CN108717433A - A kind of construction of knowledge base method and device of programming-oriented field question answering system - Google Patents
A kind of construction of knowledge base method and device of programming-oriented field question answering system Download PDFInfo
- Publication number
- CN108717433A CN108717433A CN201810454306.2A CN201810454306A CN108717433A CN 108717433 A CN108717433 A CN 108717433A CN 201810454306 A CN201810454306 A CN 201810454306A CN 108717433 A CN108717433 A CN 108717433A
- Authority
- CN
- China
- Prior art keywords
- answer
- similarity
- knowledge base
- user
- question
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 29
- 238000010276 construction Methods 0.000 title claims abstract description 20
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 29
- 238000004458 analytical method Methods 0.000 claims abstract description 28
- 230000001360 synchronised effect Effects 0.000 claims abstract description 12
- 238000012163 sequencing technique Methods 0.000 claims abstract description 10
- 238000004364 calculation method Methods 0.000 claims abstract description 8
- 241000270322 Lepidosauria Species 0.000 claims description 9
- 230000007246 mechanism Effects 0.000 claims description 5
- 238000012545 processing Methods 0.000 claims description 5
- 239000013598 vector Substances 0.000 description 25
- 239000011159 matrix material Substances 0.000 description 6
- 230000008569 process Effects 0.000 description 6
- 230000000694 effects Effects 0.000 description 4
- 238000004891 communication Methods 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000009194 climbing Effects 0.000 description 1
- 238000000205 computational method Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 239000004575 stone Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 230000017105 transposition Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/205—Parsing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
Abstract
A kind of construction of knowledge base method and device of programming-oriented field question answering system, includes the following steps, intelligent answer knowledge base is tentatively established according to the knowledge content of online exam system；Customer problem is obtained, and carries out near synonym replacement after the sentence of problem is segmented, then similarity calculation is carried out with content in knowledge base；By similarity analysis acquisition problem sentence as a result, and result is pressed sequencing of similarity, the return highest answer of similarity to user；If user's satisfaction is current to return to answer, according to question and answer synchronized update intelligent answer knowledge base, and by detecting the data in caching in real time with code Similarity algorithm；Otherwise, second-best solution is returned to user, until user is satisfied with.The problems such as this system can overcome the time difference present in traditional online answering system after class, search accuracy is low, efficiency is to be improved, effectively improves user and obtains the speed and accuracy rate of information, while optimizing the teaching efficiency of Online Judge system.
Description
Technical field
The invention belongs to microcomputer data processing fields, and in particular to a kind of programming-oriented field question answering system
Construction of knowledge base method and device.
Background technology
Intelligent Answer System is the mass data that the internet based on high speed development provides, at case study and data
Reason, the intelligence system that realistic problem to be solved is answered.This system can increase substantially the speed that user obtains information
Degree and accuracy rate.
According to the applicant understood, existing intelligent Answer System is mainly based upon huge in online question and answer knowledge base and internet
Database calculate customer problem sentence and asked with existing by being excavated to historical user's question and answer information, mass network data
The similarity between problem sentence is answered questions, the answer of the highest question and answer centering of similarity is returned into user.Simultaneously by user
The answer evaluation of feedback is used as foundation, judges answer quality, and according to the knowledge base after the foundation of user's question and answer situation synchronized update.
However, above-mentioned intelligent Answer System has a drawback in that：Since the grammer of different computer programming languages has
The answer of the problem of larger difference, system obtains under a few cases is not accurate enough.
Invention content
It is an object of the invention to：The construction of knowledge base method and dress of a kind of programming-oriented field question answering system are provided
It sets, can effectively improve user and obtain the speed and accuracy rate of information, and the teaching efficiency of Online Judge system can be optimized.
In order to reach object above, a kind of construction of knowledge base method of programming-oriented field question answering system, including such as
Lower step,
Intelligent answer knowledge base is tentatively established according to the knowledge content of online exam system；
Customer problem is obtained, and carries out near synonym replacement after the sentence of problem is segmented, then is carried out with content in knowledge base
Similarity calculation；
By similarity analysis acquisition problem sentence as a result, and result is pressed sequencing of similarity, return similarity highest
Answer to user；
If user's satisfaction is current to return to answer, according to question and answer synchronized update intelligent answer knowledge base, and by using generation
Code Similarity algorithm detects the data in caching in real time；Otherwise, second-best solution is returned to user, until user is satisfied with.
The present invention preferred embodiment be：The sentence of problem segments：
Input needs the character string segmented and the participle pattern controlled first, divides character string further according to participle pattern
Word, character string can be UTF-8 character strings or GBK character strings
Preferably, by similarity analysis obtain problem sentence as a result, and by result press sequencing of similarity, specially：
The high answer pair of matching similarity is retrieved in intelligent answer knowledge base, if being not present in knowledge base, by climbing
Worm mechanism captures the high result of similarity from network and is fed back.
Preferably, second-best solution is returned to user, specially：
According to sequencing of similarity as a result, return to time high answer of similarity, further according to feeding back to the use received after user
Family is evaluated, dissatisfied then continue to return to the high answer of next similarity, until user's satisfaction.
Preferably, according to question and answer synchronized update intelligent answer knowledge base, specially：
With code similarity algorithm, similar answer is classified as together from the knowledge content of answering system after class in real time
The different answers of one problem；If the sufficiently high question and answer pair of similarity or user are not found in original knowledge base not from original knowledge
It is found in the return answer in library when being satisfied with answer, is then crawled from network and return to the answer that is satisfied with of user and preserve, created
New question and answer pair.
The present invention also provides a kind of construction of knowledge base devices of programming-oriented field question answering system, including：
Preset repository unit, for tentatively establishing intelligent answer knowledge base according to the knowledge content of online exam system；
Problem analysis unit for obtaining customer problem, and carries out near synonym replacement after the sentence of problem is segmented, then with
Content carries out similarity calculation in knowledge base；
Answer unit is retrieved, is used for by similarity analysis acquisition problem sentence as a result, and arranging result by similarity
Sequence returns to the highest answer of similarity to user；
Updating unit is fed back, for returning to answer when user's satisfaction is current, is then known according to question and answer synchronized update intelligent answer
Know library, and by detecting the data in caching in real time with code Similarity algorithm；Otherwise, second-best solution is returned to user, until
Until user is satisfied with.
Preferably, in problem analysis unit, including:
Receiving unit, for obtaining customer problem；
Processing unit, for carrying out near synonym replacement after segmenting the sentence of problem；
Computing unit, for content in customer problem and knowledge base to be carried out similarity calculation.
Preferably, retrieval answer unit includes：
Retrieval unit, for the analysis according to the similarity generated in problem analysis unit, in intelligent answer knowledge base
Retrieve the high answer pair of matching similarity；
Reptile unit, for when the high answer pair of similarity is not present in knowledge base, being grabbed from network by reptile mechanism
The result for taking similarity high is fed back.
Preferably, feedback updating unit includes：
It is stored in unit, for customer satisfaction system optimum answer to be automatically credited in intelligent answer knowledge base, and according to question and answer
Synchronized update intelligent answer knowledge base；
Updating unit in real time will be similar from the knowledge content of answering system after class for using code similarity algorithm
Answer be classified as the different answers of the same problem；If the sufficiently high question and answer pair of similarity are not found in original knowledge base or are used
Family is not found from the return answer of original knowledge base when being satisfied with answer, then crawled from network return to user be satisfied with answer
It is preserved, creates new question and answer pair.
The present invention has the beneficial effect that：This intelligent Answer System not only possesses the function of traditional answering system, also allows student
Sentence is putd question in the form of natural language, and can automatically return to the answer of one natural language form of student, system
Knowledge base programming-oriented field, can more accurately return relevant procedures design class topic answer；Whole system gram
It taken the time difference present in traditional online answering system after class, searched for the problems such as accuracy is low, efficiency is to be improved, effectively carried
High user obtains the speed and accuracy rate of information, while optimizing the teaching efficiency of Online Judge system.
Description of the drawings
The present invention will be further described below with reference to the drawings.
Fig. 1 is the schematic process flow diagram of the present invention；
Fig. 2 is the device of the invention structural schematic diagram；
Fig. 3 is the system structural framework figure of the present invention.
Specific implementation mode
Embodiment one
The construction of knowledge base method for please referring to Fig.1 a kind of programming-oriented field question answering system, includes the following steps,
S101, the content based on Online Judge system Ability of Normal School Students communication put question to the knowledge answerred questions, preliminary to establish intelligence
The knowledge base of question and answer is used for subsequent step search problem answers；
S102, customer problem is obtained, participle operation is carried out to customer problem sentence, and carry out near synonym replacement；
S103, similarity analysis calculating is carried out to problem sentence, result is pressed into sequencing of similarity, is arranged according to similarity mode
Sequence returns to the highest answer of similarity, first returns to this answer of student；
S104, when student is dissatisfied to the answer that currently returns, suboptimum other in the same problem can be returned and answered
Case；
S105, intelligent answer knowledge base according to user's question and answer situation synchronized update, with code similarity algorithm, always
Data in caching are measured in real time.
Next each step towards programming arts intelligent answer method will be made into one with attached drawing in conjunction with the embodiments
Step is described in detail.
In step S101, it is based on existing programming Online Judge system, the wherein a large amount of teachers and students' communication of utilization
Content and the knowledge answerred questions of enquirement, tentatively establish preset knowledge base, the primary data source of answer retrieved as contingency question.
In step S102, obtain customer problem sentence after, due to it is input by user be natural language, computer itself is simultaneously
It cannot carry out effectively accurately identification.This system carries out participle operation first to problem input by user, in segmentation methods pattern
Selection on, syntype speed is fast but accuracy is not high enough, this system abandon syntype participle, selection be more suitable for text analyzing
Accurate model participle, obtain preferably segmenting effect.For example, " I wants the quick sorting algorithm generation of C++ to user's asked questions
Code ", obtain word segmentation result sentence be " I/want/C++// quicksort/algorithm/code ".After the completion of participle, nearly justice is carried out
Word is replaced, and makes the natural language of enquirement closer to the identifiable language of computer.
In step S103, similarity analysis calculating is carried out to problem sentence.After similarity analysis, result is pressed into similarity
Sequence sorts according to similarity mode and returns to the highest answer of similarity, first returns to this answer of student.Similarity is higher, this
Problem answers are closer to target answer, and error probability is smaller after returning to user, more readily satisfy user and are expected to require.And it is preset
Knowledge base is not necessarily comprehensive, and the answer retrieved even similarity is highest to be also possible to be gone out with user's expection answer
Enter.Therefore, solution is given in next step.
In step S104, the optimal answer returned after similarity analysis can not meet user's expection, can return
Previous step, selects similarity higher but non-highest answer feedback is to user.For example, " how to be write out with C++ most simple and quick
Sort algorithm？", system returns " the quick sorting algorithm program write with C Plus Plus " first.But user feels after usage
Speed it is unsatisfactory, at this time system return second-best solution " with C Plus Plus write Bubble Sort Algorithm program ".It also needs to
Bright, if the higher question and answer pair of similarity are not present in knowledge base, but it uses reptile mechanism, from possessing magnanimity information
The higher result of similarity is obtained on network to be fed back.For example, using reptile mechanism, using some search engines such as Baidu,
It searches and the higher answer of problem statement similarity.Until user is satisfied with obtained answer.
In step S105, the knowledge base of intelligent answer is according to user's question and answer situation synchronized update, with code similarity operator
Method is always measured in real time the data in caching.The knowledge base of intelligent Answer System is not unalterable in the present invention, and
That can constantly be changed with the question and answer situation of user and update, when this in knowledge base without related question and answer to information, by this intelligence
Question answering system has obtained being satisfied with after answer of customer problem, this new question and answer can be automatically stored to information in knowledge base.
Specifically, above-mentioned programming-oriented field intelligent answer method provides an example：User proposes that " I needs problem
Sentence is carried out participle operation when question answering system receives customer problem sentence by fast row's algorithm of C++ " first, participle at " I/
Need/C++// fast row/algorithm ", then the near synonym occurred in participle are replaced, problem sentence becomes after replacement：" I/it needs
Want/C++// quicksort/algorithm ".When knowledge base only has 35 question and answer to data, list preceding ten question sentences and
The similarity of each question sentence and problem sentence：
Wherein, the question sentence that number is 1 is obviously quite similar with example sentence, but since the pattern of example sentence is also unexpected with other
The question sentence of answer is similar, although so the quicksort code that can accurately have selected C++, effect is caused simultaneously non-fully to be managed
Think.When question and answer logarithm becomes 50 in knowledge base, statement similarity reaches as high as 0.998, hence it is evident that than knowledge base be 35
When effect to get well, accuracy rise.When the question and answer in knowledge base increase to 100 to quantity, statement similarity highest can
Up to 0.997, data result tends towards stability.So when the knowledge base data of this question answering system reach a certain amount of, similarity algorithm
More applicable, effect is more preferable.It is exported to use with the answer statement of the highest question and answer centering of problem statement similarity in selection knowledge base
Family.
The present embodiment also provides a kind of construction of knowledge base device of programming-oriented field intelligent Answer System, such as Fig. 2
It is shown, including：
S201, preset repository unit, what content, enquirement based on Online Judge system Ability of Normal School Students communication were answerred questions knows
Know, tentatively establishes the knowledge base of intelligent answer.The design of question and answer information table is as follows wherein in knowledge base：
Question and answer are to (number, question sentence, language are explained, code)
Specifically, the number of question and answer centering is the unique identifier of every question and answer information in knowledge base, it is the master of the relationship
Key；
Question sentence is sentence when intelligent Answer System middle school student user puts question to；
Language refers to the question and answer belong to which kind of programming language, such as Java, C/C++, Python；
Explanation is in intelligent Answer System to the explanation and description in the answer of problem；
Code is then the answer to proposing problem in intelligent Answer System.
S202, problem analysis unit are used for natural language translation into computer language to computer.This unit makes to succeed in one's scheme
Calculation machine lays the first stone it will be appreciated that the problem of User proposes for retrieval answer below.Accurate problem analysis just can guarantee
Return to the correctness of the answer of student.
Problem analysis unit includes receiving unit, processing unit, computing unit again.
Specifically, receiving unit is for receiving the problem of User proposes in question answering system in the present invention；
Processing unit, for completing the participle of problem sentence, near synonym are replaced, similarity analysis.Wherein, problem sentence
Participle process uses the scanning of the word figure based on trie tree constructions, generates the directed acyclic graph for being possible into word situation.Again
Maximum probability path is found with dynamic programming, finds the maximum cutting combination based on word frequency.Near synonym replacement process is by problem
All near synonym are substituted for defined word by the root after sentence participle according to being traversed near synonym table.Sentence similarity point
Analysis then generates the corresponding vector of sentence using the similarity calculation algorithm based on statistical vector spatial model.
S203 retrieves answer unit, including retrieval unit, reptile unit, return answer unit.
Specifically, retrieval unit screens the tuple for corresponding to language in table question and answer in database.By all tuples
Question sentence attribute carry out sentence similarity vector calculating, the sentence similarity vector transmitted with problem analysis module does angle fortune
It calculates.
It returns to answer unit the result of angle operation in retrieval unit sorts from high to low, returns to that similarity is highest asks
Answer questions tuple explanation and code to student.
Reptile unit, for when existing knowledge library can not provide and enable customer satisfaction system answer from the net for possessing magnanimity information
The higher result of similarity is obtained on network to be fed back.
S204, feedback updating unit are mainly used for the update to intelligent Answer System knowledge base.It will periodically answer questions after class and be
In the effective question and answer information deposit knowledge base generated in system.
The present invention provides knowledge base using when need the similarity algorithm flow used：
Similarity algorithm is that intelligent Answer System in the present invention uses and the important algorithm of construction of knowledge base.Next
The statement similarity algorithm and code similarity algorithm used in the present invention are illustrated.
Statement similarity algorithm, the vector T that the present invention ties up every problem sentence with a n=<T1,T2,…Tn>Carry out table
Show, wherein each Ti (1≤i≤n) represents the word separated in question sentence, the computational methods of value are：
Wherein, N TiThe number occurred in this question sentence, referred to as TF values (Term Frequency word frequency).And formula
In M be statement library in question sentence sum, m is then to contain T in statement libraryiQuestion sentence sum；IDF values (Inverse
The reverse document-frequencies of Document Frequency) it is therein：
The transposition Q ' of the n-dimensional vector of target question sentence is calculated with same method.The similarity of question sentence and target question sentence can
To calculate with the following method：
Similarity is higher after calculating is completed shows that problem sentence and object statement are more similar.
Code similarity algorithm, to improve the accuracy of similarity detection, the present invention is to original calculating sides Halstead
Method is improved.Operator and operand to go out Xiang in program are numeration object, using their occurrence number as technology mesh
Mark comes process of measurement capacity and length.Four variables first in statistics program：η1、η2、N1、N2.Variable η1It is single in program
The quantity of operator, variable η2For the quantity of single operation number in program, variable N1For the sum of all operators in program, become
Measure N2For the sum of all operands in program.Define η=η simultaneously1+η2For vocabulary, N=N is defined1+N2To execute length, obtain
The capacity for going out program is：
After obtaining aforementioned four variable, tri- variables of L, W and C are added, variables L is lines of code, and variable W is word
Number, variable C are number of characters.A feature vector H (η, N, I, L, W, C) is may be constructed by six groups of variables of η, N, I, L, W and C.Most
The Euclidean distance for calculating the feature vector H between the program to be compared afterwards shows that two programs are more similar apart from smaller.
To improve the accuracy rate of this code similarity algorithm, this code similarity algorithm be applied to the present invention in towards
When in programming arts intelligent Answer System, it is contemplated that the programming language of this system mainly include python, c/c++,
Java etc. increases the thought that weights are assigned to the variable in former Halstead methods.For different high level languages
The different role and influence power of vector element assign variable weights with certain rule, make it more suitable for certain advanced procedures language
Say the calculating of code similarity.
Following code similarity algorithm flows are obtained after making improvement to former Halstead methods in the present embodiment：
The first, program is pre-processed, first by the pre- place of null, annotation and corresponding language in original program in the present invention
Reason order (in such as C language with # beginnings include header file order) removal, the program that obtains that treated.
The second, the pretreatment of characteristic vector data, it is assumed that the program code number to be compared is n.N can so be calculated
A feature vector Hi(i=1,2 ... n).One space vector can be formed by this n feature vector
Mean normalization conversion process is carried out to data first, is become
Change formula：
Wherein the formula left side is j-th of attribute of i-th of program code in vector space after mean normalization converts
Value.
Six attributes of feature vector H (η, N, I, L, W, C) are compared in the imparting of third, feature vector weights two-by-two
Compared with aijIndicate the ratio between the influence size of ith attribute and j-th of attribute to similarity.It can obtain a comparator matrix:
A(aij)n×n,
This matrix is known as influencing similarity the judgment matrix with this six attributes.For aijValue, using Saaty
The 1-9 scales of proposition.
Wherein, the element of matrix A also meets
The maximum eigenvalue of this matrix is λmax, λmaxFor positive real number, λmaxEach member in corresponding characteristic vector W
Element is also positive real number.Finally calculate coincident indicator CI.Have
If CI<0.1, then the consistency of judgment matrix A can receive.
4th, the Euclidean distance between vector is calculated, the first row vector for normalizing later V is carried out with other row vectors
The calculating of Euclidean distance.The feature vector of original program and the calculation formula of the Euclidean distance of other feature vectors are as follows：
Wherein DiIt is the Euclidean distance of the feature vector and the feature vector of i-th of program to be compared that indicate original program.
Using the code similarity algorithm in the present invention, the similarity of fast row and the program of another quicksort are calculated
It is 99%.
The Overall Structure Design frame of intelligent Answer System where knowledge base, as shown in figure 3, including：
Entire intelligent Answer System is made of three application layer, logical layer, data Layer levels.It can be seen that entirely knowing in Fig. 3
Know the running that library supports entire intelligent Answer System.The workflow of intelligent Answer System in the present invention is stepped on from User
Record starts, and is putd question to using the system by user, and application layer provides intelligent answer interface, and problem is sent to logical layer,
The problem analysis of logical layer under the action of retrieving three answer, answer feedback modules, is completed to obtain from the knowledge base of data Layer
Answer and feedback more new knowledge base, the knowledge base to data Layer gradually improve the process for updating and terminating.
Wherein, specifically, application layer is the answering system after class and an intelligent answer machine for being similar to forum by one
Device people is constituted.
Logical layer is then made of three modules, and first module is problem analysis module, and the content of the module is mainly asked
Topic participle, near synonym replace, the calculating of sentence similarity vector and program pretreatment, program attribute counting, code similarity to
Amount calculates six functions.Second module is retrieval answer module, in the module containing retrieval question and answer to, return to two work(of answer
Can, according to the similarity analysis generated in first module as a result, to corresponding question and answer pair are retrieved in knowledge base, then by result
By sequencing of similarity, the highest answer of student's similarity is first returned to.Third module is update feedback module, and the module is main
It is responsible for the update of knowledge base.
Data Layer is then the core of whole system, is the core of intelligent answering system after class, the knowledge as whole system
Library supports the running of system.
In addition to the implementation, the present invention can also have other embodiment.It is all to use equivalent substitution or equivalent transformation shape
At technical solution, fall within the scope of protection required by the present invention.
Claims (9)
1. a kind of construction of knowledge base method of programming-oriented field question answering system, which is characterized in that include the following steps,
Intelligent answer knowledge base is tentatively established according to the knowledge content of online exam system；
Customer problem is obtained, and near synonym replacement is carried out after the sentence of problem is segmented, then is similar to content progress in knowledge base
Degree calculates；
By similarity analysis obtain problem sentence as a result, and result is pressed into sequencing of similarity, return to that similarity is highest answers
Case is to user；
If user's satisfaction is current to return to answer, according to question and answer synchronized update intelligent answer knowledge base, and by using code phase
Detect the data in caching in real time like algorithm；Otherwise, second-best solution is returned to user, until user is satisfied with.
2. a kind of construction of knowledge base method of programming-oriented field question answering system according to claim 1, feature
It is, the sentence participle of described problem is specially：
Input needs the character string segmented and the participle pattern controlled first, segments character string further according to participle pattern, word
Symbol string can be UTF-8 character strings or GBK character strings.
3. a kind of construction of knowledge base method of programming-oriented field question answering system according to claim 1, feature
Be, it is described by similarity analysis obtain problem sentence as a result, and by result press sequencing of similarity, specially：
The high answer pair of matching similarity is retrieved in intelligent answer knowledge base passes through reptile machine if being not present in knowledge base
The result that crawl similarity is high from network is made to be fed back.
4. a kind of construction of knowledge base method of programming-oriented field question answering system according to claim 1, feature
It is, the return second-best solution is to user, specially：
According to sequencing of similarity as a result, the secondary high answer of return similarity, is commented further according to the user received after user is fed back to
Valence, it is dissatisfied then continue to return to the high answer of next similarity, until user is satisfied with.
5. a kind of construction of knowledge base method of programming-oriented field question answering system according to claim 1, feature
It is, it is described according to question and answer synchronized update intelligent answer knowledge base, specially：
With code similarity algorithm, similar answer is classified as from the knowledge content of answering system after class in real time same
The different answers of problem；If the sufficiently high question and answer pair of similarity or user are not found in original knowledge base not from original knowledge base
It returns and is found in answer when being satisfied with answer, then crawled from network and return to the answer that is satisfied with of user and preserve, created new
Question and answer pair.
6. a kind of construction of knowledge base device of programming-oriented field question answering system, which is characterized in that including：
Preset repository unit, for tentatively establishing intelligent answer knowledge base according to the knowledge content of online exam system；
Problem analysis unit for obtaining customer problem, and will carry out near synonym replacement after the sentence of problem participle, then with knowledge
Content carries out similarity calculation in library；
Answer unit is retrieved, is used for by similarity analysis acquisition problem sentence as a result, and returning result by sequencing of similarity
The highest answer of similarity is returned to user；
Updating unit is fed back, for returning to answer when user's satisfaction is current, then according to question and answer synchronized update intelligent answer knowledge base,
And by detecting the data in caching in real time with code Similarity algorithm；Otherwise, second-best solution is returned to user, until user is full
It means only.
7. a kind of construction of knowledge base device of programming-oriented field question answering system according to claim 6, feature
It is, in the problem analysis unit, including:
Receiving unit, for obtaining customer problem；
Processing unit, for carrying out near synonym replacement after segmenting the sentence of problem；
Computing unit, for content in customer problem and knowledge base to be carried out similarity calculation.
8. a kind of construction of knowledge base device of programming-oriented field question answering system according to claim 6, feature
It is, the retrieval answer unit includes：
Retrieval unit is retrieved for the analysis according to the similarity generated in problem analysis unit in intelligent answer knowledge base
The high answer pair of matching similarity；
Reptile unit, for when the high answer pair of similarity is not present in knowledge base, phase to be captured from network by reptile mechanism
It is fed back like the result for spending high.
9. a kind of construction of knowledge base device of programming-oriented field question answering system according to claim 6, feature
It is, the feedback updating unit includes：
It is stored in unit, for customer satisfaction system optimum answer to be automatically credited in intelligent answer knowledge base, and is synchronized according to question and answer
Update intelligent answer knowledge base；
Updating unit is answered from the knowledge content of answering system after class by similar in real time for using code similarity algorithm
Case is classified as the different answers of the same problem；If the sufficiently high question and answer pair of similarity or user are not found in original knowledge base not
Found from the return answer of original knowledge base when being satisfied with answer, then crawled from network return to user be satisfied with answer progress
It preserves, creates new question and answer pair.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810454306.2A CN108717433A (en) | 2018-05-14 | 2018-05-14 | A kind of construction of knowledge base method and device of programming-oriented field question answering system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810454306.2A CN108717433A (en) | 2018-05-14 | 2018-05-14 | A kind of construction of knowledge base method and device of programming-oriented field question answering system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108717433A true CN108717433A (en) | 2018-10-30 |
Family
ID=63899885
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810454306.2A Pending CN108717433A (en) | 2018-05-14 | 2018-05-14 | A kind of construction of knowledge base method and device of programming-oriented field question answering system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108717433A (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109657047A (en) * | 2018-12-27 | 2019-04-19 | 四川新网银行股份有限公司 | A kind of voice automatic question-answering method and system based on crawler technology and machine learning |
| CN109710772A (en) * | 2018-11-13 | 2019-05-03 | 国云科技股份有限公司 | A kind of question and answer library Knowledge Management System and its implementation based on deep learning |
| CN109918486A (en) * | 2019-01-24 | 2019-06-21 | 平安科技（深圳）有限公司 | Corpus construction method, device, computer equipment and the storage medium of intelligent customer service |
| CN109977421A (en) * | 2019-04-15 | 2019-07-05 | 南京邮电大学 | A kind of Knowledge Base of Programming subjects answering system after class |
| CN110070872A (en) * | 2019-04-22 | 2019-07-30 | 南京邮电大学 | It is a kind of based on intelligent sound identification logistics fill out folk prescription method |
| CN110083749A (en) * | 2019-04-11 | 2019-08-02 | 艾伯资讯（深圳）有限公司 | The retrieval quickly developed for software, multiplexing, environmental structure system and method |
| CN110321420A (en) * | 2019-07-04 | 2019-10-11 | 河海大学常州校区 | The intelligent Answer System generated based on question sentence |
| CN110517019A (en) * | 2019-08-26 | 2019-11-29 | 西安文理学院 | Semantic-based registration information Security Officer authentication capability assessment method and system |
| CN111046155A (en) * | 2019-11-27 | 2020-04-21 | 中博信息技术研究院有限公司 | Semantic similarity calculation method based on FSM multi-turn question answering |
| CN111858869A (en) * | 2020-01-03 | 2020-10-30 | 北京嘀嘀无限科技发展有限公司 | Data matching method and device, electronic equipment and storage medium |
| CN111930911A (en) * | 2020-08-12 | 2020-11-13 | 杭州东方通信软件技术有限公司 | Rapid field question-answering method and device |
| CN112035638A (en) * | 2020-08-28 | 2020-12-04 | 腾讯科技（深圳）有限公司 | Information processing method, device, storage medium and equipment |
| CN114121260A (en) * | 2021-10-19 | 2022-03-01 | 上海清赟医药科技有限公司 | Medical question-answering method based on intention identification |
| CN116701579A (en) * | 2023-02-21 | 2023-09-05 | 中国人民解放军海军工程大学 | Information reply system, method and computer readable storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015058604A1 (en) * | 2013-10-21 | 2015-04-30 | 北京奇虎科技有限公司 | Apparatus and method for obtaining degree of association of question and answer pair and for search ranking optimization |
| CN107391614A (en) * | 2017-07-04 | 2017-11-24 | 重庆智慧思特大数据有限公司 | A kind of Chinese question and answer matching process based on WMD |
| CN107993724A (en) * | 2017-11-09 | 2018-05-04 | 易保互联医疗信息科技（北京）有限公司 | A kind of method and device of medicine intelligent answer data processing |
-
2018
- 2018-05-14 CN CN201810454306.2A patent/CN108717433A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015058604A1 (en) * | 2013-10-21 | 2015-04-30 | 北京奇虎科技有限公司 | Apparatus and method for obtaining degree of association of question and answer pair and for search ranking optimization |
| CN107391614A (en) * | 2017-07-04 | 2017-11-24 | 重庆智慧思特大数据有限公司 | A kind of Chinese question and answer matching process based on WMD |
| CN107993724A (en) * | 2017-11-09 | 2018-05-04 | 易保互联医疗信息科技（北京）有限公司 | A kind of method and device of medicine intelligent answer data processing |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109710772B (en) * | 2018-11-13 | 2023-03-31 | 国云科技股份有限公司 | Question-answer base knowledge management system based on deep learning and implementation method thereof |
| CN109710772A (en) * | 2018-11-13 | 2019-05-03 | 国云科技股份有限公司 | A kind of question and answer library Knowledge Management System and its implementation based on deep learning |
| CN109657047A (en) * | 2018-12-27 | 2019-04-19 | 四川新网银行股份有限公司 | A kind of voice automatic question-answering method and system based on crawler technology and machine learning |
| CN109657047B (en) * | 2018-12-27 | 2020-09-29 | 四川新网银行股份有限公司 | Voice automatic question-answering method and system based on crawler technology and machine learning |
| CN109918486A (en) * | 2019-01-24 | 2019-06-21 | 平安科技（深圳）有限公司 | Corpus construction method, device, computer equipment and the storage medium of intelligent customer service |
| CN109918486B (en) * | 2019-01-24 | 2024-03-19 | 平安科技（深圳）有限公司 | Corpus construction method and device for intelligent customer service, computer equipment and storage medium |
| CN110083749A (en) * | 2019-04-11 | 2019-08-02 | 艾伯资讯（深圳）有限公司 | The retrieval quickly developed for software, multiplexing, environmental structure system and method |
| CN109977421A (en) * | 2019-04-15 | 2019-07-05 | 南京邮电大学 | A kind of Knowledge Base of Programming subjects answering system after class |
| CN110070872A (en) * | 2019-04-22 | 2019-07-30 | 南京邮电大学 | It is a kind of based on intelligent sound identification logistics fill out folk prescription method |
| CN110321420A (en) * | 2019-07-04 | 2019-10-11 | 河海大学常州校区 | The intelligent Answer System generated based on question sentence |
| CN110321420B (en) * | 2019-07-04 | 2023-07-18 | 河海大学常州校区 | Intelligent question-answering system and method based on question generation |
| CN110517019A (en) * | 2019-08-26 | 2019-11-29 | 西安文理学院 | Semantic-based registration information Security Officer authentication capability assessment method and system |
| CN111046155A (en) * | 2019-11-27 | 2020-04-21 | 中博信息技术研究院有限公司 | Semantic similarity calculation method based on FSM multi-turn question answering |
| CN111858869A (en) * | 2020-01-03 | 2020-10-30 | 北京嘀嘀无限科技发展有限公司 | Data matching method and device, electronic equipment and storage medium |
| CN111930911A (en) * | 2020-08-12 | 2020-11-13 | 杭州东方通信软件技术有限公司 | Rapid field question-answering method and device |
| CN111930911B (en) * | 2020-08-12 | 2024-03-29 | 杭州东方通信软件技术有限公司 | Rapid field question-answering method and device thereof |
| CN112035638A (en) * | 2020-08-28 | 2020-12-04 | 腾讯科技（深圳）有限公司 | Information processing method, device, storage medium and equipment |
| CN114121260A (en) * | 2021-10-19 | 2022-03-01 | 上海清赟医药科技有限公司 | Medical question-answering method based on intention identification |
| CN116701579A (en) * | 2023-02-21 | 2023-09-05 | 中国人民解放军海军工程大学 | Information reply system, method and computer readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108717433A (en) | A kind of construction of knowledge base method and device of programming-oriented field question answering system | |
| CN108804521B (en) | Knowledge graph-based question-answering method and agricultural encyclopedia question-answering system | |
| CN110298033A (en) | Keyword corpus labeling trains extracting tool | |
| CN110298032A (en) | Text classification corpus labeling training system | |
| CN110175585B (en) | Automatic correcting system and method for simple answer questions | |
| CN110287482B (en) | Semi-automatic participle corpus labeling training device | |
| CN106649272A (en) | Named entity recognizing method based on mixed model | |
| CN112051986B (en) | Code search recommendation device and method based on open source knowledge | |
| Ahmadov et al. | Towards a hybrid imputation approach using web tables | |
| CN110008309A (en) | A kind of short phrase picking method and device | |
| CN114491082A (en) | Plan matching method based on network security emergency response knowledge graph feature extraction | |
| CN110310012B (en) | Data analysis method, device, equipment and computer readable storage medium | |
| CN116010552A (en) | Engineering cost data analysis system and method based on keyword word library | |
| CN113360582B (en) | Relation classification method and system based on BERT model fusion multi-entity information | |
| Huang et al. | Deep entity matching with adversarial active learning | |
| CN115062070A (en) | Question and answer based text table data query method | |
| Sinha et al. | NLP-based automatic answer evaluation | |
| CN111339258B (en) | University computer basic exercise recommendation method based on knowledge graph | |
| CN117494760A (en) | Semantic tag-rich data augmentation method based on ultra-large-scale language model | |
| CN116049376B (en) | Method, device and system for retrieving and replying information and creating knowledge | |
| CN116342167A (en) | Intelligent cost measurement method and device based on sequence labeling named entity recognition | |
| CN111104503A (en) | Construction engineering quality acceptance standard question-answering system and construction method thereof | |
| CN114281998A (en) | Multi-level annotator-oriented event annotation system construction method based on crowdsourcing technology | |
| CN114282875A (en) | Flow approval certainty rule and semantic self-learning combined judgment method and device | |
| CN110059314A (en) | A kind of Relation extraction method based on enhancing study |
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: 20181030 |