CN106776827A - Method for automating extension stratification ontology knowledge base - Google Patents
Method for automating extension stratification ontology knowledge base Download PDFInfo
- Publication number
- CN106776827A CN106776827A CN201611059615.7A CN201611059615A CN106776827A CN 106776827 A CN106776827 A CN 106776827A CN 201611059615 A CN201611059615 A CN 201611059615A CN 106776827 A CN106776827 A CN 106776827A
- Authority
- CN
- China
- Prior art keywords
- classification
- entity
- novel entities
- stratification
- similarity
- 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
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/22—Indexing; Data structures therefor; Storage structures
- G06F16/2228—Indexing structures
- G06F16/2246—Trees, e.g. B+trees
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/284—Relational databases
- G06F16/288—Entity relationship models
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a kind of method for automating extension stratification ontology knowledge base, step is：(1) class relations tree is built to already present stratification ontology library；(2) feature of novel entities of all categories and pre-inserted in above-mentioned stratification ontology library is extracted；(3) feature for obtaining is extracted using step (2), in the class relations tree where the stratification ontology library, the top-down similarity for calculating the novel entities and node of all categories in class relations tree, (4) novel entities and the entity in classification to the pre-inserted are processed according to one of following situations, so that the extension of implementation level ontology knowledge base.The present invention realizes the fusion of novel entities of the stratification ontology knowledge base to being continuously emerged in semantic net, advantageously form an ontology library for unified standard, preferably realize that knowledge sharing and interoperability provide help for semantic net, realize based on semantic expression and reasoning, be further to set up believable semantic net to lay a good foundation.
Description
Technical field
The concept of body (Ontology) originates from philosophy field [1] earliest, and letter is widely used in as semantic basis
The fields such as breath retrieval, artificial intelligence, semantic network, soft project, natural language processing, ecommerce and information management.This hair
The bright method for being related to be extended already present ontology library, is especially carried out dynamic, automatic to stratification ontology knowledge base
The extended method of change.
Background technology
Body illustrates [2] as the clear and definite Formal Specification of shared ideas model, is the core of Semantic Web.In order to abundant
Using already present body, many researchs concentrate on Ontology Mapping, that is, find the semantic relation between isomery body.Mapping techniques can
Two kinds [3] are mapped to be divided into element layer mapping and structure sheaf.Element layer mapping techniques ignore the relation of element and other element, enter
Only consider element in itself during row Ontology Mapping；Structure sheaf is mapping through in one big structure of analysis the relation each other between element to enter
Row mapping.Element layer mapping is often counted as the basis of structure sheaf mapping.Current element layer mapping method mainly has based on word
The technology of string is accorded with, such as compares prefix suffix, calculated [4] such as editing distance, N-gram algorithms；Technology based on language, using certain
Plant natural language (such as English) treatment technology to process input word, such as cut tail treatment, eliminate [5] such as preposition, connection words；
Using linguistics resource, shared knowledge dictionary and domain knowledge dictionary (such as WordNet) are introduced, matched using linguistic relation
[6].Above method is successfully applied in English Ontology Mapping.
But Ontology Mapping is mainly used in two already present isomery bodies, it is impossible to realize a dynamic for single ontology library
Extension, and some methods being successfully applied in English Ontology Mapping are not particularly suited for the Ontology Mapping of Chinese.
[bibliography]
[1] Zhang Xiulan, Jiang Ling Ontological concepts Review Study [J] information journals, 2007,26 (4):527-531.
[2]Thomas R G.A Translation Approach to Potable Ontology
Specification[J].Knowledge Acquisition,1993,02:199-200。
[3]PavelShvaiko,J eromeEuzenat.A Survey of Schema2based Matching
Approaches[J].Journal on Data Semantics(JoDS),IV,LNCS 3730,2005:1462171.。
[4]Do H.H.,Rahm E..COMA2a system for flexiblecombination of schema
matching approaches[J].VeryLarge DataBases Conference(VLDB),2001:610-621。
[5]Giunchiglia F.,Shvaiko P.,and Yatskevich M..S-Match:an algorithm
and an implementation of semantic matching[J].European Semantic Web Symposium
(ESWS),2004:61-75。
[6]Giunchiglia F.,YatskevichM.Element level semantic matching[D]
.ltaly:Dept.of Information andCommunication Technology University ofTrento,
2004。
The content of the invention
For problems of the prior art, the present invention proposes a kind of for automating extension stratification ontology knowledge
Automatically can be added to novel entities in already present stratification ontology library by the method in storehouse, i.e. the method, with to current semantics
The body scale constantly increased in net.
In order to solve the above-mentioned technical problem, it is proposed by the present invention a kind of for automating extension stratification ontology knowledge base
Method, comprises the following steps：
Step one, to already present stratification ontology library build class relations tree；
C of all categories in step 2, the above-mentioned stratification ontology library of extractionjAnd the novel entities e of pre-insertediFeature；
Step 3, c of all categories in the stratification ontology library that obtains is extracted using step 2jAnd the novel entities e of pre-insertedi's
Feature, it is top-down to calculate the novel entities e in the class relations tree where the stratification ontology libraryiWith class relations
C of all categories in treejThe similarity of node, and comprising one of following situations：
Once 1) find the novel entities e of the pre-insertediWith current layer classification cjSimilarity u be higher than Δ u when, Δ u values
It is 0.3, then continues to carry out the comparing of similarity u in the subclass of current layer classification, until by the novel entities of the pre-inserted
eiIt is inserted into the classification c of a leaf nodes of the similarity u higher than Δ ujIn, category cjIn original entity be respectively designated as e '；
If 2) the novel entities e of the pre-insertediWith c of all categories in current layerjSimilarity u be respectively less than or equal to Δ u, then
A classification c is set up in the parent class of current layerj, the classification cjPositioned at leaf node, by the novel entities e of the pre-insertediInsert
Enter to such cjIn, now, classification c will be inserted intojIn entity be designated as entity e '；
Step 4, the novel entities e by pre-insertediIt is inserted into a certain classification c positioned at leaf nodejAfterwards, to the pre-inserted
Novel entities eiWith classification cjIn entity e ' determination of entity name and attribute is carried out according to one of following situations, so as to realize
The extension of stratification ontology knowledge base；
If 1) the novel entities e of the pre-insertediThe classification c of leaf node is located at thisjIn entity e ' have identical reality
Body title, then to the novel entities e of the pre-insertediWith the calculating that entity e ' carries out similarity using equation below (5)；If similar
Degree s is higher than Δ s, and Δ s values are 0.5, by the novel entities e of the pre-insertediWith classification cjIn the attribute of entity e ' closed
And；Otherwise, by the novel entities e of pre-insertediEntity name rename；
If 2) the novel entities e of the pre-insertediEntity name with should be located at leaf node classification cjIn entity e '
Entity name is different, and similarity s is higher than Δ s, then by the novel entities e of the pre-insertediEntity name and classification cjIn reality
The entity name and attribute of body e ' merge respectively.
Compared with prior art, the beneficial effects of the invention are as follows：
The present invention is mainly based upon to novel entities and layer as a kind of automation extended method of stratification ontology knowledge base
The investigation of the similarity of entity contained by classification and classification in secondaryization ontology library, realizes the insertion to novel entities, i.e. body and knows
Know the automation extension in storehouse, realize the fusion of novel entities of the stratification ontology knowledge base to being continuously emerged in semantic net, meanwhile,
An ontology library for unified standard is so advantageously formed, preferably realizes that knowledge sharing and interoperability provide side for semantic net
Help, realize based on semantic expression and reasoning, be further to set up believable semantic net to lay a good foundation.
Brief description of the drawings
Fig. 1 is the classification tree example of stratification ontology library；
Fig. 2 is the process example of top-down insertion novel entities；
Fig. 3 is the process example of the similarity for calculating novel entities and a certain classification.
Specific embodiment
Technical solution of the present invention is described in further detail with specific implementation example below in conjunction with the accompanying drawings, described tool
Body embodiment is only explained to the present invention, is not intended to limit the invention.
A kind of method for automating extension stratification ontology knowledge base proposed by the present invention, comprises the following steps：
Step one, to already present stratification ontology library build class relations tree；As shown in figure 1, already present stratification
Ontology library chooses Chinese wikipedia ontology library, by the customized class label of wikipedia, plus mark wikipedia by hand
Mapping of the suffix word of classification name to classification tree node.The classification for marking " XXX sportsman " in advance should be corresponded to
" people " on classification tree, " XXX companies " should correspond to " organization " on classification tree；When running into a wikipedia entity
When, such as " Yao Ming ", it is necessary first to check its wikipedia class label, there is " basket baller " this class label, and classification
Suffix " sportsman " is marked as " people " this classification on classification tree, then wikipedia entity " Yao Ming " belongs to " people " this class
Not.
C of all categories in step 2, the above-mentioned stratification ontology library of extractionjAnd the feature of the novel entities of pre-inserted, for calculating
Similarity u in above-mentioned stratification ontology library between novel entities of all categories and pre-inserted, is existed with the novel entities for finding pre-inserted
The position that be can be inserted into above-mentioned stratification ontology library, Fig. 3 shows the process of the similarity for calculating novel entities and a certain classification, one
As for, entity is respectively provided with entity name, attribute and property value, therefore the present invention chooses feature of the entity attributes as entity,
All entity attributes set that each classification is possessed are used as such another characteristic.
Step 3, c of all categories in the stratification ontology library that obtains is extracted using step 2jAnd the novel entities e of pre-insertedi's
Feature, in the class relations tree where the stratification ontology library, Fig. 2 shows the mistake of top-down insertion novel entities
Journey, the top-down calculating novel entities eiWith c of all categories in class relations treejThe similarity of node, detailed process is as follows：
For the novel entities e of pre-insertediWith the classification c of above-mentioned stratification ontology libraryj, similarity is calculated using equation below：
First, by classification c in above-mentioned stratification ontology libraryjAnd the novel entities e of pre-insertediIt is respectively converted into vector representation
With
Wherein, wiRepresent pre-inserted novel entities eiIth attribute；
It is belonging to classification c in above-mentioned stratification ontology libraryjThe institute that is included of all entities
There is attribute, secondly, it is possible to use above-mentioned formula (2) or (3) calculate classification c in above-mentioned stratification ontology libraryjAttributeWith
The novel entities e of pre-insertediAttribute VeiThe similarity of [k], wherein formula (2) are to calculate above-mentioned stratification sheet using editing distance
Classification c in body storehousejAttributeWith the novel entities e of pre-insertediSimilarity, formula (3) is by above-mentioned stratification ontology library
Middle classification cjWith the novel entities e of pre-insertediAttributeAttribute vector be respectively converted into the expression V of term vectorkAnd Vm, so
Two term vector V are calculated afterwardskAnd VmCosine value is used as classification c in above-mentioned stratification ontology libraryjWith the novel entities e of pre-insertediBetween
Similarity.
In view of there are some attributes almost to have in each entity, without distinguishing, therefore, we assign and belonging to
State classification c in stratification ontology libraryjIn each attributeOne weightThe present invention is with above-mentioned stratification ontology library
Middle classification cjIn each attributeTF-IDF values as the attributeWeight.Wherein, the TF- of each attribute
IDF value calculating methods are as follows：
Firstly, for classification c in above-mentioned stratification ontology libraryjIn each attributeBelong to above-mentioned stratification body
Classification c in storehousejEntity ej, classification c in above-mentioned stratification ontology libraryjAttributeThe class in above-mentioned stratification ontology library
Other cjEntity ejThe number of times of middle appearance is denoted as tn, the wherein novel entities e of pre-insertediBelong to the classification c in stratification ontology libraryj,
Classification c in above-mentioned stratification ontology libraryjAttributeIn classification cjIn all entities in occur number of times be denoted as tall。
Secondly, we are all comprising attributeClassification number be denoted as dn, total classification number is denoted as dall.Next, we can
With using regular (4) computation attributeWeight
Weight except considering each attribute, the present invention is also added into priori in terms of some natural languages as punishing
Penalize a BRules(ei,cj).For example, most of entity for being possessed of classification " school " all has the name of " ... school " or " ... middle school "
Claim form, still can first collect all of entity name under a classification, participle is carried out to it and high frequency words are obtained, with reference to
Our priori, obtains entity name matching rule, then using equation below, obtains novel entities eiWith classification cjPunish
Penalize item：
Classification c in above-mentioned stratification ontology library is so obtained with by formula (1)jWith the novel entities e of pre-insertediIt
Between similarity.And comprising one of following situations：
Once 1) find the novel entities e of the pre-insertediWith current layer classification cjSimilarity u higher than Δ u ((Δ u be through
Threshold value is tested, is typically taken when 0.3), then continue to carry out the comparing of similarity u in the subclass of current layer classification, until will be described
The novel entities e of pre-insertediIt is inserted into the classification c of a leaf nodes of the similarity u higher than Δ ujIn, category cjIn original reality
Body is respectively designated as e '；
If 2) the novel entities e of the pre-insertediWith c of all categories in current layerjSimilarity u be respectively less than or equal to Δ u, then
A classification c is set up in the parent class of current layerj, the classification cjPositioned at leaf node, by the novel entities e of the pre-insertediInsert
Enter to category cjIn, now, classification c will be inserted intojIn entity be designated as entity e '；
What the generation of new category was utilized in the present invention is hierarchy clustering method.Each is not found the new reality of classification first
The attribute of body is sorted by pinyin order, then carries out participle to it, and each word is represented with the term vector for training, and obtains new reality
The vector representation of body.Each novel entities is considered as a cluster in the starting stage, merging two is immediate each time afterwards
Cluster.
Step 4, the novel entities e by pre-insertediIt is inserted into a certain classification c positioned at leaf nodejWhen, it is necessary to described pre-
The novel entities e of insertioniWith classification cjIn entity e ' judged, and carry out entity name and attribute according to one of following situations
Determination so that the extension of implementation level ontology knowledge base；
1) whether existing entity has same names with novel entities in ontology library, and refers to identical semantic content, if
In the presence of then being merged with already present entity to novel entities；Particular content is：If the novel entities e of the pre-insertediWith classification
cjIn entity e ' have identical entity name, then to the novel entities e of the pre-insertediWith entity e ', using equation below
(5) calculating of similarity is carried out；If similarity s is higher than Δ s, and (Δ s is empirical value, is typically taken 0.5), by the pre-inserted
Novel entities eiWith classification cjIn the attribute of entity e ' merge；Have with novel entities with the presence or absence of entity in ontology library identical
Title, but refer to different semantic contents, if in the presence of, disambiguation and differentiation are carried out to novel entities and already present entity, will
The novel entities e of pre-insertediEntity name rename；
If 2) the novel entities e of the pre-insertediEntity name and classification cjIn entity e ' entity name it is different, but
It is higher than Δ s to refer to identical semantic content, i.e. similarity s, then by the novel entities e of the pre-insertediEntity name and classification
Middle cjEntity e ' entity name and attribute merge respectively.
First to the novel entities e of the pre-insertediWith classification cjIn entity e ' carry out name-matches, if there is reality of the same name
Body, then calculate the novel entities e of the pre-inserted by equation belowiWith classification cjIn entity e ' between similarity：
SimEE(ei,ej)=(Si*Sj)/(||Si||×||Sj||) (5)
Wherein, the novel entities e of pre-insertediWith classification cjIn entity e ' be converted into vector representationWithFor reality
Body vector is for exampleWithIncluding entity name, entity attribute, entity attribute
Value.With v={ w1,w2,...wlRepresent vector representationWithUnion.For each word wv∈ v, use following formula
(6) each word w is calculatedvAnd eachEditing distance
WhereinRepresent wvWithBetween editing distance, | | wv| | andRepresent the length of vector
Degree.
Then, maximum s is selected as wvWithBetween semantic similarity, by computing repeatedly each element in v,
We can obtain v andBetween semantic similarity vector, be expressed as Si={ si1,si2,...sin, repeat same step
Suddenly, we can obtain v andBetween semantic similarity vector.
Although above in conjunction with accompanying drawing, invention has been described, the invention is not limited in above-mentioned specific implementation
Mode, above-mentioned specific embodiment is only schematical, and rather than restricted, one of ordinary skill in the art is at this
Under the enlightenment of invention, without deviating from the spirit of the invention, many variations can also be made, these belong to of the invention
Within protection.
Claims (1)
1. it is a kind of for automating the method for extending stratification ontology knowledge base, it is characterised in that to comprise the following steps：
Step one, to already present stratification ontology library build class relations tree；
C of all categories in step 2, the above-mentioned stratification ontology library of extractionjAnd the novel entities e of pre-insertediFeature；
Step 3, c of all categories in the stratification ontology library that obtains is extracted using step 2jAnd the novel entities e of pre-insertediFeature,
It is top-down to calculate the novel entities e in the class relations tree where the stratification ontology libraryiIt is each with class relations tree
Classification cjThe similarity of node, and comprising one of following situations：
Once 1) find the novel entities e of the pre-insertediWith current layer classification cjSimilarity u be higher than Δ u when, Δ u values are
0.3, then continue to carry out the comparing of similarity u in the subclass of current layer classification, until by the novel entities e of the pre-insertedi
It is inserted into the classification c of a leaf nodes of the similarity u higher than Δ ujIn, category cjIn original entity be respectively designated as e '；
If 2) the novel entities e of the pre-insertediWith c of all categories in current layerjSimilarity u be respectively less than or equal to Δ u, then working as
The parent class of front layer sets up a classification cj, the classification cjPositioned at leaf node, by the novel entities e of the pre-insertediIt is inserted into
Category cjIn, now, classification c will be inserted intojIn entity be designated as entity e '；
Step 4, the novel entities e by pre-insertediIt is inserted into a certain classification c positioned at leaf nodejAfterwards, to the new of the pre-inserted
Entity eiWith classification cjIn entity e ' determination of entity name and attribute is carried out according to one of following situations so that implementation level
Change the extension of ontology knowledge base；
If 1) the novel entities e of the pre-insertediThe classification c of leaf node is located at thisjIn entity e ' have identical physical name
Claim, then to the novel entities e of the pre-insertediWith the calculating that entity e ' carries out similarity using equation below (5)；If similarity s
Higher than Δ s, Δ s values are 0.5, by the novel entities e of the pre-insertediWith classification cjIn the attribute of entity e ' merge；It is no
Then, by the novel entities e of pre-insertediEntity name rename；
If 2) the novel entities e of the pre-insertediEntity name with should be located at leaf node classification cjIn entity e ' entity
Title is different, and similarity s is higher than Δ s, then by the novel entities e of the pre-insertediEntity name and classification cjIn entity e '
Entity name and attribute merge respectively.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611059615.7A CN106776827A (en) | 2016-11-24 | 2016-11-24 | Method for automating extension stratification ontology knowledge base |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611059615.7A CN106776827A (en) | 2016-11-24 | 2016-11-24 | Method for automating extension stratification ontology knowledge base |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106776827A true CN106776827A (en) | 2017-05-31 |
Family
ID=58912673
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611059615.7A Pending CN106776827A (en) | 2016-11-24 | 2016-11-24 | Method for automating extension stratification ontology knowledge base |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106776827A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107247709A (en) * | 2017-07-28 | 2017-10-13 | 广州多益网络股份有限公司 | The optimization method and system of a kind of encyclopaedia entry label |
| CN107870898A (en) * | 2017-10-11 | 2018-04-03 | 广州极天信息技术股份有限公司 | A kind of domain semanticses network modeling method of Engineering Oriented application |
| CN109960722A (en) * | 2019-03-31 | 2019-07-02 | 联想(北京)有限公司 | A kind of information processing method and device |
| CN112463788A (en) * | 2019-09-09 | 2021-03-09 | 北京国双科技有限公司 | Entity organization tree construction method, entity calling method and product in building control system |
| CN113886535A (en) * | 2021-09-18 | 2022-01-04 | 前海飞算云创数据科技(深圳)有限公司 | Knowledge graph-based question and answer method and device, storage medium and electronic equipment |
| CN116401567A (en) * | 2023-06-02 | 2023-07-07 | 支付宝(杭州)信息技术有限公司 | Clustering model training, user clustering and information pushing method and device |
-
2016
- 2016-11-24 CN CN201611059615.7A patent/CN106776827A/en active Pending
Non-Patent Citations (3)
| Title |
|---|
| 付秋实: "效应领域本体库自动填充方法研究", 《中国优秀硕士论文全文数据库》 * |
| 刘威: "基于中文文本的本体构建方法研究", 《中国优秀硕士论文全文数据库》 * |
| 陈振亚等: "利用术语本体关系扩展ＳＢＮ检索模型", 《计算机研究与发展》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107247709A (en) * | 2017-07-28 | 2017-10-13 | 广州多益网络股份有限公司 | The optimization method and system of a kind of encyclopaedia entry label |
| CN107247709B (en) * | 2017-07-28 | 2021-03-16 | 广州多益网络股份有限公司 | Encyclopedic entry label optimization method and system |
| CN107870898A (en) * | 2017-10-11 | 2018-04-03 | 广州极天信息技术股份有限公司 | A kind of domain semanticses network modeling method of Engineering Oriented application |
| CN107870898B (en) * | 2017-10-11 | 2021-09-14 | 广州极天信息技术股份有限公司 | Domain semantic web modeling method oriented to engineering application |
| CN109960722A (en) * | 2019-03-31 | 2019-07-02 | 联想(北京)有限公司 | A kind of information processing method and device |
| CN109960722B (en) * | 2019-03-31 | 2021-10-22 | 联想(北京)有限公司 | Information processing method and device |
| CN112463788A (en) * | 2019-09-09 | 2021-03-09 | 北京国双科技有限公司 | Entity organization tree construction method, entity calling method and product in building control system |
| CN113886535A (en) * | 2021-09-18 | 2022-01-04 | 前海飞算云创数据科技(深圳)有限公司 | Knowledge graph-based question and answer method and device, storage medium and electronic equipment |
| CN116401567A (en) * | 2023-06-02 | 2023-07-07 | 支付宝(杭州)信息技术有限公司 | Clustering model training, user clustering and information pushing method and device |
| CN116401567B (en) * | 2023-06-02 | 2023-09-08 | 支付宝(杭州)信息技术有限公司 | Clustering model training, user clustering and information pushing method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106776711B (en) | Chinese medical knowledge map construction method based on deep learning | |
| CN106776827A (en) | Method for automating extension stratification ontology knowledge base | |
| CN105653706B (en) | A kind of multilayer quotation based on literature content knowledge mapping recommends method | |
| CN104199972B (en) | A kind of name entity relation extraction and construction method based on deep learning | |
| CN107766585B (en) | Social network-oriented specific event extraction method | |
| Cimiano et al. | Learning concept hierarchies from text with a guided agglomerative clustering algorithm | |
| Magatti et al. | Automatic labeling of topics | |
| CN106570191B (en) | Chinese-English cross-language entity matching method based on Wikipedia | |
| Lin et al. | An integrated approach to extracting ontological structures from folksonomies | |
| CN103678275A (en) | Two-level text similarity calculation method based on subjective and objective semantics | |
| CN108647322B (en) | Method for identifying similarity of mass Web text information based on word network | |
| CN110532390B (en) | News keyword extraction method based on NER and complex network characteristics | |
| CN107679110A (en) | The method and device of knowledge mapping is improved with reference to text classification and picture attribute extraction | |
| CN110750995A (en) | File management method based on user-defined map | |
| Sarkar et al. | A comparative analysis of particle swarm optimization and K-means algorithm for text clustering using Nepali Wordnet | |
| Adhitama et al. | Topic labeling towards news document collection based on Latent Dirichlet Allocation and ontology | |
| Hanyurwimfura et al. | A centroid and relationship based clustering for organizing | |
| CN109086443A (en) | Social media short text on-line talking method based on theme | |
| Nasution | Extracting keyword for disambiguating name based on the overlap principle | |
| Tounsi Dhouib et al. | Measuring clusters of labels in an embedding space to refine relations in ontology alignment | |
| Alfarra et al. | Graph-based technique for extracting keyphrases in a single-document (gtek) | |
| CN108427769B (en) | Character interest tag extraction method based on social network | |
| CN116244497A (en) | Cross-domain paper recommendation method based on heterogeneous data embedding | |
| Muthukrishnan et al. | Simultaneous similarity learning and feature-weight learning for document clustering | |
| Faty et al. | News comments modeling for opinion mining: the case of senegalese online press |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170531 |