present invention constructs a co-occurrence network between keywords through the co-occurrence times of keywords, reveals the mutual connection and knowledge structure between scientific and technological documents, helps researchers understand the complex relationship between scientific and technological documents, and thus better understands the structure and development dynamics of a certain research field, helping researchers to make auxiliary decisions on scientific and technological innovation research topics.

[0050] (2) The present invention constructs a keyword co-occurrence network and a technology keyword co-occurrence network. The keyword co-occurrence network displays the co-occurrence relationship between keywords in the literature, revealing hot topics, research trends and the correlation between topics in the research field; the technology co-occurrence network displays the co-occurrence of different technologies or tools in the literature, identifying key technologies in the field, technology development paths and interactions between technologies. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0052] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0053] Embodiment 1:

[0054] This embodiment provides a method for assisting decision-making of scientific and technological innovation research topics based on a relationship network matrix, including the following steps:

[0055] Use web crawler technology to obtain scientific literature datasets from major scientific research paper websites;

[0056] Preprocessing of scientific literature datasets, including: cleaning data, removing duplicates, and correcting spelling errors;

[0057] The TextRank algorithm is used to extract keywords from the abstract of the paper, and together with the keywords of the paper, it forms the keyword table of its scientific literature, including the following steps:

[0058] Preprocess the article abstracts, including removing stop words, punctuation marks, and performing word segmentation; construct a vocabulary, count the frequency of each word, and assign a unique ID to each word; construct a word similarity graph; use the TextRank algorithm to iteratively calculate on this graph to determine the importance score of each word; sort the keywords according to the calculated scores, and select the top N words with the highest scores as keywords. The nodes of the similarity graph are words, and the edges represent the similarity between words. The similarity is calculated through the co-occurrence relationship of words. The specific process includes: defining a window size, sliding word by word in the text, recording the number of times each pair of words co-occurs in the window, and obtaining a co-occurrence matrix. The rows and columns of the matrix correspond to the words in the text, and each element represents the co-occurrence frequency of the corresponding word pair. The co-occurrence matrix is ​​used to calculate the Jaccard similarity between words to obtain a similarity matrix between words. Determine the importance score of each word. The specific process includes: initialize the score of each word, update the score of the current word according to the score of the adjacent word and the connection weight, and after multiple iterations, the score tends to be stable and finally converges.

[0059] Use expert experience to identify the corresponding technical keywords of each paper article from the keyword table;

[0060] Merge all keyword tables and remove duplicates to obtain a keyword library;

[0061] Merge all technical keywords and remove duplicates to obtain a technical keyword library;

[0062] Traverse each technical keyword in the technical keyword library, count the number of times each technical keyword appears in each article, and obtain a keyword relationship matrix according to the number of times each technical keyword in the technical keyword library appears in each article. The relationship matrix is ​​a two-dimensional matrix, and the numbers in the matrix are the number of times a certain keyword appears in a certain paper;

[0063] Traverse each keyword in the keyword library, count the number of times each keyword appears in each article, and obtain the technical keyword relationship matrix based on the number of times each technical keyword in the technical keyword library appears in each article; the keyword relationship matrix is:

[0064]

[0065] Among them, n is the number of keywords in the keyword library, m is the number of articles in scientific literature; a mn Indicates the number of times keyword n ​​appears in article m;

[0066] The technical keyword relationship matrix is:

[0067]

[0068] Among them, l is the number of technical keywords in the technical keyword library, b ml Indicates the number of times the technical keyword l appears in article m.

[0069] Get the keywords of interest entered by the user;

[0070] Construct a keyword co-occurrence network based on the keywords of interest and the keyword relationship matrix;

[0071] Construct a technical keyword co-occurrence network based on the relationship matrix between the keywords of interest and the technical keywords;

[0072] Constructing a keyword co-occurrence network based on the keywords of interest and the keyword relationship matrix includes the following steps:

[0073] According to the keyword of interest, traverse the keyword library and find the number k corresponding to the keyword of interest;

[0074] Traverse the keyword relationship matrix, count all the occurrences of the keyword of interest, and calculate the weight value of the keyword of interest in each row according to the number of occurrences of the keyword of interest in this column. The calculation formula is:

[0075]

[0076] Among them, w i is the weight value of the keyword of interest in the i-th row;

[0077] According to the weight value, calculate the co-occurrence value of the remaining keywords and the keyword of interest. The calculation formula is:

[0078]

[0079] Among them, 1≤j≤n,j≠k,g jk represents the co-occurrence value of keyword j and the keyword of interest;

[0080] Sort the calculated keyword co-occurrence values, and select the keywords with the largest co-occurrence values ​​before the first preset value;

[0081] Based on the selected keywords and their co-occurrence values ​​with the keywords of interest, a keyword co-occurrence network of the keywords of interest is constructed.

[0082] The method for constructing a technical keyword co-occurrence network refers to the above keyword co-occurrence network construction method.

[0083] The constructed keyword co-occurrence network has network nodes as keywords and the network center node as the keyword input by the user. The network edges have different thicknesses, which are determined by the co-occurrence values ​​between nodes. The larger the co-occurrence value, the thicker the edge. The network nodes have different sizes, which are determined by the co-occurrence values ​​between nodes. The larger the co-occurrence value, the larger the two nodes.

[0084] The keyword co-occurrence network and the technology keyword co-occurrence network are displayed to users through visualization technology, assisting users in selecting research topics for scientific and technological innovation.

[0085] The keyword co-occurrence network shows the co-occurrence relationship between keywords in the literature, which is used to reveal hot topics, research trends and correlations between topics in the research field; the technology co-occurrence network shows the co-occurrence of different technologies or tools in the literature, which is used to identify key technologies in the field, technology development paths and interactions between technologies.

[0086] Embodiment 2:

[0087] The parts not mentioned in this embodiment are the same as those in Embodiment 1.

[0088] This embodiment also provides an electronic device, including: a processor and a memory, wherein the processor is used to execute a processing program of a table in a version document stored in the memory to implement the above-mentioned method for auxiliary decision-making of scientific and technological innovation research topics based on a relationship network matrix.

[0089] This embodiment also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned method for assisting decision-making on scientific and technological innovation research topics based on the relationship network matrix is ​​implemented.

[0090] If the above functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program code.

[0091] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for assisting decision-making of scientific and technological innovation research topics based on a relationship network matrix, characterized in that: The following steps are involved: Obtain scientific literature datasets; Preprocess the scientific literature dataset; Use the TextRank algorithm to extract keywords from the abstracts of papers and articles, and together with the keywords of the papers and articles, form a keyword table for scientific literature. Use expert experience to identify the corresponding technical keywords of each paper and article from the keyword table. Merge all keyword tables and remove duplicates to obtain a keyword library; Merge all technical keywords and remove duplicates to obtain a technical keyword library; Traverse each technical keyword in the technical keyword library, count the number of times each technical keyword appears in each article, and obtain a keyword relationship matrix according to the number of times each technical keyword in the technical keyword library appears in each article. The relationship matrix is ​​a two-dimensional matrix, and the numbers in the matrix are the number of times a certain keyword appears in a certain paper; Traverse each keyword in the keyword library, count the number of times each keyword appears in each article, and obtain the technical keyword relationship matrix based on the number of times each technical keyword in the technical keyword library appears in each article; Get the keywords of interest entered by the user; Construct a keyword co-occurrence network based on the keywords of interest and the keyword relationship matrix; Construct a technical keyword co-occurrence network based on the relationship matrix between the keywords of interest and the technical keywords; The keyword co-occurrence network and the technology keyword co-occurrence network are displayed to users through visualization technology, assisting users in selecting research topics for scientific and technological innovation.

2. According to claim 1, a method for assisting decision-making of scientific and technological innovation research topics based on a relationship network matrix is ​​characterized in that: The preprocessing includes: cleaning data, removing duplicates, and correcting spelling errors.

3. The method for assisting decision-making of scientific and technological innovation research topics based on a relationship network matrix according to claim 1 is characterized in that: The method of extracting keywords from the abstract of the paper article using the TextRank algorithm specifically includes: Preprocess the article abstracts, including removing stop words, punctuation marks, and performing word segmentation; Build a vocabulary, count the frequency of each word, and assign a unique ID to each word; Construct a word similarity graph; Use the TextRank algorithm to iteratively calculate on this graph to determine the importance score of each word; Sort the keywords according to the calculated scores, and select the top N words with the highest scores as keywords.

4. The method for assisting decision-making of scientific and technological innovation research topics based on a relationship network matrix according to claim 3 is characterized in that: The nodes of the similarity graph are words, and the edges represent the similarities between words.

5. According to claim 4, a method for assisting decision-making of scientific and technological innovation research topics based on a relationship network matrix is ​​characterized in that: The similarity is calculated by the co-occurrence relationship of words, and the specific process includes: Define a window size, slide word by word in the text, record the number of times each pair of words co-occurs in the window, and obtain a co-occurrence matrix. The rows and columns of the matrix correspond to the words in the text, and each element represents the co-occurrence frequency of the corresponding word pair. Use the co-occurrence matrix to calculate the Jaccard similarity between words and obtain the similarity matrix between words.

6. The method for assisting decision-making of scientific and technological innovation research topics based on a relationship network matrix according to claim 3 is characterized in that: The specific process of determining the importance score of each word includes: initializing the score of each word, updating the score of the current word according to the scores of adjacent words and the connection weights, and after multiple iterations, the score tends to be stable and finally converges.

7. The method for assisting decision-making of scientific and technological innovation research topics based on a relationship network matrix according to claim 1 is characterized in that: The keyword relationship matrix is: Among them, n is the number of keywords in the keyword library, m is the number of articles in scientific literature; a mn Indicates the number of times keyword n ​​appears in article m; The technical keyword relationship matrix is: Among them, l is the number of technical keywords in the technical keyword library, b ml Indicates the number of times the technical keyword l appears in article m.

8. The method for assisting decision-making of scientific and technological innovation research topics based on a relationship network matrix according to claim 1 is characterized in that: The step of constructing a keyword co-occurrence network based on the keywords of interest and the keyword relationship matrix specifically includes the following steps: According to the keyword of interest, traverse the keyword library and find the number k corresponding to the keyword of interest; Traverse the keyword relationship matrix, count all the occurrences of the keyword of interest, and calculate the weight value of the keyword of interest in each row according to the number of occurrences of the keyword of interest in this column. The calculation formula is: Among them, w i is the weight value of the keyword of interest in the i-th row; According to the weight value, calculate the co-occurrence value of the remaining keywords and the keyword of interest. The calculation formula is: Among them, 1≤j≤n,j≠k,g jk represents the co-occurrence value of keyword j and the keyword of interest; Sort the calculated keyword co-occurrence values, and select the keywords with the largest co-occurrence values ​​before the first preset value; Based on the selected keywords and their co-occurrence values ​​with the keywords of interest, a keyword co-occurrence network of the keywords of interest is constructed.

9. The method for assisting decision-making of scientific and technological innovation research topics based on a relationship network matrix according to claim 1, characterized in that: The keyword co-occurrence network, the network nodes are keywords, the network center node is the keyword input by the user, the edge thickness of the network is different, which is determined by the co-occurrence value between the nodes. The larger the co-occurrence value, the thicker the edge. The node sizes of the network are different, which are determined by the co-occurrence value between the nodes. The larger the co-occurrence value, the larger the two nodes.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements the method for assisting decision-making of scientific and technological innovation research topics based on a relationship network matrix as described in any one of claims 1 to 9.

Images