CN106649394A - Fusion knowledge base processing method and device and knowledge base management system - Google Patents

Fusion knowledge base processing method and device and knowledge base management system Download PDF

Info

Publication number
CN106649394A
CN106649394A CN201510738095.1A CN201510738095A CN106649394A CN 106649394 A CN106649394 A CN 106649394A CN 201510738095 A CN201510738095 A CN 201510738095A CN 106649394 A CN106649394 A CN 106649394A
Authority
CN
China
Prior art keywords
knowledge base
knowledge
module
ontology
ontologies
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
Application number
CN201510738095.1A
Other languages
Chinese (zh)
Inventor
陈虹
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
ZTE Corp
Original Assignee
ZTE Corp
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by ZTE Corp filed Critical ZTE Corp
Priority to CN201510738095.1A priority Critical patent/CN106649394A/en
Priority to PCT/CN2016/104136 priority patent/WO2017076263A1/en
Publication of CN106649394A publication Critical patent/CN106649394A/en
Pending legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/21Design, administration or maintenance of databases
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/24Querying
    • G06F16/242Query formulation
    • G06F16/2433Query languages
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/25Integrating or interfacing systems involving database management systems
    • G06F16/254Extract, transform and load [ETL] procedures, e.g. ETL data flows in data warehouses

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Computational Linguistics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Machine Translation (AREA)

Abstract

The invention discloses a fusion knowledge base processing method and a device and a knowledge base management system. The fusion knowledge base processing method comprises the steps of extracting a knowledge classification catalogue from a common problem FAQ library and an ontology knowledge base, wherein the ontology knowledge base comprises a knowledge ontology and a ontology relation, based on the knowledge ontology, ontology relation and the knowledge classification catalogue, mount the contents of the FAQ library into the ontology knowledge base, and generating a fusion knowledge base. The fusion knowledge base processing method solves the problem that in the prior art the performance of the FAQ system can hardly be enhanced because the retrieval question answering system, the question answering system based on the knowledge base and the community cooperative question answering system have many defects in the practical application.

Description

Fusion knowledge base treating method and apparatus, and knowledge base management system
Technical field
The present invention relates to computer and communication technical field, espespecially a kind of fusion knowledge library processing method and dress Put, and knowledge base management system.
Background technology
Current question answering system is generally divided into three classes, including:Retrieval type question answering system, knowledge based storehouse The collaborative question answering system of question answering system and community.Wherein, wide variety of is retrieval type question answering system, such as Intelligent customer service system, be all based on FAQs (Frequently Asked Questions, referred to as:FAQ) Retrieval type question answering system;The question answering system in knowledge based storehouse, is generally seldom used, and application is, for example, Specialist system etc.;Social collaborative question answering system, such as Baidu is known, it is difficult to judge the accurate of question and answer Property and process question and answer result, be more, come solve problem, to be generally deficient of the mark of judgement by group intelligence Quasi-regular.
Specifically to the question and answer principle of retrieval type question answering system and the knowledge reasoning question answering system in knowledge based storehouse and Feature is analyzed.For retrieval type question answering system, primarily to overcoming the shortcomings of that search engine meets the tendency of And it is raw, the deficiency of search engine for example include cannot correct understanding and the information requirement of expression user be intended to, The simple answer that the too many non-user of information of return is wanted, be the characteristics of its is main using information retrieval and The natural language technology of shallow-layer extracts answer from text library or webpage, with not limited by field, energy Understand user view, the advantages of the answer of return is more accurate, but the question answering system there is also many shortcomings, Because the language material of retrieval type question answering system is mainly FAQ, then problems with is certainly existed:Firstth, FAQ For FAQs, draw there is the colloquial style imagination by statistics, it is not up to standard, it will usually to cause matching Not precisely, the low problem of accuracy rate;The essential technology of the secondth, system is retrieval, the return of answer according to By Similarity Measure, accuracy rate is difficult to improve;3rd, for the problem without answer, it is impossible to by pushing away Reason obtain answer, system it is intelligent poor;4th, the Knowledge Granulation of FAQ forms is thicker, for knowing For knowing management personnel, there is no hierarchical structure, management cost and threshold are higher.For knowledge based storehouse Question answering system, needs first to build knowledge base, and answer direct access can be obtained from knowledge base, can be with Obtain through reasoning on the basis of knowledge base, be knowledge more fine granularity the characteristics of the question answering system, because This can retrieve more accurately answer, and when answer is not retrieved, can also pass through reasoning Obtain the knowledge base in indirect answer, but the question answering system and be typically based on specific area, the scale of knowledge base It is limited, have impact on the use range of question answering system.
In sum, question answering system of the prior art, due to retrieval type question answering system, knowledge based storehouse The collaborative question answering system of question answering system and community all there is every defect in actual applications, and cause to ask The performance for answering system is difficult to improve.
The content of the invention
In order to solve above-mentioned technical problem, the invention provides a kind of fusion knowledge library processing method and dress Put, and knowledge base management system, to solve question answering system of the prior art, due to retrieval type question and answer System, the collaborative question answering system of the question answering system in knowledge based storehouse and community all exist in actual applications each Item defect, and cause the performance of question answering system to be difficult to improve.
In a first aspect, the present invention provides a kind of fusion knowledge library processing method, including:
Knowledge classification catalogue, the ontology knowledge base are extracted from FAQs FAQ storehouse and ontology knowledge base Include ontologies and ontological relationship;
According to the ontologies, the ontological relationship and the knowledge classification catalogue, by the FAQ The content in storehouse is mounted in the ontology knowledge base, generates fusion knowledge base.
In the first possible implementation of first aspect, it is described according to the ontologies, it is described Ontological relationship and knowledge classification catalogue are mounted to the content in the FAQ storehouses in ontology knowledge base, including:
The ontologies of the ontology knowledge base are mounted in the knowledge classification catalogue;
The content in the FAQ storehouses is mounted in the knowledge classification catalogue and the ontologies.
In second possible implementation of first aspect, it is described from FAQs FAQ storehouse and this Extract before knowledge classification catalogue in body knowledge base, also include:
Initial language material is input into, the initial language material includes corpus of text and FAQ language materials;
The structure of knowledge base is carried out according to the corpus of text and the FAQ language materials, the body is generated and is known Know storehouse.
According to second possible implementation of first aspect, in the third possible implementation, The structure that knowledge base is carried out according to the corpus of text and the FAQ language materials, generates the body and knows Know storehouse, including:
Determine body target and decimation rule, the decimation rule includes ontological relationship rule and extraction algorithm;
By the body target and the extraction algorithm from the corpus of text and the FAQ language materials Ontologies are extracted, the ontologies include domain body and general ontology;
Ontological relationship is extracted from the ontologies according to ontological relationship rule;
Ontologies and ontological relationship according to being drawn into carry out the structure of knowledge base, generate ontology knowledge Storehouse.
According to second or the third possible implementation of first aspect, in the 4th kind of possible realization In mode, after the generation ontology knowledge base, also include:
According to preset rule of inference, the extraction of knowledge inference rule is carried out in the ontology knowledge base, Generate knowledge inference rule storehouse.
According to second or the third possible implementation of first aspect, in the 5th kind of possible realization In mode, after the generation ontology knowledge base, also include:
Increment language material is input in the ontology knowledge base;
Increment body is extracted from the increment language material by the extraction algorithm;
Increment ontological relationship is extracted from the increment body according to ontological relationship rule;
Pass through be drawn into increment body and increment ontological relationship is updated to the ontology knowledge base.
Second aspect, the present invention provides a kind of fusion knowledge base processing meanss, including:
Catalogue abstraction module, for extracting knowledge classification mesh from FAQs FAQ storehouse and ontology knowledge base Record, the ontology knowledge base includes ontologies and ontological relationship;
Carry module, for according to the ontologies, the ontological relationship and the catalogue abstraction module The knowledge classification catalogue of extraction, the content in the FAQ storehouses is mounted in ontology knowledge base, generates fusion Knowledge base.
In the first possible implementation of second aspect, the carry module includes:
First carry unit, extracts for the ontologies of the ontology knowledge base to be mounted to into the catalogue In the knowledge classification catalogue that module is extracted;
Second carry unit, for by the content in the FAQ storehouses be mounted to the knowledge classification catalogue and In the ontologies.
In second possible implementation of second aspect, the fusion knowledge base processing meanss are also wrapped Include:
Input module, in the catalogue abstraction module from the FAQ storehouses and the ontology knowledge base Before extracting knowledge classification catalogue, initial language material is input into, the initial language material includes corpus of text and FAQ Language material;
Module is built, the corpus of text and the FAQ language materials for being input into according to the input module is carried out The structure of knowledge base, generates the ontology knowledge base.
According to second possible implementation of second aspect, in the third possible implementation, The structure module includes:
Determining unit, for determining body target and decimation rule, the decimation rule includes ontological relationship Rule and extraction algorithm;
Ontology extraction unit, for the body target and extraction algorithm that are determined by the determining unit from institute Ontologies, the ontologies are extracted in the corpus of text and the FAQ language materials of stating input module input Including domain body and general ontology;
Relation extraction unit, the ontological relationship for being determined according to the determining unit is regular from the body Ontological relationship is extracted in the ontologies that extracting unit is extracted;
Knowledge base signal generating unit, for the ontologies that extracted according to the ontology extraction unit and the pass It is that the ontological relationship that extracting unit is extracted carries out the structure of knowledge base, generates ontology knowledge base.
According to second or the third possible implementation of second aspect, in the 4th kind of possible realization In mode, the fusion knowledge base processing meanss also include:Rule of inference module, in the structure Module is generated after ontology knowledge base, according to preset rule of inference, is carried out in the ontology knowledge base The extraction of knowledge inference rule, generates knowledge inference rule storehouse.
According to second or the third possible implementation of second aspect, in the 5th kind of possible realization In mode, the input module is additionally operable to after the structure module generates ontology knowledge base, in institute State input increment language material in ontology knowledge base;
The ontology extraction unit, is additionally operable to the extraction algorithm that determines by the determining unit from described defeated Enter and extract in the increment language material of module input increment body;
The Relation extraction unit, is additionally operable to regular from institute according to the ontological relationship of determining unit determination State and extract in the increment body of ontology extraction unit extraction increment ontological relationship;
The fusion knowledge base processing meanss also include:Update module, for by the ontology extraction list The increment ontological relationship that the increment body and the Relation extraction unit that unit extracts is extracted is to the ontology knowledge Storehouse is updated.
The third aspect, the present invention provides a kind of knowledge base management system, including:Question and answer system and as described above Fusion knowledge base processing meanss any one of second aspect;
Wherein, the question and answer system includes:User interactive module, for obtaining the problem of user input;
Language processing module, for carrying out Language Processing, institute to the problem that the user interactive module is obtained Language Processing is stated including participle, part-of-speech tagging and Entity recognition;
Standardized module, for being standardized to the problem after language processing module process, The standardization includes wrong word correction, language conversion and standardization and words recognition;
Semantic understanding module, for carrying out semantic understanding to the problem after standardized module process, obtains Take the semantic information of the problem;
Information searching module, for the semantic information that the semantic understanding module is obtained to be known in the body Know and enter in storehouse line retrieval, obtain the answer of the problem.
In the first possible implementation of the third aspect, the fusion knowledge base processing meanss include The rule of inference module;The question and answer system also includes:Knowledge reasoning module, in described information The answer that retrieval module gets is space-time, according to the knowledge inference rule that the rule of inference module is generated Storehouse makes inferences, and obtains the answer of the problem.
According to the first possible implementation of the third aspect, in second possible implementation, Described information retrieves module, and the answer for being additionally operable to be got in the knowledge reasoning module is space-time, by institute The semantic information for stating the acquisition of semantic understanding module enters line retrieval in the FAQ storehouses, obtains the problem Answer.
In the third possible implementation of the third aspect, the question and answer system also includes:Caching is answered Case acquisition module, for carrying out to the problem after language processing module process in the standardized module After standardization, the answer of the problem is obtained from caching.
According to the third aspect, the third aspect the first in the third possible implementation any one, In the 4th kind of possible implementation, the knowledge base management system also includes managing device;
Wherein, the managing device includes:Knowledge edition module, for processing the fusion knowledge base The ontologies that device is extracted enter edlin, check and correction, and increase new ontologies;
Knowledge examination module, for carrying out multistage auditing to the ontologies in the fusion knowledge base;
Statistical analysis module, for during being putd question to and being obtained answer by the question and answer system, Problem that statistics is popular, popular knowledge, the ontologies not accessed in the fusion knowledge base, and occur One or more in the ontologies of wrong answer, and analyze by question and answer system acquisition answer Accuracy rate;
On-line testing module, for verifying to the ontologies that the knowledge edition module is edited, Obtain on-line testing result;
Rules administration module, for being managed editor to the mode for extracting and rule;
Material management module, for storing the FAQ storehouses and the initial content in the ontology knowledge base, And according to the content in the preset time storage fusion knowledge base;
Intelligent quality testing module, for determining the standard that answer is obtained by the question and answer system by sampling inspection Really rate, is additionally operable to the accuracy rate of the automatic detection answer in the course of work of the question and answer system.
The fusion knowledge base treating method and apparatus that the present invention is provided, and knowledge base management system, pass through Knowledge classification catalogue, and the knowledge sheet in ontology knowledge base are extracted from FAQ storehouses and ontology knowledge base Body, ontological relationship and the knowledge classification catalogue for having extracted, by the content in FAQ storehouses ontology knowledge base is mounted to In, fusion knowledge base is generated, the fusion knowledge base that the present embodiment is generated is applied in question answering system, it is real The answer that showed originally carries out knowledge question by FAQ storehouses and ontology knowledge base can be by the fusion Knowledge base draws, while with reference to FAQ storehouses and ontology knowledge base advantage, can avoid FAQ storehouses and The respective shortcoming of ontology knowledge base;The method that the present invention is provided solves question answering system of the prior art, Because retrieval type question answering system, the collaborative question answering system of the question answering system in knowledge based storehouse and community are in reality All there is every defect using in, and cause the performance of question answering system to be difficult to improve.
Description of the drawings
Accompanying drawing is used for providing and further understanding technical solution of the present invention, and constitutes one of description Point, it is used to explain technical scheme together with embodiments herein, do not constitute to the present invention The restriction of technical scheme.
Fig. 1 is a kind of flow chart of fusion knowledge library processing method provided in an embodiment of the present invention;
Fig. 2 is the flow chart of another kind of fusion knowledge library processing method provided in an embodiment of the present invention;
A kind of knowledge graph in the fusion knowledge library processing method that Fig. 3 is provided for embodiment illustrated in fig. 1 of the present invention Spectrum schematic diagram;
Fig. 4 is the flow chart of another fusion knowledge library processing method provided in an embodiment of the present invention;
Fig. 5 is the flow chart of another fusion knowledge library processing method provided in an embodiment of the present invention;
Fig. 6 is a kind of structural representation of fusion knowledge base processing meanss provided in an embodiment of the present invention;
Fig. 7 is the structural representation of another kind of fusion knowledge base processing meanss provided in an embodiment of the present invention;
Fig. 8 is a kind of structural representation of knowledge base management system provided in an embodiment of the present invention;
Fig. 9 is the flow chart that intelligent answer is carried out by knowledge base management system shown in Fig. 8;
Figure 10 is the structural representation of another kind of knowledge base management system provided in an embodiment of the present invention.
Specific embodiment
To make the object, technical solutions and advantages of the present invention become more apparent, below in conjunction with accompanying drawing Embodiments of the invention are described in detail.It should be noted that in the case where not conflicting, this Shen Please in embodiment and the feature in embodiment can mutual combination in any.
Can be in the computer of such as one group of computer executable instructions the step of the flow process of accompanying drawing is illustrated Perform in system.And, although show logical order in flow charts, but in some cases, Can be with the step shown or described by performing different from order herein.
Based on every shortcoming of the question answering system in retrieval type question answering system and knowledge based storehouse in prior art, It is found that the advantage of the question answering system in knowledge based storehouse can to a certain extent make up retrieval type question and answer system Accuracy rate is difficult to improve, cannot obtain the problems such as answer and thicker Knowledge Granulation by reasoning in system, and examines Cable-styled question answering system also can overcome the disadvantages that the question answering system in knowledge based storehouse, be brought due to knowledge base dimension-limited Every shortcoming.It is apparent that the excellent of both question answering systems in retrieval type question answering system and knowledge based storehouse lacks Point can form complementation, therefore, complementarity of the present invention based on above two question answering system pluses and minuses is carried Go out the technical scheme of a kind of fusion FAQ and knowledge base, and the knowledge base after fusion is applied to into question and answer Systems and management system, such that it is able to significant systematic function is improved.
Technical scheme is described in detail below by specific embodiment, the present invention is following Ontology knowledge base in each embodiment is that, based on the knowledge base of specific area, it can be based on a certain item skill Art field, or the content comprising multinomial technical field.The present invention provides following specific reality Applying example can be combined with each other, for same or analogous concept or process may no longer go to live in the household of one's in-laws on getting married in some embodiments State.
Fig. 1 is a kind of flow chart of fusion knowledge library processing method provided in an embodiment of the present invention.This enforcement The fusion knowledge library processing method that example is provided can apply in knowledge base system, to improve knowledge base system Use range, the method can perform by fusion knowledge base processing meanss, and the fusion knowledge base processes dress Put by way of hardware and software is combined to realize, the device can be integrated in the processor of terminal unit In, call for processor and use.As shown in figure 1, the method for the present embodiment can include:
S110, extracts knowledge classification catalogue, in the ontology knowledge base from FAQ storehouses and ontology knowledge base Including ontologies and ontological relationship.
Fusion knowledge library processing method provided in an embodiment of the present invention, it is therefore intended that build a kind of performance More improve and blanket knowledge base, its main thought is to combine FAQ storehouses and ontology knowledge base Advantage, and FAQ storehouses and the respective shortcoming of ontology knowledge base can be improved.Various embodiments of the present invention In ontology knowledge base be a knowledge base set up, for example, knowledge based on particular technology area Storehouse, generally has in the ontology knowledge base and there is body pass between ontologies, and each ontologies System, for example ontologies include general ontology and domain body etc., ontological relationship include generic relation, etc. Same relation, relation on attributes and component relationship etc..
The FAQ storehouses of the present embodiment include FAQ language materials, are problem and key that some Jing are often putd question to Word, with reference to FAQ language materials and the ontologies in ontology knowledge base knowledge classification catalogue is extracted, and extracting should The purpose of knowledge classification catalogue is to be easy to knowledge combing and management, for example, can adopt the upper of knowledge base Word is extracted.
S120, according to ontologies, ontological relationship and knowledge classification catalogue, the content in FAQ storehouses is hung In being downloaded to ontology knowledge base, fusion knowledge base is generated.
Knowledge classification catalogue is extracted in the present embodiment, that is, has been provided for the basis of knowledge combing and management Condition, therefore, it can the relation according to each ontologies and knowledge classification catalogue, in FAQ index databases Content respectively at knowledge classification catalogue and the relation of ontologies, and with reference to the sheet in ontology knowledge base Body relation, the content in FAQ storehouses is mounted in ontology knowledge base to construct fusion knowledge base.This reality The fusion knowledge base built in example is applied not only with the ontologies and ontological relationship in former ontology knowledge base, Also the content in FAQ storehouses is fused in ontology knowledge base in the way of knowledge classification and structural relation, structure Into fusion knowledge base;So that originally carrying out the answer of knowledge question by FAQ storehouses and ontology knowledge base Can be drawn by the fusion knowledge base.
The specific implementation of the present embodiment for another kind provided in an embodiment of the present invention as shown in Fig. 2 melt The flow chart of knowledge library processing method is closed, illustrating the mode of fusion, i.e. S120 can include: S121, the ontologies of ontology knowledge base are mounted in knowledge classification catalogue, and the step is to knowledge Body is classified;S122, by the content in FAQ storehouses knowledge classification catalogue and ontologies are mounted to On, the step is that FAQ storehouses are classified, and classification is knowledge classification catalogue and ontologies;Complete The content and structure of the knowledge base of fusion can be represented by knowledge mapping.For example, such as Fig. 3 institutes Show, a kind of knowledge mapping shows in the fusion knowledge library processing method provided for embodiment illustrated in fig. 1 of the present invention It is intended to, the knowledge classification catalogue in Fig. 3 includes catalog classification A and catalog classification B, in ontology knowledge base Ontologies be all mounted in above-mentioned knowledge classification catalogue, the content in FAQ storehouses for example include FAQ1, FAQ2 and FAQ3, can be mounted in knowledge classification catalogue or ontologies;In addition, knowledge mapping In the also behavior with ontologies and class belong to the attribute of ontologies, the root node table in knowledge mapping Show the classification on superclass, i.e. upper strata).
Below by way of an instantiation declarative knowledge body and its behavior, attribute etc., such as problem is " color How bell business is handled ", wherein, " Ring Back Tone service " is ontologies, and " handling " is event row For the expenses standard of Ring Back Tone service is attribute;Again for example, A service package and B service package are two The different set meal of rate, based on set meal, C upgrading sets can be superimposed on the basis of A service package Meal, it is known that, A service package is identity relation, A service package and C upgrading set meals with B service package For the relation of the upper and lower, i.e., the relation between ontologies can be across comparison, or vertical To contrast.
The fusion knowledge library processing method that the present embodiment is provided, by from FAQ storehouses and ontology knowledge base Middle extraction knowledge classification catalogue, and the ontologies in ontology knowledge base, ontological relationship and extracted Knowledge classification catalogue, the content in FAQ storehouses is mounted in ontology knowledge base, generate fusion knowledge Storehouse, the fusion knowledge base that the present embodiment is generated is applied in question answering system, is realized and was originally passed through FAQ storehouses and ontology knowledge base carry out the answer of knowledge question can be drawn by the fusion knowledge base, While with reference to FAQ storehouses and ontology knowledge base advantage, FAQ storehouses and ontology knowledge base can be avoided each From shortcoming;The method that the present embodiment is provided solves question answering system of the prior art, due to retrieval type Question answering system, the collaborative question answering system of the question answering system in knowledge based storehouse and community are all deposited in actual applications In every defect, and the performance of question answering system is caused to be difficult to improve.
It should be noted that having been described above the ontology knowledge base in various embodiments of the present invention in above-described embodiment For a knowledge base set up, the ontology knowledge base for example can be traditional artificial constructed knowledge Storehouse, however, because the knowledge quantity of ontology knowledge base is big, structural relation is complicated, it is artificial constructed complex And it is relatively costly.In addition, the ontology knowledge base in the present embodiment can also be to build automatically, such as Fig. 4 It is shown, it is the flow chart of another fusion knowledge library processing method provided in an embodiment of the present invention, in Fig. 4 In illustrated embodiment, the building mode of ontology knowledge base is:Also included before S110:S100, it is defeated Enter initial language material, the initial language material includes corpus of text and FAQ language materials;S101, according to corpus of text With the structure that FAQ language materials carry out knowledge base, ontology knowledge base is generated.Corpus of text in the present embodiment For the input item of knowledge base in prior art, the input item of ontology knowledge base also includes in the present embodiment FAQ language materials, can be used as the basis of follow-up fusion.
In implementing, the form for building ontology knowledge base is different from people of the prior art to the present embodiment Work builds, and its concrete mode, i.e. S101 include:S102, determines body target and decimation rule, takes out Taking rule includes ontological relationship rule and extraction algorithm;S103, by body target and extraction algorithm from Ontologies are extracted in corpus of text and FAQ language materials, the ontologies include domain body and general Body;S104, according to ontological relationship rule ontological relationship is extracted from ontologies;S105, according to institute The ontologies being drawn into and ontological relationship carry out the structure of knowledge base, generate ontology knowledge base.This enforcement The target that ontologies can first be formulated and the decimation rule for subsequently using in example, i.e., which is field sheet Body, which is general ontology, the species of ontological relationship, formulation of decimation rule etc., and these information can be with It is operator's definition, is input in terminal unit by human-computer interaction interface;Define rule and After target, taking out for the ontologies such as entity, attribute, behavior, event is carried out by predefined coding Take, predefined coding is the concrete form of extraction algorithm, extraction algorithm can for example adopt Stamford Name Entity recognition device (Named Edtity Recognition, referred to as:NER), it is also possible to from Oneself designs extraction algorithm, finally extracts general ontology and domain body, ontologies according to initial language material Node specially on knowledge mapping, as shown in figure 3, after ontologies are determined, further performing Extraction operation, this time carries out the extraction of ontological relationship according to ontological relationship rule encoding, and ontological relationship includes But generic relation, identity relation, relation on attributes, component relationship etc. are not limited to, for example can be using definition The modes such as rule are extracted, and ontological relationship extracts the side that can determine on knowledge mapping between node;Determining Behind side between node and node, the structure of knowledge base is just carried out so as to form knowledge mapping, i.e. body and know Know storehouse.
Further, after ontology knowledge base structure is completed, the method that the present embodiment is provided is in S101 Also include afterwards:S106, according to preset rule of inference, carries out knowledge reasoning in semantic knowledge-base The extraction of rule, generates knowledge inference rule storehouse.The formation in knowledge inference rule storehouse has beneficial to the knowledge base Use, when question and answer are carried out by the knowledge base, it is impossible to which the answer for directly obtaining can be according to preset Rule of inference is obtained from knowledge inference rule storehouse, further increases the performance of fusion knowledge base.
It should be noted that traditional construction of knowledge base is artificial constructed, the present embodiment adopts automatization The mode of structure forms ontology knowledge base and knowledge inference rule storehouse, on this basis, can also adopt people The mode of work check and correction monitors the content of knowledge base, to guarantee the accuracy rate and recall rate of knowledge.
Further, Fig. 5 is another fusion knowledge library processing method provided in an embodiment of the present invention Flow chart, on the basis of the various embodiments described above, after the automatic structure for completing ontology knowledge base, should be certainly The dynamic ontology knowledge base for constructing is used for and the fusion knowledge to be formed in above-described embodiment is merged in FAQ storehouses Storehouse;However, the content of the ontology knowledge base for having built is limited, it is impossible to comprising all of knowledge sheet Body, and the content in FAQ storehouses is also the FAQs summed up according to the input of user, therefore, In actually used, user is ever-increasing to the demand of knowledge base, therefore, the body in the present embodiment The content of knowledge base is not unalterable, can further increase the scope of ontology knowledge base, wherein wrapping Increase FAQ language materials are included, concrete mode is also to include after S101:S130, in ontology knowledge base Middle input increment language material;S140, by extraction algorithm increment body is extracted from increment language material; S150, according to ontological relationship rule increment ontological relationship is extracted from increment body;S160, by institute The increment body and increment ontological relationship being drawn into is updated to ontology knowledge base.
In the present embodiment, to the update mode and the mode phase for building ontology knowledge base of ontology knowledge base Seemingly, all it is that body and ontological relationship are extracted according to the language material of input, so as to the content for passing through be drawn into is entered Row builds;During renewal, increment language material can increase for newest domain knowledge, for knowledge base The knowledge content for adding, can also be enquirement record or daily record etc. of user, and purpose is exactly new to know some Know and new subject of question is added in ontology knowledge base, while FAQ language materials can be increased, i.e., simultaneously Have updated fusion knowledge base.
It should be noted that the present embodiment do not limit S130~S160 fusion knowledge library processing method in Execution sequence, for example, performed after S101, can also be what is performed after S106, Can also be what is performed after S120, Fig. 5 gives as a example by performing after S120 for S130~S160 Illustrate, i.e., the present embodiment is upgraded demand after fusion knowledge base is built to knowledge base.
Fig. 6 is a kind of structural representation of fusion knowledge base processing meanss provided in an embodiment of the present invention.This The fusion knowledge base processing meanss that embodiment is provided can apply in knowledge base management system, be known with improving Know the use range of storehouse system, the side that the fusion knowledge base processing meanss can be combined by hardware and software Realizing, the device can be integrated in the processor of terminal unit formula, call for processor and use.Such as Shown in Fig. 6, the fusion knowledge base processing meanss of the present embodiment are specifically included:Catalogue abstraction module 11 and extension Carry module 12.
Wherein, catalogue abstraction module 11, for extracting from FAQs FAQ storehouse and ontology knowledge base Knowledge classification catalogue, the ontology knowledge base includes ontologies and ontological relationship.
Fusion knowledge base processing meanss provided in an embodiment of the present invention, it is therefore intended that build a kind of performance More improve and blanket knowledge base, its main thought is to combine FAQ storehouses and ontology knowledge base Advantage, and FAQ storehouses and the respective shortcoming of ontology knowledge base can be improved;Various embodiments of the present invention In ontology knowledge base be a knowledge base set up.
Carry module 12, for what is extracted according to ontologies, ontological relationship and catalogue abstraction module 11 Knowledge classification catalogue, the content in FAQ storehouses is mounted in ontology knowledge base, generates fusion knowledge base.
The concrete mode that fusion knowledge base is generated in various embodiments of the present invention is that carry module 12 includes: First carry unit 13, takes out for the ontologies of ontology knowledge base to be mounted to into catalogue abstraction module 11 In the knowledge classification catalogue for taking;Second carry unit 14, for the content in FAQ storehouses to be mounted to into knowledge On classified catalogue and ontologies.Built by fusion knowledge base processing meanss provided in an embodiment of the present invention Fusion knowledge base equally can be represented with knowledge mapping shown in Fig. 3, knowledge mapping interior joint and side pair The content answered is in the above-described embodiments it is stated that therefore will not be described here.
Fusion knowledge base processing meanss provided in an embodiment of the present invention are used to perform Fig. 1 and Fig. 2 institutes of the present invention Show embodiment provide fusion knowledge library processing method, possess corresponding functional module, its realize principle and Technique effect is similar to, and here is omitted.
It should be noted that having been described above the ontology knowledge base in various embodiments of the present invention in above-described embodiment For a knowledge base set up, for example, can be traditional artificial constructed knowledge base, or from The dynamic ontology knowledge base for building, the concrete mode of ontology knowledge base is built in the present embodiment as shown in fig. 7, For the structural representation of another kind of fusion knowledge base processing meanss provided in an embodiment of the present invention, merge knowledge Storehouse processing meanss also include:Input module 15, in catalogue abstraction module 11 from FAQ storehouses and body Extract in knowledge base before knowledge classification catalogue, be input into initial language material, the initial language material includes corpus of text With FAQ language materials;Module 16 is built, for the corpus of text that is input into according to input module 15 and FAQ Language material carries out the structure of knowledge base, generates ontology knowledge base.
In implementing, building module 16 can include the present embodiment:Determining unit 17, for true Determine body target and decimation rule, the decimation rule includes ontological relationship rule and extraction algorithm;Body is taken out Unit 18 is taken, for the body target and extraction algorithm that determine by determining unit 17 from input module 15 Extract ontologies in the corpus of text and FAQ language materials of input, the ontologies include domain body and General ontology;Relation extraction unit 19, for determined according to determining unit 17 ontological relationship rule from Ontological relationship is extracted in the ontologies that ontology extraction unit 18 is extracted;Knowledge base signal generating unit 20, uses In the ontological relationship that the ontologies and Relation extraction unit 19 that are extracted according to ontology extraction unit 18 are extracted The structure of knowledge base is carried out, ontology knowledge base is generated.The fusion knowledge base provided by the present embodiment is processed The node and Biao Lai that the ontology knowledge base that device is generated again may be by knowledge mapping is represented.
Further, after ontology knowledge base structure is completed, the device that the present embodiment is provided also includes: Rule of inference module 21, for after building module 16 and generating ontology knowledge base, being pushed away according to preset Reason rule, carries out the extraction of knowledge inference rule in ontology knowledge base, generates knowledge inference rule storehouse. The formation in knowledge inference rule storehouse has the use beneficial to the knowledge base, and by the knowledge base question and answer are being carried out When, it is impossible to the answer for directly obtaining can be obtained according to preset rule of inference from knowledge inference rule storehouse Arrive, further increase the performance of fusion knowledge base.
Fusion knowledge base processing meanss provided in an embodiment of the present invention are used to perform enforcement shown in Fig. 4 of the present invention The fusion knowledge library processing method that example is provided, possesses corresponding functional module, and it realizes principle and technology effect Seemingly, here is omitted for fruit.
Further, after the automatic structure for completing ontology knowledge base, this is constructed the present embodiment automatically Ontology knowledge base be used for and FAQ storehouses merge the fusion knowledge base to be formed in above-described embodiment;However, The content of the ontology knowledge base for having built is limited, it is impossible to comprising all of ontologies, therefore, The content of the ontology knowledge base in the present embodiment is not unalterable, can further increase ontology knowledge The scope in storehouse, including increase FAQ language materials, concrete mode is:Input module 15, is additionally operable to Build module 16 to generate after ontology knowledge base, increment language material is input in the ontology knowledge base;Body Extracting unit 18, be additionally operable to determine by determining unit 17 extraction algorithm is input into from input module 15 Increment body is extracted in increment language material;Relation extraction unit 19, is additionally operable to be determined according to determining unit 17 Ontological relationship rule from ontology extraction unit 18 extract increment body in extract increment ontological relationship; Correspondingly, the fusion knowledge base processing meanss in the present embodiment also include:Update module, for by this The increment ontological relationship that the increment body and Relation extraction unit 19 that body extracting unit 18 is extracted is extracted is to this Body knowledge base is updated.
It should be noted that the execution sequence of the present embodiment more new knowledge base, such as known in body To know be carried out after the completion of storehouse builds and update, can also be and performed more after the completion of knowledge inference rule storehouse builds Newly, can also be it is being to perform renewal after the completion of fusion construction of knowledge base.
Fusion knowledge base processing meanss provided in an embodiment of the present invention are used to perform enforcement shown in Fig. 5 of the present invention The fusion knowledge library processing method that example is provided, possesses corresponding functional module, and it realizes principle and technology effect Seemingly, here is omitted for fruit.
Fig. 8 is a kind of structural representation of knowledge base management system provided in an embodiment of the present invention.This enforcement Example provide knowledge base management system include question and answer system 200, and shown in above-mentioned Fig. 6 and Fig. 7 it is arbitrary Fusion knowledge base processing meanss 100 in embodiment.
Wherein, the question and answer system 200 includes:User interactive module 210, for obtaining user input Problem.
Language processing module 220, for carrying out at language to the problem that user interactive module 210 is obtained Reason, Language Processing includes participle, part-of-speech tagging and Entity recognition.
Speech processes in the present embodiment are often referred to natural language processing, such as first after user input problem The natural language processings such as participle, part-of-speech tagging, Entity recognition, and traditional FAQ are first carried out to problem Unlike question and answer, general ontology and domain body are introduced in the knowledge base of the present embodiment, it is mainly made For participle, make participle more accurate, only participle is accurate, can just carry out correct semantic understanding.Lift For example, input problem is:" how day wing A8 navigator set meal charges ", using traditional participle side Formula is divided into " my god/wing/A/8/ neck/boat/set meals/how/charge " substantially, and adopts the embodiment of the present invention to generate Fusion knowledge base after, can be divided into " day wing A8 navigator's set meal/how/charge ".
Standardized module 230, for being standardized place to the problem after the process of language processing module 220 Reason, standardization includes wrong word correction, language conversion and standardization and words recognition.
Mainly by language standardization, such as wrong word correction, dialect are converted into the module in the present embodiment The standardization of standard mandarin, complete/letter/another name, can also include identification of sensitive word etc..
Semantic understanding module 240, for carrying out semantic reason to the problem after the process of standardized module 230 Solution, obtains the semantic information of problem.
After Language Processing and standardization, the correct semanteme of input problem is generally appreciated that, for example can be with It is the semantic information of the problem that gets, the semantic information mainly includes the key word for extracting and extending, extracts Tlv triple and ternary relation, omission reply, reference resolution, intention assessment, emotion based on context Analysis, problem are questioned closely.
Information searching module 250, for the semantic information that semantic understanding module 240 is obtained to be known in body Know and enter in storehouse line retrieval, obtain the answer of problem.
The semantic information that in the present embodiment, semantic understanding module 240 can be obtained, i.e. key word, Tlv triple, intention etc. enter line retrieval in fusion knowledge base, for example, can be obtained using query language and data Take agreement (Simple Protocol and RDF Query Language, referred to as:The inspection such as SPARQL) Suo Yuyan, in the answer of the problem of acquisition user can be returned to, and normally, the input of user and checking is returned The answer returned can be realized by user interactive module 210, for example, can be graphic user interface (Graphical User Interface, referred to as:GUI).
The knowledge base management system that the present embodiment is provided is used for the problem to user input in fusion knowledge base Enter line retrieval, and answer is given by merging the content of knowledge base.Use in the system that the present embodiment is provided Generated by the fusion knowledge base processing meanss that above-described embodiment is provided in the resource of retrieval, therefore had Technique effect same as the previously described embodiments, here is omitted.
Because the fusion knowledge base processing meanss 100 that the above embodiment of the present invention is provided are generating ontology knowledge Behind storehouse, moreover it is possible to knowledge inference rule storehouse is further generated, if melting above by information searching module 250 Close in knowledge base and do not obtain answer, then further according to knowledge inference rule storehouse can make inferences and Obtain answer;The question and answer system 200 that i.e. the present embodiment is provided also includes:Knowledge reasoning module 260, also Answer for getting in information searching module 250 is space-time, and what by inference rule module was generated knows Know rule of inference storehouse to make inferences, obtain the answer of the problems referred to above.
Further, if knowledge reasoning module 260 makes inferences acquisition in above-mentioned knowledge inference rule storehouse To answer be similarly space-time, the semantic information that can also be obtained semantic understanding module 240 is in FAQ Enter line retrieval in storehouse, obtain the answer of problem.In implementing, can be by the way that semantic information be existed FAQ enters line retrieval in storehouse, and retrieval result is carried out into Similarity Measure, and will finally return that in threshold value, And the answer for ranking the first is used as the answer of problem.
It should be noted that the question and answer system 200 during this enforcement is provided can also include:Caching answer is obtained Delivery block 270, for the problem after standardized module 230 is processed language processing module 220 rower is entered After quasi-ization process, the answer of problem is obtained from caching.Which is conducive to improving systematic function, can So that common problem and knowledge base are put in internal memory, after language standard on the basis of, first obtain from caching Take answer, it is to avoid substantial amounts of retrieval and calculating process.In addition, the question and answer system that the present embodiment is provided 200 can also include that answer returns module 280, for return information retrieval module 250, knowledge reasoning mould The answer that block 260 or caching answer acquisition module 270 are obtained, and by answer output to user interactive module In 210.
The knowledge base management system provided by the present embodiment carries out handling process such as Fig. 9 institutes of intelligent answer Show, be the flow chart that intelligent answer is carried out by knowledge base management system shown in Fig. 8, idiographic flow is:
S201, obtains the problem of user input.
S202, to the problem of user input natural language processing is carried out.It is usually used in processing procedure To general ontology storehouse, field ontology library, thesaurus and disable dictionary etc..
S203, to the problem of the user input language standardization's process is carried out.Lead in processing procedure Often use sensitive dictionary and dialect storehouse etc..
S204, judges whether the answer that problem can be obtained from caching.If it is not, then performing S205; If so, S210 is then performed, if the knowledge content of the problem that is stored with caching, can obtain and return The answer of the problem of returning.
S205, semantic understanding is carried out to the problem after standardization, obtains semantic information.
S206, judges whether the answer that problem can be obtained from ontology knowledge base.If it is not, then performing S207;If so, S210 is then performed, if the knowledge content of the problem that is stored with ontology knowledge base, Can obtain and return the answer of problem.
S207, judges whether the answer that problem can be obtained from knowledge inference rule storehouse.If it is not, then Perform S208;If so, S210 is then performed, if the problem can be inferred in knowledge inference rule storehouse Knowledge content, then can obtain and return the answer of problem.
S208, by semantic information line retrieval is entered in FAQ storehouses, obtains retrieval result.
S209, to retrieval result Similarity measures are carried out.
S210, obtains and returns the answer of problem.
Figure 10 is the structural representation of another kind of knowledge base management system provided in an embodiment of the present invention. On the architecture basics of the knowledge base management system that above-mentioned Fig. 8 is provided, the system that the present embodiment is provided also is wrapped Managing device 300 is included, the managing device 300 includes:Knowledge edition module 310, for merging knowledge The ontologies that storehouse processing meanss 100 are extracted enter edlin, check and correction, and increase new knowledge sheet Body.Knowledge can also manually be increased.
Knowledge examination module 320, for the ontologies in merging knowledge base multistage auditing is carried out.It is logical Often need to be audited by multistage personnel, to guarantee the correctness of knowledge.
Statistical analysis module 330, in the mistake for being putd question to and being obtained answer by question and answer system 200 Cheng Zhong, problem that statistics is popular, popular knowledge, merges the ontologies not accessed in knowledge base, and goes out One or more in the ontologies of existing wrong answer, and analyze by the acquisition answer of question and answer system 200 Accuracy rate.
On-line testing module 340, for testing the ontologies that knowledge edition module 310 is edited Card, obtains on-line testing result.
Rules administration module 350, for being managed editor to the mode for extracting and rule.For example can be with It is human-edited.
Material management module 360, for storing FAQ storehouses and the initial content in ontology knowledge base, and According to the content in preset time storage fusion knowledge base.The module can be entered to the original attachment for importing Row management, such as storage, multi version contrast, adnexa filing.
Intelligent quality testing module 370, obtain answer by question and answer system 200 for determining by sampling inspection Accuracy rate, be additionally operable to the accuracy rate of the automatic detection answer in the course of work of question and answer system 200.
It should be noted that the knowledge base management system that the present embodiment is provided can also be carried out together to knowledge Step, the mainly enquirement record by user in knowledge base management system and daily record is synchronized to KBM In system, quality inspection, real-time detection are carried out to record by manager or knowledge professional;In addition, this reality The knowledge base management system for applying example can also be by the problem without accurate answer when puing question to or asking without answer Topic, carries out Knowledge Extraction, in incremental update to fusion knowledge base, can retrieve in order to next time.Specifically Ground, the mode that synchronous mode and the mode of renewal is updated with above-mentioned fusion knowledge base processing meanss It is identical, therefore here is repeated again.
Although disclosed herein embodiment as above, described content is only to readily appreciate the present invention And the embodiment for adopting, it is not limited to the present invention.Technology people in any art of the present invention Member, without departing from disclosed herein spirit and scope on the premise of, can be in the form implemented and thin Any modification and change, but the scope of patent protection of the present invention are carried out on section, still must be with appended right The scope that claim is defined is defined.

Claims (17)

1. it is a kind of to merge knowledge library processing method, it is characterised in that to include:
Knowledge classification catalogue, the ontology knowledge base are extracted from FAQs FAQ storehouse and ontology knowledge base Include ontologies and ontological relationship;
According to the ontologies, the ontological relationship and the knowledge classification catalogue, by the FAQ storehouses Content be mounted in the ontology knowledge base, generate fusion knowledge base.
2. it is according to claim 1 to merge knowledge library processing method, it is characterised in that the basis The content in the FAQ storehouses is mounted to this by the ontologies, the ontological relationship and knowledge classification catalogue In body knowledge base, including:
The ontologies of the ontology knowledge base are mounted in the knowledge classification catalogue;
The content in the FAQ storehouses is mounted in the knowledge classification catalogue and the ontologies.
3. fusion knowledge library processing method according to claim 1, it is characterised in that described from normal See in problem FAQ storehouse and ontology knowledge base and extract before knowledge classification catalogue, also include:
Initial language material is input into, the initial language material includes corpus of text and FAQ language materials;
The structure of knowledge base is carried out according to the corpus of text and the FAQ language materials, the body is generated and is known Know storehouse.
4. it is according to claim 3 to merge knowledge library processing method, it is characterised in that the basis The corpus of text and the FAQ language materials carry out the structure of knowledge base, generate the ontology knowledge base, bag Include:
Determine body target and decimation rule, the decimation rule includes ontological relationship rule and extraction algorithm;
Taken out from the corpus of text and the FAQ language materials by the body target and the extraction algorithm Ontologies are taken, the ontologies include domain body and general ontology;
Ontological relationship is extracted from the ontologies according to ontological relationship rule;
Ontologies and ontological relationship according to being drawn into carry out the structure of knowledge base, generate ontology knowledge Storehouse.
5. the fusion knowledge library processing method according to claim 3 or 4, it is characterised in that described After generating ontology knowledge base, also include:
According to preset rule of inference, the extraction of knowledge inference rule is carried out in the ontology knowledge base, Generate knowledge inference rule storehouse.
6. the fusion knowledge library processing method according to claim 3 or 4, it is characterised in that described After generating ontology knowledge base, also include:
Increment language material is input in the ontology knowledge base;
Increment body is extracted from the increment language material by the extraction algorithm;
Increment ontological relationship is extracted from the increment body according to ontological relationship rule;
Pass through be drawn into increment body and increment ontological relationship is updated to the ontology knowledge base.
7. it is a kind of to merge knowledge base processing meanss, it is characterised in that to include:
Catalogue abstraction module, for extracting knowledge classification mesh from FAQs FAQ storehouse and ontology knowledge base Record, the ontology knowledge base includes ontologies and ontological relationship;
Carry module, for according to the ontologies, the ontological relationship and the catalogue abstraction module The knowledge classification catalogue of extraction, the content in the FAQ storehouses is mounted in ontology knowledge base, generates fusion Knowledge base.
8. it is according to claim 7 to merge knowledge base processing meanss, it is characterised in that the carry Module includes:
First carry unit, extracts for the ontologies of the ontology knowledge base to be mounted to into the catalogue In the knowledge classification catalogue that module is extracted;
Second carry unit, for the content in the FAQ storehouses to be mounted to into the knowledge classification catalogue and institute State in ontologies.
9. it is according to claim 7 to merge knowledge base processing meanss, it is characterised in that also to include:
Input module, in the catalogue abstraction module from the FAQ storehouses and the ontology knowledge base Before extracting knowledge classification catalogue, initial language material is input into, the initial language material includes corpus of text and FAQ Language material;
Module is built, the corpus of text and the FAQ language materials for being input into according to the input module is carried out The structure of knowledge base, generates the ontology knowledge base.
10. it is according to claim 9 to merge knowledge base processing meanss, it is characterised in that the structure Modeling block includes:
Determining unit, for determining body target and decimation rule, the decimation rule includes ontological relationship Rule and extraction algorithm;
Ontology extraction unit, for the body target and extraction algorithm that are determined by the determining unit from institute Ontologies, the ontologies are extracted in the corpus of text and the FAQ language materials of stating input module input Including domain body and general ontology;
Relation extraction unit, the ontological relationship for being determined according to the determining unit is regular from the body Ontological relationship is extracted in the ontologies that extracting unit is extracted;
Knowledge base signal generating unit, for the ontologies that extracted according to the ontology extraction unit and the pass It is that the ontological relationship that extracting unit is extracted carries out the structure of knowledge base, generates ontology knowledge base.
The 11. fusion knowledge base processing meanss according to claim 9 or 10, it is characterised in that also Including:Rule of inference module, after generating ontology knowledge base in the structure module, according to preset Rule of inference, carry out the extraction of knowledge inference rule in the ontology knowledge base, generate knowledge reasoning Rule base.
The 12. fusion knowledge base processing meanss according to claim 9 or 10, it is characterised in that institute Input module is stated, is additionally operable to after the structure module generates ontology knowledge base, in the ontology knowledge Increment language material is input in storehouse;
The ontology extraction unit, is additionally operable to the extraction algorithm that determines by the determining unit from described defeated Enter and extract in the increment language material of module input increment body;
The Relation extraction unit, is additionally operable to regular from institute according to the ontological relationship of determining unit determination State and extract in the increment body of ontology extraction unit extraction increment ontological relationship;
The fusion knowledge base processing meanss also include:Update module, for by the ontology extraction list The increment ontological relationship that the increment body and the Relation extraction unit that unit extracts is extracted is to the ontology knowledge Storehouse is updated.
13. a kind of knowledge base management systems, it is characterised in that include:Question and answer system and such as claim Fusion knowledge base processing meanss any one of 7~12;
Wherein, the question and answer system includes:User interactive module, for obtaining the problem of user input;
Language processing module, for carrying out Language Processing, institute to the problem that the user interactive module is obtained Language Processing is stated including participle, part-of-speech tagging and Entity recognition;
Standardized module, for being standardized to the problem after language processing module process, The standardization includes wrong word correction, language conversion and standardization and words recognition;
Semantic understanding module, for carrying out semantic understanding to the problem after standardized module process, obtains Take the semantic information of the problem;
Information searching module, for the semantic information that the semantic understanding module is obtained to be known in the body Know and enter in storehouse line retrieval, obtain the answer of the problem.
14. knowledge base management systems according to claim 13, it is characterised in that the fusion is known Knowing storehouse processing meanss includes the rule of inference module;The question and answer system also includes:Knowledge reasoning module, Answer for getting in described information retrieval module is space-time, is generated according to the rule of inference module Knowledge inference rule storehouse make inferences, obtain the answer of the problem.
15. knowledge base management systems according to claim 14, it is characterised in that described information is examined Rope module, the answer for being additionally operable to be got in the knowledge reasoning module is space-time, by the semantic understanding The semantic information that module is obtained enters line retrieval in the FAQ storehouses, obtains the answer of the problem.
16. knowledge base management systems according to claim 13, it is characterised in that the question and answer dress Putting also includes:Caching answer acquisition module, in the standardized module to the language processing module After problem after process is standardized, the answer of the problem is obtained from caching.
17. knowledge base management systems according to any one of claim 13~16, it is characterised in that Also include managing device;
Wherein, the managing device includes:Knowledge edition module, for processing the fusion knowledge base The ontologies that device is extracted enter edlin, check and correction, and increase new ontologies;
Knowledge examination module, for carrying out multistage auditing to the ontologies in the fusion knowledge base;
Statistical analysis module, for during being putd question to and being obtained answer by the question and answer system, Problem that statistics is popular, popular knowledge, the ontologies not accessed in the fusion knowledge base, and occur One or more in the ontologies of wrong answer, and analyze by question and answer system acquisition answer Accuracy rate;
On-line testing module, for verifying to the ontologies that the knowledge edition module is edited, Obtain on-line testing result;
Rules administration module, for being managed editor to the mode for extracting and rule;
Material management module, for storing the FAQ storehouses and the initial content in the ontology knowledge base, And according to the content in the preset time storage fusion knowledge base;
Intelligent quality testing module, for determining the standard that answer is obtained by the question and answer system by sampling inspection Really rate, is additionally operable to the accuracy rate of the automatic detection answer in the course of work of the question and answer system.
CN201510738095.1A 2015-11-03 2015-11-03 Fusion knowledge base processing method and device and knowledge base management system Pending CN106649394A (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201510738095.1A CN106649394A (en) 2015-11-03 2015-11-03 Fusion knowledge base processing method and device and knowledge base management system
PCT/CN2016/104136 WO2017076263A1 (en) 2015-11-03 2016-10-31 Method and device for integrating knowledge bases, knowledge base management system and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510738095.1A CN106649394A (en) 2015-11-03 2015-11-03 Fusion knowledge base processing method and device and knowledge base management system

Publications (1)

Publication Number Publication Date
CN106649394A true CN106649394A (en) 2017-05-10

Family

ID=58661667

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510738095.1A Pending CN106649394A (en) 2015-11-03 2015-11-03 Fusion knowledge base processing method and device and knowledge base management system

Country Status (2)

Country Link
CN (1) CN106649394A (en)
WO (1) WO2017076263A1 (en)

Cited By (20)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108595471A (en) * 2018-03-07 2018-09-28 中山大学 A kind of knowledge acquisition method based on intelligent planning
CN108959290A (en) * 2017-05-18 2018-12-07 日本电气株式会社 The processing method and equipment of knowledge data
CN108959488A (en) * 2018-06-22 2018-12-07 阿里巴巴集团控股有限公司 Safeguard the method and device of Question-Answering Model
CN109284383A (en) * 2018-10-09 2019-01-29 北京来也网络科技有限公司 Text handling method and device
CN109388793A (en) * 2017-08-03 2019-02-26 阿里巴巴集团控股有限公司 Entity mask method, intension recognizing method and corresponding intrument, computer storage medium
CN109460452A (en) * 2018-10-10 2019-03-12 长沙师范学院 Intelligent customer service system based on ontology
CN109460448A (en) * 2018-08-31 2019-03-12 厦门快商通信息技术有限公司 It is a kind of can autonomous configuration FAQ service framework
CN109460459A (en) * 2018-10-31 2019-03-12 神思电子技术股份有限公司 A kind of conversational system automatic optimization method based on log study
CN109582798A (en) * 2017-09-29 2019-04-05 阿里巴巴集团控股有限公司 Automatic question-answering method, system and equipment
CN110083608A (en) * 2019-04-26 2019-08-02 北京零秒科技有限公司 The contents management method and device in knowledge based library
CN110110065A (en) * 2019-05-08 2019-08-09 北京颢云信息科技股份有限公司 A kind of natural language question answering system of knowledge based map
CN110209589A (en) * 2019-06-05 2019-09-06 北京百度网讯科技有限公司 Knowledge base system test method, device, equipment and medium
CN110298445A (en) * 2019-05-30 2019-10-01 合肥阿拉丁智能科技有限公司 Deep learning autonomous operation method
CN110309509A (en) * 2019-06-28 2019-10-08 神思电子技术股份有限公司 A kind of semantic knowledge base construction method
CN111949855A (en) * 2020-07-31 2020-11-17 国网上海市电力公司 Knowledge map-based engineering technology knowledge retrieval platform and method thereof
CN112015919A (en) * 2020-09-15 2020-12-01 重庆广播电视大学重庆工商职业学院 Dialogue management method based on learning auxiliary knowledge graph
WO2021044246A1 (en) * 2019-09-03 2021-03-11 International Business Machines Corporation Resolving queries using structured and unstructured data
CN112612866A (en) * 2020-12-29 2021-04-06 北京奇艺世纪科技有限公司 Knowledge base text synchronization method and device, electronic equipment and storage medium
CN116383345A (en) * 2023-06-05 2023-07-04 中国医学科学院医学信息研究所 Method, device, electronic equipment and storage medium for body fusion
CN116860893A (en) * 2023-07-14 2023-10-10 浪潮智慧科技有限公司 Water conservancy data management method and system

Families Citing this family (31)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110019709B (en) * 2017-11-22 2024-05-17 上海智臻智能网络科技股份有限公司 Automatic question and answer method for robot and automatic question and answer system for robot
CN110837563B (en) * 2018-08-17 2023-05-30 阿里巴巴集团控股有限公司 Case judge method, device and system
CN109598384A (en) * 2018-12-06 2019-04-09 同方知网(北京)技术有限公司 A kind of agricultural industry innovation service map construction system
CN110580339B (en) * 2019-08-21 2023-04-07 华东理工大学 Method and device for perfecting medical term knowledge base
CN110619038A (en) * 2019-09-20 2019-12-27 上海氦豚机器人科技有限公司 Method, system and electronic equipment for vertically guiding professional consultation
CN110807603B (en) * 2019-11-12 2023-08-08 中国舰船研究设计中心 Knowledge application system and method for ship living cabin arrangement design
CN111104501A (en) * 2019-12-23 2020-05-05 中国银行股份有限公司 Knowledge determination method and device for test service
CN111159045B (en) * 2019-12-31 2024-04-19 中国银行股份有限公司 Compatibility problem detection method, device and storage medium
CN111414331B (en) * 2020-03-26 2023-08-08 北京字节跳动网络技术有限公司 Document importing method, device, storage medium and equipment of online collaborative knowledge base
CN111506719B (en) * 2020-04-20 2023-09-12 深圳追一科技有限公司 Associated question recommending method, device and equipment and readable storage medium
CN111708869B (en) * 2020-05-12 2023-07-14 北京明略软件系统有限公司 Processing method and device for man-machine conversation
CN111914550B (en) * 2020-07-16 2023-12-15 华中师范大学 Knowledge graph updating method and system oriented to limited field
CN111916110B (en) * 2020-08-06 2024-04-26 龙马智芯(珠海横琴)科技有限公司 Voice quality inspection method and device
CN111782825A (en) * 2020-08-20 2020-10-16 支付宝(杭州)信息技术有限公司 Knowledge base construction method and device
CN112163094B (en) * 2020-08-25 2022-10-14 中国科学院计算机网络信息中心 Scientific and technological resource convergence and continuous service method and device
CN112015915A (en) * 2020-09-01 2020-12-01 哈尔滨工业大学 Question-answering system and device based on knowledge base generated by questions
CN112015920A (en) * 2020-09-15 2020-12-01 重庆广播电视大学重庆工商职业学院 Intelligent auxiliary learning system based on knowledge graph and edge calculation
CN112463986A (en) * 2020-12-08 2021-03-09 北京明略软件系统有限公司 Information storage method and device
CN112612902B (en) * 2020-12-23 2023-07-14 国网浙江省电力有限公司电力科学研究院 Knowledge graph construction method and device for power grid main equipment
CN112528046B (en) * 2020-12-25 2023-09-15 网易(杭州)网络有限公司 New knowledge graph construction method and device and information retrieval method and device
CN112801492B (en) * 2021-01-22 2023-07-25 中国平安人寿保险股份有限公司 Knowledge-hierarchy-based data quality inspection method and device and computer equipment
CN113268604B (en) * 2021-05-19 2024-06-07 国网辽宁省电力有限公司 Knowledge base self-adaptive expansion method and system
CN113515643A (en) * 2021-05-19 2021-10-19 思必驰科技股份有限公司 User identity verification method and device
CN113282698A (en) * 2021-06-07 2021-08-20 中国科学院地理科学与资源研究所 Construction method of ecological civilization geographical knowledge standardized knowledge base
CN113568998A (en) * 2021-06-18 2021-10-29 武汉理工数字传播工程有限公司 Knowledge service resource processing method, device, equipment and storage medium
CN113535685B (en) * 2021-07-28 2024-05-17 深圳供电局有限公司 Method for constructing event knowledge base of intelligent power grid dispatching
CN114638230B (en) * 2022-03-16 2023-05-05 四川智胜慧旅科技有限公司 Internet big data analysis method and system
CN114399006B (en) * 2022-03-24 2022-07-12 山东省计算中心(国家超级计算济南中心) Multi-source abnormal composition image data fusion method and system based on super-calculation
CN116775847B (en) * 2023-08-18 2023-11-28 中国电子科技集团公司第十五研究所 Question answering method and system based on knowledge graph and large language model
CN117077792B (en) * 2023-10-12 2024-01-09 支付宝(杭州)信息技术有限公司 Knowledge graph-based method and device for generating prompt data
CN118036753A (en) * 2024-03-25 2024-05-14 行至智能(北京)技术有限公司 Large language model reasoning method and system based on multistage knowledge retrieval enhancement

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101373532A (en) * 2008-07-10 2009-02-25 昆明理工大学 FAQ Chinese request-answering system implementing method in tourism field
CN103902652A (en) * 2014-02-27 2014-07-02 深圳市智搜信息技术有限公司 Automatic question-answering system
CN104850539A (en) * 2015-05-28 2015-08-19 宁波薄言信息技术有限公司 Natural language understanding method and travel question-answering system based on same

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7543232B2 (en) * 2004-10-19 2009-06-02 International Business Machines Corporation Intelligent web based help system
US8832064B2 (en) * 2005-11-30 2014-09-09 At&T Intellectual Property Ii, L.P. Answer determination for natural language questioning

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101373532A (en) * 2008-07-10 2009-02-25 昆明理工大学 FAQ Chinese request-answering system implementing method in tourism field
CN103902652A (en) * 2014-02-27 2014-07-02 深圳市智搜信息技术有限公司 Automatic question-answering system
CN104850539A (en) * 2015-05-28 2015-08-19 宁波薄言信息技术有限公司 Natural language understanding method and travel question-answering system based on same

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
徐立恒 等: "基于多特征表示的本体概念挂载", 《中文信息学报》 *
朱旺南: "基于本体的自动问答客服系统研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 *

Cited By (31)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108959290A (en) * 2017-05-18 2018-12-07 日本电气株式会社 The processing method and equipment of knowledge data
CN109388793A (en) * 2017-08-03 2019-02-26 阿里巴巴集团控股有限公司 Entity mask method, intension recognizing method and corresponding intrument, computer storage medium
CN109388793B (en) * 2017-08-03 2023-04-07 阿里巴巴集团控股有限公司 Entity marking method, intention identification method, corresponding device and computer storage medium
CN109582798A (en) * 2017-09-29 2019-04-05 阿里巴巴集团控股有限公司 Automatic question-answering method, system and equipment
CN108595471A (en) * 2018-03-07 2018-09-28 中山大学 A kind of knowledge acquisition method based on intelligent planning
CN108595471B (en) * 2018-03-07 2022-08-02 中山大学 Knowledge acquisition method based on intelligent planning
CN108959488B (en) * 2018-06-22 2021-12-07 创新先进技术有限公司 Method and device for maintaining question-answering model
CN108959488A (en) * 2018-06-22 2018-12-07 阿里巴巴集团控股有限公司 Safeguard the method and device of Question-Answering Model
CN109460448A (en) * 2018-08-31 2019-03-12 厦门快商通信息技术有限公司 It is a kind of can autonomous configuration FAQ service framework
CN109284383A (en) * 2018-10-09 2019-01-29 北京来也网络科技有限公司 Text handling method and device
CN109460452A (en) * 2018-10-10 2019-03-12 长沙师范学院 Intelligent customer service system based on ontology
CN109460459A (en) * 2018-10-31 2019-03-12 神思电子技术股份有限公司 A kind of conversational system automatic optimization method based on log study
CN110083608A (en) * 2019-04-26 2019-08-02 北京零秒科技有限公司 The contents management method and device in knowledge based library
CN110083608B (en) * 2019-04-26 2021-10-15 北京零秒科技有限公司 Content management method and device based on knowledge base
CN110110065A (en) * 2019-05-08 2019-08-09 北京颢云信息科技股份有限公司 A kind of natural language question answering system of knowledge based map
CN110298445A (en) * 2019-05-30 2019-10-01 合肥阿拉丁智能科技有限公司 Deep learning autonomous operation method
CN110209589A (en) * 2019-06-05 2019-09-06 北京百度网讯科技有限公司 Knowledge base system test method, device, equipment and medium
CN110209589B (en) * 2019-06-05 2022-11-18 北京百度网讯科技有限公司 Knowledge base system test method, device, equipment and medium
CN110309509A (en) * 2019-06-28 2019-10-08 神思电子技术股份有限公司 A kind of semantic knowledge base construction method
WO2021044246A1 (en) * 2019-09-03 2021-03-11 International Business Machines Corporation Resolving queries using structured and unstructured data
CN114391142A (en) * 2019-09-03 2022-04-22 国际商业机器公司 Parsing queries using structured and unstructured data
GB2603323A (en) * 2019-09-03 2022-08-03 Ibm Resolving queries using structured and unstructured data
US11841883B2 (en) 2019-09-03 2023-12-12 International Business Machines Corporation Resolving queries using structured and unstructured data
CN111949855A (en) * 2020-07-31 2020-11-17 国网上海市电力公司 Knowledge map-based engineering technology knowledge retrieval platform and method thereof
CN112015919A (en) * 2020-09-15 2020-12-01 重庆广播电视大学重庆工商职业学院 Dialogue management method based on learning auxiliary knowledge graph
CN112612866A (en) * 2020-12-29 2021-04-06 北京奇艺世纪科技有限公司 Knowledge base text synchronization method and device, electronic equipment and storage medium
CN112612866B (en) * 2020-12-29 2023-07-21 北京奇艺世纪科技有限公司 Knowledge base text synchronization method and device, electronic equipment and storage medium
CN116383345A (en) * 2023-06-05 2023-07-04 中国医学科学院医学信息研究所 Method, device, electronic equipment and storage medium for body fusion
CN116383345B (en) * 2023-06-05 2023-08-22 中国医学科学院医学信息研究所 Method, device, electronic equipment and storage medium for body fusion
CN116860893A (en) * 2023-07-14 2023-10-10 浪潮智慧科技有限公司 Water conservancy data management method and system
CN116860893B (en) * 2023-07-14 2024-03-08 浪潮智慧科技有限公司 Water conservancy data management method and system

Also Published As

Publication number Publication date
WO2017076263A1 (en) 2017-05-11

Similar Documents

Publication Publication Date Title
CN106649394A (en) Fusion knowledge base processing method and device and knowledge base management system
JP4920023B2 (en) Inter-object competition index calculation method and system
US11347803B2 (en) Systems and methods for adaptive question answering
US9092789B2 (en) Method and system for semantic analysis of unstructured data
US8812411B2 (en) Domains for knowledge-based data quality solution
CN109446341A (en) The construction method and device of knowledge mapping
CN107944025A (en) Information-pushing method and device
CN111538844B (en) Target field knowledge base generation and problem solution method and device
US20130117202A1 (en) Knowledge-based data quality solution
CN107077486A (en) Affective Evaluation system and method
US20130117219A1 (en) Architecture for knowledge-based data quality solution
CN105912645B (en) A kind of intelligent answer method and device
CN109063000A (en) Question sentence recommended method, customer service system and computer readable storage medium
CN105912629A (en) Intelligent question and answer method and device
EP3855320A1 (en) Systems and methods for adaptive question answering related applications
CN110348719A (en) A kind of risk control method based on user information knowledge mapping, device and electronic equipment
US20170249555A1 (en) Concepts and link discovery system
CN108009229A (en) Method, terminal device and the storage medium that public sentiment event data is found
CN111177653B (en) Credit evaluation method and device
CN113326381A (en) Semantic and knowledge graph analysis method, platform and equipment based on dynamic ontology
CN113407734B (en) Method for constructing knowledge graph system based on real-time big data
CN115510247A (en) Method, device, equipment and storage medium for constructing electric carbon policy knowledge graph
US11972356B2 (en) System and/or method for an autonomous linked managed semantic model based knowledge graph generation framework
CN115329078A (en) Text data processing method, device, equipment and storage medium
US20150100515A1 (en) Customer data unification

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
RJ01 Rejection of invention patent application after publication

Application publication date: 20170510