CN105912600A - Question-answer knowledge base and establishing method thereof, intelligent question-answering method and system - Google Patents
Question-answer knowledge base and establishing method thereof, intelligent question-answering method and system Download PDFInfo
- Publication number
- CN105912600A CN105912600A CN201610207928.6A CN201610207928A CN105912600A CN 105912600 A CN105912600 A CN 105912600A CN 201610207928 A CN201610207928 A CN 201610207928A CN 105912600 A CN105912600 A CN 105912600A
- Authority
- CN
- China
- Prior art keywords
- speech
- semantic
- question
- word
- answer
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/903—Querying
- G06F16/9032—Query formulation
- G06F16/90332—Natural language query formulation or dialogue systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/284—Lexical analysis, e.g. tokenisation or collocates
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Databases & Information Systems (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention brings forward a question-answer knowledge base and an establishing method thereof, an intelligent question-answering method and system. The question-answer knowledge base comprises multiple knowledge points. Each knowledge point comprises one or more questions. At least a number of problems are with semantic expressions. Each semantic expression comprises one or more first-stage word classes. The first-stage word class comprises multiple related words. A part of first-stage word classes in semantic expressions of questions in at least partial number further comprises semantic segments. Each semantic segment comprises more than one second-stage word class including multiple related words. The question-answer knowledge base and the establishing method thereof, the intelligent question-answering method and system have following beneficial effects: time cost for semantic expansion when knowledge engineers write semantic expressions so that information search efficiency of the intelligent question-answering system is improved.
Description
Technical field
The present invention relates to information search technique field, particularly relate to a kind of question and answer knowledge base and method for building up thereof,
Intelligent answer method and system.
Background technology
In the middle of the word of Chinese, a lot of words are all by multiple different word synthesis, such as " payment " just
Synthesized by " payment " and " fund ".Compound word has a lot of synonym, same such as " payment "
Justice word has " payment ", " paying " etc..In this case, the mark in knowledge engineer is knowledge base
Standard is asked or extends when asking the expression formula writing correspondence, it is contemplated that compound word and synonym thereof, it is necessary to
Write out the multiple semantic formulas being made up of the different terms that implication is identical, such as: for the mark in knowledge base
Standard asks " when paying the bill ", and the expression formula that knowledge engineer writes out has: 1) " when "+" pay
Money ", 2) " when "+" payment "+" fund ", 3) " when "+" paying ", etc..Can
Seeing, the workload of knowledge engineer is the biggest.Follow-up, after user inputs user request information, by user
When solicited message and above-mentioned numerous semantic formula carry out Similarity Measure one by one, owing to the difference of weight is led
The Similarity value causing same semantic different expression formulas is different, extends the time of information search, reduces information
The efficiency of search.
Summary of the invention
The technical problem to be solved in the present invention is to provide a kind of question and answer knowledge base and method for building up thereof, intelligence is asked
Answer method and system, reduce information search and process the consumption of resource.
The technical solution used in the present invention is to provide a kind of question and answer knowledge base, including multiple knowledge points, each
Knowledge point includes one or more problem, and the problem of at least part of number is set up semantic formula, institute's predicate
Justice expression formula includes that one or more first order part of speech, described first order part of speech include multiple relevant word, extremely
Part first order part of speech in the semantic formula of the problem of small part number also includes semantic segment, institute's predicate
Justice fragment includes that more than one second level part of speech, described second level part of speech include multiple relevant word.
Further, described semantic formula also includes: semantic rule word.
Further, described second level part of speech also includes more than one third level part of speech, described third level word
Class includes multiple relevant word.
Further, at least some of second level part of speech in described semantic segment is and column selection from each other
Relation.
Further, it is unordered expression between each second level part of speech in same semantic segment.
The present invention also provides for the method for building up of a kind of question and answer knowledge base, and described question and answer knowledge base includes multiple knowledge
Point, each knowledge point includes one or more problem, and the problem of at least part of number is set up semantic formula,
Described method includes:
Part of speech data base is provided;Part of speech in described part of speech data base includes multiple relevant word;
When the semantic formula set up for problem includes first order part of speech, it is judged that described first order part of speech is
No comprise compound word;
When described first order part of speech comprises compound word, it is judged that whether described compound word has word belong to institute's predicate
Part of speech in class data base;
When compound word there being word belong to the part of speech in described part of speech data base, institute's predicate is replaced with institute's predicate
Affiliated part of speech is using as second level part of speech.
Further, in first order part of speech is compound word and described compound word, there is arbitrary word and be belonging respectively to word
During the different part of speech of two or more in class data base, using the two or more part of speech belonging to institute's predicate all as
Two grades of parts of speech and be each other and the relation of column selection.
The present invention also provides for a kind of intelligent answer method, including:
Thering is provided above-mentioned question and answer knowledge base, described knowledge point also includes answer;
Obtain user request information;
Described user request information is carried out Semantic Similarity Measurement with the problem in described question and answer knowledge base, and
Answer corresponding for problem the highest for semantic similitude angle value is supplied to user.
Further, when carrying out Semantic Similarity Measurement, the used problem in described question and answer knowledge base
In first order part of speech described in semantic formula, each word is identical with the weight of each second level part of speech.
The present invention also provides for the system of setting up of a kind of question and answer knowledge base, and described question and answer knowledge base includes multiple knowledge
Point, each knowledge point includes one or more problem, and the problem of at least part of number is set up semantic formula,
Described system includes:
First provides module, is used for providing part of speech data base, and the part of speech in described part of speech data base includes multiple
Relevant word;
First judge module, for when the semantic formula set up for problem includes first order part of speech, it is judged that
Whether described first order part of speech comprises compound word;
Second judge module, for when described first order part of speech comprises compound word, it is judged that in described compound word
Word whether is had to belong to described part of speech data base;
Replacement module, for when there being word to belong to the part of speech in described part of speech data base, by described in compound word
Word replaces with the part of speech belonging to institute's predicate using as second level part of speech.
Further, described replacement module is additionally operable to: in first order part of speech is compound word and described compound word
When there is the different part of speech of two or more that arbitrary word is belonging respectively in part of speech data base, belonging to institute's predicate
Two or more part of speech is all as second level part of speech and be each other and the relation of column selection.
The present invention also provides for a kind of Intelligent Answer System, including:
Second provides module, is used for providing above-mentioned question and answer knowledge base, and described knowledge point also includes answer;
Acquisition module, is used for obtaining user request information;
Computing module, for carrying out semanteme by described user request information with the problem in described question and answer knowledge base
Similarity Measure obtains semantic similitude angle value, and answer corresponding for problem the highest for semantic similitude angle value is provided
To user.
Further, when described computing module carries out Semantic Similarity Measurement, used described question and answer knowledge
Described in the semantic formula of the problem in storehouse, in first order part of speech, each word is identical with the weight of each second level part of speech.
Using technique scheme, the present invention at least has the advantage that
Question and answer knowledge base of the present invention and method for building up, intelligent answer method and system, due to according to conjunction
Become resolution characteristic and the relatedness of part of speech of word, the semantic formula of the problem in question and answer knowledge base is carried out
The most careful comprehensive rewriting, enumerates the grammatical category information corresponding to each word forming this compound word, at word
Under the support of class data base, the part of speech of part of speech data base can be quoted by level, makes the semantic formula of problem
Included in the parts of speech at different levels information that comprised of reality more substantial, also save simultaneously and write for problem
The number of semantic formula, thus decrease knowledge engineer and carry out semantic extension when writing semantic formula
Time cost.
The advantage of the question and answer knowledge base of the present invention is: simple in construction, internal memory are few, information is more complete.Adopting
When responding user request information by the question and answer knowledge base of the present invention, carry out the institute used in Semantic Similarity Measurement
State described in the semantic formula of problem in question and answer knowledge base each word and each second level part of speech in first order part of speech
Weight identical so that the Similarity value of the expression formula expressing same semanteme is identical, improves intelligent answer system
The information search efficiency of system.
Accompanying drawing explanation
Fig. 1 is the method for building up flow chart of the question and answer knowledge base of second embodiment of the invention;
Fig. 2 is the intelligent answer method flow diagram of third embodiment of the invention;
Fig. 3 is system of the setting up composition structural representation of the question and answer knowledge base of fourth embodiment of the invention;
Fig. 4 is the Intelligent Answer System composition structural representation of fifth embodiment of the invention.
Detailed description of the invention
By further illustrating the technological means and effect that the present invention taked by reaching predetermined purpose, below tie
Close accompanying drawing and preferred embodiment, after the present invention is described in detail such as.
First introduce the question and answer knowledge base of this area, the most original and simplest shape in the knowledge point in knowledge base
Formula is exactly " problem and the answer FAQ " commonly used at ordinary times, and general form is that " ask-answer " is right.Such as, " color
The rate of bell " it is exactly to express standard clearly to ask description.Here " asking " should be narrowly interpreted as " asking
Ask ", and should broadly understand one " input " i.e. user request information, " input " should have the " defeated of correspondence
Go out ".Such as, for for the semantics recognition of control system, an instruction of user, such as " open
Radio " also should be understood to be one " asking ", now corresponding " answering " could be for performing phase
Calling of the control program that should control.
User is when inputting to machine, and optimal situation is that use standard is asked, then the intelligent semantic of machine is known
Other system is at once it will be appreciated that the meaning of user.But, user the most not uses standard to ask, but
The form of some deformation that standard is asked.Such as, if the standard form of asking switched for wireless radio station is " to change
One radio station ", then the order that user may use is " switching a radio station ", and machine is also required to know
What other user expressed is the same meaning.
Therefore, for intelligent semantic identification, the extension that the standard that needs in knowledge base is asked is asked, this extension
Ask and ask that expression-form has slight difference with standard, but express identical implication.
In an embodiment of the present invention, standard asked and extend and ask the problem being referred to as in question and answer knowledge base.
Introducing the concept of the part of speech of this area again, part of speech is to carry out dividing, one according to the semanteme of word
The phrase that group is relevant is woven in together the part of speech data base forming a tree.
Part of speech is to collect one group of related term, is made up of part of speech title and one group of related term.Related term can
Think synonym, it is also possible to for semantic dissimilar identity set word.Part of speech name is to have in this group related term
There is the representative of the word of label effect, i.e. part of speech.Including at least two words in one part of speech.When a word not
When belonging to any part of speech, then it is word, as:.
Next introducing the concept of semantic formula, a part of problem in question and answer knowledge base is to there being language
Justice expression formula, semantic formula is mainly made up of word, part of speech and their "or" relation, and its core depends on
Lai Yu " part of speech ".Between semantic formula and user request information, relation is the value (similarity) by quantifying
Representing, this value quantified makes the similarity between the problem in knowledge base and user request information simultaneously
Can compare.Due to semantic formula Similarity Measure to be participated in, so there to be enough abilities to express
Semantic.
In semantic formula, for distinguishing the word in expression formula and part of speech, square brackets [] can be used to express word
Class, is presented herein below the example of some simple semantic formulas: [Fetion] [how] [open-minded], [introduction] [multimedia message] [industry
Business], [login] [method], [call reminding] of [Fetion] [how] [charge].At multiple words in a different order
Permutation and combination together after in the case of expressed semanteme is the same meaning, how such as: " " " handling " is " color
Bell " and how " CRBT " " " " handling " expressed by semanteme be all CRBT handle method, can be by language
How justice expression formula is write as [] [handling] [CRBT], and this semantic formula comprises above-mentioned two kind way to put questions, this
In semantic formula, each part of speech can be lack of alignment.Exist in multiple words permutation and combination in a different order
In the case of semanteme expressed after together is the different meaning, such as: " dollar " " exchange " " RMB " " converges
Rate " and " RMB " " exchange " " dollar " " exchange rate " formed with same word, but expressed language
Justice is but different, now, can only each part of speech ordered arrangement in semantic formula, or be that [dollar] [is converted
Change] [RMB] [exchange rate], or be [RMB] [exchange] [dollar] [exchange rate], express correct implication.
It should be noted that above-mentioned symbol is only for example, it is not intended to protection scope of the present invention.
First embodiment of the invention, a kind of question and answer knowledge base, including multiple knowledge points, each knowledge point includes
One or more problems, the problem of at least part of number is set up semantic formula, described semantic formula bag
Including one or more first order part of speech, described first order part of speech includes multiple relevant word, at least part of number
Problem semantic formula in part first order part of speech also include that semantic segment, described semantic segment include
More than one second level part of speech, described second level part of speech includes multiple relevant word.
Concrete, in the present embodiment, semantic segment can comprise more than part of speech, may also include word.First
Level part of speech includes word and semantic segment;Semantic segment includes again second level part of speech, it is also possible to include word.
For problem more complicated in question and answer knowledge base, described semantic formula also includes: semantic rule word,
Such as:, pass through etc.;For problem better simply in question and answer knowledge base, in its semantic formula
Can only comprise part of speech, and semantic rule word need not be comprised.
Preferably, described second level part of speech can also include more than one third level part of speech, the described third level
Part of speech includes multiple relevant word.
In some cases, at least some of second level part of speech in described semantic segment is arranged side by side from each other
The relation selected.Such as: in first order part of speech is compound word and described compound word, there is arbitrary independent word pair
Should be when the part of speech that two or more is different, using two or more part of speech corresponding for described independent word all as the second level
Part of speech and be each other and the relation of column selection.
Generally, in ordered arrangement state between the second level part of speech in same semantic segment.Optionally, same
It is unordered expression between each second level part of speech in semantic segment.
The present embodiment is exemplified below:
When one problem " is paid the bill ", the mode of writing of existing semantic formula be intended to write to
Few following two semantic formulas:
[when] [pay the bill]
[when] [pay] [fund]
Visible, the number of semantic formula to be write much more, and due to the weight of [payment] and [payment] [fund]
Weighted, cause two semantic formulas and same user request information to carry out Semantic Similarity Measurement
Time, Similarity value may be different.
The embodiment of the present invention, creates a semantic segment below [payment] this first order part of speech: [payment],
[fund], actually [pays], [fund] is second level part of speech, then have only to this problem above
Write a semantic formula: [when] [payment].
Semantic segment [payment] [fund], therefore [what time is included due in now first order part of speech [payment]
Wait] [payment] this semantic formula also includes [when] [payment] [fund] this semantic formula.
During additionally, due to carry out Similarity Measure, the weight of first order part of speech [payment] be exactly the most each word or
The weight of second level part of speech, the weight of [payment] [fund] now is identical with the weight of [payment], thus with existing
When having technology to compare, [] [payment] is identical with the Similarity value of [when] [payment] [fund].
It should be noted that word corresponding to semantic segment is compound word, the most plural word, wherein
Can be second level part of speech with each word, it is also possible to only partial words is second level part of speech (i.e. remaining word
For word), it is all within protection scope of the present invention.
Above-mentioned example is unordered expression between second level part of speech [payment] and [fund], thus user request information
" fund when paid " also can be matched [when] [payment] this semantic formula.
And in the prior art, when not existing [when] [payment] [fund] this expression formula, instant [payment]
Include " payment " this compound word, then " fund when paid " this user asks
Information also will not match [when] [payment] this semantic formula.
It should be noted that in other embodiments of the invention semantic segment also may be used between the part of speech of the second level
Thinking expression in order, it is not intended to protection scope of the present invention.
The semantic formula that this problem is corresponding above can also present lack of alignment between the part of speech of each second level
State, it may be assumed that [payment] [fund] [when].
When there being customer problem " when fund pays " to input, it is possible to by by unordered semanteme
The semantic formula [payment] [fund] [when] of fragment composition is captured.
Second embodiment of the invention, the method for building up of a kind of question and answer knowledge base, described question and answer knowledge base includes many
Individual knowledge point, each knowledge point includes one or more problem, and the problem of at least part of number is set up semanteme
Expression formula, as it is shown in figure 1, described method includes:
Step S101, it is provided that part of speech data base;Part of speech in described part of speech data base includes multiple relevant word;
Step S102, when the semantic formula set up for problem includes first order part of speech, it is judged that described first
Whether level part of speech comprises compound word;
Step S103, when described first order part of speech comprises compound word, it is judged that whether have word in described compound word
Belong to the part of speech in described part of speech data base;
Step S104, when there being word to belong to the part of speech in described part of speech data base in compound word, replaces institute's predicate
It is changed to the part of speech belonging to institute's predicate using as second level part of speech.
Concrete, in first order part of speech is compound word and described compound word, there is arbitrary word is belonging respectively to part of speech
During the different part of speech of two or more in data base, using the two or more part of speech belonging to institute's predicate all as second
Level part of speech and be the relation of also column selection each other.
Such as: " fund " word in " payment " this compound word belongs simultaneously in part of speech data base
[fund] and [banknote] two parts of speech, " payment " belongs to and [pays] this part of speech in part of speech data base, the most now exist
The second level part of speech set up under [payment] this first order part of speech is [payment] [fund | banknote], " | " therein
Represent coordination.The method for building up using the question and answer knowledge base in the present embodiment can be set up as first implements
Question and answer knowledge base described in example.
In addition it should be noted that in the question and answer knowledge base of first embodiment of the invention, in semantic formula
Part first order part of speech also includes that semantic segment, described semantic segment include more than one second level part of speech,
I.e. using a semantic segment to instead of one or more word, a first order part of speech is permissible in the present embodiment
Including one or more semantic segments.
Due to the present embodiment resolution characteristic according to compound word and the relatedness of part of speech, in question and answer knowledge base
The semantic formula of problem carried out the most careful comprehensive rewriting, enumerate each list forming this compound word
The solely grammatical category information corresponding to word, under the support of part of speech data base, the part of speech of part of speech data base can be with tegillum
Level is quoted, and the information making the reality of the parts of speech at different levels included in the semantic formula of problem be comprised is more substantial,
Also save the number of the semantic formula write for problem simultaneously, thus decrease knowledge engineer and writing
The time cost of semantic extension is carried out during semantic formula.
Third embodiment of the invention, a kind of intelligent answer method, use the question and answer knowledge base in first embodiment
Or using the question and answer knowledge base set up by the second embodiment, the knowledge point in question and answer knowledge base also includes answer;
As in figure 2 it is shown, described method includes:
Step S201, obtains user request information;
Step S202, carries out semantic similitude by described user request information with the problem in described question and answer knowledge base
Degree calculates, and answer corresponding for problem the highest for semantic similitude angle value is supplied to user.
Concrete, when carrying out Semantic Similarity Measurement, the language of the used problem in described question and answer knowledge base
Described in justice expression formula, in first order part of speech, each word is identical with the weight of each second level part of speech.
The present embodiment, owing to have employed above-mentioned question and answer knowledge base, therefore can significantly improve the standard of intelligent answer
Really rate.
Fourth embodiment of the invention, corresponding with the second embodiment, the present embodiment introduces a kind of question and answer knowledge base
Setting up system, described question and answer knowledge base includes that multiple knowledge point, each knowledge point include one or more problem,
At least partly the problem of number is set up has semantic formula, as it is shown on figure 3, described system includes consisting of
Part:
Module 301 is provided, is used for providing part of speech data base, the part of speech in described part of speech data base to include multiple phase
The word closed;
First judge module 302, for when the semantic formula set up for problem includes first order part of speech, sentences
Whether disconnected described first order part of speech comprises compound word;
Second judge module 303, for when described first order part of speech comprises compound word, it is judged that described compound word
In whether have word to belong to described part of speech data base;
Replacement module 304, for when there being word to belong to the part of speech in described part of speech data base, by institute in compound word
Predicate replaces with the part of speech belonging to institute's predicate using as second level part of speech.
Concrete, described replacement module 304, it is additionally operable to: when first order part of speech is compound word and described compound word
Middle when there is the different part of speech of two or more that arbitrary word is belonging respectively in part of speech data base, belonging to institute's predicate
Two or more part of speech all as second level part of speech and be each other and the relation of column selection.
Fifth embodiment of the invention, corresponding with the 3rd embodiment, the present embodiment introduces a kind of Intelligent Answer System,
Using above-mentioned question and answer knowledge base, described knowledge point also includes answer;As shown in Figure 4, described system includes:
Acquisition module 401, is used for obtaining user request information;
Computing module 402, for carrying out language by described user request information with the problem in described question and answer knowledge base
Justice Similarity Measure obtains semantic similitude angle value, and answer corresponding for problem the highest for semantic similitude angle value is carried
Supply user.
Further, when computing module 402 carries out Semantic Similarity Measurement, used described question and answer knowledge
Described in the semantic formula of the problem in storehouse, in first order part of speech, each word is identical with the weight of each second level part of speech.
Intelligent Answer System in the embodiment of the present invention uses the question and answer knowledge base response user of the embodiment of the present invention
During solicited message, carry out the semantic table of problem in the described question and answer knowledge base used in Semantic Similarity Measurement
Reach each word in first order part of speech described in formula identical with the weight of each second level part of speech, improve intelligent answer system
The information search efficiency of system.
By the explanation of detailed description of the invention, it should can be to the present invention by reaching the technology that predetermined purpose is taked
Means and effect are able to more deeply and concrete understanding, but appended diagram is only to provide reference and explanation
With, not it is used for the present invention is any limitation as.
Claims (13)
1. a question and answer knowledge base, including multiple knowledge points, each knowledge point includes one or more problem,
At least partly the problem of number is set up has semantic formula, described semantic formula to include one or more first
Level part of speech, described first order part of speech includes multiple relevant word, it is characterised in that asking of at least part of number
Part first order part of speech in the semantic formula of topic also includes that semantic segment, described semantic segment include one
Above second level part of speech, described second level part of speech includes multiple relevant word.
Question and answer knowledge base the most according to claim 1, it is characterised in that described semantic formula is also wrapped
Include: semantic rule word.
Question and answer knowledge base the most according to claim 1, it is characterised in that described second level part of speech also wraps
Including more than one third level part of speech, described third level part of speech includes multiple relevant word.
Question and answer knowledge base the most according to claim 1, it is characterised in that in described semantic segment extremely
Few a part of second level part of speech is and the relation of column selection from each other.
Question and answer knowledge base the most according to claim 1, it is characterised in that each in same semantic segment
It it is unordered expression between the part of speech of the second level.
6. a method for building up for question and answer knowledge base, described question and answer knowledge base includes multiple knowledge point, Mei Gezhi
Knowing point and include one or more problem, the problem of at least part of number is set up semantic formula, and its feature exists
In, described method includes:
Thering is provided part of speech data base, the part of speech in described part of speech data base includes multiple relevant word;
When the semantic formula set up for problem includes first order part of speech, it is judged that described first order part of speech is
No comprise compound word;
When described first order part of speech comprises compound word, it is judged that whether described compound word has word belong to institute's predicate
Part of speech in class data base;
When compound word there being word belong to the part of speech in described part of speech data base, institute's predicate is replaced with institute's predicate
Affiliated part of speech is using as second level part of speech.
The method for building up of question and answer knowledge base the most according to claim 6, it is characterised in that work as the first order
Part of speech is to there is arbitrary word in compound word and described compound word to be belonging respectively to the two or more in part of speech data base not
With part of speech time, using the two or more part of speech belonging to institute's predicate all as second level part of speech and be also each other
The relation of column selection.
8. an intelligent answer method, it is characterised in that including:
Thering is provided the question and answer knowledge base as according to any one of claim 1 to 5, described knowledge point also includes answering
Case;
Obtain user request information;
Described user request information is carried out Semantic Similarity Measurement with the problem in described question and answer knowledge base, and
Answer corresponding for problem the highest for semantic similitude angle value is supplied to user.
Intelligent answer method the most according to claim 8, it is characterised in that carry out semantic similarity meter
During calculation, each in first order part of speech described in the semantic formula of the used problem in described question and answer knowledge base
Word is identical with the weight of each second level part of speech.
10. a system of setting up for question and answer knowledge base, described question and answer knowledge base includes multiple knowledge point, each
Knowledge point includes one or more problem, and the problem of at least part of number is set up semantic formula, its feature
Being, described system includes:
First provides module, is used for providing part of speech data base, and the part of speech in described part of speech data base includes multiple
Relevant word;
First judge module, for when the semantic formula set up for problem includes first order part of speech, it is judged that
Whether described first order part of speech comprises compound word;
Second judge module, for when described first order part of speech comprises compound word, it is judged that in described compound word
Word whether is had to belong to described part of speech data base;
Replacement module, for when there being word to belong to the part of speech in described part of speech data base, by described in compound word
Word replaces with the part of speech belonging to institute's predicate using as second level part of speech.
11. question and answer knowledge bases according to claim 10 set up system, it is characterised in that described replace
Die change block, is additionally operable to: there is arbitrary word in first order part of speech is compound word and described compound word and is belonging respectively to
During the different part of speech of two or more in part of speech data base, using the two or more part of speech belonging to institute's predicate all as
Second level part of speech and be each other and the relation of column selection.
12. 1 kinds of Intelligent Answer Systems, it is characterised in that including:
Second provides module, for providing the question and answer knowledge base as according to any one of claim 1 to 5,
Described knowledge point also includes answer;
Acquisition module, is used for obtaining user request information;
Computing module, for carrying out semanteme by described user request information with the problem in described question and answer knowledge base
Similarity Measure obtains semantic similitude angle value, and answer corresponding for problem the highest for semantic similitude angle value is provided
To user.
13. Intelligent Answer Systems according to claim 12, it is characterised in that described computing module enters
During row Semantic Similarity Measurement, described in the semantic formula of the used problem in described question and answer knowledge base
In first order part of speech, each word is identical with the weight of each second level part of speech.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610207928.6A CN105912600B (en) | 2016-04-05 | 2016-04-05 | Question and answer knowledge base and its method for building up, intelligent answer method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610207928.6A CN105912600B (en) | 2016-04-05 | 2016-04-05 | Question and answer knowledge base and its method for building up, intelligent answer method and system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105912600A true CN105912600A (en) | 2016-08-31 |
CN105912600B CN105912600B (en) | 2019-08-16 |
Family
ID=56744743
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610207928.6A Active CN105912600B (en) | 2016-04-05 | 2016-04-05 | Question and answer knowledge base and its method for building up, intelligent answer method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105912600B (en) |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106547741A (en) * | 2016-11-21 | 2017-03-29 | 江苏科技大学 | A kind of Chinese language text auto-collation based on collocation |
CN107256258A (en) * | 2017-06-12 | 2017-10-17 | 上海智臻智能网络科技股份有限公司 | Semantic formula generation method and device |
CN107798123A (en) * | 2017-11-10 | 2018-03-13 | 上海智臻智能网络科技股份有限公司 | Knowledge base and its foundation, modification, intelligent answer method, apparatus and equipment |
CN108090177A (en) * | 2017-12-15 | 2018-05-29 | 上海智臻智能网络科技股份有限公司 | The generation methods of more wheel question answering systems, equipment, medium and take turns question answering system more |
CN108628906A (en) * | 2017-03-24 | 2018-10-09 | 北京京东尚科信息技术有限公司 | Short text template method for digging, device, electronic equipment and readable storage medium storing program for executing |
CN109033262A (en) * | 2018-07-09 | 2018-12-18 | 北京寻领科技有限公司 | Question and answer knowledge base update method and device |
CN109783506A (en) * | 2018-12-05 | 2019-05-21 | 北京国电通网络技术有限公司 | A kind of spoken meaning of one's words understanding method, device and the electronic equipment of intelligent Answer System |
CN109947908A (en) * | 2017-11-22 | 2019-06-28 | 上海智臻智能网络科技股份有限公司 | The building method and construction system of robot knowledge base |
CN110019739A (en) * | 2017-11-30 | 2019-07-16 | 上海智臻智能网络科技股份有限公司 | Answering method and device, computer equipment and storage medium based on necessary condition |
CN110019723A (en) * | 2017-12-21 | 2019-07-16 | 上海智臻智能网络科技股份有限公司 | Answering method, terminal device and storage medium based on question answering system |
CN110019731A (en) * | 2017-12-25 | 2019-07-16 | 上海智臻智能网络科技股份有限公司 | Intelligent interactive method, device, terminal device and storage medium |
CN110147358A (en) * | 2017-11-22 | 2019-08-20 | 上海智臻智能网络科技股份有限公司 | The building method and construction system of automatic question answering knowledge base |
CN111400458A (en) * | 2018-12-27 | 2020-07-10 | 上海智臻智能网络科技股份有限公司 | Automatic generalization method and device |
US11341422B2 (en) | 2017-12-15 | 2022-05-24 | SHANGHAI XIAOl ROBOT TECHNOLOGY CO., LTD. | Multi-round questioning and answering methods, methods for generating a multi-round questioning and answering system, and methods for modifying the system |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102859515A (en) * | 2010-02-12 | 2013-01-02 | 谷歌公司 | Compound splitting |
CN103491205A (en) * | 2013-09-30 | 2014-01-01 | 北京奇虎科技有限公司 | Related resource address push method and device based on video retrieval |
US20150120718A1 (en) * | 2013-10-24 | 2015-04-30 | International Business Machines Corporation | Smart question routing and answerer growing for online community question-answer services |
CN104933084A (en) * | 2015-05-04 | 2015-09-23 | 上海智臻网络科技有限公司 | Method, apparatus and device for acquiring answer information |
-
2016
- 2016-04-05 CN CN201610207928.6A patent/CN105912600B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102859515A (en) * | 2010-02-12 | 2013-01-02 | 谷歌公司 | Compound splitting |
CN103491205A (en) * | 2013-09-30 | 2014-01-01 | 北京奇虎科技有限公司 | Related resource address push method and device based on video retrieval |
US20150120718A1 (en) * | 2013-10-24 | 2015-04-30 | International Business Machines Corporation | Smart question routing and answerer growing for online community question-answer services |
CN104933084A (en) * | 2015-05-04 | 2015-09-23 | 上海智臻网络科技有限公司 | Method, apparatus and device for acquiring answer information |
Cited By (19)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106547741A (en) * | 2016-11-21 | 2017-03-29 | 江苏科技大学 | A kind of Chinese language text auto-collation based on collocation |
CN108628906A (en) * | 2017-03-24 | 2018-10-09 | 北京京东尚科信息技术有限公司 | Short text template method for digging, device, electronic equipment and readable storage medium storing program for executing |
CN107256258B (en) * | 2017-06-12 | 2019-09-06 | 上海智臻智能网络科技股份有限公司 | Semantic formula generation method and device |
CN107256258A (en) * | 2017-06-12 | 2017-10-17 | 上海智臻智能网络科技股份有限公司 | Semantic formula generation method and device |
US10796096B2 (en) | 2017-06-12 | 2020-10-06 | Shanghai Xiaoi Robot Technology Co., Ltd. | Semantic expression generation method and apparatus |
CN107798123A (en) * | 2017-11-10 | 2018-03-13 | 上海智臻智能网络科技股份有限公司 | Knowledge base and its foundation, modification, intelligent answer method, apparatus and equipment |
CN107798123B (en) * | 2017-11-10 | 2022-01-11 | 上海智臻智能网络科技股份有限公司 | Knowledge base and establishing, modifying and intelligent question and answer methods, devices and equipment thereof |
CN109947908A (en) * | 2017-11-22 | 2019-06-28 | 上海智臻智能网络科技股份有限公司 | The building method and construction system of robot knowledge base |
CN110147358A (en) * | 2017-11-22 | 2019-08-20 | 上海智臻智能网络科技股份有限公司 | The building method and construction system of automatic question answering knowledge base |
CN110147358B (en) * | 2017-11-22 | 2024-05-17 | 上海智臻智能网络科技股份有限公司 | Construction method and construction system of automatic question-answering knowledge base |
CN110019739A (en) * | 2017-11-30 | 2019-07-16 | 上海智臻智能网络科技股份有限公司 | Answering method and device, computer equipment and storage medium based on necessary condition |
CN108090177A (en) * | 2017-12-15 | 2018-05-29 | 上海智臻智能网络科技股份有限公司 | The generation methods of more wheel question answering systems, equipment, medium and take turns question answering system more |
CN108090177B (en) * | 2017-12-15 | 2020-05-05 | 上海智臻智能网络科技股份有限公司 | Multi-round question-answering system generation method, equipment, medium and multi-round question-answering system |
US11341422B2 (en) | 2017-12-15 | 2022-05-24 | SHANGHAI XIAOl ROBOT TECHNOLOGY CO., LTD. | Multi-round questioning and answering methods, methods for generating a multi-round questioning and answering system, and methods for modifying the system |
CN110019723A (en) * | 2017-12-21 | 2019-07-16 | 上海智臻智能网络科技股份有限公司 | Answering method, terminal device and storage medium based on question answering system |
CN110019731A (en) * | 2017-12-25 | 2019-07-16 | 上海智臻智能网络科技股份有限公司 | Intelligent interactive method, device, terminal device and storage medium |
CN109033262A (en) * | 2018-07-09 | 2018-12-18 | 北京寻领科技有限公司 | Question and answer knowledge base update method and device |
CN109783506A (en) * | 2018-12-05 | 2019-05-21 | 北京国电通网络技术有限公司 | A kind of spoken meaning of one's words understanding method, device and the electronic equipment of intelligent Answer System |
CN111400458A (en) * | 2018-12-27 | 2020-07-10 | 上海智臻智能网络科技股份有限公司 | Automatic generalization method and device |
Also Published As
Publication number | Publication date |
---|---|
CN105912600B (en) | 2019-08-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105912600A (en) | Question-answer knowledge base and establishing method thereof, intelligent question-answering method and system | |
CN110377716A (en) | Exchange method, device and the computer readable storage medium of dialogue | |
CN107329967B (en) | Question answering system and method based on deep learning | |
Destradi | Empire, hegemony, and leadership: Developing a research framework for the study of regional powers | |
CN102236709B (en) | Flex-based monitoring page display method and system | |
CN109753560B (en) | Information processing method and device of intelligent question-answering system | |
CN102983999B (en) | Method for parameter configuration, the system of a kind of monitor supervision platform system and device cluster | |
CN106487603A (en) | A kind of response test method and device | |
CN108279954A (en) | A kind of method and device of application program sequence | |
CN102436440A (en) | Information display apparatus and computer-readable storage medium | |
CN111625629B (en) | Task type dialogue robot response method and device, robot and storage medium | |
CN105653528B (en) | Service field multi-state display method and device | |
CN109120708A (en) | Construction method, call method and the device of business module based on micro services component | |
Forsberg et al. | The potential for learning from good RRI practices and implications for the usefulness of RRI as an umbrella concept | |
CN102567796A (en) | Method and device for importing data | |
CN105630960B (en) | The method and apparatus of testing field Task conversational system | |
CN112348403B (en) | Wind control model construction method and device and electronic equipment | |
CN107545035A (en) | A kind of information search method and device | |
CN113886427A (en) | Conversation processing method and device and electronic equipment | |
CN106357519A (en) | Method for prompting information pushing | |
CN107784548A (en) | Order processing method and apparatus | |
CN107248097A (en) | A kind of food and drink retail promotions engine based on dynamic script language | |
CN103297865B (en) | The method and system of intelligent network service off-line reproduction | |
CN114330339B (en) | Text processing method, computer-readable storage medium and computer device | |
Yadav et al. | Effects of the industrial 4.0 transition on consumer behavior: A systematic overview |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
PE01 | Entry into force of the registration of the contract for pledge of patent right |
Denomination of invention: Question answering knowledge base and its establishment method, intelligent question answering method and system Effective date of registration: 20230223 Granted publication date: 20190816 Pledgee: China Construction Bank Corporation Shanghai No.5 Sub-branch Pledgor: SHANGHAI XIAOI ROBOT TECHNOLOGY Co.,Ltd. Registration number: Y2023980033272 |
|
PE01 | Entry into force of the registration of the contract for pledge of patent right |