CN109101583A - A kind of knowledge mapping construction method and system for non-structured text - Google Patents
A kind of knowledge mapping construction method and system for non-structured text Download PDFInfo
- Publication number
- CN109101583A CN109101583A CN201810812091.7A CN201810812091A CN109101583A CN 109101583 A CN109101583 A CN 109101583A CN 201810812091 A CN201810812091 A CN 201810812091A CN 109101583 A CN109101583 A CN 109101583A
- Authority
- CN
- China
- Prior art keywords
- knowledge mapping
- feature database
- knowledge
- supervision
- structured text
- 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
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
Inventive network field of communication technology is related to a kind of knowledge mapping construction method and system for non-structured text.The method of the present invention includes: S1, extracts characteristic relation, establishes feature database；S2, creation supervision sample；S3 establishes knowledge base by training relational model；S4 calls institutional data in distributed data base to form knowledge mapping.The present invention is directed to a large amount of unstructured information, therefrom extracts effective structured message, and analyze the relationship between them, so that knowledge mapping is formed, the efficiency that promotion obtains effective information and technical staff solves the problems, such as.
Description
Technical field
The invention belongs to data analysis technique field more particularly to a kind of knowledge mapping buildings for non-structured text
Method and system.
Background technique
With the rapid development of digital Age, various data are all flooded with around people all the time, it be from this
The information interested to us is extracted in a little data becomes the big problem of present one, and data mining is exactly rich and varied from these
Data in obtain concern information.And in flight-line maintenance field, most mantenance data is all non-structured data, from
The structured messages such as associated trouble unit, problem, reason, solution are obtained in past data, can be practitioner
Very big help is provided, for reach TB (Trillionbyte, terabyte, terabyte) even PB (Petabyte, ten million
Hundred million bytes, petabyte) rank unstructured fault information data carry out key message extraction, feature extraction and relationship analysis
Deng, the knowledge base about fault message is established, knowledge mapping is formed, provides support for technical staff's handling failure problem, promotion
Troubleshooting efficiency.
Therefore, the demand for constructing the knowledge mapping of such data is more more and more intense.
Application No. is the domestic patents of invention of CN201710109316.8 to disclose a kind of knowledge mapping construction method and dress
It sets, the method for specifically disclosing includes: based on object language, and building is directed to the rudimental knowledge map of object, object language
It is less than the lightweight data interchange format of RDF language for complexity, includes various needed for semantic understanding in rudimental knowledge map
Key element, various key elements are stored in same file；From at least one data source, collection and various key elements
At least one matched industry data of key element；Industry data is added at least one of rudimental knowledge map key element
The position of instruction obtains the object knowledge map of object.Due to being less than the lightweight object language of RDF language based on complexity
Knowledge mapping is constructed, therefore knowledge mapping readability and maintainability are preferable, can promote the chat effect of chat robots.This
Outside, the various key elements in same file are contained needed for semantic understanding and be stored in knowledge mapping, are carried out convenient for unified
Management.The patent of invention is explained in detail how to construct knowledge mapping, obtains data source and matches with key request indicating positions,
Knowledge mapping is obtained, but is only applicable to the lightweight data that complexity is less than RDF language, while to the knowledge of non-structured text
Map can not construct well, so, it would be desirable to one kind is designed to a large amount of unstructured information, is therefrom extracted effective
Structured message, and the relationship between them is analyzed, to form the method and system of knowledge mapping.
Summary of the invention
In view of the problems of the existing technology the present invention, proposes a kind of knowledge mapping building for non-structured text
Method and system.
The technical scheme is that:
A kind of knowledge mapping construction method for non-structured text, comprising the following steps:
S1 extracts characteristic relation, establishes feature database；
S2, creation supervision sample；
S3 establishes knowledge base by training relational model；
S4 calls institutional data in distributed data base to form knowledge mapping.
As the preferred of the technical program, include: before the step S1
S0 obtains a large amount of unstructured text datas.
As the preferred of the technical program, the step S1 includes:
S1.1 segments the unstructured text data, part-of-speech tagging, name entity mark, dependence point
Analysis；
S1.2 is extracted according to the result of entity mark in terms of extracting trouble unit, problem, reason with sentence dependence
Crucial phrase or phrase, the markup information between analysis of key phrase；
S1.3, in conjunction with the markup information, analysis obtains characteristic relation, and feature database is established in extraction.
As the preferred of the technical program, the step S2 includes:
S2.1 calls the feature database；
S2.2 automatically creates supervision sample using the method far supervised；
S2.3 analyzes the supervision sample, is marked to form positive and negative sample set according to dependence.
As the preferred of the technical program, the step S3 includes:
S3.1 imports the feature database in factor graph model；
S3.2 calls the positive and negative sample set to exercise supervision；
S3.3, training form knowledge base and are stored in distributed data base.
A kind of knowledge mapping building system for non-structured text, comprising:
Feature database establishes module, for extracting characteristic relation, establishes feature database；
Sample creation module is supervised, for creating supervision sample；
Knowledge base establishes module, for establishing knowledge base by training relational model；
Knowledge mapping forms module, for calling institutional data in distributed data base to form knowledge mapping.
As the preferred of the technical program, the feature database establishes module and includes: before
Unstructured text data obtains module, for obtaining a large amount of unstructured text datas.
As the preferred of the technical program, the feature database establishes module and includes:
Participle unit, for being segmented to the unstructured text data, part-of-speech tagging, name entity mark, according to
Deposit relationship analysis；
Unit is marked, the result for marking according to entity is extracted and sentence dependence extracts trouble unit, problem, original
Markup information because of the crucial phrase or phrase of aspect, between analysis of key phrase；
Feature database establishes unit, and for analyzing and obtaining characteristic relation in conjunction with the markup information, feature database is established in extraction.
As the preferred of the technical program, the supervision sample creation module includes:
Call unit, for calling the feature database；
Creating unit, for automatically creating supervision sample using the method far supervised；
Marking unit marks to form positive and negative sample set according to dependence for analyzing the supervision sample.
As the preferred of the technical program, the knowledge base establishes module and includes:
Import unit, for importing the feature database in factor graph model；
Supervision unit, for calling the positive and negative sample set to exercise supervision；
Training unit forms knowledge base for training and is stored in distributed distributed data base.
The technical program has the beneficial effect that
The creation that knowledge base is carried out for mass data, by establishing feature database, creation supervision sample, training relational model
Knowledge base is established, whole process realizes full-automation, is finally stored in the data information of the structuring of extraction with relation information
Among distributed database, the creation of knowledge mapping is easily realized.
Detailed description of the invention
Fig. 1 is a kind of flow chart of the knowledge mapping construction method for non-structured text of the present invention；
Fig. 2 is a kind of flow chart of the knowledge mapping construction method step S1 for non-structured text of the present invention；
Fig. 3 is a kind of flow chart of the knowledge mapping construction method step S2 for non-structured text of the present invention；
Fig. 4 is a kind of flow chart of the knowledge mapping construction method step S3 for non-structured text of the present invention；
Fig. 5 is the block diagram that a kind of knowledge mapping for non-structured text of the present invention constructs system；
Fig. 6 is that a kind of knowledge mapping for non-structured text of the present invention constructs the frame that module is established in system features library
Figure；
Fig. 7 is the frame that a kind of knowledge mapping for non-structured text of the present invention constructs system monitor sample creation module
Figure；
Fig. 8 is that a kind of knowledge mapping for non-structured text of the present invention constructs the frame that system knowledge base establishes module
Figure；
Fig. 9 is the example schematic of knowledge mapping of the present invention.
Specific implementation method
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.It should be appreciated that described herein, specific examples are only used to explain the present invention, not
For limiting the present invention.Based on the embodiments of the present invention, those of ordinary skill in the art are not before making creative work
Every other embodiment obtained is put, shall fall within the protection scope of the present invention.In addition, the technical side between each embodiment
Case can be combined with each other, but must be based on can be realized by those of ordinary skill in the art, when the combination of technical solution
Conflicting or cannot achieve when occur will be understood that the combination of this technical solution is not present, also not the present invention claims guarantor
Within the scope of shield.
Following is a specific embodiment of the present invention in conjunction with the accompanying drawings, technical scheme of the present invention will be further described,
However, the present invention is not limited to these examples.
Embodiment 1
As shown in Figure 1, for a kind of flow chart of the knowledge mapping construction method for non-structured text of the present invention.
Reach the TB even unstructured fault information data of PB rank to solve flight-line maintenance FIELD Data amount, establishes and close
In the knowledge base of fault message, knowledge mapping is formed, a kind of knowledge mapping construction method for non-structured text is devised.
Its method the following steps are included:
S1 extracts characteristic relation, establishes feature database.
S2, creation supervision sample.
S3 establishes knowledge base by training relational model.
S4 calls institutional data in distributed data base to form knowledge mapping.
For a large amount of non-structured text information, effective structured message is therefrom extracted, and analyze between them
Relationship, to form knowledge mapping.
Knowledge mapping, also referred to as mapping knowledge domains are known as knowledge domain visualization or ken in books and information group
Map is mapped, is a series of a variety of different figures of explicit knowledge's development process and structural relation, is described with visualization technique
Knowledge resource and its carrier, excavation, analysis, building, drafting and the correlation between explicit knowledge and knowledge.At present, pass through
Knowledge mapping is constructed, and the knowledge mapping of building is applied into mechanical breakdown and determines and find method for maintaining, is had become
A kind of way that those skilled in the art take extensively.
A large amount of non-structured texts are imported into distributed data base, non-structured text number is read from distributed data base
It is segmented according to and to unstructured text data, part-of-speech tagging, name entity mark, dependency analysis, according to entity mark
Infuse structure extraction trouble unit, problem, the key messages such as reason, then combine part-of-speech tagging between key message, entity mark,
Dependence annotation results establish feature database.The cause and effect dependence of feature, analysis creation supervision sample are extracted according to feature database.
The feature database transduced element graph model is trained, sample monitoring uses the fault message text of structuring, finally, passing through
The model trained completes the creation of knowledge base and the presentation of knowledge mapping.
S0 obtains a large amount of unstructured text datas.
The present embodiment carries out failure primarily with respect to the unstructured fault information data of the high-volume such as vehicle, ship or aircraft
The extraction of problem, failure cause, solution.Before establishing professional domain dictionary and mark dictionary, it is necessary first to obtain big
It measures unstructured text data and carries out subsequent analysis operating process in this, as basis.
As shown in Fig. 2, for a kind of process of the knowledge mapping construction method step S1 for non-structured text of the present invention
Figure.
S1 extracts characteristic relation, establishes feature database.
It establishes the dictionary of professional domain and carries out related mark, it is ensured that the professional phrase of fault message can be separated by participle
With subsequent mark, use stanford corenlp (being a set of tool with processing natural language by Stanford University's open source)
A large amount of unstructured text datas are segmented, part-of-speech tagging, name entity mark, dependency analysis, in participle process
It is middle to need to load the professional domain dictionary created just now and mark dictionary, by the result analyzed according to the correspondence of participle and mark
Relationship is stored among distributed data base, is extracted according to the result of entity mark and sentence dependence is extracted trouble unit, asked
Topic, the crucial phrase in terms of reason or phrase, the markup information between analysis of key phrase, including part-of-speech tagging, entity mark
These markup informations are combined together to form feature according to certain rules by the information of note and dependence mark, construct institute
Show feature database.
The step S1 includes:
S1.1 segments the unstructured text data, part-of-speech tagging, name entity mark, dependence point
Analysis.
S1.2 is extracted according to the result of entity mark in terms of extracting trouble unit, problem, reason with sentence dependence
Crucial phrase or phrase, the markup information between analysis of key phrase.
S1.3, in conjunction with the markup information, analysis obtains characteristic relation, and feature database is established in extraction.
As shown in Table 1
Table one
The first row in table one indicates one section of non-structured text about vehicle failure maintenance, we therefrom obtain useful
Information needs just to be able to achieve after running through understanding for whole section.Now we it is segmented using stanford corenlp, word
Property mark, name entity mark, interdependent syntactic analysis, to the component entities such as engine, nozzle tip be labeled as PARTS (token name
Title can be made by oneself), the words such as " shake ", " sending " are labeled as Q (question, problem), and the words such as " obstruction " are labeled as R
(reason, reason), the verbs such as " cleaning " are labeled as S (solve, settling mode), the generation side of crucial phrase are described below
Method.
The verb for finding the expression failure problems labeled as Q, utilizes nsubj (nominal subject), the dobj of interdependent syntax
Syntactic relations such as (dynamic guest's phrases) are combined, such as: engine shake makes a sound；Find the expression failure original labeled as R
The verb of cause is combined also according to syntactic relations such as nsubj (nominal subject), dobj (dynamic guest's phrase), such as oil spout
Head obstruction；The verb for finding the expression solution labeled as S, is combined according to mentioned above principle, such as: cleaning nozzle tip.
The problem of thus extracting expression fault message, reason, the crucial phrase of three aspects of solution.
Find the word that name entity indicia is Q and R, using the markup information of sentence between them establish problem and reason it
Between feature be Q-R relationship, such as the sentence of italic overstriking in table, pass through markup information between " sending " and " obstruction " two words
Feature is established, takes the markup information near the two words within three words combine establishing feature, only generally in order to reduce spy
Levy the size and calculation amount in library.Similarly, the word that name entity indicia is R and S is found, such as " obstruction " and " cleaning ", with identical
The feature established between problem and settling mode of mode be Q-S relationship.
The feature database is collectively formed in finally all Q-R relation datas and Q-S relation data.
As shown in figure 3, for a kind of process of the knowledge mapping construction method step S2 for non-structured text of the present invention
Figure.
S2, creation supervision sample.
It is exercised supervision the automatically creating of sample using the method far supervised.The label of data set is in distributed data base
In unreliable situation (unreliable can be here marks incorrect, a variety of labels, marks insufficient, local flag etc.), far
Supervised learning method is more applicable for when supervision message is imperfect or indefinite object.
The step S2 includes:
S2.1 calls the feature database.
S2.2 automatically creates supervision sample using the method far supervised.
S2.3 analyzes the supervision sample, is marked to form positive and negative sample set according to dependence.
Exercised supervision the automatically creating of sample using the method far supervised, name entity indicia between Q and the word of R if
When there are the words such as " ", " because ", " due to ", we are as the positive sample in failure problems and reason this pair of of relationship
This, the too big label of other differences is；Name entity indicia between R and the word of S if there is " solution ", " reality
When row ", " executions ", the words such as " normal ", we as failure cause and solution this to the positive sample in relationship,
It completes incoherent sentence and is labeled as negative sample.The rule citing for only establishing supervision sample above, can also add more
Rule establish positive and negative sample set, these rules are generally realized by writing the mode of script (shell, python etc.).
As shown in figure 4, for a kind of process of the knowledge mapping construction method step S3 for non-structured text of the present invention
Figure.
S3 establishes knowledge base by training relational model.
Training pattern uses factor graph model here, by the feature database transduced element artwork obtained in the step S1
Type is trained, and is exercised supervision using positive and negative sample set obtained in the step S2, relational model finally can be obtained, foundation is known
Know library.
The step S3 includes:
S3.1 imports the feature database in factor graph model.
S3.2 calls the positive and negative sample set to exercise supervision.
S3.3, training form knowledge base and are stored in distributed data base.
Unstructured text data is imported into relational model by pretreatment, the Q-R relation data of structuring can be obtained
With R-S relation data, final relation data is stored in distributed database, and here it is the relationship knowledge established
Library.
S4 calls institutional data in distributed data base to form knowledge mapping.
Finally the structural data in database is called to be presented, knowledge mapping can be formed, as shown in Figure 9.
Embodiment 2
For based on the system on 1 basis of embodiment.
As shown in figure 5, constructing the block diagram of system for a kind of knowledge mapping for non-structured text of the present invention.
A kind of knowledge mapping building system for non-structured text, comprising:
Feature database establishes module, for extracting characteristic relation, establishes feature database.
Sample creation module is supervised, for creating supervision sample.
Knowledge base establishes module, for establishing knowledge base by training relational model.
Knowledge mapping forms module, for calling institutional data in distributed data base to form knowledge mapping.
The feature database establishes module
Unstructured text data obtains module, for obtaining a large amount of unstructured text datas.
As shown in figure 5, constructing system features library for a kind of knowledge mapping for non-structured text of the present invention establishes mould
The block diagram of block.
Feature database establishes module, for extracting characteristic relation, establishes feature database.
The feature database establishes module
Participle unit, for being segmented to the unstructured text data, part-of-speech tagging, name entity mark, according to
Deposit relationship analysis.
Unit is marked, the result for marking according to entity is extracted and sentence dependence extracts trouble unit, problem, original
Markup information because of the crucial phrase or phrase of aspect, between analysis of key phrase.
Feature database establishes unit, and for analyzing and obtaining characteristic relation in conjunction with the markup information, feature database is established in extraction.
As shown in fig. 7, constructing the creation of system monitor sample for a kind of knowledge mapping for non-structured text of the present invention
The block diagram of module.
Sample creation module is supervised, for creating supervision sample.
The supervision sample creation module includes:
Call unit, for calling the feature database.
Creating unit, for automatically creating supervision sample using the method far supervised.
Marking unit marks to form positive and negative sample set according to dependence for analyzing the supervision sample.
As shown in figure 8, constructing system knowledge base for a kind of knowledge mapping for non-structured text of the present invention establishes mould
The block diagram of block.
Knowledge base establishes module, for establishing knowledge base by training relational model.
The knowledge base establishes module
Import unit, for importing the feature database in factor graph model；
Supervision unit, for calling the positive and negative sample set to exercise supervision；
Training unit forms knowledge base for training and is stored in distributed distributed data base.
Knowledge mapping forms module, for calling institutional data in distributed data base to form knowledge mapping.
Specific embodiment described herein is only an example for the spirit of the invention.The neck of technology belonging to the present invention
The technical staff in domain can make various modifications or additions to the described embodiments or be replaced using similar method
In generation, however, it does not deviate from the spirit of the invention or beyond the scope of the appended claims.
Claims (10)
1. a kind of knowledge mapping construction method for non-structured text, which comprises the following steps:
S1 extracts characteristic relation, establishes feature database；
S2, creation supervision sample；
S3 establishes knowledge base by training relational model；
S4 calls institutional data in distributed data base to form knowledge mapping.
2. a kind of knowledge mapping construction method for non-structured text according to claim 1, which is characterized in that institute
Include: before stating step S1
S0 obtains a large amount of unstructured text datas.
3. a kind of knowledge mapping construction method for non-structured text according to claim 2, which is characterized in that institute
Stating step S1 includes:
S1.1 segments the unstructured text data, part-of-speech tagging, name entity mark, dependency analysis；
S1.2 extracts the key in terms of extracting trouble unit, problem, reason with sentence dependence according to the result of entity mark
Phrase or phrase, the markup information between analysis of key phrase；
S1.3, in conjunction with the markup information, analysis obtains characteristic relation, and feature database is established in extraction.
4. a kind of knowledge mapping construction method for non-structured text according to claim 3, which is characterized in that institute
Stating step S2 includes:
S2.1 calls the feature database；
S2.2 automatically creates supervision sample using the method far supervised；
S2.3 analyzes the supervision sample, is marked to form positive and negative sample set according to dependence.
5. a kind of knowledge mapping construction method for non-structured text according to claim 4, which is characterized in that institute
Stating step S3 includes:
S3.1 imports the feature database in factor graph model；
S3.2 calls the positive and negative sample set to exercise supervision；
S3.3, training form knowledge base and are stored in distributed data base.
6. a kind of knowledge mapping for non-structured text constructs system characterized by comprising
Feature database establishes module, for extracting characteristic relation, establishes feature database；
Sample creation module is supervised, for creating supervision sample；
Knowledge base establishes module, for establishing knowledge base by training relational model；
Knowledge mapping forms module, for calling institutional data in distributed data base to form knowledge mapping.
7. a kind of knowledge mapping for non-structured text according to claim 6 constructs system, which is characterized in that institute
It states before feature database establishes module and includes:
Unstructured text data obtains module, for obtaining a large amount of unstructured text datas.
8. a kind of knowledge mapping for non-structured text according to claim 7 constructs system, which is characterized in that institute
It states feature database and establishes module and include:
Participle unit for being segmented to the unstructured text data, part-of-speech tagging, names entity mark, interdependent pass
System's analysis；
Unit is marked, the result for marking according to entity is extracted and sentence dependence extracts trouble unit, problem, reason side
The crucial phrase or phrase in face, the markup information between analysis of key phrase；
Feature database establishes unit, and for analyzing and obtaining characteristic relation in conjunction with the markup information, feature database is established in extraction.
9. a kind of knowledge mapping for non-structured text according to claim 8 constructs system, which is characterized in that institute
Stating supervision sample creation module includes:
Call unit, for calling the feature database；
Creating unit, for automatically creating supervision sample using the method far supervised；
Marking unit marks to form positive and negative sample set according to dependence for analyzing the supervision sample.
10. a kind of knowledge mapping for non-structured text according to claim 9 constructs system, which is characterized in that
The knowledge base establishes module
Import unit, for importing the feature database in factor graph model；
Supervision unit, for calling the positive and negative sample set to exercise supervision；
Training unit forms knowledge base for training and is stored in distributed distributed data base.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810812091.7A CN109101583A (en) | 2018-07-23 | 2018-07-23 | A kind of knowledge mapping construction method and system for non-structured text |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810812091.7A CN109101583A (en) | 2018-07-23 | 2018-07-23 | A kind of knowledge mapping construction method and system for non-structured text |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109101583A true CN109101583A (en) | 2018-12-28 |
Family
ID=64847225
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810812091.7A Pending CN109101583A (en) | 2018-07-23 | 2018-07-23 | A kind of knowledge mapping construction method and system for non-structured text |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109101583A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110135598A (en) * | 2019-05-16 | 2019-08-16 | 兰州交通大学 | A kind of high-speed railway electricity business maintenance auxiliary system of knowledge based service |
| CN110543630A (en) * | 2019-08-21 | 2019-12-06 | 北京仝睿科技有限公司 | Method and device for generating text structured representation and computer storage medium |
| CN111209472A (en) * | 2019-12-24 | 2020-05-29 | 中国铁道科学研究院集团有限公司电子计算技术研究所 | Railway accident fault association and accident fault reason analysis method and system |
| CN111435366A (en) * | 2019-01-14 | 2020-07-21 | 阿里巴巴集团控股有限公司 | Equipment fault diagnosis method and device and electronic equipment |
| CN112163681A (en) * | 2020-10-15 | 2021-01-01 | 珠海格力电器股份有限公司 | Equipment fault cause determination method, storage medium and electronic equipment |
| CN112380354A (en) * | 2020-11-13 | 2021-02-19 | 哈尔滨工业大学 | Knowledge acquisition method and device for overall design of spacecraft and storage medium |
-
2018
- 2018-07-23 CN CN201810812091.7A patent/CN109101583A/en active Pending
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111435366A (en) * | 2019-01-14 | 2020-07-21 | 阿里巴巴集团控股有限公司 | Equipment fault diagnosis method and device and electronic equipment |
| CN110135598A (en) * | 2019-05-16 | 2019-08-16 | 兰州交通大学 | A kind of high-speed railway electricity business maintenance auxiliary system of knowledge based service |
| CN110543630A (en) * | 2019-08-21 | 2019-12-06 | 北京仝睿科技有限公司 | Method and device for generating text structured representation and computer storage medium |
| CN111209472A (en) * | 2019-12-24 | 2020-05-29 | 中国铁道科学研究院集团有限公司电子计算技术研究所 | Railway accident fault association and accident fault reason analysis method and system |
| CN111209472B (en) * | 2019-12-24 | 2023-08-18 | 中国铁道科学研究院集团有限公司电子计算技术研究所 | Railway accident fault association and accident fault cause analysis method and system |
| CN112163681A (en) * | 2020-10-15 | 2021-01-01 | 珠海格力电器股份有限公司 | Equipment fault cause determination method, storage medium and electronic equipment |
| CN112380354A (en) * | 2020-11-13 | 2021-02-19 | 哈尔滨工业大学 | Knowledge acquisition method and device for overall design of spacecraft and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109101583A (en) | A kind of knowledge mapping construction method and system for non-structured text | |
| CN106777275B (en) | Entity attribute and property value extracting method based on more granularity semantic chunks | |
| CN107608949B (en) | A kind of Text Information Extraction method and device based on semantic model | |
| WO2018036239A1 (en) | Method, apparatus and system for monitoring internet media events based on industry knowledge mapping database | |
| CN107391677B (en) | Method and device for generating Chinese general knowledge graph with entity relation attributes | |
| CN110516256A (en) | A kind of Chinese name entity extraction method and its system | |
| WO2018000272A1 (en) | Corpus generation device and method | |
| Cunningham | A definition and short history of Language Engineering | |
| CN109947952A (en) | Search method, device, equipment and storage medium based on english knowledge map | |
| CN106528863B (en) | A kind of training of CRF identifier and technology and its attribute-name relationship are to abstracting method | |
| CN108763195A (en) | A kind of non-limiting type relation excavation method based on interdependent syntax and pattern rules | |
| CN115080694A (en) | Power industry information analysis method and equipment based on knowledge graph | |
| CN111159356A (en) | Knowledge graph construction method based on teaching content | |
| CN112733547A (en) | Chinese question semantic understanding method by utilizing semantic dependency analysis | |
| CN111966792A (en) | Text processing method and device, electronic equipment and readable storage medium | |
| CN114911893A (en) | Method and system for automatically constructing knowledge base based on knowledge graph | |
| CN111597302B (en) | Text event acquisition method and device, electronic equipment and storage medium | |
| Vea Escarza | Split and unified functions in the formation of Old English nouns and adjectives | |
| CN117251455A (en) | Intelligent report generation method and system based on large model | |
| Anchiêta et al. | A rule-based AMR parser for Portuguese | |
| Nguyen et al. | A vietnamese question answering system | |
| CN106407271B (en) | Intelligent customer service system and updating method of intelligent customer service knowledge base thereof | |
| CN104281695A (en) | Combination theory based quasi natural language semantic information extraction method and system | |
| CN115757720A (en) | Project information searching method, device, equipment and medium based on knowledge graph | |
| CN114417008A (en) | Construction engineering field-oriented knowledge graph construction method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20201203 Address after: Room 10242, No. 260, Jiangshu Road, Xixing street, Binjiang District, Hangzhou City, Zhejiang Province Applicant after: Hangzhou Jiji Intellectual Property Operation Co., Ltd Address before: 201616 Shanghai city Songjiang District Sixian Road No. 3666 Applicant before: Phicomm (Shanghai) Co.,Ltd. |
|
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |