CN109582961A - A kind of efficient robot data similarity calculation algorithm - Google Patents
A kind of efficient robot data similarity calculation algorithm Download PDFInfo
- Publication number
- CN109582961A CN109582961A CN201811433367.7A CN201811433367A CN109582961A CN 109582961 A CN109582961 A CN 109582961A CN 201811433367 A CN201811433367 A CN 201811433367A CN 109582961 A CN109582961 A CN 109582961A
- Authority
- CN
- China
- Prior art keywords
- word
- term
- similarity
- mapping
- ontology
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/284—Lexical analysis, e.g. tokenisation or collocates
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/237—Lexical tools
- G06F40/247—Thesauruses; Synonyms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
- G06F40/295—Named entity recognition
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Machine Translation (AREA)
Abstract
The phenomenon that a kind of efficient robot data similarity calculation algorithm is claimed in the present invention, is related to field there are multiple and different ontologies, and then cause the Interoperability between isomery ontology.This method defines ontology pretreatment, maps candidate generation, the merging of the Word similarity of concept term, multiple mapping and mapping result output.In all kinds of mapping methods, the concept similarity based on word constitutive characteristic calculates other corpus resources not needed in addition to word itself substantially and supports that calculating has the characteristics that direct and agility.But there are still for semantic identical but write the problems such as not fully consistent allosome, synonym similarity calculation be difficult and the allocation strategy of concept term composition term weighing to be matched is not perfect for presently relevant method.For these problems, the invention proposes a kind of efficient robot data similarity calculation algorithms, to promote Ontology Mapping resultant effect.
Description
Technical field
The invention belongs to computer information processing fields more particularly to a kind of efficient robot data similarity calculation to calculate
Method.
Background technique
Ontology Mapping Method can generally be summarized as following 4 kinds:
(1) by concept similarity calculation method, compare the similarity penetrated between object, to find between isomery ontology
Connection, such as Rodr í guez etc. propose a kind of method for calculating Concept Similarity using concept definition, by the concept in ontology
It is divided into three parts of semantic relation collection for indicating the synset of concept, the feature set for portraying concept, concept, and utilizes this three
Divide and carries out similarity calculation, the final mapping relations determined between concept.
(2) similitude between isomery ontology in structure is analyzed, finds mapping relations by writing mapping ruler.
Sunna etc. proposes a kind of method for using bulk junction composition as contextual information to realize Ontology Mapping.This method is in addition to examining
Consider outside node self-information, reference is also made to the multi-level information such as its father node, child nodes, grandchild node.
(3) by the example in ontology, the mapping relations between ontology are found using technologies such as machine learning.It is typical
Example is the GLUE system of the propositions such as the Doan of University of Washington.This method comprehensively considers the various Heterogeneities of ontology, passes through
Machine learning classifies to the example of concept, and the Joint Distribution probability that then occurs in concept using example calculates concept
Between similarity and combine domain constraint and heuristic knowledge finally to determine mapping relations.
(4) a variety of methods are carried out to comprehensive Ontology Mapping Method.Shailendra etc. develops a hybrid ontology and reflects
Engine is penetrated, by the Ontology Mapping algorithm based on syntax, the Ontology Mapping algorithm based on background knowledge and structure-based ontology
Mapping algorithm is integrated, and multi-angle is mapped from many aspects, while playing various method advantages, also compensates for difference
The shortcoming of method.High bright wait improves similarity calculating method in terms of Ontological concept title, structure, example, attribute 4,
And propose the similarity calculating method of fusion.Li Jia etc. proposes the side using Hownet, calculated in conjunction with a variety of Lexical Similarities
Method realizes the ELOMC system of Chinese Ontology Mapping.
In above-mentioned Ontology Mapping Method, find that the method contacted between ontology is generally only borrowed by concept similarity calculating
Help word, phrase of expression Ontological concept etc. that can carry out similarity calculation as input, it is of less demanding to design conditions.This
Outside, such method not only can individually carry out Ontology Mapping calculating, but also can easily be counted with other types method integration
It calculates, use is more flexible, therefore is widely used.The core of such Ontology Mapping Method is the similarity based on word
It calculates.There are some similarity calculating methods based on word at present, following 4 major class can be divided into:
(1) cosine value method.Cosine similarity (Cosine Similarity) algorithm by calculate have n dimension two to
Cosine angle between amount reflects estimating for similarity degree between vector.Such algorithm has widely applicable, realization simply
The features such as, it is one of the Words similarity algorithm being most widely used.
(2) edit distance approach.In this type of method, the Smith- that Smith and Wa-terma was proposed in 1981
Waterman algorithm is most representative.There is no directly two sequences of calculating othernesses on the whole for the algorithm, but by
Different location insertion space, multiple calculate take optimal method in shorter sequence, find out two sequence highest similarity value conducts
Final result.The Jaro-Winkler Distance algorithm that Winkler is proposed is also a kind of more typical editing distance calculation
Method, he improves on the basis of JaroDistance algorithm, the quantity factor of character match is not only allowed for, also by character
Matched positional factor takes into account.
(3) similarity calculating method based on word role.The word of Ontological concept title is constituted, that is, in title
There is different similarities to contribute by centre word (Head) and qualifier part (Modifiers), can be used as that estimate word similar
The important symbol of property.Each term is expressed as by Nenadic etc. in carrying out field of biomedicine Similarity of Term calculating task
Bipartite word description scheme, i.e., the modification composition part of key vocabularies and term in term；Then one is used
The Dice coefficients (Dicelike Coefficient) of a weighting compare the word description scheme of two terms.Those are shared more
The long term for collectively constituting part will obtain more score values；If two term keywords having the same, additional point
Value will increase in similarity measure result.
(4) a variety of concept similarities calculate the method combined.Noess-ner etc. proposes the integrated side of CODI Words similarity
Method, by cosine similarity algorithm, Levenshtein similarity algorithm, Jaro-Winkler Dis-tance similarity algorithm,
Smith-Waterman similarity algorithm, Over-lap Coefficient similarity algorithm and Jaccard similarity algorithm
It is integrated with specified weight, to promote the concept term similarity calculation effect during Ontology Matching.
In the above-mentioned experiment for calculating progress domain body mapping by concept similarity of application, find existing method two
A aspect still has deficiency:
(1) for allosome word as similar " Mutation " (variation) and " Variation " (variation), synonym, closely
Adopted word hardly results in correct mapping relations using the existing similarity calculating method based on word.Especially when these words
When the centre word of the term as expression Ontological concept, counting loss will lead to corresponding Ontological concept matching fault.
(2) existing algorithm distributes the similar weight of qualifier in term in average mode, this makes similarity calculation
As a result discrimination is bad.
The present invention proposes a kind of efficient robot data similarity calculation algorithm, introduces the synonym of WordNet, nearly justice
Word and search and editing distance calculate, to solve the problems, such as that related algorithm is difficult to allosome word, synonym, near synonym；By again
The Weight Value Distributed Methods for designing term centre word and term qualifier, keep centre word, qualifier weight distribution more reasonable, in turn
Effectively promote the resultant effect of Ontological concept mapping.
Summary of the invention
Present invention seek to address that the above problem of the prior art.Propose it is a kind of needed for condition it is few, dispose quick and easy, energy
It is enough effectively to promote the efficient robot data similarity calculation algorithm for calculating effect.Technical scheme is as follows:
A kind of efficient robot data similarity calculation algorithm comprising following steps:
A. ontology to be matched is imported, ontology to be matched is pre-processed；
B. mapping is candidate generates, including between the similar situation in pretreatment ontology concept characteristic, conceptual example
Factor is analyzed, and selects concept mapping candidate right；
C. synonymous using the editing distance similarity calculation, the Word Net based on centre word that constitute word based on term
The method that word, near synonym retrieval and the weight automatic Assignment based on term centre word, qualifier combine, between ontology
The term description collections that combination of two is formed are to progress similarity calculation；
D. knot of the highest concept of mapping of similarity to set, as Ontology Mapping is selected from qualified concept group
Fruit is exported, and is formatted output and storage to Ontology Mapping result.
Further, in the step a ontology pretreatment specifically include: import ontology to be matched, in ontology include class
Concept, attribute, the term for reading including example, attribute instance parsed, feature extraction and formatted storage, be subsequent
With ready for operation.
Further, candidate generate is mapped in the step b to have used for reference Huber etc. and use in CODI Ontology Mapping Method
Mapping candidate's generation method, have main steps that: factors such as similar situation in 1. betweens of ontologies concept characteristic, conceptual example point
Analysis；2. the possible concept mapping of selection and combination is candidate right.
Further, described using the editing distance similarity calculation for constituting word based on term, it specifically includes: 3.1) art
Language constitutes the similarity calculation of word level, on the basis of establishing the word one-to-one similarity calculation in term, therefore, leads to
The mode of two term word matrixes of building is crossed, the best match corresponding relationship in term between word is found, in the list of two terms
In set of words matching process, in order to find optimum matching relation, following formula is proposed:
sim(wi,wj)=dω(wi,wj)if dω(wi,wj)≥0.8
Wherein, sim (wi,wj) indicating similarity in two terms between any word pair, similarity threshold takes 0.8, dω
(wi,wj) indicate to calculate the word w obtained by editing distance formulaiAnd wjBetween editing distance.
Further, the similarity calculation between the word pair is divided into three kinds of situations:
1. if the d of two wordsωValue is less than 0.8, then it is assumed that mismatches between them, similarity 0；
2. if the d of two wordsωValue is more than or equal to 0.8, then it is assumed that mismatches between them, similarity dω；
3. if first retrieving Word Net two words are the centre word of respectively place term and judging whether the two is same each other
Adopted word, near synonym or antonym, if it is return value is 1, and otherwise word calculates similarity by two.
Further, the Words similarity weight distribution specifically includes: parsing, is found out in analytic tree to term
The most deep noun in position, and using the word as term centre word, other words are as qualifier, to term centre word and qualifier
It carries out after correctly distinguishing, carries out the distribution of similarity weight to each word using following formula；
Wherein, wt1iAnd wt2jRespectively indicate term t1In i-th of word and term t2In j-th of word, d (wt1i) indicate word
Language wt1iWith t1The distance between subject term in term, Weight (wt1i,wt2j) indicate that matching word is to similar in progress in two terms
Weight distribution when degree calculates.
Further, the Words similarity COMPREHENSIVE CALCULATING formula are as follows:
Wherein, t1And t2For the term pair for needing to calculate Words similarity, sim (wt1i,wt2j) indicate term t1In i-th
Word and term t2In similarity between j-th of word, Weight (wt1i,wt2j) indicate according to and subject term in respective term it
Between distance, be wt1iAnd wt2jThe similarity weight of distribution, in denominator the value range of l be [0, Max (| t1|,|t2|) -1], really
Entire calculated result has been protected to be in always in [0,1] range.
Further, multiple mapping merges in the step d and mapping result exports, and specifically includes: according to experimental data
Analysis sets optimal trusted degree threshold value, filters out the concept pair for having more than this threshold value, is being more than the ontology to be matched of threshold value
Concept centering will form the mapping relations of 1:n, m:1 or m:n between concept, in order to be translated into 1:1 mapping relations, so as to
Mapping result collection is manually marked with domain expert to compare, and needs to select similarity from these qualified concept groups again
To set, the result as Ontology Mapping is exported highest concept of mapping；And Ontology Mapping result is formatted
Output and storage.
It advantages of the present invention and has the beneficial effect that:
Innovative point of the invention is to propose a kind of efficient robot data similarity calculation algorithm, in Ontology Mapping
Concept similarity calculates existing deficiency, proposes a kind of improved method.By the synonym of Word Net, near synonym retrieval and editor
Distance algorithm introduces the similarity deterministic process between term centre word, and by new automatic Weight Value Distributed Methods in term
Heart word and term modification Word similarity are integrated.It comprises the concrete steps that: ontology pretreatment, the candidate generation of mapping, Words similarity
Calculating, multiple mapping merge and mapping result output, wherein in Word similarity, by constituting centre word to term
Carry out Word Net synonym, near synonym retrieval, and to the centre word and all qualifiers that can not retrieve edited away from
From calculating, the correct matching relationship between two terms composition word is found, and is each section by similarity Weight Value Distributed Methods
With relation allocation weight, similarity calculation result between term is obtained eventually by COMPREHENSIVE CALCULATING formula.The advantages of this method is: tool
Have required condition it is few, deployment it is quick and easy, can effectively promote calculating effect
Detailed description of the invention
Fig. 1 is that the present invention provides Ontology Mapping procedure chart of the preferred embodiment based on Word similarity；
Fig. 2 is Word similarity flow chart of the invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, detailed
Carefully describe.Described embodiment is only a part of the embodiments of the present invention.
The technical solution that the present invention solves above-mentioned technical problem is:
The invention proposes a kind of efficient robot data similarity calculation algorithms, it is characterised in that: by editing distance
The composition word match that algorithm is applied to Ontological concept term calculates, and the synonym of WordNet, near synonym retrieval are applied to art
Similarity judgement between language centre word, and the Weight Value Distributed Methods of Design Technology language centre word and term qualifier again, to mention
Rise Ontology Mapping resultant effect.Hereinafter reference will be made to the drawings and invention is further described in detail in conjunction with example.
As shown in Figure 1, the Ontology Mapping procedure chart of the invention based on Word similarity, is specifically realized in
:
1. ontology pre-processes, it is responsible for importing ontology to be matched, to the concept of class in ontology, attribute, reads example, attribute instance
Equal terms are parsed, feature extraction and formatted storage, are subsequent match ready for operation.
It is generated 2. mapping is candidate, by between the factor analyses such as similar situation in ontology concept characteristic, conceptual example, choosing
It selects and to combine the mapping of possible concept candidate right.
3. Word similarity, using constituting the editing distance similarity calculation of word based on term, be based on centre word
Word Net synonym, near synonym retrieval and the weight automatic Assignment based on term centre word, qualifier combine
Method, the term description collections formed between combination of two ontology are to similarity calculation is carried out, and result is as judging ontology
The important evidence of middle concept mapping, the specific implementation process is as follows step:
3.1) term constitutes the similarity calculation of word level, it should establish the one-to-one similarity of word in term
On the basis of calculating.Therefore, it is necessary to find the best match in term between word by way of constructing two term word matrixes
Corresponding relationship.In the set of letters matching process of two terms, in order to find optimum matching relation, following formula is proposed:
sim(wi,wj)=dω(wi,wj)if dω(wi,wj)≥0.8
Wherein, sim (wi,wj) indicate similarity in two terms between any word pair, it is determined here by Germicidal efficacy
Similarity threshold take it is 0.8 more appropriate.dω(wi,wj) indicate to calculate the word w obtained by editing distance formulaiAnd wjIt
Between editing distance.Similarity calculation between word pair is divided into three kinds of situations:
1. if the d of two wordsωValue is less than 0.8, then it is assumed that mismatches between them, similarity 0；
2. if the d of two wordsωValue is more than or equal to 0.8, then it is assumed that mismatches between them, similarity dω；
3. if first retrieving Word Net two words are the centre word of respectively place term and judging whether the two is same each other
Adopted word, near synonym or antonym, if it is return value is 1, and otherwise word calculates similarity by two.
3.2) Words similarity weight distribution designs, and parses to term, finds out the name that position is most deep in analytic tree
Word, and using the word as term centre word, other words are as qualifier.It is correctly distinguished to term centre word and qualifier
Afterwards, the distribution of similarity weight is carried out to each word using following formula.
Wherein, wt1iAnd wt2jRespectively indicate term t1In i-th of word and term t2In j-th of word, d (wt1i) indicate word
Language wt1iWith t1The distance between subject term in term, Weight (wt1i,wt2j) indicate that matching word is to similar in progress in two terms
Weight distribution when degree calculates.
3.3) Words similarity COMPREHENSIVE CALCULATING designs
Wherein, t1And t2For the term pair for needing to calculate Words similarity, sim (wt1i,wt2j) indicate term t1In i-th
Word and term t2In similarity between j-th of word, Weight (wt1i,wt2j) indicate according to and subject term in respective term it
Between distance, be wt1iAnd wt2jThe similarity weight of distribution, in denominator the value range of l be [0, Max (| t1|,|t2|) -1], really
Entire calculated result has been protected to be in always in [0,1] range.
4. multiple mapping merges and mapping result output, according to analysis of experimental data, optimal trusted degree threshold value, screening are set
The concept pair of this threshold value is had more than out.In the Ontological concept pair to be matched for being more than threshold value, it may will form between concept
The mapping relations of 1:n, m:1 or m:n, in order to be translated into 1:1 mapping relations, manually to mark mapping knot with domain expert
Fruit collection compares, and needs to select the highest concept of mapping of similarity to set from these qualified concept groups again, make
It is exported for the result of Ontology Mapping；And output and storage are formatted to Ontology Mapping result, in order to subsequent benefit
With and calculate effect evaluation and test.
As shown in Fig. 2, Word similarity flow chart of the invention, Fig. 2 is done for Word similarity in Fig. 1
Further refinement.Specifically it is achieved in that
1. obtaining two words first from ontology library；
2. two words that will acquire carry out judging whether it is synonym；
3. if two words are synonym, return value 1；
4. if obtaining the distance of two words by calculating two words are not synonyms；
5. the similarity of two words is 0 when if the distance of two words is less than 0.8；
6. the similarity of two words is d when if the distance of two words is greater than or equal to 0.8w(between two words away from
From value).
The above embodiment is interpreted as being merely to illustrate the present invention rather than limit the scope of the invention.?
After the content for having read record of the invention, technical staff can be made various changes or modifications the present invention, these equivalent changes
Change and modification equally falls into the scope of the claims in the present invention.
Claims (8)
1. a kind of efficient robot data similarity calculation algorithm, which comprises the following steps:
A. ontology to be matched is imported, ontology to be matched is pre-processed；
B. mapping is candidate generates, by between the factor including the similar situation in pretreatment ontology concept characteristic, conceptual example
It is analyzed, selects concept mapping candidate right；
C. the editing distance similarity calculation of word, the Word Net synonym based on centre word, close is constituted using based on term
The method that adopted word and search and weight automatic Assignment based on term centre word, qualifier combine, between ontology two-by-two
The term description collections formed are combined to progress similarity calculation；
D. select the highest concept of mapping of similarity to set from qualified concept group, as Ontology Mapping result into
Row output, and output and storage are formatted to Ontology Mapping result.
2. a kind of efficient robot data similarity calculation algorithm according to claim 1, which is characterized in that the step
Ontology pretreatment specifically includes in rapid a: ontology to be matched is imported, to including the concept of class in ontology, attribute, read example, attribute
Term including example parsed, feature extraction and formatted storage, is subsequent match ready for operation.
3. a kind of efficient robot data similarity calculation algorithm according to claim 1, which is characterized in that the step
Candidate generate of mapping has used for reference Huber etc. and map candidate's generation method used in CODI Ontology Mapping Method in rapid b, mainly
Step is: the similar situation factor analysis in 1. betweens of ontologies concept characteristic, conceptual example；2. the possible concept of selection and combination
Mapping is candidate right.
4. a kind of efficient robot data similarity calculation algorithm according to claim 1, which is characterized in that described to adopt
With the editing distance similarity calculation for constituting word based on term, specifically include: 3.1) term constitutes the similarity of word level
It calculates, on the basis of establishing the word one-to-one similarity calculation in term, therefore, passes through two term word matrixes of building
Mode finds the best match corresponding relationship in term between word, in the set of letters matching process of two terms, in order to
Optimum matching relation is enough found, proposes following formula:
sim(wi,wj)=dω(wi,wj)if dω(wi,wj)≥0.8
Wherein, sim (wi,wj) indicating similarity in two terms between any word pair, similarity threshold takes 0.8, dω(wi,wj)
It indicates to calculate the word w obtained by editing distance formulaiAnd wjBetween editing distance.
5. a kind of efficient robot data similarity calculation algorithm according to claim 4, which is characterized in that the list
Similarity calculation between word pair is divided into three kinds of situations:
1. if the d of two wordsωValue is less than 0.8, then it is assumed that mismatches between them, similarity 0；
2. if the d of two wordsωValue is more than or equal to 0.8, then it is assumed that mismatches between them, similarity dω；
3. if first retrieving Word Net judges whether the two is synonymous each other two words are the centre word of respectively place term
Word, near synonym or antonym, if it is return value is 1, and otherwise word calculates similarity by two.
6. a kind of efficient robot data similarity calculation algorithm according to claim 4, which is characterized in that institute's predicate
Language similarity weight distribution specifically includes: parsing to term, finds out the noun that position is most deep in analytic tree, and by the word
As term centre word, other words are as qualifier, and after correctly distinguish to term centre word and qualifier, use is following
Formula carries out the distribution of similarity weight to each word；
Wherein, wt1iAnd wt2jRespectively indicate term t1In i-th of word and term t2In j-th of word, d (wt1i) indicate word
wt1iWith t1The distance between subject term in term, Weight (wt1i,wt2j) indicate that matching word is in progress similarity in two terms
Weight distribution when calculating.
7. a kind of efficient robot data similarity calculation algorithm according to claim 6, which is characterized in that institute's predicate
Language similarity COMPREHENSIVE CALCULATING formula are as follows:
Wherein, t1And t2For the term pair for needing to calculate Words similarity, sim (wt1i,wt2j) indicate term t1In i-th of word
With term t2In similarity between j-th of word, Weight (wt1i,wt2j) indicate according to between subject term in respective term
Distance is wt1iAnd wt2jThe similarity weight of distribution, in denominator the value range of l be [0, Max (| t1|,|t2|) -1], it is ensured that
Entire calculated result is in always in [0,1] range.
8. a kind of efficient robot data similarity calculation algorithm according to claim 7, which is characterized in that the step
Multiple mapping merges in rapid d and mapping result exports, and specifically includes: according to analysis of experimental data, optimal trusted degree threshold value is set,
The concept pair for having more than this threshold value is filtered out, in the Ontological concept pair to be matched for being more than threshold value, will form between concept
The mapping relations of 1:n, m:1 or m:n, in order to be translated into 1:1 mapping relations, manually to mark mapping knot with domain expert
Fruit collection compares, and needs to select the highest concept of mapping of similarity to set from these qualified concept groups again, make
It is exported for the result of Ontology Mapping；And output and storage are formatted to Ontology Mapping result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811433367.7A CN109582961A (en) | 2018-11-28 | 2018-11-28 | A kind of efficient robot data similarity calculation algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811433367.7A CN109582961A (en) | 2018-11-28 | 2018-11-28 | A kind of efficient robot data similarity calculation algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109582961A true CN109582961A (en) | 2019-04-05 |
Family
ID=65924735
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811433367.7A Pending CN109582961A (en) | 2018-11-28 | 2018-11-28 | A kind of efficient robot data similarity calculation algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109582961A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110704522A (en) * | 2019-09-04 | 2020-01-17 | 南京航空航天大学 | Concept data model automatic conversion method based on semantic analysis |
| CN111768821A (en) * | 2020-05-29 | 2020-10-13 | 上海森亿医疗科技有限公司 | Distributed patient record matching method, system and terminal |
| WO2021051763A1 (en) * | 2019-09-16 | 2021-03-25 | 深圳中兴网信科技有限公司 | Term matching method and apparatus, terminal, and computer readable storage medium |
| CN113326686A (en) * | 2020-02-28 | 2021-08-31 | 株式会社斯库林集团 | Similarity calculation device, recording medium, and similarity calculation method |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104699767A (en) * | 2015-02-15 | 2015-06-10 | 首都经济贸易大学 | Large-scale ontology mapping method for Chinese languages |
| CN107885749A (en) * | 2016-09-30 | 2018-04-06 | 南京理工大学 | Ontology extends the process knowledge search method with collaborative filtering Weighted Fusion |
-
2018
- 2018-11-28 CN CN201811433367.7A patent/CN109582961A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104699767A (en) * | 2015-02-15 | 2015-06-10 | 首都经济贸易大学 | Large-scale ontology mapping method for Chinese languages |
| CN107885749A (en) * | 2016-09-30 | 2018-04-06 | 南京理工大学 | Ontology extends the process knowledge search method with collaborative filtering Weighted Fusion |
Non-Patent Citations (1)
| Title |
|---|
| 徐健 等: "一种基于词语相似度计算的本体映射方法", 《现代图书情报技术》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110704522A (en) * | 2019-09-04 | 2020-01-17 | 南京航空航天大学 | Concept data model automatic conversion method based on semantic analysis |
| CN110704522B (en) * | 2019-09-04 | 2023-04-07 | 南京航空航天大学 | Concept data model automatic conversion method based on semantic analysis |
| WO2021051763A1 (en) * | 2019-09-16 | 2021-03-25 | 深圳中兴网信科技有限公司 | Term matching method and apparatus, terminal, and computer readable storage medium |
| CN113326686A (en) * | 2020-02-28 | 2021-08-31 | 株式会社斯库林集团 | Similarity calculation device, recording medium, and similarity calculation method |
| CN113326686B (en) * | 2020-02-28 | 2024-05-10 | 株式会社斯库林集团 | Similarity calculation device, recording medium, and similarity calculation method |
| CN111768821A (en) * | 2020-05-29 | 2020-10-13 | 上海森亿医疗科技有限公司 | Distributed patient record matching method, system and terminal |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Grishman | Information extraction | |
| CN109582961A (en) | A kind of efficient robot data similarity calculation algorithm | |
| Kashyap et al. | TaxaMiner: an experimentation framework for automated taxonomy bootstrapping | |
| Sarkar et al. | A new approach to keyphrase extraction using neural networks | |
| CN106844658A (en) | A kind of Chinese text knowledge mapping method for auto constructing and system | |
| US20110208776A1 (en) | Method and apparatus of semantic technological approach based on semantic relation in context and storage media having program source thereof | |
| US9659001B2 (en) | Query evaluation using ancestor information | |
| Lin et al. | An integrated approach to extracting ontological structures from folksonomies | |
| CN101685455A (en) | Method and system of data retrieval | |
| Thijs | Using neural-network based paragraph embeddings for the calculation of within and between document similarities | |
| Janik et al. | Training-less ontology-based text categorization | |
| CN114997288A (en) | Design resource association method | |
| Babekr et al. | Personalized semantic retrieval and summarization of web based documents | |
| Ramachandran et al. | A Novel Method for Text Summarization and Clustering of Documents | |
| Luong et al. | Ontology learning using word net lexical expansion and text mining | |
| Chifu et al. | Unsupervised semantic annotation of Web service datatypes | |
| Bounhas et al. | A hierarchical approach for semi-structured document indexing and terminology extraction | |
| Chen | English translation template retrieval based on semantic distance ontology knowledge recognition algorithm | |
| Wang et al. | A graph-based approach for semantic similar word retrieval | |
| Abel et al. | The impact of multifaceted tagging on learning tag relations and search | |
| Boese et al. | Semantic document networks to support concept retrieval | |
| Yu et al. | Performance analysis and assessment of a tf-idf based archetype-SNOMED-CT binding algorithm | |
| Lauser | Semi-automatic ontology engineering and ontology supported document indexing in a multilingual environment | |
| Pernelle | Semantic enrichment of data: annotation and data linking | |
| Mousavi et al. | Mining semantics structures from syntactic structures in web document corpora |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190405 |
|
| RJ01 | Rejection of invention patent application after publication |