Method, device, electronic device and readable storage medium for determining authors with the same name

Abstract

The invention discloses a method, device, electronic device and readable storage medium for determining authors with the same name, and relates to the technical field of big data and data processing. In the method for determining authors with the same name, the invention applies big data technology to search out all papers with the name of a target author, classifies them according to the units that publish the papers, retrieves the basic words of the papers, and applies the word frequency-inverse document word frequency method to extract the core words of the papers; then, a word frequency vector of the core words of the papers is established, and the similarity between different papers is calculated, thereby establishing a similarity autocorrelation matrix of papers of the same unit and a similarity mutual correlation matrix of papers of different units; finally, through the similarity autocorrelation matrix and the similarity mutual correlation matrix, the autocorrelation matrix similarity index and the mutual correlation matrix similarity index of the paper similarity are calculated respectively, thereby determining whether the authors of papers published by different units with the name of the target author are the same person.

Description

Technical Field

[0001] The present invention relates to the field of big data and data processing technology, and in particular to a method, device, electronic device and readable storage medium for determining authors with the same name. Background Art

[0002] With the development of big data and artificial intelligence technology, the accuracy and effectiveness of information retrieval have received more and more attention. However, in actual information retrieval, how to obtain real and effective data as much as possible is a challenging task.

[0003] Especially in the subject evaluation and talent assessment, one of the important indicators is the publication of academic papers, which needs to be searched based on the author's name. However, the author's name, especially the name of the Chinese author, is often pinyin in foreign journals, which will result in a large number of duplicate names, making the search more difficult. Therefore, how to determine whether the author with the same name in papers published by different units is the same person is a very difficult but necessary task. Summary of the invention

[0004] The embodiments of the present invention provide a method, device, electronic device and readable storage medium for determining authors with the same name, so as to determine whether the authors with the same name of papers published by different units are the same person.

[0005] In order to solve the above problem, the first aspect of the embodiment of the present invention discloses a method for determining authors with the same name, the method comprising:

[0006] Taking the unit of published papers as the classification granularity, classify all papers with the searched target author's name, and establish a basic vocabulary of all the papers;

[0007] According to the basic word library, the core words of each paper in all the papers are extracted, all the core words of all the papers are statistically obtained, and according to all the core words of all the papers, normalized word frequency vectors of the core words of different papers in all the papers are established;

[0008] According to the normalized word frequency vectors of the core words of the different papers, the similarities of the different papers in all the papers are calculated, and a similarity autocorrelation matrix of the papers of the same unit and a similarity cross-correlation matrix of the papers of different units are established;

[0009] According to the similarity autocorrelation matrix and the similarity mutual correlation matrix, the autocorrelation matrix similarity index and the mutual correlation matrix similarity index of the paper similarity are calculated respectively, and it is determined whether the authors of the papers published by different units under the name of the target author are the same person.

[0010] Based on the same inventive concept, a second aspect of an embodiment of the present invention discloses a device for determining authors of the same name, the device comprising:

[0011] A basic word library building module is used to classify all papers with the searched target author's name based on the unit of published papers as the classification granularity, and to build a basic word library for all the papers;

[0012] A core word normalized word frequency vector acquisition module is used to extract the core words of each paper in all the papers according to the basic word library, obtain all the core words of all the papers by statistics, and establish the core word normalized word frequency vectors of different papers in all the papers according to all the core words of all the papers;

[0013] A similarity matrix building module is used to calculate the similarity of different papers in all the papers based on the normalized word frequency vectors of the core words of the different papers, and to build a similarity autocorrelation matrix of papers of the same unit and a similarity cross-correlation matrix of papers of different units;

[0014] The module for determining authors with the same name is used to calculate the autocorrelation matrix similarity index and the cross-correlation matrix similarity index of the paper similarity according to the similarity autocorrelation matrix and the similarity cross-correlation matrix, and to determine whether the authors of the papers published by different units under the name of the target author are the same person.

[0015] Based on the same inventive concept, the third aspect of an embodiment of the present invention discloses a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps in the method for determining authors with the same name as described in the first aspect of an embodiment of the present invention are implemented.

[0016] Based on the same inventive concept, the fourth aspect of an embodiment of the present invention discloses an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps in the method for determining authors with the same name as described in the first aspect of an embodiment of the present invention are implemented.

[0017] The embodiments of the present invention include the following advantages:

[0018] The embodiment of the present invention uses big data technology to search for all papers with the name of a target author, and classifies them according to the units that publish the papers, retrieves the basic words of the papers, and uses the word frequency-inverse document word frequency method in natural language processing technology to extract the core words of the papers; then, a word frequency vector of the core words of the papers is established, and the similarity between different papers is calculated, thereby establishing a similarity autocorrelation matrix of papers from the same unit and a similarity mutual correlation matrix of papers from different units; finally, the autocorrelation matrix similarity index and the mutual correlation matrix similarity index of the paper similarity are calculated respectively through the similarity autocorrelation matrix and the similarity mutual correlation matrix, thereby determining whether the authors of papers published by different units under the name of the target author are the same person. This method greatly reduces the difficulty of determining whether the authors with the same name of papers published by different units are the same person among a large number of duplicate names, and has achieved good results through actual testing. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is a flowchart of the steps of a method for determining authors with the same name proposed in one embodiment of the present invention;

[0020] Figure 2 It is a functional module diagram of a device for determining authors of the same name proposed in one embodiment of the present invention. DETAILED DESCRIPTION

[0021] 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 are within the scope of protection of the present invention.

[0022] In the subject evaluation and talent assessment, one of the important indicators is the publication of academic papers, which needs to be searched based on the author's name. However, the author's name, especially the name of the Chinese author, is often pinyin in foreign journals, and there will be a large number of duplicate names. For example: the author of the paper published by unit A in a foreign journal is named "Li Xiang", the author of the paper published by unit B in a foreign journal is named "Li Xiang", the author of the paper published by unit C in a foreign journal is named "Li Xiang", and the author of the paper published by unit D in a foreign journal is named "Li Xiang"... Among them, "Li Xiang" of unit C and "Li Xiang" of unit D are the same person, but because the pinyin of the authors of the papers published by the above-mentioned units A, B, C, and D are all "LiXang", which is expressed as "XangLi" in English papers, and this is only a part of the vast number of names that are all pinyin "LiXang", this leads to a large number of duplicate names in the search, which makes it a very difficult but necessary task to determine whether the authors with the same name in papers published by different units are the same person.

[0023] In view of this, the present invention proposes a method, device, electronic device and readable storage medium for determining authors with the same name, aiming to determine whether authors with the same name are the same person based on the core word similarity index of papers published by different units. Verification has proved that this method has achieved good results.

[0024] refer to Figure 1 , Figure 1 is a flowchart of the steps of a method for determining authors with the same name proposed in an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:

[0025] Step S101: using the unit of published papers as the classification granularity, classifying all papers with the searched target author's name, and establishing a basic vocabulary of all the papers;

[0026] In the present invention, all papers with the target author's name can be searched in a large amount of paper data by big data technology, wherein the target author's name is a given author's name, that is, the author's name to be searched, such as "LiXang".

[0027] In practice, a large number of authors with the same name who publish papers may belong to different units, such as an author with the same name "LiXang" who belongs to multiple different units such as A, B, C, and D. Therefore, the present invention can use the unit as a classification unit to classify all papers with the target author's name searched out by unit. For example, after classification, the author with the name "LiXang" in unit A has published X papers, the author with the name "LiXang" in unit B has published Y papers, and the author with the name "LiXang" in unit C has published Z papers, etc. Of course, there may be multiple authors with the same name with the target author's name in the same unit, such as the author with the name "LiXang" in unit A actually has "Li Xiang" and "Li Xiang", among which "Li Xiang" has published X1 papers, and "Li Xiang" has published X2 papers. It can be seen that all the papers with the same name as the target author's name searched out are still massive, so the present invention first uses the unit where the paper is published as the classification granularity to classify all the papers with the target author's name searched out.

[0028] While classifying all papers with the searched target author's name with units as classification granularity, the present invention also establishes a basic word library of all papers published with the target author's name.

[0029] In one embodiment of the present invention, the basic word library of all papers can be established by the method disclosed in sub-steps S101-1 to S101-2:

[0030] Sub-step S101-1: extracting basic words of each paper in all the papers by a text retrieval method;

[0031] Among them, the text retrieval method is an information retrieval method, which can retrieve, classify, filter, etc. a text collection based on the text content, such as keywords, semantics, etc.

[0032] Among them, basic words refer to entity words in the paper except function words, conjunctions, modal particles, etc. The entity words can specifically be some main words that indicate the core content of the paper, such as communication, signal, algorithm, etc.

[0033] Sub-step S101 - 2: Filter the stop words in the basic words of each paper, and then merge the basic words of all the papers to obtain the basic word library of all the papers.

[0034] Among them, stop words refer to certain characters or words that are automatically filtered out before or after processing natural language data (or text) in information retrieval to save storage space and improve search efficiency. In the present invention, stop words can specifically be some common words in papers, such as computer, program and other common words in the subject, as well as of, with, and and other common words in language.

[0035] After filtering the stop words in the basic words of each paper, this embodiment merges all the papers with the stop words filtered out to obtain the basic word library of all the papers. Compared with the method of directly filtering the stop words in the paper and then extracting the basic words, this embodiment adopts the method of first extracting the basic words and then filtering the stop words in the basic words. On the one hand, it can greatly reduce the filtering amount of stop words and reduce the data processing burden; on the other hand, it can also avoid the influence of filtering the stop words first on the judgment and extraction of the main words of the paper, and can have a higher probability of obtaining a large number of basic words that can accurately indicate the main content of the paper.

[0036] Step S102: extracting the core words of each paper in all the papers according to the basic word library, obtaining the core words of all the papers by statistics, and establishing the normalized word frequency vectors of the core words of different papers in all the papers according to all the core words of all the papers;

[0037] In one embodiment of the present invention, the word frequency-inverse document word frequency method can be used to extract the core words of each paper in all the papers, which can specifically include the following sub-steps:

[0038] Sub-step S102-1: Calculate the total number of times each basic word in the basic word library appears in all the papers, and divide the number of times a target basic word appears in all the papers by the number of times the basic word that appears most frequently in all the papers to obtain a normalized word frequency of the target basic word; wherein the target basic word is any basic word in the basic word library;

[0039] After obtaining the basic word library, this embodiment can first count the total number of times each basic word appears in all papers of the target author's name, and then perform normalization processing on each basic word in the basic word library, that is, the normalized word frequency of the target basic word = the number of times the target basic word appears in all papers / the number of times the basic word with the most appearances in all papers appears. By performing normalization processing on each basic word in the basic word library, this embodiment can obtain the normalized word frequency of all basic words in the basic word library.

[0040] For example, the target basic word is "communication", and the total number of times it appears in all papers of the target author's name is 200 times, and the basic word with the largest number of appearances in all papers of the target author's name is "analysis", which appears 4000 times. The two are divided, that is, 200 / 4000=0.05, indicating that the normalized word frequency of the basic word "communication" is 0.05.

[0041] Sub-step S102-2: Count the subjects involved in all the papers and the number of papers corresponding to each subject, and take the logarithm of the quotient of the number of papers corresponding to each subject and the number of papers containing the target basic word in all the papers, to obtain the inverse document frequency of the target basic word in the subject;

[0042] At present, the international publications include philosophy, economics, law, education, literature, history, science, engineering, agriculture, medicine, military science, management and art, and each major category has several first-level disciplines, such as science, which has 14 first-level disciplines such as mathematics, physics, and chemistry, and a first-level discipline has multiple second-level disciplines, such as "0701 Mathematics", which has "Basic Mathematics", "Computational Mathematics", "Probability Theory and Mathematical Statistics", "Applied Mathematics" and "Operations Research and Control Theory" and other second-level disciplines. Therefore, considering the complexity of the discipline, the discipline in this embodiment can refer to a first-level discipline and / or a second-level discipline. When the number of first-level disciplines involved in all the papers is lower than a preset value (the preset value can be set to 6), the second-level discipline can be used as the statistical granularity to count the number of papers corresponding to the second-level disciplines involved in all the papers. On the contrary, when the number of first-level disciplines involved in all the papers is greater than the preset value, the first-level discipline can be used as the statistical granularity to count the number of papers corresponding to the first-level disciplines involved in all the papers.

[0043] Next, taking all papers with the target author's name as the statistical scope, divide the number of papers corresponding to each discipline by the number of papers containing the target basic word in all papers, and then take the logarithm to obtain the inverse document frequency of the target basic word in each discipline. For example, the number of papers in the subject "applied mathematics" among all papers with the target author's name is 1, the target basic word is "communication", and the number of papers in all papers with the target author's name containing the basic word "communication" is 10,000. If the logarithm with base 10 is taken, the inverse document frequency of the target basic word in the subject "applied mathematics" can be calculated, which is in the range of 0-4.

[0044] By counting the inverse document word frequencies of the target basic word in all the disciplines involved in all the papers, a set of inverse document word frequencies of the target basic word can be obtained.

[0045] Sub-step S102-3: determining the normalized term frequency-inverse document term frequency of the target basic word in each subject according to the product of the normalized term frequency of the target basic word and the inverse document term frequency of the target basic word in each subject, so as to obtain a normalized term frequency-inverse document term frequency set of the target basic word, wherein the normalized term frequency-inverse document term frequency set of the target basic word includes a plurality of normalized term frequencies-inverse document term frequencies, and different normalized term frequencies-inverse document term frequencies correspond to different subjects;

[0046] Similarly, since the target basic word may appear in multiple disciplines involved in all the papers, by calculating the normalized word frequency of the target basic word and the inverse document word frequency of the target basic word in different disciplines one by one, the normalized word frequency-inverse document word frequency of the target basic word in multiple different disciplines can be obtained, that is, the normalized word frequency-inverse document word frequency set of the target basic word can be obtained.

[0047] Sub-step S102-4: Based on the normalized word frequency-inverse document word frequency set of each basic word in the basic word library, the basic words in each paper in all the papers whose normalized word frequency-inverse document word frequency is greater than a preset threshold are determined as the core words of the paper.

[0048] According to the normalized term frequency-inverse document term frequency set of the target basic word, in each paper containing the target basic word under each subject, the normalized term frequency-inverse document term frequency corresponding to the subject is compared with a preset threshold. When the normalized term frequency-inverse document term frequency corresponding to the subject is greater than the preset threshold, the target basic word is determined as the core word of the paper.

[0049] In the related art, the keywords of the paper abstract are usually used to indicate the content of the paper, but in reality, firstly, many English papers do not have keywords, and secondly, the number of keywords in the paper abstract is limited, which cannot reflect all the characteristics of the paper. And a single word frequency cannot represent the importance of a word, because there may be a situation where a certain word appears many times in a certain paper (the word frequency is very high), but it cannot be said that this word represents the characteristics of this paper. For example, common words often appear with a high word frequency. Considering this, the present invention adopts the word frequency-inverse document word frequency method to extract the core words of each paper in all the papers. When the number of appearances is large, the normalized word frequency and the inverse document word frequency are also large, this word is important to the target author's paper. After calculating the normalized word frequency and the inverse document word frequency, the present application can determine the importance of this word in this paper.

[0050] After extracting the core words of each paper in all the papers of the target author, the core words of all the papers can be counted, wherein the number of core words of each paper can be multiple. Optionally, the number of core words can be 100 to 200 depending on the number of all the papers.

[0051] In one embodiment of the present invention, the step of establishing normalized frequency vectors of core words of different papers in all the papers according to all the core words of all the papers includes:

[0052] Sub-step S102-5: Calculate the number of times each core word appears in each of the papers according to all the core words of all the papers, and divide the number of times a target core word appears in the current paper by the number of times the core word with the largest number of occurrences in the current paper to obtain the normalized word frequency of the target core word in the current paper, wherein the target core word is any one of the core words of all the papers;

[0053] After obtaining all the core words of all papers of the target author's name, this embodiment first calculates the number of times each core word appears in each paper, and then performs normalization processing on each core word. That is, the normalized word frequency of the target core word in this paper = the number of times the core word appears in this paper / the number of times the core word with the most occurrences in this paper appears.

[0054] Sub-step S102-6: Count the normalized word frequencies of all core words in each of the papers in all the papers to obtain normalized word frequency vectors of core words in different papers in all the papers.

[0055] The normalized frequency vector of the core words in the paper is the set of normalized frequencies of all the core words in the paper. Based on the method shown in substeps S102-5 to S102-6, by normalizing the different core words of each paper one by one, this embodiment can obtain the normalized frequency of all the core words in the paper, that is, the normalized frequency vector of the core words.

[0056] For example, if a paper has a total of w core words, then the normalized frequency vector of the core words of the paper = [normalized frequency of core word 1, normalized frequency of core word 2, ..., normalized frequency of core word w]. In this way, the normalized frequency vectors of core words in different papers are obtained, that is, the normalized frequency vector of the core word of paper k = [normalized frequency of core word 1 of paper k, normalized frequency of core word 2 of paper k, ..., normalized frequency of core word n ​​of paper k] = [k1, k2, ..., k w ].

[0057] Step S103: Calculate the similarity of different papers in all the papers based on the normalized word frequency vectors of the core words of the different papers, and establish a similarity autocorrelation matrix of papers of the same unit and a similarity cross-correlation matrix of papers of different units;

[0058] In one embodiment of the present invention, the similarity S of different papers in all papers can be calculated based on the normalized word frequency vectors of the core words of the different papers using the following formula: kl :

[0059]

[0060] In the above formula, k g , l g are the normalized word frequencies of the g-th core words in paper k and paper l among all the papers, g=1,2,…,w.

[0061] From formula (1), it can be seen that for any two papers with the target author’s name, since there are multiple core words in the papers, the calculated similarity will also have multiple values. Therefore, the similarity S of any two papers is kl It is actually a series matrix.

[0062] After obtaining the similarities of different papers in all the papers, the step S103 of establishing a similarity autocorrelation matrix of papers of the same unit may include:

[0063] Sub-step S103-1: Based on the similarities of different papers in all the papers, establish the similarity between the papers published by the target author in the same unit. If the target author has published m papers in the i-th unit, then the similarity autocorrelation matrix Z of the papers in the same unit is i for:

[0064]

[0065] After obtaining the similarities of different papers in all the papers, the step S103 of establishing a similarity inter-correlation matrix of papers from different units may include:

[0066] Sub-step S103-2: Based on the similarities of different papers of the target author name, establish the similarities between papers published by the target author name in different units. If the target author name has published m and n papers in the i-th unit and the j-th unit respectively, then the similarity cross-correlation matrix H of the papers in different units is ij for:

[0067]

[0068] Step S104: Calculate the autocorrelation matrix similarity index and the cross-correlation matrix similarity index of the paper similarity respectively according to the similarity autocorrelation matrix and the similarity cross-correlation matrix, and determine whether the authors of the papers published by different units under the name of the target author are the same person.

[0069] In an embodiment of the present invention, the implementation steps of step S104 may include:

[0070] Sub-step S104-1: Based on the similarity autocorrelation matrix Z i , get the similarity between two papers in the number m of papers published by the same unit, remove the similarity value of each paper in the number m of papers published by the same unit, accumulate and average the similarity between each paper and other papers, and calculate the autocorrelation matrix similarity index z of the similarity of papers published by the same unit s :

[0071]

[0072] In this embodiment, removing the similarity value of each paper itself means removing the similarity autocorrelation matrix Z i S 11 , S 22 ......S mm , and then Z i The remaining S 12 With S 21 , S 1m、Sm1、Sm2、S 2m By summing up the similarities and then averaging them, we can calculate the similarity index z of the autocorrelation matrix of the similarity of papers published by the same unit. s The present invention can normalize and average the similarities between any two papers in the number m of papers published in the same unit by removing the similarity value of each paper in the number m of papers published in the same unit, thereby ensuring the accuracy of the calculation of the autocorrelation matrix similarity index.

[0073] Sub-step S104-2: Based on the similarity mutual correlation matrix H ij , we get the similarity between papers m and n published by different units, accumulate and average the similarities, and calculate the similarity index h of the cross-correlation matrix of the similarities of papers published by different units s :

[0074]

[0075] In this embodiment, the similarity correlation matrix H of different unit papers is ij All similarities S' mn By summing up and averaging, we can calculate the similarity index h of the cross-correlation matrix of the similarity of papers published by the same unit. s .

[0076] Sub-step S104-3: When the mutual correlation matrix similarity index h s The similarity index z with the autocorrelation matrix s When the ratio of is greater than a preset threshold, it is determined that the author of the paper published by the i-th unit under the target author's name and the author of the paper published by the j-th unit under the target author's name are the same person.

[0077] Among them, the preset threshold can be set to 0.6.

[0078] In this embodiment, after obtaining the mutual correlation matrix similarity index h through the above steps, s The similarity index z with the autocorrelation matrix s Afterwards, the similarity index h of the cross-correlation matrix s The similarity index z with the autocorrelation matrix s The ratio of is calculated and compared with the preset threshold, so that the author with the same name in the papers published by the i-th unit and the j-th unit can be determined to be the same person. s / z s > 0.6, it indicates that the author of the paper published by the i-th unit under the target author's name and the author of the paper published by the j-th unit under the target author's name are the same person.

[0079] It should be noted that, if multiple units publish papers with the same author, it is necessary for the units to compare and judge each other using the method shown in step S103 to step S104.

[0080] Based on the same inventive concept, Figure 2 , Figure 2 is a functional module diagram of a device for determining authors of the same name according to an embodiment of the present invention. Figure 2 As shown, the device may include:

[0081] The basic word library building module 201 is used to classify all papers with the searched target author's name based on the unit of published papers as the classification granularity, and to build a basic word library for all the papers;

[0082] The core word normalized word frequency vector acquisition module 202 is used to extract the core words of each paper in the whole papers according to the basic word library, obtain all the core words of the whole papers by statistics, and establish the core word normalized word frequency vectors of different papers in the whole papers according to all the core words of the whole papers;

[0083] A similarity matrix building module 203 is used to calculate the similarity of different papers in all the papers based on the normalized word frequency vectors of the core words of the different papers, and to build a similarity autocorrelation matrix of papers of the same unit and a similarity cross-correlation matrix of papers of different units;

[0084] The module 204 for determining authors with the same name is used to calculate the autocorrelation matrix similarity index and the cross-correlation matrix similarity index of the paper similarity according to the similarity autocorrelation matrix and the similarity cross-correlation matrix, and to determine whether the authors of the papers published by different units under the name of the target author are the same person.

[0085] In one embodiment of the present invention, the basic word library building module 201 includes:

[0086] A basic word retrieval submodule, used for extracting basic words of each paper in all the papers by a text retrieval method;

[0087] The basic word merging submodule is used to filter the stop words in the basic words of each paper, and then merge the basic words of all the papers to obtain the basic word library of all the papers.

[0088] In one embodiment of the present invention, the core word normalized word frequency vector obtaining module 202 includes:

[0089] A basic word normalized frequency acquisition submodule is used to calculate the total number of times each basic word in the basic word library appears in all the papers, and divide the number of times a target basic word appears in all the papers by the number of times the basic word that appears most frequently in all the papers to obtain the normalized frequency of the target basic word; wherein the target basic word is any basic word in the basic word library;

[0090] The basic word inverse document frequency acquisition submodule is used to count the disciplines involved in all the papers and the number of papers corresponding to each discipline, and in all the papers, take the logarithm of the quotient of the number of papers corresponding to each discipline and the number of papers containing the target basic word in all the papers, to obtain the inverse document frequency of the target basic word in the discipline;

[0091] A word frequency-inverse document word frequency set acquisition submodule is used to determine the normalized word frequency-inverse document word frequency of the target basic word in each subject according to the product of the normalized word frequency of the target basic word and the inverse document word frequency of the target basic word in each subject, so as to obtain the normalized word frequency-inverse document word frequency set of the target basic word, wherein the normalized word frequency-inverse document word frequency set of the target basic word includes a plurality of normalized word frequencies-inverse document word frequencies, and different normalized word frequencies-inverse document word frequencies correspond to different subjects;

[0092] The core word determination submodule is used to determine the basic words whose normalized word frequency-inverse document word frequency in each paper in all the papers is greater than a preset threshold as the core words of the paper based on the normalized word frequency-inverse document word frequency set of each basic word in the basic word library.

[0093] In one embodiment of the present invention, the core word normalized word frequency vector obtaining module 202 further includes:

[0094] A core word normalized frequency acquisition submodule is used to calculate the number of times each core word appears in each of the papers according to all the core words in the papers, and divide the number of times a target core word appears in the paper by the number of times the core word with the largest number of occurrences in the paper to obtain the normalized frequency of the target core word in the paper, wherein the target core word is any one of the core words in all the papers;

[0095] The core word normalized frequency statistics submodule is used to count the normalized frequency of all core words in each of the papers in all the papers, and obtain the normalized frequency vectors of the core words in different papers in all the papers.

[0096] In one embodiment of the present invention, the similarity matrix establishing module 203 includes:

[0097] The paper similarity calculation submodule is used to calculate the similarity S of different papers in all the papers according to the normalized word frequency vectors of the core words of the different papers using the following formula: kl :

[0098]

[0099] In the above formula, k g , l g are the normalized word frequencies of the g-th core words in paper k and paper l among all the papers, g=1,2,…,w.

[0100] In one embodiment of the present invention, the similarity matrix establishing module 203 further includes:

[0101] The similarity autocorrelation matrix establishment submodule is used to establish the similarity between the papers published by the target author in the same unit according to the similarity of different papers in all the papers. If the target author has published m papers in the i-th unit, then the similarity autocorrelation matrix Z of the papers in the same unit is i for:

[0102]

[0103] The similarity mutual correlation matrix establishment submodule is used to establish the similarity between the papers published by the target author in different units according to the similarity of different papers of the target author. If the target author has published m and n papers in the i-th unit and the j-th unit respectively, the similarity mutual correlation matrix H of the papers in different units is ij for:

[0104]

[0105] In one embodiment of the present invention, the same-name author determination module 204 includes:

[0106] The autocorrelation matrix similarity index calculation submodule is used to calculate the similarity index of the autocorrelation matrix Z according to the similarity i , get the similarity between two papers in the number m of papers published by the same unit, remove the similarity value of each paper in the number m of papers published by the same unit, accumulate and average the similarity between each paper and other papers, and calculate the autocorrelation matrix similarity index z of the similarity of papers published by the same unit s :

[0107]

[0108] The mutual correlation matrix similarity index calculation submodule is used to calculate the similarity mutual correlation matrix H according to the similarity mutual correlation matrix H ij, we get the similarity between papers m and n published by different units, accumulate and average the similarities, and calculate the similarity index h of the cross-correlation matrix of the similarities of papers published by different units s :

[0109]

[0110] The authors of the same name determine the submodule for the similarity index h in the cross-correlation matrix s The similarity index z with the autocorrelation matrix s When the ratio of is greater than a preset threshold, it is determined that the author of the paper published by the i-th unit under the target author's name and the author of the paper published by the j-th unit under the target author's name are the same person.

[0111] Based on the same inventive concept, an embodiment of the present invention further discloses a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps in the method for determining authors with the same name as described in the embodiment of the present invention are implemented.

[0112] Based on the same inventive concept, an embodiment of the present invention further discloses an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps in the method for determining authors with the same name as described in the embodiment of the present invention are implemented.

[0113] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0114] It will be appreciated by those skilled in the art that the embodiments of the present invention may be provided as methods, devices, or computer program products. Therefore, the embodiments of the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the embodiments of the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0115] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of the methods, terminal devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0116] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0117] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0118] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. "And / or" means that either one of the two can be selected, or both can be selected. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or terminal device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal device. In the absence of more restrictions, the elements defined by the sentence "including one..." do not exclude the existence of other identical elements in the process, method, article or terminal device including the elements.

[0119] The above is a detailed introduction to the method, device, electronic device and readable storage medium for determining authors of the same name provided by the present invention. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scopes. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A method for determining authors with the same name, characterized in that: The method comprises: Taking the unit of published papers as the classification granularity, classify all papers with the searched target author's name, and establish a basic vocabulary of all the papers; According to the basic word library, the core words of each paper in all the papers are extracted, all the core words of all the papers are statistically obtained, and according to all the core words of all the papers, normalized word frequency vectors of the core words of different papers in all the papers are established; According to the normalized word frequency vectors of the core words of the different papers, the similarities of the different papers in all the papers are calculated, and a similarity autocorrelation matrix of the papers of the same unit and a similarity cross-correlation matrix of the papers of different units are established; According to the similarity autocorrelation matrix and the similarity mutual correlation matrix, respectively calculating the autocorrelation matrix similarity index and the mutual correlation matrix similarity index of the paper similarity, and judging whether the authors of the papers published by different units under the name of the target author are the same person; The step of calculating the autocorrelation matrix similarity index and the cross-correlation matrix similarity index of the paper similarity respectively according to the similarity autocorrelation matrix and the similarity cross-correlation matrix, and judging whether the authors of the papers published by different units under the name of the target author are the same person comprises: According to the similarity autocorrelation matrix Z i , get the similarity between two papers in the number m of papers published by the same unit, remove the similarity value of each paper in the number m of papers published by the same unit, accumulate and average the similarity between each paper and other papers, and calculate the autocorrelation matrix similarity index z of the similarity of papers published by the same unit s : According to the similarity cross-correlation matrix H ij , we get the similarity between papers m and n published by different units, accumulate and average the similarities, and calculate the similarity index h of the cross-correlation matrix of the similarities of papers published by different units s : When the cross-correlation matrix similarity index h s The similarity index z with the autocorrelation matrix s When the ratio of is greater than a preset threshold, it is determined that the author of the paper published by the i-th unit under the target author's name and the author of the paper published by the j-th unit under the target author's name are the same person.

2. The method for determining authors with the same name according to claim 1, characterized in that: The steps of establishing the basic vocabulary of all the papers include: Extracting basic words of each paper in all the papers by text retrieval method; The stop words in the basic words of each paper are filtered, and then the basic words of all the papers are merged to obtain the basic word library of all the papers.

3. The method for determining authors with the same name according to claim 2, characterized in that: The step of extracting the core words of each paper in all the papers according to the basic word library comprises: Calculate the total number of times each basic word in the basic word library appears in all the papers, and divide the number of times a target basic word appears in all the papers by the number of times the basic word that appears most frequently in all the papers to obtain a normalized word frequency of the target basic word; wherein the target basic word is any basic word in the basic word library; Count the subjects involved in all the papers and the number of papers corresponding to each subject, and among all the papers, take the logarithm of the quotient of the number of papers corresponding to each subject and the number of papers containing the target basic word in all the papers, to obtain the inverse document frequency of the target basic word in the subject; According to the product of the normalized word frequency of the target basic word and the inverse document word frequency of the target basic word in each subject, the normalized word frequency-inverse document word frequency of the target basic word in each subject is determined to obtain a normalized word frequency-inverse document word frequency set of the target basic word, wherein the normalized word frequency-inverse document word frequency set of the target basic word includes a plurality of normalized word frequencies-inverse document word frequencies, and different normalized word frequencies-inverse document word frequencies correspond to different subjects; According to the normalized word frequency-inverse document word frequency set of each basic word in the basic word library, the basic words in each paper of all the papers whose normalized word frequency-inverse document word frequency is greater than a preset threshold are determined as the core words of the paper.

4. The method for determining authors with the same name according to claim 3, characterized in that: The step of establishing normalized word frequency vectors of core words of different papers in all the papers according to all the core words of all the papers includes: According to all the core words of all the papers, the number of times each core word appears in each of the papers is calculated, and the number of times the target core word appears in the current paper is divided by the number of times the core word with the largest number of appearances in the current paper, to obtain the normalized word frequency of the target core word in the current paper, wherein the target core word is any one of all the core words of all the papers; The normalized word frequencies of all core words in each of the papers in all the papers are counted to obtain the normalized word frequency vectors of the core words in different papers in all the papers.

5. The method for determining authors with the same name according to claim 4, characterized in that: The step of calculating the similarity of different papers in all the papers according to the normalized word frequency vectors of the core words of the different papers includes: Based on the normalized word frequency vectors of the core words of the different papers, the similarity S of the different papers in all the papers is calculated using the following formula: kl : In the above formula, k g , l g are the normalized word frequencies of the g-th core words in paper k and paper l among all the papers, g=1,2,…,w.

6. The method for determining authors with the same name according to claim 5, characterized in that: The steps to establish the similarity autocorrelation matrix of papers from the same unit include: According to the similarity of different papers in all the papers, the similarity between the papers published by the target author in the same unit is established. If the target author has published m papers in the i-th unit, the similarity autocorrelation matrix Z of the papers in the same unit is i for: The steps to establish the similarity correlation matrix of papers from different units include: According to the similarity of different papers of the target author name, the similarity between papers published by the target author name in different units is established. If the target author name has published m and n papers in the i-th unit and the j-th unit respectively, the similarity cross-correlation matrix H of the papers in different units is ij for:

7. A device for determining authors of the same name, characterized in that: The device comprises: A basic word library building module is used to classify all papers with the searched target author's name based on the unit of published papers as the classification granularity, and to build a basic word library for all the papers; A core word normalized word frequency vector acquisition module is used to extract the core words of each paper in all the papers according to the basic word library, obtain all the core words of all the papers by statistics, and establish the core word normalized word frequency vectors of different papers in all the papers according to all the core words of all the papers; A similarity matrix building module is used to calculate the similarity of different papers in all the papers based on the normalized word frequency vectors of the core words of the different papers, and to build a similarity autocorrelation matrix of papers of the same unit and a similarity cross-correlation matrix of papers of different units; A module for determining authors with the same name, for calculating the autocorrelation matrix similarity index and the cross-correlation matrix similarity index of the paper similarity according to the similarity autocorrelation matrix and the similarity cross-correlation matrix, and determining whether the authors of the papers published by different units under the name of the target author are the same person; Wherein, the module for determining authors with the same name includes: The autocorrelation matrix similarity index calculation submodule is used to calculate the similarity index of the autocorrelation matrix Z according to the similarity i , get the similarity between two papers in the number m of papers published by the same unit, remove the similarity value of each paper in the number m of papers published by the same unit, accumulate and average the similarity between each paper and other papers, and calculate the autocorrelation matrix similarity index z of the similarity of papers published by the same unit s : The mutual correlation matrix similarity index calculation submodule is used to calculate the similarity mutual correlation matrix H according to the similarity mutual correlation matrix H ij , we get the similarity between papers m and n published by different units, accumulate and average the similarities, and calculate the similarity index h of the cross-correlation matrix of the similarities of papers published by different units s : The authors of the same name determine the submodule for the similarity index h in the cross-correlation matrix s The similarity index z with the autocorrelation matrix s When the ratio of is greater than a preset threshold, it is determined that the author of the paper published by the i-th unit under the target author's name and the author of the paper published by the j-th unit under the target author's name are the same person.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps in the method according to any one of claims 1 to 6 are implemented.

9. An electronic device comprising a memory, a processor, and a device stored in the memory and capable of being processed A computer program running on a device, characterized in that When the processor executes the computer program, Implement the steps of the method according to any one of claims 1 to 6.

Images