A method for managing terms based on a tuple model

Abstract

The application discloses a kind of based on tuple model's term management method, it includes the following steps: using the term representation structure model based on six tuples, the term in terminology database is described one by one, and the step of establishing the first mapping index structure of word in term definition information to the term in terminology database based on terminology database;The step of searching and checking duplicate management based on the terminology database to new input term information;The step of collaborative management based on the terminology database to the term in terminology database and newly added term.The application can effectively find the repeatedly added term, effectively solve the non-standard problems such as the same name synonym, different name synonym and the like existing in terminology entry and maintenance management, on the other hand, can effectively improve the efficiency of terminology management, facilitate effective citation of terminology, and effectively support the entry, writing and standard use of terminology.

Description

Technical Field

[0001] This invention relates to a natural language processing method, specifically a terminology management method. Background Technology

[0002] The writing of ISO standards and related documents requires the definition and reference of a large number of terms. With the emergence of a large number of ISO standards, more and more terms have appeared.

[0003] Currently, there is no standardized and intelligent terminology management method, which makes terminology resources difficult to manage. For example, many terms are defined repeatedly, resulting in a large number of terms with the same name and meaning, the same name but different meanings, and different names but the same meaning. Furthermore, terms are difficult to share and use, leading to problems such as duplicate definitions of terms in many ISO standards.

[0004] To address the aforementioned issues, this invention proposes and implements a multi-collaborative terminology management method. Summary of the Invention

[0005] Purpose of the invention: In order to overcome the shortcomings of the prior art, the purpose of this invention is to provide a term management method based on a tuple model.

[0006] Technical Solution: To solve the above-mentioned technical problems, this invention provides a terminology management method based on a tuple model, which includes the following steps:

[0007] The method employs a terminology representation structure model based on six tuples to describe each term in the terminology database, and establishes a first mapping index structure from words in the terminology definition information to terms in the terminology database based on the terminology database.

[0008] Based on the terminology database, the steps include retrieving and deduplicating newly input terminology information;

[0009] The steps involve collaboratively managing the terms in the terminology database and newly added terms.

[0010] Preferably, in the steps of describing each term in the terminology database using a six-tuple-based terminology representation structure model, and establishing a first mapping index structure from words in the terminology definition information to terms in the terminology database based on the terminology database, the six-tuple-based terminology representation structure model is as follows:

[0011] Term=(ID,TermName,TermEnName,Definition,DomainSet,Status)

[0012] in:

[0013] ID is a unique identifier for the term;

[0014] TermName is the Chinese name of the term;

[0015] TermEnName is the English name of the term;

[0016] Definition provides the definition information for terms;

[0017] DomainSet is a collection of domains to which terms belong;

[0018] Status is the status value of a term; the range of the term status value is the status set StatusSet = {Private, Public, Lock, Del}, where: Private indicates that the term is a private term, public indicates that the term is a public term and is in an unlocked state, Lock indicates that the term is a public term and is in a locked state, and Del indicates that the term is in a deleted state.

[0019] Preferably, the steps of using a six-tuple-based terminology representation structure model to describe each term in the terminology database, and establishing a first mapping index structure from words in the terminology definition information to terms in the terminology database based on the terminology database, specifically include the following steps:

[0020] 11) Establish a terminology representation structure model based on six-tuples;

[0021] 12) Based on the terminology representation structure model of the six-tuple, describe each term in the terminology database TermDic one by one;

[0022] 13) Based on the terminology database TermDic, a word segmentation algorithm is used to segment the definition information of each term, and a first mapping index structure wordToTerm is established from the words in the terminology definition information to the corresponding terms in the terminology database.

[0023] Preferably, the step of retrieving and deduplicating newly input terminology information based on the terminology database includes the following steps;

[0024] 21) If the newly input term information includes its Chinese name but does not include the term definition information and the English name of the term, iterate through each term in the term database TermDic to judge the similarity of the Chinese name of the term, and put the terms whose Chinese name similarity is greater than the first preset threshold and whose status is Public or Lock into the similar term set Result, until all terms in the term database TermDic have been judged, and jump to step 25);

[0025] 22) If the newly input term information includes its term definition information but does not include the Chinese name or English name of the term, use a word segmentation algorithm to segment the term definition information to obtain {W1,…,W n For each word Wi, the word W is found using the first mapping index structure wordToTerm. i The mapping relationship to terms in the terminology database is used to find terms in the terminology database whose definition information contains W1, W2, ..., or W. n The first set of terms;

[0026] Iterate through each term in the first term set, perform term definition information similarity judgment, and put terms whose term definition information similarity is greater than the second preset threshold and whose term status is Public or Lock into the similar term set Result, until all terms in the first term set have been judged, then jump to step 25);

[0027] 23) If the newly input term information includes its English name but does not include the Chinese name or definition information, iterate through each term in the terminology database TermDic to determine the similarity of the English name. Then, add terms whose English name similarity is greater than the third preset threshold and whose status is Public or Lock to the similar term set Result. Continue until all terms in the terminology database TermDic have been determined, and then proceed to step 25).

[0028] 24) If the newly input term information contains two or more of the following: Chinese name of the term, definition information of the term, and English name of the term, iterate through each term in the term database TermDic to perform a comprehensive similarity judgment. Then, put terms with a comprehensive similarity greater than the fourth preset threshold and a status value of Public or Lock into the similar term set Result. Continue until all terms in the term database TermDic have been judged, and then proceed to step 25).

[0029] 25) Combine the similarity score with the domain set DomainSet to sort the terms in the similar terms set Result, and output the ordered identical or similar terms as the automatic deduplication results, thereby realizing automatic deduplication of terms.

[0030] More preferably, in step 21), the similarity between the Chinese name of the term "Name" in the newly input terminology information and the Chinese name of a term "TermName" in the terminology database TermDic is:

[0031]

[0032] More preferably, in step 22), the similarity between the term definition information Def contained in the newly input term information and the term definition information Definition of a certain term in the first term set is:

[0033]

[0034] More preferably, in step 23), the similarity between the English name of the term EnName contained in the newly input term information and the English name TermEnName of a term in the terminology database TermDic is obtained by using edit distance similarity, which is Sim(EnName,TermEnName).

[0035] Further preferred, the edit distance between the English names of the two terms used to calculate similarity is no more than 3.

[0036] Further preferably, in step 24), based on two or more of the information contained in the newly input terminology information—the Chinese name (Name), the definition (Def), and the English name (EnName)—the comprehensive terminology similarity between the newly input terminology information Term1 and a certain term Term2 in the terminology database TermDic is:

[0037] Sim(Term1,Term2)=w1*Sim(Name,TermName)+w2*Sim(Def,Definition)+w3*Sim(EnName,TermEnName)

[0038] in:

[0039]

[0040]

[0041] Where: Sim(EnName,TermEnName) is obtained using edit distance similarity;

[0042] Wherein: Name is the Chinese name of the term contained in the newly entered term information Term1; TermName is the Chinese name of the term Term2 in the terminology database TermDic; Def is the definition information of the term contained in the newly entered term information Term1; Definition is the definition information of the term Term2 in the terminology database TermDic; EnName is the English name of the term contained in the newly entered term information Term1; TermEnName is the English name of the term Term2 in the terminology database TermDic.

[0043] Where w1, w2, and w3 are all weighting coefficients; when the newly input term information does not contain the Chinese name of the term, w1 = 0; when the newly input term information does not contain the term definition information, w2 = 0; when the newly input term information does not contain the English name of the term, w3 = 0; w1 + w2 + w3 = 1.

[0044] More preferably, in step 25), the terms in the similar term set Result are sorted by combining the similarity score and the domain set of the terms, specifically including:

[0045] For each term in the similar term set Result, it is first sorted in descending order of its similarity to the newly input term information; for several terms with the same similarity value, they are sorted in order according to the DomainSet classification; or

[0046] For each term in the similar term set Result, it is first sorted according to the domain set; for several terms in the same domain set, they are sorted in descending order of their similarity to the newly input term information.

[0047] More preferably, step 25) further includes: displaying sequentially arranged identical or similar terms as reference prompts for the user to determine whether to add them as new terms in the terminology entry or update interface.

[0048] Preferably, the step of collaboratively managing the terms in the terminology database and newly added terms based on the terminology database includes the following steps:

[0049] For newly input terminology, when the user selects it as a new term based on the automatic deduplication results output in step 2), the new term is added to the terminology database TermDic, and the terminology database is updated; and

[0050] For terms that are in the editing state of entry or update, set the term's status value to Lock and lock the term's current editing management permissions to the current user's steps; and

[0051] The steps to set the status value of a term to Public and release its editing and management permissions after it has been entered or updated and the editor has exited the editing process.

[0052] Further preferably, the step of collaboratively managing terms in the terminology database and newly added terms based on the terminology database also includes: setting the status value of the term to Private; and setting different user editing and management permissions for existing terms and / or newly added terms in the terminology database.

[0053] Beneficial Effects: The terminology management method based on a tuple model provided by this invention, compared with existing technologies, can effectively identify duplicate terms, thus resolving issues such as homonyms and synonyms with different names during terminology entry and maintenance. Furthermore, it can effectively save time and improve efficiency in terminology management, facilitating the effective use of terms, such as providing related information based on similarity during term retrieval and entry. It provides effective support for terminology entry, writing, and standardized use. Overall, it can be widely and efficiently applied to standard writing or terminology management platforms. Attached Figure Description

[0054] Figure 1 This is a flowchart illustrating the terminology management method based on a tuple model provided for an embodiment. Detailed Implementation

[0055] The present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The following embodiments do not constitute a limitation on the present invention.

[0056] This invention provides a terminology management method based on a tuple model, the method as follows: Figure 1 The steps shown are as follows:

[0057] The method employs a terminology representation structure model based on six tuples to describe each term in the terminology database, and establishes a first mapping index structure from words in the terminology definition information to terms in the terminology database based on the terminology database.

[0058] Based on the terminology database, the newly input terminology information (corresponding to...) Figure 1 The steps for retrieving and managing duplicate information (new terminology information);

[0059] The steps involve collaboratively managing the terms in the terminology database and newly added terms.

[0060] This invention provides a terminology management method based on a tuple model, which employs a six-tuple-based terminology representation structure model to describe each term in the terminology database, and establishes a first mapping index structure from words in the terminology definition information to terms in the terminology database based on the terminology database. Specifically, the method includes the following steps:

[0061] 11) Establish a terminology representation structure model based on six-tuples, as follows:

[0062] Term=(ID,TermName,TermEnName,Definition,DomainSet,Status)

[0063] in:

[0064] ID is a unique identifier for the term;

[0065] TermName is the Chinese name of the term;

[0066] TermEnName is the English name of the term;

[0067] Definition provides the definition information for terms;

[0068] DomainSet is a set of domains to which a term belongs, namely: {Domain1, Domain2, ...};

[0069] Status is the status value of a term; the range of the term status value is the status set StatusSet = {Private, Public, Lock, Del}, where: Private indicates that the term is a private term, public indicates that the term is a public term and is in an unlocked state, Lock indicates that the term is a public term and is in a locked state, and Del indicates that the term is in a deleted state.

[0070] Example: For the term "sports equipment", use ("5296-2008-gb-e-300_5000", "sports equipment", "various instruments and apparatus used in various sports", "{industry, sports}", "Public").

[0071] 12) Based on the above six-tuple term representation structure model, each term in the terminology database TermDic is described one by one; this embodiment specifically includes:

[0072] Establish the terminology and its storage structure, such as Figure 1 As shown, the terminology information from the TermDic terminology database is read in, a memory structure is established, and the information is represented and stored in a six-tuple format as follows:

[0073]

[0074]

[0075] In this embodiment, the Term structure is used to store the relevant information of each term in the terminology library; wherein, termVec is an abstract array representation structure that stores the structural information of all terms in the terminology library, or it can be said to be an abstract array representation structure of the terminology library.

[0076] 13) Based on the terminology database TermDic, a word segmentation algorithm is used to segment the definition information of each term, and a first mapping index structure wordToTerm is established from the words in the terminology definition information to the corresponding terms in the terminology database; the first mapping index structure wordToTerm established in this embodiment is specifically as follows:

[0077] map <Word,Set <termid>>wordToTerm;

[0078] Word is the definition information of a term, and each word is a word after word segmentation. Set <termid>It is the set of IDs of terms containing the word "Word" in the definition information within the TermVec structure.

[0079] The present invention provides a terminology management method based on a tuple model, wherein the step of retrieving and deduplicating newly input terminology information based on the terminology database specifically includes the following steps;

[0080] 21) If the newly input term information includes its Chinese name but does not include term definition information or English name, iterate through each term in the terminology database TermDic and perform a similarity test on the Chinese name. Add terms with a similarity score greater than the first preset threshold α and a status value of Public or Lock to the similar term set Result. Continue this process until all terms in the terminology database TermDic have been evaluated, then proceed to step 25); where:

[0081] The similarity (i.e., term name similarity) between the Chinese name of a term in the newly entered term information and the Chinese name of a term in the terminology database TermDic is as follows:

[0082]

[0083] Where: Name is the Chinese name of the term contained in the newly entered term information; TermName is the Chinese name of a term in the terminology database TermDic;

[0084] In this embodiment, based on the set first preset threshold α (the larger α is, the higher the similarity), terms that satisfy Sim(term,TermName)>α and whose state value is Public or Lock are placed into the similar term set Result.

[0085] 22) If the newly input term information includes its term definition information but does not include the Chinese name or English name of the term, use a word segmentation algorithm to segment the term definition information to obtain {W1,…,W n For each word Wi, the word W is found using the first mapping index structure wordToTerm. i The mapping relationship to terms in the terminology database is used to find terms in the terminology database whose definition information contains W1, W2, ..., or W. n The first set of terms;

[0086] Iterate through each term in the first term set, perform term definition information similarity judgment, and add terms whose term definition information similarity is greater than the second preset threshold β and whose state value is Public or Lock to the similar term set Result, until all terms in the first term set have been judged, then jump to step 25); where:

[0087] The similarity (i.e., term definition similarity) between the term definition information Def contained in the newly input term information and the term definition information Definition of a certain term in the first term set is:

[0088]

[0089] Where: Def is the term definition information contained in the newly input term information; Definition is the term definition information of a certain term in the first term set;

[0090] In this embodiment, based on the set second preset threshold β (the larger β is, the higher the similarity), terms that satisfy Sim(Def,Definition)>β and whose state value is Public or Lock are placed into the similar term set Result.

[0091] In the terminology management method based on the tuple model provided by this invention, when the newly input terminology information contains its terminology definition information but does not contain the Chinese name and English name of the terminology, the similarity of terminology definition information is judged by directly traversing the first terminology set instead of traversing the entire terminology database, which can further effectively improve the efficiency of terminology retrieval and deduplication management.

[0092] 23) If the newly input terminology information includes its English name but not its Chinese name or definition, iterate through each term in the terminology database TermDic and perform a similarity test on its English name. Add terms with an English name similarity greater than the third preset threshold γ and a status of Public or Lock to the similar term set Result. Continue this process until all terms in the terminology database TermDic have been evaluated, then proceed to step 25); where:

[0093] The similarity between the English name of the term EnName in the newly input term information and the English name of a term TermEnName in the terminology database TermDic is Sim(EnName,TermEnName), obtained by using edit distance similarity.

[0094] In this embodiment, the edit distance between the English names of the two terms used to calculate similarity does not exceed 3;

[0095] In this embodiment, based on the set third preset threshold γ (the larger γ is, the higher the similarity), terms that satisfy Sim(EnName,TermEnName)>γ and whose state value is Public or Lock are placed into the similar term set Result.

[0096] 24) If the newly input terminology information contains two or more of the following: the Chinese name of the term, the definition of the term, and the English name of the term, iterate through each term in the terminology database TermDic and perform a comprehensive term similarity judgment. Add terms with a comprehensive similarity greater than the fourth preset threshold θ and a status value of Public or Lock to the similar term set Result. Continue this process until all terms in the terminology database TermDic have been judged, then proceed to step 25); where:

[0097] Based on two or more of the following information contained in the newly input terminology information: the Chinese name (Name), the definition (Def), and the English name (EnName), the overall terminology similarity between the newly input terminology information Term1 and a term Term2 in the terminology database TermDic is:

[0098] Sim(Term1,Term2)=w1*Sim(Name,TermName)+w2*Sim(Def,Definition)+w3*Sim(EnName,TermEnName)

[0099] in:

[0100]

[0101]

[0102] Where: Sim(EnName,TermEnName) is obtained using edit distance similarity;

[0103] Wherein: Name is the Chinese name of the term contained in the newly entered term information Term1; TermName is the Chinese name of the term Term2 in the terminology database TermDic; Def is the definition information of the term contained in the newly entered term information Term1; Definition is the definition information of the term Term2 in the terminology database TermDic; EnName is the English name of the term contained in the newly entered term information Term1; TermEnName is the English name of the term Term2 in the terminology database TermDic.

[0104] Where: w1 is the weighted coefficient of the similarity between Chinese terms (Sim(Name,TermName)), w2 is the weighted coefficient of the similarity between Chinese terms (Sim(Def,Definition)), and w3 is the weighted coefficient of the similarity between Chinese terms (Sim(EnName,TermEnName)). When the newly input term information does not contain the Chinese term name, w1 = 0; when the newly input term information does not contain the term definition information, w2 = 0; when the newly input term information does not contain the English term name, w3 = 0; w1 + w2 + w3 = 1.

[0105] The weighting coefficients w1, w2, and w3 in this document can be flexibly set according to the actual application scenario. In this embodiment, when the newly input terminology information includes the Chinese name, definition, and English name of the terminology, w1 = 0.4, w2 = 0.3, and w3 = 0.3; when the newly input terminology information includes the Chinese name and definition but not the English name, w1 = 0.5, w2 = 0.5, and w3 = 0; when the newly input terminology information includes the Chinese name and English name but not the definition, w1 = 0.5, w2 = 0, and w3 = 0.5; when the newly input terminology information includes the definition and English name but not the Chinese name, w1 = 0, w2 = 0.5, and w3 = 0.5.

[0106] In this embodiment, based on the set fourth preset threshold θ (the larger θ is, the higher the similarity), terms that satisfy Sim(Term1,Term2)>θ and whose state value is Public or Lock are placed into the similar term set Result.

[0107] The first preset threshold α, the second preset threshold β, the third preset threshold γ, and the fourth preset threshold θ can be flexibly set according to the actual application scenario. In this embodiment, the first preset threshold α = 0.8, the second preset threshold β = 0.75, the third preset threshold γ = 0.7, and the fourth preset threshold θ = 0.65.

[0108] The newly input terminology information mentioned in the text refers to new terminology information to be searched for plagiarism, containing at least one of the following: the Chinese name of the term, the definition of the term, and the English name of the term. In other words, the newly input terminology information contains at least one of these three pieces of information. Alternatively, it can be stated that the newly input terminology information to be searched for plagiarism contains the Chinese name of the term and / or the definition of the term and / or the English name of the term.

[0109] 25) Combine the similarity score with the domain set DomainSet to sort the terms in the similar terms set Result, and output the ordered identical or similar terms as the automatic deduplication results, thereby realizing automatic deduplication of terms.

[0110] In step 25) above, the terms in the similar term set Result are sorted by combining the similarity and the domain set of the term. In this embodiment, this specifically includes: for each term in the similar term set Result, firstly, sorting them in descending order of their similarity to the newly input term information; for several terms with the same similarity value, sorting them in order according to the domain set classification.

[0111] In some embodiments, step 25) above, which combines the similarity magnitude and the domain set of the term to sort the terms in the similar term set Result, specifically includes: sorting each term in the similar term set Result according to the domain set classification; and sorting several terms in the same domain set according to their similarity to the newly input term information from largest to smallest.

[0112] In this embodiment, step 25) above also includes: displaying sequentially arranged identical or similar terms as reference prompts for the user to determine whether to add them as new terms in the terminology input or update interface.

[0113] This invention provides a terminology management method based on a tuple model, wherein the step of collaboratively managing terms in the terminology database and newly added terms specifically includes the following steps:

[0114] For newly input terminology, when the user selects it as a new term based on the automatic deduplication results output in step 2), the new term is added to the terminology database TermDic, and the terminology database is updated; and

[0115] For terms that are in the editing state of entry or update, set the term's status value to Lock and lock the term's current editing management permissions to the current user's steps; and

[0116] The steps to set the status value of a term to Public and release its editing and management permissions after it has been entered or updated and the editor has exited the editing process.

[0117] In some embodiments of the terminology management method based on the tuple model, the step of collaboratively managing terms in the terminology library and newly added terms based on the terminology library further includes: setting the state value of the term to Private; and setting different user editing and management permissions for existing terms and / or newly added terms in the terminology library.

[0118] Experiment: This experiment first labeled 1500 pairs of similar and identical terms based on a six-tuple model to establish a terminology database, including 500 pairs each of synonyms, terms with different names but different meanings, and similar terms. Then, multiple rounds of testing and verification were conducted by inputting new terminology information and performing deduplication and retrieval management on the newly input terminology information. Experimental data shows that the method achieves a recall rate of over 87% and an accuracy rate of over 88.5% for retrieving new terminology information input by the user. The experiment demonstrates that the terminology management method based on the tuple model provided by this invention can effectively manage terms through deduplication and retrieval; simultaneously, the state management method in this method can effectively manage terms collaboratively.

[0119] Therefore, the terminology management method based on the tuple model provided in this embodiment is very conducive to the effective use of terms. On the one hand, it provides a related display by sorting terms according to similarity during term retrieval and entry, which is flexible and easy to use. On the other hand, the provided similarity calculation is comprehensive, flexible, accurate and efficient, and suitable for application in standard writing or terminology management platforms, providing effective support for terminology entry, writing and standardized use.

[0120] The above are merely preferred embodiments of the present invention. It should be noted that the above embodiments do not constitute a limitation on the present invention. Various changes and modifications made by those skilled in the art without departing from the technical concept of the present invention all fall within the protection scope of the present invention.< / termid> < / termid>

Claims

1. A terminology management method based on a tuple model, characterized in that... The method includes the following steps: The method employs a terminology representation structure model based on six tuples to describe each term in the terminology database, and establishes a first mapping index structure from words in the terminology definition information to terms in the terminology database based on the terminology database. Based on the terminology database, the steps include retrieving and deduplicating newly input terminology information; Based on the terminology database, the steps include collaborative management of terms in the terminology database and newly added terms; In the steps of describing each term in the terminology database using a six-tuple-based terminology representation structure model, and establishing a first mapping index structure from words in the terminology definition information to terms in the terminology database based on the terminology database, the six-tuple-based terminology representation structure model is as follows: Term=(ID,TermName,TermEnName,Definition,DomainSet,Status) in: ID is a unique identifier for the term; TermName is the Chinese name of the term; TermEnName is the English name of the term; Definition provides the definition information for terms; DomainSet is a collection of domains to which terms belong; Status is the status value of a term; the range of the term status value is the status set StatusSet = {Private, Public, Lock, Del}, where: Private indicates that the term is a private term, public indicates that the term is a public term and is in an unlocked state, Lock indicates that the term is a public term and is in a locked state, and Del indicates that the term is in a deleted state. The steps of using a six-tuple-based terminology representation structure model to describe each term in the terminology database, and establishing a first mapping index structure from words in the terminology definition information to terms in the terminology database, specifically include the following steps: 11) Establish a terminology representation structure model based on six-tuples; 12) Based on the terminology representation structure model of the six-tuple, describe each term in the terminology database TermDic one by one; 13) Based on the terminology database TermDic, a word segmentation algorithm is used to segment the definition information of each term, and a first mapping index structure wordToTerm is established from the words in the terminology definition information to the corresponding terms in the terminology database. The step of retrieving and deduplicating newly input terminology information based on the terminology database includes the following steps; 21) If the newly input term information includes its Chinese name but does not include the term definition information and the English name of the term, iterate through each term in the term database TermDic to judge the similarity of the Chinese name of the term, and put the terms whose Chinese name similarity is greater than the first preset threshold and whose status is Public or Lock into the similar term set Result, until all terms in the term database TermDic have been judged, and jump to step 25); 22) If the newly input term information contains its term definition information but does not contain the Chinese name or English name of the term, the term definition information is segmented into {W1,…,Wn} using a word segmentation algorithm. For each word Wi, the mapping relationship between word Wi and the term in the term database is found through the first mapping index structure wordToTerm. Thus, the first set of terms in the term database that contains W1, or W2,…, or Wn in the term definition information is found. Iterate through each term in the first term set, perform term definition information similarity judgment, and put terms whose term definition information similarity is greater than the second preset threshold and whose term status is Public or Lock into the similar term set Result, until all terms in the first term set have been judged, then jump to step 25); 23) If the newly input term information includes its English name but does not include the Chinese name or definition information, iterate through each term in the terminology database TermDic to determine the similarity of the English name. Then, add terms whose English name similarity is greater than the third preset threshold and whose status is Public or Lock to the similar term set Result. Continue until all terms in the terminology database TermDic have been determined, and then proceed to step 25). 24) If the newly input term information contains two or more of the following: Chinese name of the term, definition information of the term, and English name of the term, iterate through each term in the term database TermDic to perform a comprehensive similarity judgment. Then, put terms with a comprehensive similarity greater than the fourth preset threshold and a status value of Public or Lock into the similar term set Result. Continue until all terms in the term database TermDic have been judged, and then proceed to step 25). 25) Combine the similarity score with the domain set DomainSet to sort the terms in the similar terms set Result, and output the ordered identical or similar terms as the automatic deduplication result, thereby realizing automatic deduplication of terms; In step 21), the similarity between the Chinese name of the term "Name" in the newly input terminology information and the Chinese name of a term "TermName" in the terminology database TermDic is as follows: In step 22), the similarity between the term definition information Def contained in the newly input term information and the term definition information Definition of a certain term in the first term set is: In step 23), the similarity between the English name of the term EnName contained in the newly input term information and the English name of a term TermEnName in the terminology database TermDic is obtained by using edit distance similarity, which is Sim(EnName,TermEnName). In step 24), based on two or more of the following information contained in the newly input terminology information: the Chinese name of the term (Name), the term definition information (Def), and the English name of the term (EnName), the comprehensive terminology similarity between the newly input terminology information Term1 and a certain term Term2 in the terminology database TermDic is: Sim(Term1,Term2)=w1*Sim(Name,TermName)+w2*Sim(Def,Definition)+w3*Sim(EnName,TermEnName) in: Where: Sim(EnName,TermEnName) is obtained using edit distance similarity; Wherein: Name is the Chinese name of the term contained in the newly entered term information Term1; TermName is the Chinese name of the term Term2 in the terminology database TermDic; Def is the definition information of the term contained in the newly entered term information Term1; Definition is the definition information of the term Term2 in the terminology database TermDic; EnName is the English name of the term contained in the newly entered term information Term1; TermEnName is the English name of the term Term2 in the terminology database TermDic. Where: w1, w2, and w3 are all weighting coefficients; when the newly input term information does not contain the Chinese name of the term, w1 = 0; when the newly input term information does not contain the term definition information, w2 = 0; when the newly input term information does not contain the English name of the term, w3 = 0; w1 + w2 + w3 = 1; Step 25) involves sorting the terms in the similarity term set Result by combining the similarity score and the domain set of the terms, specifically including: For each term in the similar term set Result, it is first sorted in descending order of its similarity to the newly input term information; for several terms with the same similarity value, they are sorted in order according to the DomainSet classification; or For each term in the similar term set Result, it is first sorted according to the domain set; for several terms in the same domain set, they are sorted in descending order of their similarity to the newly input term information. Step 25) further includes: displaying sequentially arranged identical or similar terms as reference prompts for users to determine whether to add them as new terms in the terminology entry or update interface. The step of collaboratively managing terms in the terminology database and newly added terms based on the terminology database includes the following steps: For newly input terminology, when the user selects it as a new term based on the automatic deduplication results output in step 2), the new term is added to the terminology database TermDic, and the terminology database is updated; and For terms that are in the editing state of entry or update, set the term's status value to Lock and lock the term's current editing management permissions to the current user's steps; and The steps to set the status value of a term to Public and release its editing and management permissions after it has been entered or updated and the editor has exited the editing process.

Images