CN106372055B - A kind of semanteme similar processing method and system in man-machine natural language interaction - Google Patents
A kind of semanteme similar processing method and system in man-machine natural language interaction Download PDFInfo
- Publication number
- CN106372055B CN106372055B CN201610709517.7A CN201610709517A CN106372055B CN 106372055 B CN106372055 B CN 106372055B CN 201610709517 A CN201610709517 A CN 201610709517A CN 106372055 B CN106372055 B CN 106372055B
- Authority
- CN
- China
- Prior art keywords
- user
- read statement
- sentence
- search database
- preliminary search
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3344—Query execution using natural language analysis
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Audiology, Speech & Language Pathology (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Machine Translation (AREA)
Abstract
The present invention relates to the similar processing method and system of semanteme in a kind of man-machine natural language interaction, are related to natural language human-machine interactions field.Purpose is to solve the problem of that existing human-computer interaction technology causes human-computer interaction not realize normally there are accuracy rate is low on semantic understanding.This method realizes process are as follows: S1, establishes preliminary search database and receives user's read statement；S2, the sentence in preliminary search database is screened according to the format of user's read statement；S3, the sentence filtered out in preliminary search database and user's read statement are subjected to semantic comparison, and export final result.The format that the present invention passes through user's read statement first carries out preliminary screening to the sentence in database, then the similitude in user's read statement and database between problem sentence is compared by the comparison of Semantic Similarity, so that robot is improved 10% to 25% to the accuracy of semantic understanding, interactive process is made to become more natural, smooth.
Description
Technical field
The present invention relates to natural language human-machine interactions fields.
Background technique
At present in field of human-computer interaction, when comparing the similitude of two words, clause is not handled, is not concerned in sentence
Relationship between word and word, or even it is not concerned with function word.Such as input " you and Xiao Ming are more severe than whom " and " Xiao Ming in robot
With you than more severe ", robot is the difference of this two word of can not distinguishing one from the other.For some function words, robot also can not be into
Row is distinguished, such as the difference between " your What for " and " your What for ".
But in customer service field, in terms of robot question and answer, as long as robot cannot accurately distinguish the meaning of two words, machine
Device people cannot accurate understanding user be intended to, customer satisfaction system answer cannot be given.With current technology, present semantic understanding is put down
Equal accuracy only has 64%, is also much unable to reach the man-machine purpose normally interacted.
Summary of the invention
Technical problem to be solved by the invention is to provide the similar processing sides of semanteme in a kind of man-machine natural language interaction
Method and system, it is therefore intended that solve existing human-computer interaction technology there are accuracys rate on semantic understanding it is low, cause human-computer interaction without
The problem of method is normally realized.
The technical scheme to solve the above technical problems is that the semanteme in a kind of man-machine natural language interaction is similar
Processing method, it is achieved in the following ways:
S1, it establishes preliminary search database and receives user's read statement；
S2, the sentence in preliminary search database is screened according to the format of user's read statement；
S3, the sentence filtered out in preliminary search database and user's read statement are subjected to semantic comparison, and exported most
Terminate fruit.
Further, S2 specific implementation process includes:
Subject, predicate and object in S21, extraction user's input language；
The master of all sentences in S22, the subject by user's input language, predicate and object and preliminary search database
Language, predicate and object compare；
S23, the language with subject identical as user's input language, predicate and object is filtered out in preliminary search database
Sentence.
Further, the S3 specific implementation process includes:
S31, user's read statement is subjected to phrase fractionation；
S32, by phrases all in user's read statement respectively with wrapped in the sentence that is filtered out in preliminary search database
The phrase contained compares；
S33, it is compared between acquisition every two sentence according to the phrase of user's read statement and preliminary search database
Semantic similar value, and final result is exported according to the result of semantic similar value.
The acquisition process of the semanteme similar value are as follows: by each sentence pair in user's read statement and preliminary search database
Than identical phrase number later divided by phrase number all in user's read statement as semanteme similar value.
The beneficial effects of the present invention are: the present invention pass through first the format of user's read statement to the sentence in database into
Then row preliminary screening compares the phase in user's read statement and database between problem sentence by the comparison of Semantic Similarity
Like property, optimum is exported makes robot improve 10% to 25% to the accuracy of semantic understanding to user, makes human-computer dialogue
Process become more natural, smooth.
A kind of similar processing system of semanteme in man-machine natural language interaction, the system include:
Database module, for establishing preliminary search database and receiving user's read statement；
Sentence screening module, for being sieved according to the format of user's read statement to the sentence in preliminary search database
Choosing；
Semantic contrast module, sentence and user's read statement for will filter out in preliminary search database carry out semantic
Comparison, and export final result.
Further, the sentence screening module includes:
Sentence extraction module, for extracting subject, predicate and object in user's input language；
Format contrast module, for will be in subject, predicate and the object and preliminary search database in user's input language
Subject, predicate and the object of all sentences compare；
The selection result obtains module, has master identical as user's input language for filtering out in preliminary search database
The sentence of language, predicate and object.
Further, the semantic contrast module includes:
Phrase splits module, for user's read statement to be carried out phrase fractionation；
Phrase contrast module, for filtering out phrases all in user's read statement with preliminary search database respectively
Sentence included in phrase compare；
As a result output module, for obtaining every two according to the comparison of the phrase of user's read statement and preliminary search database
Semantic similar value between a sentence, and final result is exported according to the result of semantic similar value.
Further, the acquisition process of the semantic similar value are as follows: will be every in user's read statement and preliminary search database
Identical phrase number is semantic similar value divided by phrase number all in user's read statement after a sentence comparison.
Detailed description of the invention
Fig. 1 is the flow chart of the similar processing method of semanteme in man-machine natural language interaction described in the embodiment of the present invention；
Fig. 2 is according to the format of user's read statement described in the embodiment of the present invention to the sentence in preliminary search database
The flow chart screened；
Fig. 3 be described in the embodiment of the present invention by the sentence filtered out in preliminary search database and user's read statement into
The flow chart of the semantic comparison of row；
Fig. 4 is the principle signal of the similar processing system of semanteme in man-machine natural language interaction described in the embodiment of the present invention
Figure；
Fig. 5 is the schematic illustration of sentence screening module 2 described in the embodiment of the present invention；
Fig. 6 is the schematic illustration of semantic contrast module 3 described in the embodiment of the present invention.
In attached drawing, parts list represented by the reference numerals are as follows:
1, Database module, 2, sentence screening module, 3, semantic contrast module, 4, sentence extraction module, 5, format
Contrast module, 6, the selection result acquisition module, 7, phrase fractionation module, 8, phrase contrast module, 9, result output module.
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.
Embodiment 1
As shown in Figure 1, the present embodiment proposes the similar processing method of semanteme in a kind of man-machine natural language interaction, it is
It is accomplished by the following way:
S1, it establishes preliminary search database and receives user's read statement；
S2, the sentence in preliminary search database is screened according to the format of user's read statement；
S3, the sentence filtered out in preliminary search database and user's read statement are subjected to semantic comparison, and exported most
Terminate fruit.
In the present embodiment, in the beginning handled user's read statement, first to the format for read statement into
Row extracts, and the trunk portion by extracting problem compares so that carrying out the first step deletes choosing；Detailed process is as shown in Figure 2:
Subject, predicate and object in S21, extraction user's input language；
The master of all sentences in S22, the subject by user's input language, predicate and object and preliminary search database
Language, predicate and object compare；
S23, the language with subject identical as user's input language, predicate and object is filtered out in preliminary search database
Sentence.
After carrying out preliminary screening, can may also there are the Subject, Predicate and Object and user's read statement of many problems in database
Subject, Predicate and Object is identical, but in comparison, and sentence corresponding with user's read statement is very in preliminary search database
It is few, it is then then non-by the workload for comparing the phrase of the phrase of each sentence filtered out and user's read statement
Often small, versus speed is also very fast, and detailed process is as shown in Figure 3:
S31, user's read statement is subjected to phrase fractionation；
S32, by phrases all in user's read statement respectively with wrapped in the sentence that is filtered out in preliminary search database
The phrase contained compares；
S33, it is compared between acquisition every two sentence according to the phrase of user's read statement and preliminary search database
Semantic similar value, and final result is exported according to the result of semantic similar value.
Wherein, the acquisition process of semantic similar value are as follows: by each sentence in user's read statement and preliminary search database
Identical phrase number is semantic similar value divided by phrase number all in user's read statement after comparison.
There are ten words: A1+A2+A3+A4+A5+A6+A7+A8+A9+
A0, and there are five word and user's read statement are complete with an identical sentence with its Subject, Predicate and Object in preliminary search database
It is exactly the same: A2+A3+A4+A5+A6, since the clause of the two sentences is identical, then it is assumed that the similitude of the two sentences is
50%, if there are four word is identical, similitude 40%, and so on.According to user's read statement and preliminary search data
Sentence carries out semantic comparison in library, filters out the highest sentence of semantic similar value, then the sentence is final output sentence.
Embodiment 2
As shown in figure 4, the present embodiment proposes the similar processing system of semanteme in a kind of man-machine natural language interaction, this is
System includes:
Database module 1, for establishing preliminary search database and receiving user's read statement；
Sentence screening module 2, for being carried out according to the format of user's read statement to the sentence in preliminary search database
Screening；
Semantic contrast module 3, sentence and user's read statement for will filter out in preliminary search database carry out language
Justice comparison, and export final result.
Preferably, as shown in figure 5, the sentence screening module 2 includes:
Sentence extraction module 4, for extracting subject, predicate and object in user's input language；
Format contrast module 5, for will be in subject, predicate and the object and preliminary search database in user's input language
Subject, predicate and the object of all sentences compare；
The selection result obtains module 6, for being filtered out in preliminary search database with identical as user's input language
The sentence of subject, predicate and object.
Preferably, as shown in fig. 6, the semanteme contrast module 3 includes:
Phrase splits module 7, for user's read statement to be carried out phrase fractionation；
Phrase contrast module 8, for screening phrases all in user's read statement with preliminary search database respectively
Phrase included in sentence out compares；
As a result output module 9, it is every for being obtained according to the comparison of the phrase of user's read statement and preliminary search database
Semantic similar value between two sentences, and final result is exported according to the result of semantic similar value.
Preferably, the acquisition process of the semantic similar value are as follows: will be every in user's read statement and preliminary search database
Identical phrase number is semantic similar value divided by phrase number all in user's read statement after a sentence comparison.
The format that the present embodiment passes through user's read statement first carries out preliminary screening to the sentence in database, then leads to
The similitude in the comparison comparison user's read statement and database of Semantic Similarity between problem sentence is crossed, optimum is defeated
Out to user, so that robot is improved 10% to 25% to the accuracy of semantic understanding, become interactive process more certainly
So, smooth.
The foregoing is merely presently preferred embodiments of the present invention, is not intended to limit the invention, it is all in spirit of the invention and
Within principle, any modification, equivalent replacement, improvement and so on be should all be included in the protection scope of the present invention.
Claims (2)
1. the similar processing method of semanteme in a kind of man-machine natural language interaction, which is characterized in that it is real in the following manner
Existing:
S1, it establishes preliminary search database and receives user's read statement；
S2, the sentence in preliminary search database is screened according to the format of user's read statement；
S3, the sentence filtered out in preliminary search database and user's read statement are subjected to semantic comparison, and export and most terminates
Fruit；
Wherein, the S2 specific implementation process includes:
Subject, predicate and object in S21, extraction user's read statement；
The subject of all sentences, meaning in S22, the subject by user's read statement, predicate and object and preliminary search database
Language and object compare；
S23, the sentence with subject identical as user's read statement, predicate and object is filtered out in preliminary search database；
Wherein, the S3 specific implementation process includes:
S31, user's read statement is subjected to phrase fractionation；
S32, by phrases all in user's read statement respectively and included in the sentence that is filtered out in preliminary search database
Phrase compares；
S33, the semanteme obtained between every two sentence is compared according to the phrase of user's read statement and preliminary search database
Similar value, and final result is exported according to the result of semantic similar value；
The acquisition process of the semanteme similar value are as follows: each sentence in user's read statement and preliminary search database is compared it
Identical phrase number is semantic similar value divided by phrase number all in user's read statement afterwards.
2. the similar processing system of semanteme in a kind of man-machine natural language interaction, which is characterized in that it includes:
Database module (1), for establishing preliminary search database and receiving user's read statement；
Sentence screening module (2), for being sieved according to the format of user's read statement to the sentence in preliminary search database
Choosing；
Semantic contrast module (3), sentence and user's read statement for will filter out in preliminary search database carry out semantic
Comparison, and export final result；
Wherein, the sentence screening module (2) includes:
Sentence extraction module (4), for extracting subject, predicate and object in user's read statement；
Format contrast module (5), for by institute in subject, predicate and the object and preliminary search database in user's read statement
There are the subject, predicate and object of sentence to compare；
The selection result obtains module (6), has master identical as user's read statement for filtering out in preliminary search database
The sentence of language, predicate and object；
Wherein, the semantic contrast module (3) includes:
Phrase splits module (7), for user's read statement to be carried out phrase fractionation；
Phrase contrast module (8), for filtering out phrases all in user's read statement with preliminary search database respectively
Sentence included in phrase compare；
As a result output module (9), for obtaining every two according to the comparison of the phrase of user's read statement and preliminary search database
Semantic similar value between a sentence, and final result is exported according to the result of semantic similar value；
The acquisition process of the semanteme similar value are as follows: each sentence in user's read statement and preliminary search database is compared it
Identical phrase number is semantic similar value divided by phrase number all in user's read statement afterwards.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610709517.7A CN106372055B (en) | 2016-08-23 | 2016-08-23 | A kind of semanteme similar processing method and system in man-machine natural language interaction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610709517.7A CN106372055B (en) | 2016-08-23 | 2016-08-23 | A kind of semanteme similar processing method and system in man-machine natural language interaction |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106372055A CN106372055A (en) | 2017-02-01 |
| CN106372055B true CN106372055B (en) | 2019-10-29 |
Family
ID=57879031
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610709517.7A Active CN106372055B (en) | 2016-08-23 | 2016-08-23 | A kind of semanteme similar processing method and system in man-machine natural language interaction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106372055B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109815484B (en) * | 2018-12-21 | 2022-03-15 | 平安科技（深圳）有限公司 | Semantic similarity matching method and matching device based on cross attention mechanism |
| CN110019688A (en) * | 2019-01-23 | 2019-07-16 | 艾肯特公司 | The method that robot is trained |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1928864A (en) * | 2006-09-22 | 2007-03-14 | 浙江大学 | FAQ based Chinese natural language ask and answer method |
| CN101286161A (en) * | 2008-05-28 | 2008-10-15 | 华中科技大学 | Intelligent Chinese request-answering system based on concept |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2008253551A (en) * | 2007-04-05 | 2008-10-23 | Toshiba Corp | Image reading report search apparatus |
-
2016
- 2016-08-23 CN CN201610709517.7A patent/CN106372055B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1928864A (en) * | 2006-09-22 | 2007-03-14 | 浙江大学 | FAQ based Chinese natural language ask and answer method |
| CN101286161A (en) * | 2008-05-28 | 2008-10-15 | 华中科技大学 | Intelligent Chinese request-answering system based on concept |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106372055A (en) | 2017-02-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106528532B (en) | Text error correction method, device and terminal | |
| CN107393554B (en) | Feature extraction method for fusion inter-class standard deviation in sound scene classification | |
| CN109241266B (en) | Method and device for creating extended question based on standard question in man-machine interaction | |
| CN104504150A (en) | News public opinion monitoring system | |
| CN106557508A (en) | A kind of text key word extracting method and device | |
| CN105893524B (en) | A kind of intelligent answer method and device | |
| CN105653620B (en) | Log analysis method and device of intelligent question-answering system | |
| CN106777261A (en) | Data query method and device based on multi-source heterogeneous data set | |
| KR20210106372A (en) | New category tag mining method and device, electronic device and computer-readable medium | |
| CN105469789A (en) | Voice information processing method and voice information processing terminal | |
| CN105931637A (en) | User-defined instruction recognition speech photographing system | |
| CN104142831B (en) | Application program searching method and device | |
| CN106372055B (en) | A kind of semanteme similar processing method and system in man-machine natural language interaction | |
| CN109817206A (en) | A kind of voice interaction device and method for automatic terminal equipment | |
| CN106550268B (en) | Video processing method and video processing device | |
| JP2011257817A (en) | Patent specification analyzer and text analyzer | |
| CN104751856A (en) | Voice sentence recognizing method and device | |
| US20190213486A1 (en) | Virtual Adaptive Learning of Financial Articles Utilizing Artificial Intelligence | |
| CN110532551A (en) | Method, equipment and the storage medium that text key word automatically extracts | |
| CN110321557A (en) | A kind of file classification method, device, electronic equipment and storage medium | |
| CN109241438A (en) | Across channel focus incident discovery method, apparatus and storage medium based on element | |
| JP2011123565A (en) | Faq candidate extracting system and faq candidate extracting program | |
| CN106446046B (en) | A method of quickly analysis records in time in relational database | |
| Kumala et al. | Indonesian speech emotion recognition using cross-corpus method with the combination of MFCC and teager energy features | |
| Ganoun et al. | Performance analysis of spoken arabic digits recognition techniques |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | 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 |