CN1512396A - Analytic method of open type natural language template - Google Patents
Analytic method of open type natural language template Download PDFInfo
- Publication number
- CN1512396A CN1512396A CNA021592411A CN02159241A CN1512396A CN 1512396 A CN1512396 A CN 1512396A CN A021592411 A CNA021592411 A CN A021592411A CN 02159241 A CN02159241 A CN 02159241A CN 1512396 A CN1512396 A CN 1512396A
- Authority
- CN
- China
- Prior art keywords
- semantic
- natural language
- language template
- groove
- chunk
- 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
Images
Landscapes
- Machine Translation (AREA)
Abstract
The analysis process of opened natural language template includes the following steps: extracting semantic channel information from the natural language template according to preset opened natural language template compiling format; and replacing all the semantic channels in each semantic block layer one layer by one layer with specific information values the system provides to create natural language. The said analysis process of opened natural language template may be used in analyzing natural language template created in the opened natural language template creating process. No matter how complicated and flexible the grammar may be, the language template can be converted to corresponding humanized natural language intelligently without needing any coding of the user.
Description
Technical field
The present invention relates to treatment technology, particularly a kind of analytic method of open natural language template to natural language.
Background technology
At present, along with popularizing of computing machine, people are on the increase the demand of intelligent man-machine interaction system, machine translation system and other customizing messages disposal systems, system related information is handled and in addition intelligent and presenting of hommization becomes one by the field of extensive concern, and present as the intellectuality of one of the most basic information interchange means of mankind natural language, then be a focus in this field.
In the existing intelligent man-machine interaction system, man-to-man often mode.General situation is to realize the various logic of languages of appointing with the program language of computing machine, this needs good fully by appointment logic of language to come the coding language, require simultaneously in the man-machine interaction process, must import natural language by good by appointment logic of language, this method implementation procedure is loaded down with trivial details and can't adapt to flexible and changeable mutual guiding demand.
In the existing machine translation system, can only reconfigure the vocabulary after translating according to the grammer described in the program, semanteme mostly, and form the final objective language.
Though there is minority system in forming the target language process, also to introduce certain language template notion, provided fixing language template, be that basic mode is resolved to fill a vacancy, form target language.The analytic method of fixedly language template like this has certain effect for the fixedly language of simple specific transactions or technical translator.But, present analytic method can only be resolved at the fixedly language template of some specific professional or specific translation speciality, can not resolve different language template, this just need carry out program development respectively to each language template, thereby causes the wasting of resources of great amount of manpower and material resources.
A kind of open natural language template establishment method has appearred at present, this method can be created and modification natural language template according to business demand is autonomous by the user, its basic process is that the information variable that will extract from the natural language original shape is as semantic groove, with natural language original shape combination according to the different business information extraction, the semantic groove and the natural language original shape that may repeat are combined into semantic chunk, or different semantic chunks are carried out modes such as nested combination create the natural language template with the natural language original shape.Since this method have use simple, flexibly, the naturalness advantages of higher, have broad application prospects, yet the fixing analytic method of language template at present can't be realized the parsing of the open language template created by this method.
Summary of the invention
In view of this, the object of the present invention is to provide a kind of analytic method of open natural language template, can resolve the natural language template that adopts open natural language template establishment method to create.
For achieving the above object, technical scheme of the present invention specifically is achieved in that
A kind of analytic method of open natural language template, this method may further comprise the steps:
1) according to predefined open natural language template compiling form, extracts the semantic groove information in the natural language template;
2) successively each the semantic groove in every layer of semantic chunk is replaced with the specifying information value that system provides, all replaced, generate natural language up to all semantic grooves of current natural language template.
Wherein, this method step 2) can replace by carry out semantic groove from internal layer to outer field order, also can carry out semantic groove and replace by the order from the skin to the internal layer.
This method step 1) may further include: add up the element of open natural language template, and preserve each semantic groove information.
This method step 2) may further include: resolve counting for this semantic chunk is provided with circulation, and read piece cycle index in the semantic chunk of current parsing, relevant with the piece cycle index semantic groove with the every replacement of specifying information value in the system once in this semantic chunk, circulation is resolved counting and is added one, resolve counting up to circulation and equate step 2 with the piece cycle index) can further include: the semantic groove that has nothing to do with the piece cycle index in to the semantic chunk of current parsing, keep this semanteme groove under one deck resolve.Step 2) can further include: the semantic chunk that reads is replaced with the specifying information value in the system with the semantic groove in the semantic chunk of layer.
This method step 2) may further include: replace with the specifying information value of system being arranged in the outer semantic groove of semantic chunk in the open natural language template.
This method can further be resolved the natural language that the back generates with the language template of user's input, exports to the user with the form of tabulation.
As seen from the above technical solutions, the analytic method of this open natural language template of the present invention, can resolve the natural language template that adopts open natural language template establishment method to create, no matter how complicated its grammer is, flexible, and not needing the user to carry out any coding can both intelligentizedly be converted into corresponding with it hommization natural language with language template.
Description of drawings
Fig. 1 is the analytic method process flow diagram of open natural language template of the present invention;
Fig. 2 is to the detailed process process flow diagram of basic semantic chunk dissection process among Fig. 1.
Embodiment
For making the purpose, technical solutions and advantages of the present invention clearer, below in conjunction with embodiment and accompanying drawing, the present invention is described in more detail.
Fig. 1 is the analytic method process flow diagram of open natural language template of the present invention, and as shown in Figure 1, the process of the analytic method of open natural language template of the present invention is:
Compiling form and rule detection are carried out to open natural language template in step 102～105, judge whether it is wrong, if wrong demonstration error message stops resolving.If there is not mistake, then extract the semantic groove in the natural language template, extract the information such as nest relation between title, position and each semantic chunk of semantic groove and preserve.
The innermost layer semantic chunk in the current outermost layer semantic chunk semantic chunk is searched and read to step 106 for nested semantic chunk,, and present embodiment is by resolving to outer field order from internal layer, also can resolving by the order from the skin to the internal layer.
Fig. 2 be among Fig. 1 to the detailed process process flow diagram of basic semantic chunk dissection process, as shown in Figure 2, the process that basic semantic chunk is handled is:
As seen from Figure 2, before carrying out basic semantic chunk parsing, at first need to be provided with the piece circulation numeration " LoopCount " of current semantic chunk acquiescence, need determine the current replacement information variate-value that should in system, choose according to the mark " LoopCount " of current circulation parsing number of times for the dynamic semantics groove relevant with cycle index; For with the circulation incoherent dynamic semantics groove then can in current semantic chunk resolving, be left intact, keeps original information, to last layer or on which floor resolve.
It can also be seen that by Fig. 2,, except above-mentioned replacement method, in resolving, also follow a kind of cyclic process semantic groove carrying out basic semantic chunk when resolving.The every circulation of semantic chunk resolved once all to be needed semantic grooves all in the semantic chunk is handled once according to above-mentioned replacement method, until circulation resolve number of times mark " LoopCount " cycle index corresponding with current semantic chunk when equal till.
The most basic semantic chunk of the above-mentioned replacement method of integrated use, round-robin method resolve to only contain the natural language literal and with the current semantic chunk text message that uncorrelated semantic groove forms that circulates, even only contain the text message of natural language.
If this semantic chunk is semantic chunk unique in the language template then only to be needed to carry out simple substitution to all remaining semantic grooves and just formed final hommization natural language information, otherwise it will be resolved according to semantic chunk analytic method shown in Figure 1 as the part in the newly-generated semantic chunk.
By above-mentioned step as seen, no matter to complexity, comprise many groups nested combinations, still do not contain semantic chunk or only contain the open natural language template of the simple semantic chunk of an individual layer, can both adopt above-mentioned analytic method to resolve, and finally generate the natural language of hommization.Its basic skills is a unit with the semantic chunk exactly, respectively semantic chunk is resolved, and resolves the outer semantic groove of semantic chunk at last; Also can resolve the outer semantic groove of semantic chunk earlier, resolve semantic chunk again.For the parsing of nested semantic chunk, according to order from the inside to the outside, successively the semantic chunk that is in internal layer is resolved, for then carrying out successively, resolve the outermost layer semantic chunk at last with reference to the appearance order with layer semantic chunk.Certainly, also can resolve by skin to internal layer.All may form a new semantic chunk after every layer of semantic chunk resolved, resolving and other semantic chunks of the new semantic chunk that forms are in full accord.
According to the above-mentioned longitudinal and transverse analytic method that interweaves, the language template of complexity has been realized parsing layer by layer from inside to outside, and all linked with one another, whole resolving is organically constituted an integral body, until forming the final natural language of forming by text message fully.
Below in conjunction with the application example of open natural language template in voice-mail system, the detailed process that open natural language template is resolved describes targetedly.The voice-mail system here is meant the system that the interactive voice by phone and system carries out the mail read and write.
Realize interactive voice, need before voice-mail system comes into operation, in system, set up the natural language template, when voice-mail system uses, the natural language template is resolved the generation natural language play to the user by phone and system.The natural language template being resolved the generation natural language play to the user, is that voice-mail system is realized an important step by phone and system's dialogue.
For example, voice-mail system need play to the user with the situation of mail in the subscriber mailbox, the natural language template that it utilized can for: " you have [AllMailNum] envelope mail now; { [UserMailNum] seal from [UserName]; { [] sealed and to be themed as [title], is an envelope [importencevalue] mail.[@UserMailNum]}，[@UserNum]}”。The definition of each semantic groove is referring to table one.
| Semantic slot name | Definition |
| [AllMailNum] | Mail sum in the mailbox |
| [@UserNum] | The different numbers of posting a letter |
| [UserName] | The name of posting a letter |
| [UserMailNum]、[@UserMailNum] | The mail number that each addresser sends |
| [@loopcount] | The cycle count of current block |
| [title] | The mail matter topics of every envelope mail correspondence |
| [importencevalue] | Important attribute (relevant) with every envelope mail |
| Important attribute (relevant) with the addresser |
Table one
Wherein [] is semantic groove, and { } is semantic chunk.[@loopcount] be cycle count groove, [@UserMailNum] and [@UserNum] be the cycle index groove.
This is a complex language template that contains nested semantic chunk, can be referring to Fig. 1, Fig. 2 to its process of resolving.
At first, add up the template key element of this language template, wherein [with], with all be paired appearance, meet the redaction rule of open natural language template.
Then, wherein all semantic grooves are extracted, the information such as nest relation in the title of extraction tank, position, each semantic chunk are also preserved.
Then, search and read current innermost layer semantic chunk, in this template be [@loopcount] envelope themes as [title], is an envelope [importencevalue] mail.[@UserMailNum] } this semantic chunk.Semantic chunk hereto can have following two kinds of methods to resolve according to the actual conditions of nest relation in the semantic chunk of above-mentioned steps extraction tank and user mail.
First kind: relevant if user's wherein semantic groove when creating this open natural language template is set to every envelope mail, resolve to directly then that " first envelope themes as meeting, is an envelope important email.Second envelope themes as tourism, is an envelope important email.First envelope themes as test, is the inessential mail of an envelope ".
Second kind: if the user create should be open during the natural language template wherein semantic groove be set to relevantly with the addresser, then can be divided into two and go on foot parsing, resolve to earlier that " first envelope themes as meeting, is an envelope [importencevalue] mail.Second envelope themes as tourism, is an envelope [importencevalue] mail ".First envelope themes as test, is an envelope [importencevalue] mail.At this moment, [importencevalue] is uncorrelated semantic groove with the piece cycle index.And then resolve to that " first envelope themes as meeting, is the private important email of an envelope.Second envelope themes as tourism, is the private important email of an envelope.First envelope themes as test, is the common office mail of an envelope ".
This semantic chunk does not have other semantic chunks with layer, so next outer semantic chunk is resolved, promptly { [UserMailNum] seals from [UserName], [] } this one deck resolves, analytic method is the same, and analysis result can be " 2 envelopes are from Zhang San, and 1 envelope is from Li Si ".
At this moment, the semantic chunk of this language template is just resolved and is finished, system finds [AllMailNum] the semantic groove that also has in " you have [AllMailNum] envelope mail now " and does not also resolve, replace [AllMailNum] semantic groove with the sum of this user mail in the system this moment, and whole like this language template is resolved and finished.This template can resolve to according to being provided with finally of different user that " you have 3 envelope mails, and 2 envelopes are from Zhang San, and first envelope themes as meeting, are envelope important email.Second envelope themes as tourism, is an envelope important email.1 envelope themes as test from Li Si, is the inessential mail of an envelope ".Or " you have 3 envelope mails, and 2 envelopes are from Zhang San, and first envelope themes as meeting, are the private important email of an envelope.Second envelope themes as tourism, is the private important email of an envelope.1 envelope is from Li Si, and first envelope themes as test, is the common office mail of an envelope ".
By the above embodiments as seen, the analytic method of this open natural language template of the present invention, can resolve the natural language template that adopts open natural language template establishment method to create, no matter how complicated its grammer is, flexible, and not needing the user to carry out any coding can both intelligentizedly be converted into corresponding with it hommization natural language with language template.
Claims (9)
1, a kind of analytic method of open natural language template is characterized in that, this method may further comprise the steps:
1) according to predefined open natural language template compiling form, extracts the semantic groove information in the natural language template;
2) successively each the semantic groove in every layer of semantic chunk is replaced with the specifying information value that system provides, all replaced, generate natural language up to all semantic grooves of current natural language template.
2, analytic method as claimed in claim 1 is characterized in that, this method step 2) replace by carry out semantic groove from internal layer to outer field order.
3, analytic method as claimed in claim 1 is characterized in that, this method step 2) carry out semantic groove replacement by the order from the skin to the internal layer.
4, analytic method as claimed in claim 1 is characterized in that, this method step 1) further comprise: add up the element of open natural language template, and preserve the information of each semantic groove.
5, analytic method as claimed in claim 1, it is characterized in that, this method step 2) further comprises: resolve counting for this semantic chunk is provided with circulation, and read piece cycle index in the semantic chunk of current parsing, relevant with the piece cycle index semantic groove with the every replacement of specifying information value in the system once in this semantic chunk, circulation is resolved counting and is added one, resolves counting up to circulation and equates with the piece cycle index.
6, analytic method as claimed in claim 5 is characterized in that, this method step 2) further comprise: to the semantic groove that has nothing to do with the piece cycle index in the semantic chunk of current parsing, keep semantic groove and resolve to following one deck.
7, analytic method as claimed in claim 1 is characterized in that, this method step 2) further comprise: the semantic chunk that reads is replaced with the specifying information value in the system with the semantic groove in the semantic chunk of layer.
8, analytic method as claimed in claim 1 is characterized in that, this method step 2) further comprise: replace with the specifying information value of system being arranged in the outer semantic groove of semantic chunk in the open natural language template.
9, analytic method as claimed in claim 1 is characterized in that, this method further comprises: the language template of user's input is resolved the natural language that the back generates, export to the user with the form of tabulation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNA021592411A CN1512396A (en) | 2002-12-27 | 2002-12-27 | Analytic method of open type natural language template |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNA021592411A CN1512396A (en) | 2002-12-27 | 2002-12-27 | Analytic method of open type natural language template |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN1512396A true CN1512396A (en) | 2004-07-14 |
Family
ID=34237381
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNA021592411A Pending CN1512396A (en) | 2002-12-27 | 2002-12-27 | Analytic method of open type natural language template |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1512396A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100399329C (en) * | 2005-01-19 | 2008-07-02 | 结信网络技术服务（上海）有限公司 | Intelligent movable guiding engine systems |
| CN105138575A (en) * | 2015-07-29 | 2015-12-09 | 百度在线网络技术（北京）有限公司 | Analysis method and device of voice text string |
| CN107450725A (en) * | 2017-07-31 | 2017-12-08 | 科大讯飞股份有限公司 | Man-machine interaction application platform, method and storage medium |
| CN110232189A (en) * | 2019-06-11 | 2019-09-13 | 上海证大喜马拉雅网络科技有限公司 | Semantic analytic method, device, equipment and storage medium |
-
2002
- 2002-12-27 CN CNA021592411A patent/CN1512396A/en active Pending
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100399329C (en) * | 2005-01-19 | 2008-07-02 | 结信网络技术服务（上海）有限公司 | Intelligent movable guiding engine systems |
| CN105138575A (en) * | 2015-07-29 | 2015-12-09 | 百度在线网络技术（北京）有限公司 | Analysis method and device of voice text string |
| CN105138575B (en) * | 2015-07-29 | 2017-09-05 | 百度在线网络技术（北京）有限公司 | The analysis method and device of speech text string |
| CN107450725A (en) * | 2017-07-31 | 2017-12-08 | 科大讯飞股份有限公司 | Man-machine interaction application platform, method and storage medium |
| CN107450725B (en) * | 2017-07-31 | 2020-09-11 | 科大讯飞股份有限公司 | Man-machine interaction application platform, method and storage medium |
| CN110232189A (en) * | 2019-06-11 | 2019-09-13 | 上海证大喜马拉雅网络科技有限公司 | Semantic analytic method, device, equipment and storage medium |
| CN110232189B (en) * | 2019-06-11 | 2023-06-02 | 上海喜马拉雅科技有限公司 | Semantic analysis method, device, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Partee | Noun phrase interpretation and type-shifting principles | |
| CN1159661C (en) | System for Chinese tokenization and named entity recognition | |
| CN111061862B (en) | Method for generating abstract based on attention mechanism | |
| CN107463553A (en) | For the text semantic extraction, expression and modeling method and system of elementary mathematics topic | |
| CN1340804A (en) | Automatic new term fetch method and system | |
| CN101114298A (en) | Method for gaining oral vocabulary entry, device and input method system thereof | |
| CN1945691A (en) | Method for improving template sentence synthetic effect in voice synthetic system | |
| CN108874791B (en) | Semantic analysis and Chinese-English sequencing method and system based on minimum semantic block | |
| CN101539910A (en) | A sentence taking method for computer aided translation and system thereof | |
| Zarisheva et al. | Dialog act annotation for twitter conversations | |
| Linderman | Computer content analysis and manual coding techniques: A comparative analysis | |
| CN106227575A (en) | Method for generating and analyzing text file | |
| CN1512396A (en) | Analytic method of open type natural language template | |
| CN1776673A (en) | Method for converting PDF file to XML file | |
| CN1773453A (en) | System constituting method based on data definition | |
| CN1512395A (en) | Establishing method for open type natural language | |
| CN1270363A (en) | Database management method | |
| CN101034394A (en) | System and method for enhancing translation efficiency | |
| CN102929700B (en) | Method for importing word test library to interactive teaching platform | |
| CN115292347A (en) | Active SQL algorithm performance checking device and method based on rules | |
| CN114419645A (en) | Contract intelligent analysis method based on AI | |
| CN1220971C (en) | Organizing and identifying method for natural language | |
| CN1226692C (en) | Machine translation system based on semanteme and its method | |
| CN106846444A (en) | A kind of animation system and the method for making animation | |
| CN1417707A (en) | Natural language semantic information united-coding method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |