An author contribution quantification and distribution method and device based on an academic network

Abstract

A method and device for quantifying and distributing author contribution based on an academic network, the method comprising: S1: writing a network crawler program and collecting paper data, citation relationship data between papers, and author data of the papers; S2: constructing a citation network of a paper p0 for which author contribution score points need to be calculated, the nodes of the network being scientific papers and the edges being citation relationships between the papers; the network structure being divided into two layers, the first layer being a reference literature layer of the paper p0, and the second layer being citations of the reference literature layer, including the paper p0 and other citations; S3: calculating the contribution score of each author in the paper p0 in each paper in the citation layer, and constructing an author contribution matrix A of the paper p0; S4: calculating the number of times that the paper p0 and each paper in the citation layer including the paper p0 jointly cite the papers in the reference literature layer, and constructing a common citation intensity matrix S; and S5: multiplying the contribution matrix and the common citation intensity matrix to obtain the final contribution score C of the authors of the paper p0.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of science of science and the field of data mining, and in particular, to an author contribution quantification and distribution method and device based on an academic network. BACKGROUND

[0002] Science of science is a discipline that observes and studies scientific activities themselves and puts forward quantitative understanding of scientific structure and its development. It aims to have a deep understanding and analysis of factors that promote scientific success, so as to put forward more effective strategies to promote scientific progress. In the research process of science of science, how to quantify the scientific research process and its output results of each discipline is an especially important direction. The evaluation of scientific researchers is the premise of talent resource management and use, and is an important breakthrough to stimulate the innovation vitality of talents. This makes it increasingly important to find an index that can reasonably evaluate the scientific research work of authors.

[0003] With the development of contemporary society, a large amount of culture, knowledge and technology has far exceeded the range that an independent individual can master. With the continuous progress and development of society, more and more emerging disciplines are emerging, and different professional fields are showing a diversified trend. In this case, when a person needs to complete a scientific research, the knowledge or technology he needs may far exceed his own mastery, which makes the scientific researcher tend to consult or cooperate with field experts. Therefore, in the early twentieth century, scientific research turned from theoretical deduction based on independent work to large-scale experimental research, and the case of scientific researchers conducting scientific research as individuals gradually decreased, and more and more independent research was replaced by scientific team cooperation or large research institutions. The complexity of scientific teams leads to the diversity of contribution distribution methods in the team. The original reward mechanism for the academic contribution of self-sufficient independent scientific researchers is no longer reasonable in the current team-dominated scientific development system, so scientific researchers begin to seek more reasonable distribution methods.

[0004] In the initial contribution distribution process, most scientific researchers will treat all co-authors of a paper equally, give them equal scores, or use an average distribution method to distribute author contributions. However, in a paper, the specific contributions of each author are not the same, so the contribution scores given to different authors should be different. The difference in the author's contribution to academic achievements is usually reflected in the signature of the academic paper. In the field of science, there are several different methods for distributing contributions based on the author's signature position. Existing author contribution distribution methods based on academic networks, such as a method for calculating the author score of a multi-signature academic paper disclosed in publication number CN106095934A, propose a specific quantitative calculation method for the score of each type of author such as the "first author", "corresponding author", "co-first author", "co-corresponding author", second author, third author, etc. of the paper, which changes the previous method of calculating the paper score simply by the number of first authors or corresponding authors. However, this contribution allocation method only considers the specific contributions of authors in a single paper and assigns contribution scores according to the order in which the authors are signed in the paper. It does not consider the contributions of the paper authors in the entire academic network, and does not distinguish between the different contributions of elite authors and young authors in teamwork. This leads to certain errors in the contribution allocation process and unfair evaluation of the authors involved in the paper creation. Summary of the Invention

[0005] The present invention aims to overcome the above-mentioned shortcomings of the existing methods and provides a method and device for quantifying and allocating author contributions based on an academic network.

[0006] This method prioritizes the number of publications by authors involved in a paper to differentiate between elite and younger authors in their academic network contributions. It also combines the collective contributions of authors within the academic network with their specific contributions to individual papers to calculate their contribution scores. This improves the accuracy and efficiency of author evaluation.

[0007] Specifically, the technical solutions adopted by the present invention to achieve the above-mentioned object are as follows:

[0008] A method for quantifying and allocating author contributions based on academic networks, comprising the following steps:

[0009] S1: Collect paper data, citation relationship data between papers, and paper author data through web crawler methods.

[0010] S2: Construct a citation network for paper p0, for which author contribution scores need to be calculated. The nodes of the network are scientific papers, and the edges are citations between papers. The network structure is divided into two layers: the first layer is the reference layer of paper p0, and the second layer is the citations in the reference layer, which includes paper p0 and other citations.

[0011] S3: Calculate the contribution score of each author in the paper p0 in each paper in the citation layer, and build the paper p0 author contribution matrix A.

[0012] S4: Calculate the number of papers in the co-citation reference layer that are co-cited by the paper p0 and each paper in the citation layer including p0, and build the co-citation intensity matrix S.

[0013] S5: Multiply the contribution matrix and the co-citation intensity matrix to get the final contribution score C of the authors of the paper p0.

[0014] Further, the step S1 includes the following steps:

[0015] S1.1: Obtain the interface information of the required database website, and use the crawler technology to write the code. The program can determine the required parameters by input, such as the data content to be crawled, the data page number to be crawled, the number of data to be crawled, and the email address for receiving data, etc. And the program can show the progress of crawling data and the number of crawled data.

[0016] S1.2: Collect the paper p0 whose author contribution score needs to be calculated, and collect all the author information of the paper through the crawler program, including the author's name, the order of signature in the paper, and the number of papers in the entire career.

[0017] S1.3: Collect all the references of the paper p0.

[0018] S1.4: Collect the citations of all references in the reference layer, and obtain the author information of all papers of the above citations, including the author's name, the order of signature in the paper, and the number of papers in the entire career.

[0019] S1.5: Clean the existing data. First, delete the papers with incomplete article information and author information, second, delete the single-author articles in the data set, and set a reasonable time window according to the publication time of the paper p0 to prevent the publication time of the paper from being too early or too late, and the incomplete reference data and citation data queried.

[0020] Further, the step S2 includes the following steps:

[0021] S2.1: Collect the reference literature of the paper p0 in the data set as the first layer of the citation network, i.e. the reference layer.

[0022] S2.2: All papers that cite the reference literature are taken as the second layer of the citation network, i.e. the citation layer.

[0023] Further, the step S3 specifically operates as follows:

[0024] S3.1: Calculate the specific contribution of the author of paper p0 in the papers he participated in writing based on the author signature order and the total number of works in each paper in the citation layer. ij , and finally the contribution score matrix A is obtained. The contribution matrix calculation steps are divided into three steps, and its core formula is as follows:

[0025] Calculate the ratio of the total number of publications of author i to the total number of publications of all authors of paper j:

[0026]

[0027] Where n is the number of authors in a paper, w i is the number of papers published by author i in his career.

[0028] The harmonic counting formula is used to calculate the score obtained by author i in paper j according to the order of signature:

[0029]

[0030] PaperRatio ij With Harmonic ij Multiply them together to get the final contribution score A of author i in paper j ij :

[0031] A ij =PaperRatio ij ×Harmonic ij (3)

[0032] S3.2: Use the calculated contribution A ij Construct the contribution matrix A, where the rows are the authors of paper p0 and the columns are all the papers in the citation layer.

[0033] Furthermore, the step S4 specifically includes:

[0034] Calculate the number of times each paper in the citation layer, including p0, appears in the entire citation network, which is recorded as the co-citation intensity S j (0≤j≤m), m is the total number of articles included in the citation layer. j Construct the co-citation intensity matrix S, where the rows are all papers in the citation layer and the columns are paper p0.

[0035] Furthermore, the step S5 specifically includes:

[0036] S5.1: Contribute A ij With the co-citation strength S j Multiply them together to get the contribution score of author i in paper p0. The specific calculation formula is as follows:

[0037]

[0038] S5.2: Calculate the contribution score c i Normalization is performed to obtain the normalized contribution score C of each author in the paper p0 i .

[0039] The second aspect of the application relates to an academic network-based author contribution quantification and distribution method device, comprising a memory and one or more processors, wherein the memory stores executable code, and the one or more processors execute the executable code to implement the academic network-based author contribution quantification and distribution method of the application.

[0040] The third aspect of the application relates to a computer-readable storage medium having a program stored thereon, wherein the program is executed by a processor to implement the academic network-based author contribution quantification and distribution method of the application.

[0041] The technical concept of the application is an academic network-based author contribution quantification and distribution method. In step S1, a crawler program for obtaining required data on a data website is written, and data required for constructing a citation network is crawled and cleaned. In steps S2 and S3, the participation degree and contribution of an author participating in the creation of a paper p0 in the entire academic network are calculated by constructing a citation network of the references of the paper p0 and analyzing the co-citation relationship between the focus paper and the remaining citations of the references, thereby quantifying the contribution of the author in the scientific community. In steps S4 and S5, the specific contribution of the author in the paper is quantified by taking into account the author's order of authorship in a focus paper, and the contribution of the author in the entire scientific community is further highlighted by calculating the ratio of the number of papers published by the author in his career to the total number of papers published by all authors in the paper, and the contribution allocation error caused by the age gap between senior authors and young authors is reduced.

[0042] The application provides an academic network-based author contribution quantification and distribution method. By combining the contribution of an author in the academic network and the specific contribution of the author in a single paper, a more fair and reasonable contribution allocation of the author in a team collaboration is achieved. The main innovations are as follows:

[0043] (1) A contribution allocation index combining the contribution of an author in the entire academic network and in a single paper is constructed, and the specific contribution of the author in the paper creation process recognized by the team members is emphasized by considering the author's order of authorship in the focus paper.

[0044] (2) By calculating the ratio of the number of papers of a specific author in the paper to the sum of the number of papers of the remaining authors, a contribution matrix A is constructed. This approach further highlights the author's contribution to the entire scientific community and takes into account the fact that young authors may have certain disadvantages compared to senior authors in the year of publication of the target paper, leading to an unfair distribution of contributions in the quantification process based on the entire academic network, reducing the contribution allocation error caused by the age gap between senior and young authors.

[0045] The benefits of the present application are: 1) combining the specific contribution of the author in the focus paper with the contribution in the entire academic network, a more fair and reasonable contribution allocation of the authors in team cooperation is made. 2) Combining the number of papers in the entire career of the author, the contribution allocation error between senior and relatively young authors is reduced. 3) Compared with the existing contribution allocation method, it has higher accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0046] Figure 1 is a block diagram of the method of the present application.

[0047] Figure 2 is a citation network diagram of the present application.

[0048] Figure 3 is a schematic diagram of the device of the present application. DETAILED DESCRIPTION

[0049] The specific embodiments of the present application will be further described in detail below in conjunction with the drawings of the specification.

[0050] Example 1

[0051] Referring to Figure 1 A method for quantifying and allocating author contributions based on an academic network includes the following steps:

[0052] S1: The reference data, citation data and author data are crawled on the Openalex official website through network crawler, and the final data set is obtained by cleaning the crawled data.

[0053] S2: The citation network of the focus paper whose author's contribution needs to be allocated is constructed by the data in the final data set.

[0054] S3: Calculate the contribution of the focus paper author in all papers in the citation network, understand the specific contribution of the author in a single paper.

[0055] S4: Analyze the citation relationship of the focus paper in the citation network, calculate the co-citation strength. By quantifying the participation of the focus paper author in the entire academic network, the contribution of the author in the scientific community is evaluated.

[0056] S5: Obtain the final contribution score of the author in the focus paper by combining the specific contribution of the author in each paper in which the author participates in cooperation and the contribution of the author in the entire academic network.

[0057] S6: Test the method on the Nobel Prize-winning paper dataset, compare its accuracy with the method COCD proposed by Zeng et al., and verify the effectiveness and superiority of the method.

[0058] Further, the step S1 comprises the following steps:

[0059] S1.1: Obtain the paper p0 requiring author contribution allocation from the Nobel Prize-winning dataset, and use a crawler to crawl the article information and all author information of the paper p0 in the Openalex dataset on the official website of Openalex, including the publication year of the paper, the name of the author, the order of signature in the paper, and the number of papers in the entire career.

[0060] S1.2: Obtain all references of the paper p0 in the Openalex dataset.

[0061] S1.3: Obtain all references in the reference layer in the Openalex dataset, and obtain the author information of all papers of the above references, including the name of the author, the order of signature in the paper, and the number of papers in the entire career.

[0062] S1.4: Clean the existing data. First, delete the papers with incomplete article information and author information, second, delete the papers with one author in the paper p0 dataset, set a time window according to the publication time of the focus paper, retain the Nobel Prize-winning papers published from 1960 to 2000, and delete the data beyond the time window to ensure the integrity of the reference data. Finally, set the publication time window of the papers in the citation layer to 1960-2010 to ensure the integrity of the citation data.

[0063] Further, with reference to Figure 2 , the step S2 specifically comprises:

[0064] S2.1: The focus paper p0 has n authors {a i}(0≤i≤n-1), all references cited by the paper p0 are defined as a set R≡{r1, r2,..., r k}, which is the first layer of the citation network, i.e. the reference layer.

[0065] S2.2: Define all citations of the papers in the reference set R as a set P≡{p j}(0≤j≤m-1), m is the total number of articles contained in the citation layer. This set is taken as the second layer of the citation network, i.e., the citation layer.

[0066] Further, the step S3 comprises the following steps:

[0067] S3.1: According to the author's name order in each paper in the set P, the total number of works, the specific contribution a of the author of the paper p0 in the paper he participated in is calculated ij , specifically, A ij represents the contribution score of author a i in paper p j . Finally, the contribution score matrix A is obtained. The steps of contribution matrix calculation are divided into three steps, and the core formula is as follows:

[0068] Calculate the proportion of the total number of papers of author i to the sum of the total number of papers of all authors in paper j:

[0069]

[0070] Where n is the number of authors in a paper, w i is the number of works published by author i in his career.

[0071] Harmonic counting formula is used to calculate the score obtained by author a i in paper p j according to the order of signature:

[0072]

[0073] Multiply PaperRatio ij and Harmonic ij to get the final contribution score A ij of author i in paper j:

[0074] A ij = PaperRatio ij ×Harmonic ij (3)

[0075] S3.2: Use the calculated contribution A ij to construct the contribution matrix A, the rows of the matrix are the authors of the paper p0, and the columns are all the papers in the citation layer.

[0076] Further, the step S4 specifically comprises:

[0077] Calculate the number of times each paper in the citation layer including p0 appears in the entire citation network, denoted as co-citation strength S j . Get S jA co-citation strength matrix S is constructed, the matrix acts on all papers in the citation layer, and the columns are all papers p0.

[0078] Further, the step S5 specifically comprises:

[0079] S5.1: multiply the contribution A ij with the co-citation strength S j to obtain the contribution score of the author i in the paper p0. The specific calculation formula is as follows:

[0080]

[0081] S5.2: calculate C i using the contribution score matrix C, the specific calculation formula is as follows:

[0082] C = A x S (5)

[0083] S5.3: normalize the content in the contribution score matrix C to obtain the normalized contribution score C i of each author in the paper p0.

[0084] The method provided by the application quantifies the specific contribution of the author in the paper by taking into account the author's signing order in the paper, quantifies the contribution of the author in the focus paper in the academic network based on the analysis of the citation network of the paper, and finally highlights the contribution of the author in the entire scientific community by calculating the ratio of the number of papers published by the author in the career to the total number of papers published by all authors in the paper, and reduces the contribution allocation error caused by the age gap between senior authors and young authors. Through the fair and just contribution allocation of the scientific research laborer in the team cooperation, the scientific research laborer can obtain fair treatment and be given a contribution share matching the responsibility he bears, which is conducive to improving the creativity and scientific research motivation of the scientific research laborer, and finally can promote the long-term healthy development of the scientific system.

[0085] Embodiment 2

[0086] The embodiment relates to an author contribution quantification and allocation method based on an academic network, which comprises a memory and one or more processors, the memory stores executable codes, and the one or more processors execute the executable codes to realize the author contribution quantification and allocation method based on the academic network in the embodiment 1.

[0087] Embodiment 3

[0088] The embodiment relates to a computer readable storage medium, which stores a program, and the program is executed by a processor to realize the author contribution quantification and allocation method based on the academic network in the embodiment 1.

[0089] The embodiments of the present specification are merely illustrative of the implementation forms of the inventive concept, and the protection scope of the present application should not be regarded as being limited to the specific forms presented in the embodiments, and the protection scope of the present application also encompasses equivalent technical means that can be thought of by those skilled in the art according to the inventive concept.

Claims

1. An author contribution quantification and distribution method based on academic networks, characterized in that: The method comprises the following steps: S1: confirming a paper p0 requiring author contribution allocation calculation, obtaining corresponding reference data, citation data and author data through a network crawler, and cleaning the obtained data to obtain a final data set; S2: constructing a citation network of the focus paper requiring contribution allocation based on the data in the final data set; S3: calculating the contribution of the author of the paper in all papers in which the author participates in cooperation in the citation network, and understanding the specific contribution of the author in a single paper; Specifically comprising: S3.1: Calculate the specific contribution a of the focus paper author in the papers he participated in according to the author's signature order in each paper in the set P, the total number of works ij , Specifically, A ij represents the contribution score of author ai in paper p j ; the final contribution score matrix A is obtained; the steps of contribution matrix calculation are divided into three steps, and the core formula is as follows: calculating the proportion of the total number of papers published by the author i to the sum of the total number of papers published by all authors in the paper j: where n is the number of authors in a paper, w i is the number of works published by author i in his / her career; Harmonic count formula for calculating the score that author ai gets in paper p j in the order of the authorship: PaperRatio ij is multiplied by Harmonic ij to obtain the final contribution score A of author i in paper j ij : A ij = PaperRatio ij ×Harmonic ij (3) S3.2: Compute the contribution A ij The contribution matrix A is constructed with the rows focusing on the authors of the papers and the columns on all papers in the citation layer. S4: analyzing the citation relationship of the paper in the citation network, calculating the co-citation strength, quantifying the participation degree of the author in the entire academic network, and evaluating the contribution of the author in the scientific community; S5: combining the specific contribution of the author in each paper in the citation network and the contribution of the author in the entire academic network to obtain the final contribution score of the author in the paper.

2. The method of quantifying and distributing authorship based on academic networks of claim 1, wherein: The step S1 specifically comprises: S1.1: obtaining the interface information of the required database website, and writing codes by using the crawler technology; the program can determine the required parameters by input, including: the data content to be crawled, the data page number to be crawled, the data number to be crawled, and the email address for logging in the account; and the program can display the progress of the crawled data and the number of the crawled data; S1.2: collecting the paper p0 requiring calculation of the author contribution score, and obtaining the article information and all author information of the paper, including the name of the author, the signature order of the author in the paper and the number of papers published in the entire career of the author; S1.3: collecting all references of the paper p0, then collecting all citations of the references of the paper p0, and obtaining all paper information of the citations, including the publication year of the paper, the name of the author of the paper, the signature order of the author in the paper and the number of papers published in the entire career of the author; S1.4: cleaning the existing data; first, deleting the papers with incomplete article information and author information, second, deleting the single-author papers in the data set, and setting a reasonable time window according to the publication time of the paper p0 to prevent the reference data and citation data queried from being incomplete due to too early or too late publication time of the paper.

3. The method of quantifying and distributing authorship based on academic networks of claim 1, wherein: The step S2 specifically comprises: S2.1: The paper p0 has n authors {a i}(0≤i≤n-1), define all the references cited by the paper p0 as a set R≡{r1, r2,..., r k} as the first layer of the citation network, i.e. the reference layer; S2.2: Define all citations of papers in the citation set R as a set P≡{p j}(0≤j≤m-1), m is the total number of articles contained in the citation layer; this set is defined as the second layer of the citation network, i.e. the citation layer.

4. The method of claim 1, wherein: The step S4 specifically comprises: The number of times each paper in the citation layer including p0 appears in the entire citation network is counted, denoted as co-citation strength S j S is obtained j A co-citation strength matrix S is constructed, with rows representing all papers in the citation layer and columns representing the focus paper p0.

5. The method of quantifying and distributing authorship based on academic networks of claim 1, wherein: The step S5 specifically comprises: S5.1: the contribution A ij is multiplied by the co-citation strength S j to obtain the contribution score of author i in paper p0; the specific calculation formula is as follows: S5.2: Compute C i The final contribution score matrix C for all authors in the paper p0 is constructed with the following formula: C=A× S (5) S5.3: Normalize the content in the contribution score matrix C to obtain the normalized contribution score C of each author in the paper p0 i .

6. An author contribution quantification and allocation device based on an academic network, characterized in that: The memory stores executable code, and the one or more processors execute the executable code to implement the author contribution quantification and allocation method based on the academic network in any one of claims 1-5.

7. A computer readable storage medium characterized in that, The program is stored on the memory and is executed by the processor to implement the author contribution quantification and allocation method based on the academic network in any one of claims 1-5.

Images