CN107644012B - Electronic device, problem identification confirmation method and computer readable storage medium - Google Patents
Electronic device, problem identification confirmation method and computer readable storage medium Download PDFInfo
- Publication number
- CN107644012B CN107644012B CN201710754550.6A CN201710754550A CN107644012B CN 107644012 B CN107644012 B CN 107644012B CN 201710754550 A CN201710754550 A CN 201710754550A CN 107644012 B CN107644012 B CN 107644012B
- Authority
- CN
- China
- Prior art keywords
- feature words
- probability
- text
- predetermined
- 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.)
- Active
Links
- 238000012790 confirmation Methods 0.000 title claims abstract description 28
- 238000000034 method Methods 0.000 title claims abstract description 24
- 238000003860 storage Methods 0.000 title claims abstract description 14
- 230000011218 segmentation Effects 0.000 claims abstract description 48
- 238000009826 distribution Methods 0.000 claims abstract description 39
- 238000013507 mapping Methods 0.000 claims abstract description 37
- 238000012545 processing Methods 0.000 claims abstract description 25
- 238000004364 calculation method Methods 0.000 claims description 14
- 238000012795 verification Methods 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 8
- 238000010586 diagram Methods 0.000 description 10
- 238000005520 cutting process Methods 0.000 description 6
- 238000006243 chemical reaction Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 239000004744 fabric Substances 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007257 malfunction Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000000007 visual effect 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
-
- 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
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/26—Speech to text systems
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computational Linguistics (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Multimedia (AREA)
- Acoustics & Sound (AREA)
- Human Computer Interaction (AREA)
- Machine Translation (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention discloses a kind of electronic device, problem identification confirmation method and computer readable storage medium, wherein this method comprises: receiving the problem of user issues voice, carries out speech recognition to voice the problem of reception, generates question text;Word segmentation processing is carried out according to predetermined word segmentation regulation to text the problem of generation, obtains the corresponding participle of described problem text;If containing predetermined Feature Words in the participle obtained, then according to the probability distribution between Feature Words and problem, the problem of determining the predetermined Feature Words corresponding maximum probability, and according to the mapping relations between predetermined question and answer, the problem of determining the maximum probability corresponding answer;By determining answer feedback to user.Technical solution of the present invention improves intelligent customer service robot, intelligent customer service answering system feed back to user answer accuracy.
Description
Technical field
The present invention relates to intelligent sound technical field, in particular to a kind of electronic device, problem identification confirmation method and meter
Calculation machine readable storage medium storing program for executing.
Background technique
Currently, in order to which the waiting situation of customer service is effectively reduced, promotion service quality, improves the convenient of customer service
Property, the company (for example, mobile with quotient, insurance company, financial institution etc.) of many service types uses intelligent customer service machine under line
Device people (for example, entity Administrative Area in be arranged intelligent customer service robot) and/or line on intelligent customer service answering system (for example,
Intelligent voice response system) it is that client services.Intelligent customer service response system in this kind of Xian Xia intelligent customer service robot and/or line
The existing scheme generallyd use of uniting is: being pre-configured with the mapping relations data of typical problem and model answer;When receiving client
After the typical problem of proposition, according to the mapping relations data of preconfigured typical problem and model answer, determine received
The corresponding model answer of typical problem, and the model answer determined is fed back into client.For non-standard the asking of user's proposition
Topic, this existing scheme will be difficult to give answer feedback.
Although there is a kind of improvement project for solving non-standard issue on the market at present: when non-standard issue can not be found
When corresponding answer, non-standard issue and each typical problem are subjected to similarity calculation, and the standard of maximum similarity is asked
Corresponding model answer is inscribed to be fed back as the corresponding answer of non-standard issue.It is non-standard but due in most cases
Similarity between problem and typical problem be all because of some words unrelated with sentence meaning (for example, " ", " ") and produce
Raw, therefore, the accuracy of this improvement project is very low, often malfunctions, causes to give an irrelevant answer.
Summary of the invention
The main object of the present invention is to provide a kind of problem identification confirmation method, it is intended to promote intelligent customer service system to nonstandard
The accuracy of quasi- problem identification confirmation, thus accuracy of the lift pins to the feedback answer of non-standard issue.
To achieve the above object, electronic device proposed by the present invention includes memory, processor, is stored on the memory
There is the problem of can running on the processor recognition and verification system, described problem recognition and verification system is executed by the processor
Shi Shixian following steps:
S1, the problem of user issues voice is received, speech recognition is carried out to voice the problem of reception, generates question text;
S2, word segmentation processing is carried out according to predetermined word segmentation regulation to text the problem of generation, obtains described problem text
This corresponding participle;
If containing predetermined Feature Words in S3, the participle obtained, according to the probability between Feature Words and problem point
Cloth, the problem of determining the predetermined Feature Words corresponding maximum probability, and according between predetermined question and answer
Mapping relations, the problem of determining the maximum probability corresponding answer;
S4, by determining answer feedback to user.
Preferably, the step S3 replaces with following steps:
If containing predetermined Feature Words in the participle obtained, according to the probability distribution between Feature Words and problem,
Determine that the predetermined Feature Words correspond to the probability of each problem;
Sequence from big to small according to probability is that each problem is ranked up, and determines the preceding preset quantity that sorts
Determining each candidate problem is provided or is broadcasted and selected to user as candidate problem by problem;
After user has selected a problem, according to the mapping relations between predetermined question and answer, determining should
The corresponding answer of problem.
Preferably, the predetermined word segmentation regulation is priority of long word word segmentation regulation.
Preferably, the probability distribution between the Feature Words and problem determines in accordance with the following steps:
The implicit theme of preset quantity is added between Feature Words and problem;
The problem text of pending training is obtained, and word segmentation processing is carried out respectively to text the problem of acquisition, is obtained each
The corresponding participle of question text;
According to the mapping relations of predetermined implicit theme and Feature Words, the spy that each implicit theme contains is determined respectively
The first quantity for levying word, determines the second quantity of implicit theme belonging to each Feature Words, according to corresponding first quantity respectively
Determine each Feature Words to the first choice probability of each implicit theme with the second quantity;
According to the mapping relations of predetermined implicit theme and question text, determine what each question text contained respectively
The third quantity of implicit theme determines the 4th quantity of problem text belonging to each implicit theme respectively, according to corresponding the
Three quantity and the 4th quantity determine each implicit theme to the second select probability of each question text;
Corresponding first choice probability and the second select probability are substituted into predetermined calculation formula to calculate, calculated
Each Feature Words are to the third select probability of each question text out, and calculated each Feature Words are respectively to each question text
Third select probability be probability distribution between Feature Words and problem.
Preferably, the predetermined calculation formula are as follows:
P3=P1*P2, wherein P1 represents first choice probability, and P2 represents the second select probability, and it is general that P3 represents third selection
Rate.
The present invention also proposes a kind of problem identification confirmation method, which is characterized in that the method comprising the steps of:
S1, the problem of user issues voice is received, speech recognition is carried out to voice the problem of reception, generates question text;
S2, word segmentation processing is carried out according to predetermined word segmentation regulation to text the problem of generation, obtains described problem text
This corresponding participle;
If containing predetermined Feature Words in S3, the participle obtained, according to the probability between Feature Words and problem point
Cloth, the problem of determining the predetermined Feature Words corresponding maximum probability, and according between predetermined question and answer
Mapping relations, the problem of determining the maximum probability corresponding answer;
S4, by determining answer feedback to user.
Preferably, the step S3 replaces with following steps:
If containing predetermined Feature Words in the participle obtained, according to the probability distribution between Feature Words and problem,
Determine that the predetermined Feature Words correspond to the probability of each problem;
Sequence from big to small according to probability is that each problem is ranked up, and determines the preceding preset quantity that sorts
Determining each candidate problem is provided or is broadcasted and selected to user as candidate problem by problem;
After user has selected a problem, according to the mapping relations between predetermined question and answer, determining should
The corresponding answer of problem.
Preferably, the predetermined word segmentation regulation is priority of long word word segmentation regulation.
Preferably, the probability distribution between the Feature Words and problem determines in accordance with the following steps:
The implicit theme of preset quantity is added between Feature Words and problem;
The problem text of pending training is obtained, and word segmentation processing is carried out respectively to text the problem of acquisition, is obtained each
The corresponding participle of question text;
According to the mapping relations of predetermined implicit theme and Feature Words, the spy that each implicit theme contains is determined respectively
The first quantity for levying word, determines the second quantity of implicit theme belonging to each Feature Words, according to corresponding first quantity respectively
Determine each Feature Words to the first choice probability of each implicit theme with the second quantity;
According to the mapping relations of predetermined implicit theme and question text, determine what each question text contained respectively
The third quantity of implicit theme determines the 4th quantity of problem text belonging to each implicit theme respectively, according to corresponding the
Three quantity and the 4th quantity determine each implicit theme to the second select probability of each question text;
Corresponding first choice probability and the second select probability are substituted into predetermined calculation formula to calculate, calculated
Each Feature Words are to the third select probability of each question text out, and calculated each Feature Words are respectively to each question text
Third select probability be probability distribution between Feature Words and problem.
The present invention also proposes a kind of computer readable storage medium, the problematic knowledge of computer-readable recording medium storage
Not Que Ren system, described problem recognition and verification system can be executed by least one processor, so that at least one described processor
Execute problem identification confirmation method described in any of the above embodiments.
Technical solution of the present invention is by dividing question text for after the problematic text of speech recognition the problem of user
Word, obtains the theme for being able to reflect customer problem that contains or the Feature Words in semantic direction in word segmentation result, and by Feature Words with
Probability distribution between problem, so that the problem of finding out maximum probability (i.e. most probable problem), then determines maximum probability
The corresponding answer of problem, to feed back to user;Since in the technical program, Feature Words are able to reflect the theme or language of customer problem
Therefore the right way of conduct is taken compared to the prior art to the corresponding answer that the problem of, maximum probability corresponding by Feature Words is found
Entire problem and typical problem are subjected to similarity-rough set, in a manner of obtaining the corresponding answer of most like problem for, this case
The accuracy for feeding back to the answer of user significantly improves.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
The structure shown according to these attached drawings obtains other attached drawings.
Fig. 1 is the flow diagram of one embodiment of problem identification confirmation method of the present invention;
Fig. 2 is the flow diagram of two embodiment of problem identification confirmation method of the present invention;
Fig. 3 is to determine that the process of the probability distribution between Feature Words and problem is illustrated in problem identification confirmation method of the present invention
Figure;
Fig. 4 is the running environment schematic diagram that problem identification of the present invention confirms system preferred embodiment;
Fig. 5 is the structural schematic diagram that problem identification of the present invention confirms one embodiment of system;
Fig. 6 is the structural schematic diagram that problem identification of the present invention confirms two embodiment of system;.
The embodiments will be further described with reference to the accompanying drawings for the realization, the function and the advantages of the object of the present invention.
Specific embodiment
The principle and features of the present invention will be described below with reference to the accompanying drawings, and the given examples are served only to explain the present invention, and
It is non-to be used to limit the scope of the invention.
The present invention proposes a kind of problem identification confirmation method, is mainly used for intelligent customer service answering system or intelligent customer service machine
The intelligent customer services product such as people.
As shown in FIG. 1, FIG. 1 is the flow diagrams of one embodiment of problem identification confirmation method of the present invention.
In the present embodiment, which includes:
Step S1 receives the problem of user issues voice, carries out speech recognition to voice the problem of reception, generates problem text
This;
When user puts question to intelligent customer service voice system or intelligent customer service robot, problem identification confirms that system receives and uses
The problem of family issues when puing question to voice identifies the problem of receiving voice and the problem text that speech production the problem of identification is corresponding
This.
Step S2 carries out word segmentation processing according to predetermined word segmentation regulation to text the problem of generation, asks described in acquisition
Inscribe the corresponding participle of text;
The problem of will receive after the problematic text of speech recognition conversion, problem identification confirms system according to predetermined
Word segmentation regulation to the question text carry out word segmentation processing then obtain the corresponding participle of the question text after word segmentation processing.
In the present embodiment, the participle includes word and word, such as: described problem text can be that " safety is proposed the macro life product of honor
", the result after participle is " safety ", " release ", " ", " the macro life of honor ", " product ", " ".
Step S3, if containing predetermined Feature Words in the participle obtained, according to general between Feature Words and problem
Rate distribution, the problem of determining the predetermined Feature Words corresponding maximum probability, and according to predetermined question and answer
Between mapping relations, the problem of determining the maximum probability corresponding answer;
There is preset Feature Words (for example, " the macro life of honor ", " safety " etc.) in system, Feature Words, which are able to reflect, asks
Inscribe theme or the semanteme direction of problem corresponding to text;Also there is the probability between predetermined Feature Words and problem in system
Distribution, i.e., each Feature Words are respectively provided with probability value corresponding with each the problem of prestoring, and contain each feature word problem text
This may be the probability of each problem;System is additionally provided with the mapping table between preset question and answer.System is obtaining
After obtaining the corresponding participle of described problem text, analyze in the participle of acquisition whether contain predetermined Feature Words;When analyzing
Predetermined Feature Words are not contained in the participle of acquisition, then user is prompted to put question to or prompt again not identifying and asked a question
Deng processing.When containing predetermined Feature Words in the participle for analyzing acquisition, then according to general between Feature Words and problem
Rate distribution, the problem of determining the predetermined Feature Words that contain corresponding maximum probability, i.e., most possible problem, in problem
After determination, then according to the mapping relations between predetermined question and answer, the problem of obtaining the maximum probability, is corresponding
Answer.
Step S4, by determining answer feedback to user.
System after obtaining determining answer, by determining answer by voice broadcast or be sent to display equipment show or
It is sent to the modes such as the default terminal of user and feeds back to user.
The present embodiment technical solution is by dividing question text for after the problematic text of speech recognition the problem of user
Word, obtains the theme for being able to reflect customer problem that contains or the Feature Words in semantic direction in word segmentation result, and by Feature Words with
Probability distribution between problem, so that the problem of finding out maximum probability (i.e. most probable problem), then determines maximum probability
The corresponding answer of problem, to feed back to user;Since in the technical program, Feature Words are able to reflect the theme or language of customer problem
Therefore the right way of conduct is taken compared to the prior art to the corresponding answer that the problem of, maximum probability corresponding by Feature Words is found
Entire problem and typical problem are subjected to similarity-rough set, in a manner of obtaining the corresponding answer of most like problem for, this case
The accuracy of the answer of feedback user significantly improves.
Preferably, in the present embodiment, the predetermined word segmentation regulation is priority of long word word segmentation regulation.The priority of long word
Word segmentation regulation refers to: the phrase T1 for needing to segment for one finds out one from the dictionary prestored first since first character A
Then a longest word X1 originated by A rejects X1 from T1 and is left T2, then uses identical cutting principle to T2, after cutting
Result be " X1/X2/,,, ";For example, in the dictionary prestored include " safety ", " release ", " ", " the macro life of honor ",
When " product ", " ", the cutting result of phrase " safety is proposed the macro life product of honor " be " safety "/" release "/" "/
" the macro life of honor "/" product "/" ".
As shown in Fig. 2, Fig. 2 is the flow diagram of two embodiment of problem identification confirmation method of the present invention, the present embodiment side
Case replaces with following steps on the basis of first embodiment, by the step S3:
Step S301, if containing predetermined Feature Words in the participle obtained, according between Feature Words and problem
Probability distribution determines that the predetermined Feature Words correspond to the probability of each problem;
After containing predetermined Feature Words in the participle for analyze acquisition, according to Feature Words predetermined in system
It is each to determine that the predetermined Feature Words contained in the participle for obtaining the acquisition respectively correspond for probability distribution between problem
The probability of problem.
Step S302, the sequence from big to small according to probability are that each problem is ranked up, and determine to sort preceding
The problem of preset quantity, is used as candidate problem, and determining each candidate problem is provided or broadcasted and is selected to user;
It is right after the predetermined Feature Words contained in the participle for obtaining the acquisition respectively correspond the probability of each problem
Each problem carries out descending sort according to obtained probability, then extracts the preceding present count of sequence in the problem sequence after sequence
The candidate problem of extraction is fed back to user, so that user selects as candidate problem by the problem of measuring (such as 3,4)
It selects.Wherein, the mode that candidate problem feeds back to user can be with are as follows: 1, voice broadcast;2, selection interface is provided, candidate problem is shown
In selection interface (it is selected for example, generating problem selection interface for user, which may include candidate problem list,
Corresponding " determination " button of each of described list candidate's problem, user can click the button and select corresponding ask
Topic);Deng.
Step S303 is closed after user has selected a problem according to the mapping between predetermined question and answer
System, determines the corresponding answer of the problem.
After user is made a choice based on the candidate problem of system feedback, system receives the problem of user selects, then root
According to the mapping relations between question and answer predetermined in system, determine that the problem of user received selects is corresponding
Answer.
As shown in figure 3, the probability distribution between the Feature Words and problem determines in accordance with the following steps:
Step S51 adds the implicit theme of preset quantity between Feature Words and problem;
Firstly, the implicit theme of predicted quantity (for example, 50) is added between this two layers of Feature Words and problem, as in
Interbed, to constitute problem preference pattern;Wherein, the implicit theme is virtual, and there is no real meanings;Each implicit master
Topic generally comprises multiple Feature Words, and each problem generally comprises multiple implicit themes again.
Step S52 obtains the problem text of pending training, and carries out word segmentation processing respectively to text the problem of acquisition,
Obtain the corresponding participle of each question text;
After forming problem preference pattern, obtain pending training problem text (question text be prepare in advance
), word segmentation processing is carried out respectively to each question text of acquisition, to obtain the corresponding word segmentation result of each question text.
Step S53 determines each implicit theme according to the mapping relations of predetermined implicit theme and Feature Words respectively
First quantity of the Feature Words contained determines the second quantity of implicit theme belonging to each Feature Words, according to corresponding respectively
First quantity and the second quantity determine each Feature Words to the first choice probability of each implicit theme;
According to the mapping relations of implicit theme and Feature Words predetermined in system, each implicit theme is determined respectively
In implicit theme belonging to the first quantity of Feature Words for containing and each Feature Words the second quantity, further according to corresponding first
Quantity and the second quantity respectively obtain each Feature Words to the first choice probability of each implicit theme;For example, belonging to Feature Words Y
The second quantity of implicit theme be X2, the first quantity of the Feature Words that an implicit theme contains is X1, then Y pairs of the specific word
The select probability of the implicit theme are as follows: 1/ (X1*X2).
Step S54 determines each problem text according to the mapping relations of predetermined implicit theme and question text respectively
Originally the third quantity of the implicit theme contained determines the 4th quantity of problem text belonging to each implicit theme respectively, according to
Corresponding third quantity and the 4th quantity determine each implicit theme to the second select probability of each question text;
According to the mapping relations of implicit theme and question text predetermined in system, each question text is determined respectively
In problem text belonging to the third quantity of implicit theme that contains and each implicit theme the 4th quantity, further according to corresponding
Third quantity and the 4th quantity respectively obtain each implicit theme to the second select probability of each question text;For example, implicit
4th quantity of problem text belonging to theme K is J2, and the third quantity for the implicit theme that a question text contains is J1, then
Select probability of the implicit theme K to the question text are as follows: 1/ (J1*J2).In the present embodiment, the step S54 and step S53
Sequence interchangeable.
Corresponding first choice probability and the second select probability are substituted into predetermined calculation formula and carried out by step S55
It calculates, calculates each Feature Words to the third select probability of each question text, calculated each Feature Words are respectively to each
The third select probability of a question text is the probability distribution between Feature Words and problem.
The second choosing according to Feature Words to the first choice probability distribution and implicit theme of implicit theme to question text
Probability distribution is selected, further you can get it, and Feature Words are distributed the third select probability of question text.Specifically, by that will correspond to
First choice probability and the second select probability substitute into predetermined calculation formula and calculate, obtain each Feature Words respectively to each
The select probability of a question text is to get the probability distribution arrived between Feature Words and problem.In the present embodiment, this is predetermined
Calculation formula are as follows: P3=P1*P2, wherein P1 represents first choice probability, and P2 represents the second select probability, and P3 represents third choosing
Select probability.For example, first choice probability of the Feature Words Y to implicit theme K are as follows: 1/ (X1*X2) implies theme K to question text W
The second select probability are as follows: 1/ (J1*J2), then Feature Words Y is then 1/ (X1*X2) * (J1* to the select probability of question text W
J2)。
The present invention also proposes a kind of problem identification confirmation system.
Referring to Fig. 4, being the running environment schematic diagram of problem identification confirmation 10 preferred embodiment of system of the present invention.
In the present embodiment, problem identification confirmation system 10 is installed and is run in electronic device 1.Electronic device 1 can be with
It is that desktop PC, notebook, palm PC and server etc. calculate equipment.The electronic device 1 may include, but not only limit
In memory 11, processor 12 and display 13.Fig. 3 illustrates only the electronic device 1 with component 11-13, it should be understood that
Be, it is not required that implement all components shown, the implementation that can be substituted is more or less component.
Memory 11 is a kind of computer storage medium, can be the storage inside of electronic device 1 in some embodiments
Unit, such as the hard disk or memory of the electronic device 1.Memory 11 is also possible to electronic device 1 in further embodiments
The plug-in type hard disk being equipped on External memory equipment, such as electronic device 1, intelligent memory card (Smart Media Card,
SMC), secure digital (Secure Digital, SD) blocks, flash card (Flash Card) etc..Further, memory 11 may be used also
With the internal storage unit both including electronic device 1 or including External memory equipment.Memory 11 is installed on electronics for storing
The application software and Various types of data of device 1, such as the program code etc. of problem identification confirmation system 10.Memory 11 can also be used
In temporarily storing the data that has exported or will export.
Processor 12 can be in some embodiments a central processing unit (Central Processing Unit,
CPU), microprocessor or other data processing chips, program code or processing data for being stored in run memory 11, example
Such as executive problem recognition and verification system 10.
Display 13 can be in some embodiments light-emitting diode display, liquid crystal display, touch-control liquid crystal display and
OLED (Organic Light-Emitting Diode, Organic Light Emitting Diode) touches device etc..Display 13 is for being shown in
The information that is handled in electronic device 1 and for showing visual user interface, such as business customizing interface etc..Electronic device
1 component 11-13 is in communication with each other by system bus.
Referring to Fig. 5, being the structural schematic diagram of problem identification confirmation 10 1 embodiment of system of the present invention.In the present embodiment
In, problem identification confirmation system 10 can be divided into one or more modules, one or more module is stored in storage
In device 11, and it is performed by one or more processors (the present embodiment is processor 12), to complete the present invention.For example, in Fig. 5
In, problem identification confirmation system 10 can be divided into identification module 101, word segmentation module 102, determining module 103 and feedback mould
Block 104.The so-called module of the present invention is the series of computation machine program instruction section for referring to complete specific function, more suitable than program
Together in the implementation procedure of description problem identification confirmation system 10 in the electronic apparatus 1, in which:
Identification module 101 carries out speech recognition to voice the problem of reception for receiving the problem of user issues voice,
Generate question text;
When user puts question to intelligent customer service voice system or intelligent customer service robot, problem identification confirms that system receives and uses
The problem of family issues when puing question to voice identifies the problem of receiving voice and the problem text that speech production the problem of identification is corresponding
This.
Word segmentation module 102 is obtained for carrying out word segmentation processing according to predetermined word segmentation regulation to text the problem of generation
Obtain the corresponding participle of described problem text;
The problem of will receive after the problematic text of speech recognition conversion, problem identification confirms system according to predetermined
Word segmentation regulation to the question text carry out word segmentation processing then obtain the corresponding participle of the question text after word segmentation processing.
In the present embodiment, the participle includes word and word, such as: described problem text can be that " safety is proposed the macro life product of honor
", the result after participle is " safety ", " release ", " ", " the macro life of honor ", " product ", " ".
Determining module 103, after containing predetermined Feature Words in the participle of acquisition, according to Feature Words and problem
Between probability distribution, the problem of determining the predetermined Feature Words corresponding maximum probability, and being asked according to predetermined
The corresponding answer of the problem of inscribing the mapping relations between answer, determining the maximum probability;
There is preset Feature Words (for example, " the macro life of honor ", " safety " etc.) in system, Feature Words, which are able to reflect, asks
Inscribe theme or the semanteme direction of problem corresponding to text;Also there is the probability between predetermined Feature Words and problem in system
Distribution, i.e., each Feature Words are respectively provided with probability value corresponding with each the problem of prestoring, and contain each feature word problem text
This may be the probability of each problem;System is additionally provided with the mapping table between preset question and answer.System is obtaining
After obtaining the corresponding participle of described problem text, analyze in the participle of acquisition whether contain predetermined Feature Words;When analyzing
When containing predetermined Feature Words in the participle of acquisition, then according to the probability distribution between Feature Words and problem, determination contains
Predetermined Feature Words corresponding maximum probability the problem of, i.e., most possible problem, after problem determines, then basis
Mapping relations between predetermined question and answer, the problem of obtaining the maximum probability corresponding answer.In addition, determining
Module 103, without containing after predetermined Feature Words, prompts user to put question to or mention again also in the participle for analyzing acquisition
The processing such as ask a question can not be identified by showing.
Feedback module 104, for by the answer feedback determined to user.
System after obtaining determining answer, by determining answer by voice broadcast or be sent to display equipment show or
It is sent to the modes such as the default terminal of user and feeds back to user.
The present embodiment technical solution is by dividing question text for after the problematic text of speech recognition the problem of user
Word, obtains the theme for being able to reflect customer problem that contains or the Feature Words in semantic direction in word segmentation result, and by Feature Words with
Probability distribution between problem, so that the problem of finding out maximum probability (i.e. most probable problem), then determines maximum probability
The corresponding answer of problem, to feed back to user;Since in the technical program, Feature Words are able to reflect the theme or language of customer problem
Therefore the right way of conduct is taken compared to the prior art to the corresponding answer that the problem of, maximum probability corresponding by Feature Words is found
Entire problem and typical problem are subjected to similarity-rough set, in a manner of obtaining the corresponding answer of most like problem for, this case
The accuracy of the answer of feedback user significantly improves.
Preferably, in the present embodiment, the predetermined word segmentation regulation is priority of long word word segmentation regulation.The priority of long word
Word segmentation regulation refers to: the phrase T1 for needing to segment for one finds out one from the dictionary prestored first since first character A
Then a longest word X1 originated by A rejects X1 from T1 and is left T2, then uses identical cutting principle to T2, after cutting
Result be " X1/X2/,,, ";For example, in the dictionary prestored include " safety ", " release ", " ", " the macro life of honor ",
When " product ", " ", the cutting result of phrase " safety is proposed the macro life product of honor " be " safety "/" release "/" "/
" the macro life of honor "/" product "/" ".
As shown in fig. 6, Fig. 6 is the structural schematic diagram that problem identification of the present invention confirms two embodiment of system, the present embodiment side
Case replaces with following module on the basis of first embodiment, by the determining module 103:
First determines submodule 105, after containing predetermined Feature Words in the participle of acquisition, according to Feature Words
Probability distribution between problem determines that the predetermined Feature Words correspond to the probability of each problem;
After containing predetermined Feature Words in the participle for analyze acquisition, according to Feature Words predetermined in system
It is each to determine that the predetermined Feature Words contained in the participle for obtaining the acquisition respectively correspond for probability distribution between problem
The probability of problem.
Second determines submodule 106, is that each problem is ranked up for the sequence from big to small according to probability, determines
The problem of preceding preset quantity that sorts out, is used as candidate problem, and by determining each candidate problem provide or broadcast to
Family is selected;
It is right after the predetermined Feature Words contained in the participle for obtaining the acquisition respectively correspond the probability of each problem
Each problem carries out descending sort according to obtained probability, then extracts the preceding present count of sequence in the problem sequence after sequence
The candidate problem of extraction is fed back to user, so that user selects as candidate problem by the problem of measuring (such as 3,4)
It selects.Wherein, the mode that candidate problem feeds back to user can be with are as follows: 1, voice broadcast;2, selection interface is provided, candidate problem is shown
In selection interface (it is selected for example, generating problem selection interface for user, which may include candidate problem list,
Corresponding " determination " button of each of described list candidate's problem, user can click the button and select corresponding ask
Topic);Deng.
Third determines submodule 107, for after user has selected a problem, according to predetermined question and answer
Between mapping relations, determine the corresponding answer of the problem.
After user is made a choice based on the candidate problem of system feedback, system receives the problem of user selects, then root
According to the mapping relations between question and answer predetermined in system, determine that the problem of user received selects is corresponding
Answer.
Preferably, in the present embodiment, the probability distribution between the Feature Words and problem determines in accordance with the following steps:
1, the implicit theme of preset quantity is added between Feature Words and problem;
Firstly, the implicit theme of predicted quantity (for example, 50) is added between this two layers of Feature Words and problem, as in
Interbed, to constitute problem preference pattern;Wherein, the implicit theme is virtual, and there is no real meanings;Each implicit master
Topic generally comprises multiple Feature Words, and each problem generally comprises multiple implicit themes again.
2, the problem text of pending training is obtained, and word segmentation processing is carried out respectively to text the problem of acquisition, is obtained each
The corresponding participle of a question text;
After forming problem preference pattern, obtain pending training problem text (question text be prepare in advance
), word segmentation processing is carried out respectively to each question text of acquisition, to obtain the corresponding word segmentation result of each question text.
3, according to the mapping relations of predetermined implicit theme and Feature Words, determine what each implicit theme contained respectively
First quantity of Feature Words determines the second quantity of implicit theme belonging to each Feature Words respectively, according to corresponding first number
Amount and the second quantity determine each Feature Words to the first choice probability of each implicit theme;
According to the mapping relations of implicit theme and Feature Words predetermined in system, each implicit theme is determined respectively
In implicit theme belonging to the first quantity of Feature Words for containing and each Feature Words the second quantity, further according to corresponding first
Quantity and the second quantity respectively obtain each Feature Words to the first choice probability of each implicit theme;For example, belonging to Feature Words Y
The second quantity of implicit theme be X2, the first quantity of the Feature Words that an implicit theme contains is X1, then Y pairs of the specific word
The select probability of the implicit theme are as follows: 1/ (X1*X2).
4, according to the mapping relations of predetermined implicit theme and question text, determine that each question text contains respectively
Implicit theme third quantity, the 4th quantity of problem text belonging to each implicit theme is determined respectively, according to corresponding
Third quantity and the 4th quantity determine each implicit theme to the second select probability of each question text;
According to the mapping relations of implicit theme and question text predetermined in system, each question text is determined respectively
In problem text belonging to the third quantity of implicit theme that contains and each implicit theme the 4th quantity, further according to corresponding
Third quantity and the 4th quantity respectively obtain each implicit theme to the second select probability of each question text;For example, implicit
4th quantity of problem text belonging to theme K is J2, and the third quantity for the implicit theme that a question text contains is J1, then
Select probability of the implicit theme K to the question text are as follows: 1/ (J1*J2).
5, corresponding first choice probability and the second select probability are substituted into predetermined calculation formula to calculate, is counted
Each Feature Words are calculated to the third select probability of each question text, calculated each Feature Words are respectively to each problem text
This third select probability is the probability distribution between Feature Words and problem.
The second choosing according to Feature Words to the first choice probability distribution and implicit theme of implicit theme to question text
Probability distribution is selected, further you can get it, and Feature Words are distributed the third select probability of question text.Specifically, by that will correspond to
First choice probability and the second select probability substitute into predetermined calculation formula and calculate, obtain each Feature Words respectively to each
The select probability of a question text is to get the probability distribution arrived between Feature Words and problem.In the present embodiment, this is predetermined
Calculation formula are as follows: P3=P1*P2, wherein P1 represents first choice probability, and P2 represents the second select probability, and P3 represents third choosing
Select probability.For example, first choice probability of the Feature Words Y to implicit theme K are as follows: 1/ (X1*X2) implies theme K to question text W
The second select probability are as follows: 1/ (J1*J2), then Feature Words Y is then 1/ (X1*X2) * (J1* to the select probability of question text W
J2)。
The present invention also proposes a kind of computer readable storage medium, the problematic identification of the computer-readable recording medium storage
Confirmation system, described problem recognition and verification system can be executed by least one processor, so that at least one described processor is held
Any of the above-described problem identification confirmation method as described in the examples of row.
The above description is only a preferred embodiment of the present invention, is not intended to limit the scope of the invention, all at this
Under the inventive concept of invention, using equivalent structure transformation made by description of the invention and accompanying drawing content, or directly/use indirectly
It is included in other related technical areas in scope of patent protection of the invention.
Claims (8)
1. a kind of electronic device, which is characterized in that the electronic device includes memory, processor, is stored on the memory
There is the problem of can running on the processor recognition and verification system, described problem recognition and verification system is executed by the processor
Shi Shixian following steps:
S1, the problem of user issues voice is received, speech recognition is carried out to voice the problem of reception, generates question text;
S2, word segmentation processing is carried out according to predetermined word segmentation regulation to text the problem of generation, obtains described problem text pair
The participle answered;
If containing predetermined Feature Words in S3, the participle obtained, according to the probability distribution between Feature Words and problem, really
The problem of fixed predetermined Feature Words corresponding maximum probability, and according to the mapping between predetermined question and answer
Relationship, the problem of determining the maximum probability corresponding answer;
S4, by determining answer feedback to user;
Probability distribution between the Feature Words and problem determines in accordance with the following steps:
The implicit theme of preset quantity is added between Feature Words and problem;
The problem text of pending training is obtained, and word segmentation processing is carried out respectively to text the problem of acquisition, obtains each problem
The corresponding participle of text;
According to the mapping relations of predetermined implicit theme and Feature Words, the Feature Words that each implicit theme contains are determined respectively
The first quantity, the second quantity of implicit theme belonging to each Feature Words is determined respectively, according to corresponding first quantity and
Two quantity determine each Feature Words to the first choice probability of each implicit theme;
According to the mapping relations of predetermined implicit theme and question text, determine that each question text contains implicit respectively
The third quantity of theme determines the 4th quantity of problem text belonging to each implicit theme, according to corresponding third number respectively
Amount and the 4th quantity determine each implicit theme to the second select probability of each question text;
Corresponding first choice probability and the second select probability are substituted into predetermined calculation formula to calculate, calculated every
A Feature Words are to the third select probability of each question text, and calculated each Feature Words are respectively to the of each question text
Three select probabilities are the probability distribution between Feature Words and problem.
2. electronic device as described in claim 1, which is characterized in that the step S3 replaces with following steps:
If containing predetermined Feature Words in the participle obtained, according to the probability distribution between Feature Words and problem, determine
The predetermined Feature Words correspond to the probability of each problem;
According to probability sequence from big to small be each problem be ranked up, determine to sort preceding preset quantity the problem of
As candidate problem, and determining each candidate problem is provided or broadcasted and is selected to user;
After user has selected a problem, according to the mapping relations between predetermined question and answer, the problem is determined
Corresponding answer.
3. electronic device as described in claim 1, which is characterized in that the predetermined word segmentation regulation is priority of long word point
Word rule.
4. electronic device as described in claim 1, which is characterized in that the predetermined calculation formula are as follows:
P3=P1*P2, wherein P1 represents first choice probability, and P2 represents the second select probability, and P3 represents third select probability.
5. a kind of problem identification confirmation method, which is characterized in that the method comprising the steps of:
S1, the problem of user issues voice is received, speech recognition is carried out to voice the problem of reception, generates question text;
S2, word segmentation processing is carried out according to predetermined word segmentation regulation to text the problem of generation, obtains described problem text pair
The participle answered;
If containing predetermined Feature Words in S3, the participle obtained, according to the probability distribution between Feature Words and problem, really
The problem of fixed predetermined Feature Words corresponding maximum probability, and according to the mapping between predetermined question and answer
Relationship, the problem of determining the maximum probability corresponding answer;
S4, by determining answer feedback to user;
Probability distribution between the Feature Words and problem determines in accordance with the following steps:
The implicit theme of preset quantity is added between Feature Words and problem;
The problem text of pending training is obtained, and word segmentation processing is carried out respectively to text the problem of acquisition, obtains each problem
The corresponding participle of text;
According to the mapping relations of predetermined implicit theme and Feature Words, the Feature Words that each implicit theme contains are determined respectively
The first quantity, the second quantity of implicit theme belonging to each Feature Words is determined respectively, according to corresponding first quantity and
Two quantity determine each Feature Words to the first choice probability of each implicit theme;
According to the mapping relations of predetermined implicit theme and question text, determine that each question text contains implicit respectively
The third quantity of theme determines the 4th quantity of problem text belonging to each implicit theme, according to corresponding third number respectively
Amount and the 4th quantity determine each implicit theme to the second select probability of each question text;
Corresponding first choice probability and the second select probability are substituted into predetermined calculation formula to calculate, calculated every
A Feature Words are to the third select probability of each question text, and calculated each Feature Words are respectively to the of each question text
Three select probabilities are the probability distribution between Feature Words and problem.
6. problem identification confirmation method as claimed in claim 5, which is characterized in that the step S3 replaces with following steps:
If containing predetermined Feature Words in the participle obtained, according to the probability distribution between Feature Words and problem, determine
The predetermined Feature Words correspond to the probability of each problem;
According to probability sequence from big to small be each problem be ranked up, determine to sort preceding preset quantity the problem of
As candidate problem, and determining each candidate problem is provided or broadcasted and is selected to user;
After user has selected a problem, according to the mapping relations between predetermined question and answer, the problem is determined
Corresponding answer.
7. problem identification confirmation method as claimed in claim 5, which is characterized in that the predetermined word segmentation regulation is length
The preferential word segmentation regulation of word.
8. a kind of computer readable storage medium, which is characterized in that the problematic identification of computer-readable recording medium storage
Confirmation system, described problem recognition and verification system can be executed by least one processor, so that at least one described processor is held
Problem recognition and verification method of the row as described in any one of claim 5-7.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710754550.6A CN107644012B (en) | 2017-08-29 | 2017-08-29 | Electronic device, problem identification confirmation method and computer readable storage medium |
PCT/CN2017/108763 WO2019041517A1 (en) | 2017-08-29 | 2017-10-31 | Electronic device, question recognition and confirmation method, and computer-readable storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710754550.6A CN107644012B (en) | 2017-08-29 | 2017-08-29 | Electronic device, problem identification confirmation method and computer readable storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107644012A CN107644012A (en) | 2018-01-30 |
CN107644012B true CN107644012B (en) | 2019-03-01 |
Family
ID=61110870
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710754550.6A Active CN107644012B (en) | 2017-08-29 | 2017-08-29 | Electronic device, problem identification confirmation method and computer readable storage medium |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN107644012B (en) |
WO (1) | WO2019041517A1 (en) |
Families Citing this family (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108388558B (en) * | 2018-02-07 | 2022-04-19 | 平安普惠企业管理有限公司 | Question matching method and device, customer service robot and storage medium |
CN108595519A (en) * | 2018-03-26 | 2018-09-28 | 平安科技(深圳)有限公司 | Focus incident sorting technique, device and storage medium |
CN109697228A (en) * | 2018-12-13 | 2019-04-30 | 平安科技(深圳)有限公司 | Intelligent answer method, apparatus, computer equipment and storage medium |
CN109829048B (en) * | 2019-01-23 | 2023-06-23 | 平安科技(深圳)有限公司 | Electronic device, interview assisting method, and computer-readable storage medium |
CN110211704A (en) * | 2019-05-05 | 2019-09-06 | 平安科技(深圳)有限公司 | The engine method and server of matter of opening |
CN110414006B (en) * | 2019-07-31 | 2023-09-08 | 京东方科技集团股份有限公司 | Text theme labeling method and device, electronic equipment and storage medium |
CN110968669B (en) * | 2019-11-30 | 2023-07-28 | 南京森林警察学院 | Intelligent video analysis police test question classification and recommendation method |
CN111881694A (en) * | 2020-08-05 | 2020-11-03 | 科大讯飞股份有限公司 | Chapter point detection method, device, equipment and storage medium |
CN112399019A (en) * | 2020-09-16 | 2021-02-23 | 中国农业银行股份有限公司河北省分行 | Intelligent outbound call method, terminal equipment and readable storage medium |
CN115394293A (en) * | 2022-08-08 | 2022-11-25 | 湖北星纪时代科技有限公司 | Dialog system and method for implementing a dialog |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104424290A (en) * | 2013-09-02 | 2015-03-18 | 佳能株式会社 | Voice based question-answering system and method for interactive voice system |
CN106550156A (en) * | 2017-01-23 | 2017-03-29 | 苏州咖啦魔哆信息技术有限公司 | A kind of artificial intelligence's customer service system and its implementation based on speech recognition |
CN106847279A (en) * | 2017-01-10 | 2017-06-13 | 西安电子科技大学 | Man-machine interaction method based on robot operating system ROS |
Family Cites Families (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8990400B2 (en) * | 2007-08-17 | 2015-03-24 | The Invention Science Fund I, Llc | Facilitating communications among message recipients |
US8434066B2 (en) * | 2008-06-12 | 2013-04-30 | Oracle International Corporation | Enabling a business entity to validate support for specific activities in a CRM system before migration from earlier versions |
KR101173561B1 (en) * | 2010-10-25 | 2012-08-13 | 한국전자통신연구원 | Question type and domain identifying apparatus and method |
CN102622413A (en) * | 2012-02-17 | 2012-08-01 | 清华大学 | Method and device for answering natural language questions |
US9443005B2 (en) * | 2012-12-14 | 2016-09-13 | Instaknow.Com, Inc. | Systems and methods for natural language processing |
US9916298B2 (en) * | 2014-09-03 | 2018-03-13 | International Business Machines Corporation | Management of content tailoring by services |
CN104657346A (en) * | 2015-01-15 | 2015-05-27 | 深圳市前海安测信息技术有限公司 | Question matching system and question matching system in intelligent interaction system |
CN105094641A (en) * | 2015-07-15 | 2015-11-25 | 百度在线网络技术(北京)有限公司 | Question-answer interaction method and device based on voice |
CN106445905B (en) * | 2015-08-04 | 2019-05-21 | 阿里巴巴集团控股有限公司 | Question and answer data processing, automatic question-answering method and device |
CN105955976B (en) * | 2016-04-15 | 2019-05-14 | 中国工商银行股份有限公司 | A kind of automatic answering system and method |
CN106601237B (en) * | 2016-12-29 | 2020-02-07 | 上海智臻智能网络科技股份有限公司 | Interactive voice response system and voice recognition method thereof |
-
2017
- 2017-08-29 CN CN201710754550.6A patent/CN107644012B/en active Active
- 2017-10-31 WO PCT/CN2017/108763 patent/WO2019041517A1/en active Application Filing
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104424290A (en) * | 2013-09-02 | 2015-03-18 | 佳能株式会社 | Voice based question-answering system and method for interactive voice system |
CN106847279A (en) * | 2017-01-10 | 2017-06-13 | 西安电子科技大学 | Man-machine interaction method based on robot operating system ROS |
CN106550156A (en) * | 2017-01-23 | 2017-03-29 | 苏州咖啦魔哆信息技术有限公司 | A kind of artificial intelligence's customer service system and its implementation based on speech recognition |
Also Published As
Publication number | Publication date |
---|---|
WO2019041517A1 (en) | 2019-03-07 |
CN107644012A (en) | 2018-01-30 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107644012B (en) | Electronic device, problem identification confirmation method and computer readable storage medium | |
US11386893B2 (en) | Human-computer interaction processing system, method, storage medium, and electronic device | |
CN107204184B (en) | Audio recognition method and system | |
CN107680588B (en) | Intelligent voice navigation method, device and storage medium | |
CN108491433A (en) | Chat answer method, electronic device and storage medium | |
US11741094B2 (en) | Method and system for identifying core product terms | |
CN112380870A (en) | User intention analysis method and device, electronic equipment and computer storage medium | |
CN111783471B (en) | Semantic recognition method, device, equipment and storage medium for natural language | |
CN109445818A (en) | Method for updating pages and device, electronic equipment, storage medium based on big data | |
CN112559687A (en) | Question identification and query method and device, electronic equipment and storage medium | |
CN113127621A (en) | Dialogue module pushing method, device, equipment and storage medium | |
CN113435859A (en) | Letter processing method and device, electronic equipment and computer readable medium | |
CN112364145A (en) | Work order processing method and device, electronic equipment and storage medium | |
CN113051380A (en) | Information generation method and device, electronic equipment and storage medium | |
CN116955561A (en) | Question answering method, question answering device, electronic equipment and storage medium | |
CN112906368B (en) | Industry text increment method, related device and computer program product | |
CN113705224A (en) | Voice recognition scheduling service voice interaction method and system | |
CN113886577A (en) | Text classification method, device, equipment and storage medium | |
CN111783424B (en) | Text sentence dividing method and device | |
CN110826991B (en) | Electronic receipt processing system and method | |
CN116796758A (en) | Dialogue interaction method, dialogue interaction device, equipment and storage medium | |
CN114880498B (en) | Event information display method and device, equipment and medium | |
CN116048463A (en) | Intelligent recommendation method and device for content of demand item based on label management | |
CN113890948B (en) | Resource allocation method based on voice outbound robot dialogue data and related equipment | |
CN112328871B (en) | Reply generation method, device, equipment and storage medium based on RPA module |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |