CN109727591A - A kind of method and device of phonetic search - Google Patents
A kind of method and device of phonetic search Download PDFInfo
- Publication number
- CN109727591A CN109727591A CN201811645648.9A CN201811645648A CN109727591A CN 109727591 A CN109727591 A CN 109727591A CN 201811645648 A CN201811645648 A CN 201811645648A CN 109727591 A CN109727591 A CN 109727591A
- Authority
- CN
- China
- Prior art keywords
- resource
- phonetic
- search
- search result
- module
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000004458 analytical method Methods 0.000 claims abstract description 12
- 235000013399 edible fruits Nutrition 0.000 claims description 7
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000013480 data collection Methods 0.000 claims description 3
- 230000011218 segmentation Effects 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 5
- 230000007257 malfunction Effects 0.000 abstract description 3
- 230000006870 function Effects 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 230000003993 interaction Effects 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 239000002131 composite material Substances 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000010485 coping Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Machine Translation (AREA)
Abstract
The present invention provides a kind of method and devices of phonetic search, which comprises pre-establishes the inverted index of resource, while the temperature of more new resources；Acquire user voice data；The user voice data is identified, is converted into text data；Semantic analysis is carried out to the text data, obtains core keyword；According to the length of the core keyword, using different phonetic search strategies, initial search result is obtained；And initial search result is ranked up by the temperature of resource, obtain final phonetic search result.The present invention solves the problems, such as that existing voice identification often malfunctions, search speed is slow, user speech search experience is poor, it proposes that phonetic search technical solution can fast and accurately export search result in centainly misrecognition range, and can greatly improve user speech search experience effect.
Description
Technical field
The invention belongs to phonetic search technical field more particularly to a kind of method and devices of phonetic search.
Background technique
With smart machine, human-computer interaction and the development of AI technology, to phonetic search the relevant technologies, particularly in speed
It is increasingly stringenter with the requirement in terms of precision.Currently, phonetic search will generally be searched by speech recognition, semantic analysis, resource
Four rope, sort result steps.Since polyphone, the accent of user and the factors such as dialect and environment noise cause voice to know
It does not often malfunction, search space is caused to expand, search result and user are intended to the problems such as far from each other, search speed is slow, greatly
Influence user speech search experience.Phonetic search technology proposed in this paper can be fast and accurately in centainly misrecognition range
Search result is exported, user speech search experience effect can be greatly improved.
Summary of the invention
In consideration of it, the embodiment of the invention provides a kind of method and device of phonetic search, to solve existing voice identification
Often error, search speed be slow, problem of user speech search experience difference, and realize in centainly misrecognition range can it is quick,
Accurately output search result, and can greatly improve user speech search experience effect.
In a first aspect, providing a kind of method of phonetic search, which comprises
Pre-establish the inverted index of resource, while the temperature of more new resources；
Acquire user voice data；
The user voice data is identified, is converted into text data；
Semantic analysis is carried out to the text data, obtains core keyword；
According to the length of the core keyword, using different phonetic search strategies, initial search result is obtained；And it presses
The temperature of resource is ranked up initial search result, obtains final phonetic search result.
Further, the inverted index for pre-establishing resource specifically includes:
To all resource numbers to be searched, unique ID of generation resource index；
According to resource content, extract keyword and sentence for retrieving the resource；
It is segmented, is individual Chinese character by the keyword and sentence cutting, obtains set of keywords；
According to the size of the set of keywords, inverted index is established by relative strategy.
Further, the size according to the set of keywords, establishes inverted index by relative strategy and specifically includes:
Judge the value of the size K of the set of keywords；
As K≤2, the pronunciation of each word in the set of keywords is obtained, obtains pronunciation set { a1, a2 ..., an },
And corresponding phonetic resource index is established to each pronunciation in pronunciation set；
As K > 2, obtain the coding of each word in the set of keywords, obtain character code set c1, c2 ...,
Cn }, and establish corresponding character resource to each character in character code set and index.
Further, the length according to the core keyword using different phonetic search strategies, obtains tentatively
Search result；And initial search result is ranked up by the temperature of resource, final phonetic search is obtained as a result, specifically including:
The core keyword obtained described in cutting obtains user's search key set；
Judge the value of the size M of user's search key set；
As M≤2, the phonetic resource index is searched for, initial search result is obtained；And by the temperature of resource to tentatively searching
Hitch fruit is ranked up, and obtains final phonetic search result；
As M > 2, searches for the character resource index and by the value length of M using different fault-tolerant strategies, obtain preliminary
Search result；And initial search result is ranked up by the temperature of resource, obtain final phonetic search result.
Second aspect, provides a kind of device of phonetic search, and described device includes:
Preprocessing module, for pre-establishing the inverted index of resource, while the temperature of more new resources；
Speech data collection module, for acquiring user voice data；
Speech recognition module, for being identified to the user voice data, being converted into text data；
Semantic module obtains core keyword for carrying out semantic analysis to the text data；
Phonetic search and result output module are searched for the length according to the core keyword using different voices
Rope strategy obtains initial search result；And initial search result is ranked up by the temperature of resource, obtain final phonetic search
As a result.
Further, the preprocessing module device further include:
ID generation module is indexed, for unique ID to all resource numbers to be searched, generation resource index；
Crucial words and phrases abstraction module, for according to resource content, extract keyword and sentence for retrieving the resource；
The keyword and sentence cutting are individual Chinese character, obtain set of keywords by word segmentation module for being segmented；
Inverted index establishes module, for the size according to the set of keywords, establishes inverted index by relative strategy.
Further, the inverted index is established module and is specifically included:
The size judgment module of set of keywords, the value of the size K for judging the set of keywords；
Phonetic resource index establishes module, for obtaining the pronunciation of each word in the set of keywords, obtaining as K≤2
Corresponding phonetic resource index is established to pronunciation set { a1, a2 ..., an }, and to each pronunciation in pronunciation set；
Character resource index establishes module, for obtaining the coding of each word in the set of keywords, obtaining as K > 2
Corresponding character resource rope is established to character code set { c1, c2 ..., cn }, and to each character in character code set
Draw.
Further, the phonetic search and result output module specifically include:
Cutting core key word modules, the core keyword for obtaining described in cutting obtain user's search key collection
It closes；
The size judgment module of user's search key set, for judging the size of user's search key set
The value of M；
First phonetic search object module, for searching for the phonetic resource index, obtaining initial search knot as M≤2
Fruit；And initial search result is ranked up by the temperature of resource, obtain final phonetic search result；
Second phonetic search object module, for as M > 2, searching for the character resource index and being adopted by the value length of M
With different fault-tolerant strategies, initial search result is obtained；And initial search result is ranked up by the temperature of resource, it obtains most
Whole phonetic search result.
Compared with prior art, the present invention is by pre-establishing the inverted index of resource at calculating center, while updating money
The temperature in source；User voice data is acquired using equipment；The user voice data is identified first, is converted into textual data
According to；And semantic analysis is carried out to the text data, obtain core keyword；Further according to the length of the core keyword, adopt
With different phonetic search strategies, obtain initial search result；And initial search result is ranked up by the temperature of resource, it obtains
To final phonetic search result；To propose a kind of new phonetic search mode, the energy in centainly misrecognition range is realized
Search result is fast and accurately exported, and can greatly improve user speech search experience effect.
Detailed description of the invention
Fig. 1 is the implementation flow chart of the method for phonetic search provided in an embodiment of the present invention.
Fig. 2 is the composite structural diagram of the device of phonetic search provided in an embodiment of the present invention.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.
The embodiment of the present invention is by pre-establishing the inverted index of resource, while the temperature of more new resources at calculating center；
User voice data is acquired using equipment；The user voice data is identified first, is converted into text data；And to institute
It states text data and carries out semantic analysis, obtain core keyword；Further according to the length of the core keyword, using different languages
Sound search strategy obtains initial search result；And initial search result is ranked up by the temperature of resource, obtain final voice
Search result solves the problems, such as that existing voice identification often malfunctions, search speed is slow, user speech search experience is poor.This hair
Bright embodiment additionally provides corresponding device, and which will be described in detail below.
Fig. 1 shows the implementation flow chart of the method for phonetic search provided in an embodiment of the present invention.
In embodiments of the present invention, the method for the phonetic search is applied to computer, smart machine, AI equipment, machine
People etc. needs to input in the human-computer interaction terminal of voice.
Refering to fig. 1, the method for the phonetic search includes:
In step sl, the inverted index of resource, while the temperature of more new resources are pre-established.
In embodiments of the present invention, need to be pre-processed at calculating center, pretreated purpose be pre-establish it is all
Resource inverted index, facilitates resource quickly to position, and is generally completed by following 4 steps:
S101, to all resource numbers to be searched, generate resource index unique ID.
S102, according to resource content, extract keyword and sentence for retrieving the resource.
S103, it is segmented, is individual Chinese character by the keyword and sentence cutting, obtains set of keywords.
S104, according to the size of the set of keywords, establish inverted index by relative strategy.Wherein, inverted index
(inverted index): also referred to as reverse indexing, merging archives or reversed archives are a kind of indexing means, are used to store
The mapping of storage location of some word in a document or one group of document under full-text search.
When it is implemented, assuming that the size of the set of keywords is K, the step S104 further include:
Judge the value of the size K of the set of keywords；
As K≤2, the pronunciation of each word in the set of keywords is obtained, obtains pronunciation set { a1, a2 ..., an },
And corresponding phonetic resource index is established to each pronunciation in pronunciation set；
As K > 2, obtain the coding of each word in the set of keywords, obtain character code set c1, c2 ...,
Cn }, and establish corresponding character resource to each character in character code set and index.
In the above manner, pre-establishing the inverted index of all resources in calculating center, including phonetic resource index
With character resource index, because combine character and phonetic progress resource searching, compare traditional technical solution, more can easily by
Phonetic search resource is quickly positioned.
S2, acquisition user voice data.Pass through the human-computer interaction terminals such as computer, smart machine, AI equipment, robot
Speech ciphering equipment collects the voice of user.
S3, the user voice data is identified, is converted into text data.The user speech of acquisition is known
, also referred to as automatic speech recognition, voice do not turn text identification, target be with the calculating center of the intelligent terminals such as computer from
It is dynamic that the voice content of the mankind is converted into corresponding text.Different from Speaker Identification and speaker verification, the latter attempts identification
Or confirm the speaker for issuing voice rather than vocabulary content included in it.
S4, semantic analysis is carried out to the text data, obtains core keyword.I.e. intelligent terminal carries out text
Semantic analysis refers to a lot of text or content, from the abstract and general idea for wherein analyzing this paragraph.Even more into one
Step sorts out the Wen Yi of entire article to come.To judge that user is intended to, output user searches for core keyword.
S5, according to the length of the core keyword, using different phonetic search strategies, obtain initial search result；
And initial search result is ranked up by the temperature of resource, obtain final phonetic search result.
The present invention is when it is implemented, character type resource index carries out different fault-tolerant strategies by the length of core keyword.
Specifically, the step S5 is specifically sequentially included the following steps:
The core keyword obtained described in S501, cutting obtains user's search key set.
S502, judge user's search key set size M value.
S503, as M≤2, search for the phonetic resource index, obtain initial search result；And press the temperature pair of resource
Initial search result is ranked up, and obtains final phonetic search result.
Specifically, it as M≤2, executes according to the following steps:
S5031, the pronunciation for first obtaining each keyword；
S5032 and according to phonetic resource index, searches the corresponding resource collection of each pronunciation；
S5033, it seeks common ground again to the corresponding resource collection of each pronunciation；
S5034, judge whether the intersection is empty set；
If S5035, the intersection are empty sets, it is transferred to progress character resource search in step S5041；If being otherwise transferred to
S5036；
S5036, the intersection are not empty set, in the intersection resource, by S1 calculate center update resource temperature carry out
Sequence, and obtain final phonetic search result.
In the following, being illustrated with M=2:
First obtain the pronunciation of each word；
It is assumed to be a1, a2, b1, b2, wherein a1, a2 are the pronunciation of first Chinese character, and b1, b2 are the pronunciation of second word,
According to phonetic resource index, the corresponding resource collection S1 of a1, a2, the corresponding resource collection S2 of b1, b2 are obtained；
Seek S1, S2 intersection of sets collection S0；
Judge whether set S0 is empty set；
If set S0 is empty set, it is transferred to progress character resource search in step S5041, if set S0 is not empty
Collection is then ranked up the resource in set S0, the resource temperature that calculates by S1 center update, and inputs final phonetic search knot
Fruit.
S504, as M > 2, search for character resource index and by the value length of M using different fault-tolerant strategies, obtain
To initial search result；And initial search result is ranked up by the temperature of resource, obtain final phonetic search result.
When it is implemented, being executed according to the following steps as M > 2:
S5041, the corresponding character code of each keyword is first obtained；
S5042 is simultaneously indexed according to character resource, and lookup obtains the corresponding resource collection of each character；
S5043, whether judge M value≤5；If then seeking common ground to the corresponding resource collection of each character, and it is transferred to
S5046；If it is not, being then transferred to S5044.Specifically, in M≤5, allow to lose a Chinese character, i.e. a resource in the intersection
All occur in 4 or more resource collections above, is just added into the intersection.
S5044, whether judge M value≤9；If then seeking common ground to the corresponding resource collection of each character, and it is transferred to
S5046；If it is not, being then transferred to S5045.Specifically, in M≤9, allow to lose two Chinese characters, i.e. a resource in the intersection
All occur in 7 or more resource collections above, is just added into the intersection.
When S5045, M > 9, seek common ground to the corresponding resource collection of each character, and be transferred to S5046.Specifically, in M > 9
When, allow to lose three Chinese characters in the intersection, it, will as soon as that is, resource all occurs in 6 or more resource collections above
The intersection is added in it.
S5046, judge whether intersection is empty set, if then using strategy of revealing all the details；If being otherwise transferred to S5047；
S5047, in the intersection resource, calculate the resource temperature that center updates by S1 and be ranked up, and obtain final
Phonetic search result.
In the following, being illustrated with M=9:
The corresponding character code of each keyword is first obtained, c1, c2 ..., c9 are obtained；
And indexed according to character resource, lookup obtains each character corresponding resource collection S1, S2 ..., S9；
Seek S1, S2 ..., S9 intersection of sets collection SN.Due to allowing to lose two Chinese characters when regulation K=9, it may be assumed that a resource
All occur in 7 or more set above, is just added into SN.
Judge whether intersection SN is empty set, if returning in single word temperature in resource collection then using revealing all the details strategy
Highest resource；If otherwise in the intersection resource, calculate the resource temperature that center updates by S1 and be ranked up, and obtain most
Whole phonetic search result.
In conclusion phonetic search scheme provided in an embodiment of the present invention combines character and phonetic carries out resource searching,
More traditional scheme has two o'clock promotion:
1, it is searched for relative to simple characters, searching method provided herein is adopted when user's search key set is less than 2
The mode searched for phonetic, the problems such as can solve cacoepy when short input, dialect mistake；When user's search key collection
When closing greater than 2, and in such a way that fault-tolerant and character match combines, the input keyword letter of user can be efficiently used
Breath does not influence search speed while covering identification mistake again.
2, it is searched for relative to simple pronunciation, searching method provided herein is adopted when user's search key length is greater than 2
It is scanned for character and the fault-tolerant mode combined, user is made full use of to input information, so that search space greatly reduces,
It searches for more accurate, rapid.
It should be understood that in the above-described embodiments, the size of the serial number of each step is not meant that the order of the execution order, it is each to walk
Rapid execution sequence should be determined by its function and internal logic, and the implementation process without coping with the embodiment of the present invention constitutes any limit
It is fixed.
Fig. 2 shows the composite structural diagrams of the device of phonetic search provided in an embodiment of the present invention, for ease of description, only
Show part related to the embodiment of the present invention.
In embodiments of the present invention, described device for realizing above-mentioned Fig. 1 phonetic search as described in the examples method,
It can be the unit for being built in computer, the software unit of server, hardware cell or software and hardware combining.
Referring to Fig.2, described device includes:
Preprocessing module 21, for pre-establishing the inverted index of resource, while the temperature of more new resources；
Speech data collection module 22, for acquiring user voice data；
Speech recognition module 23, for being identified to the user voice data, being converted into text data；
Semantic module 24 obtains core keyword for carrying out semantic analysis to the text data；
Phonetic search and result output module 25, for the length according to the core keyword, using different voices
Search strategy obtains initial search result；And initial search result is ranked up by the temperature of resource, it obtains final voice and searches
Hitch fruit.
Further, the preprocessing module device further include:
ID generation module is indexed, for unique ID to all resource numbers to be searched, generation resource index；
Crucial words and phrases abstraction module, for according to resource content, extract keyword and sentence for retrieving the resource；
The keyword and sentence cutting are individual Chinese character, obtain set of keywords by word segmentation module for being segmented；
Inverted index establishes module, for the size according to the set of keywords, establishes inverted index by relative strategy.
The inverted index is established module and is specifically included:
The size judgment module of set of keywords, the value of the size K for judging the set of keywords；
Phonetic resource index establishes module, for obtaining the pronunciation of each word in the set of keywords, obtaining as K≤2
Corresponding phonetic resource index is established to pronunciation set { a1, a2 ..., an }, and to each pronunciation in pronunciation set；
Character resource index establishes module, for obtaining the coding of each word in the set of keywords, obtaining as K > 2
Corresponding character resource rope is established to character code set { c1, c2 ..., cn }, and to each character in character code set
Draw.
The phonetic search and result output module specifically include:
Cutting core key word modules, the core keyword for obtaining described in cutting obtain user's search key collection
It closes；
The size judgment module of user's search key set, for judging the size of user's search key set
The value of M；
First phonetic search object module, for searching for the phonetic resource index, obtaining initial search knot as M≤2
Fruit；And initial search result is ranked up by the temperature of resource, obtain final phonetic search result；
Second phonetic search object module, for as M > 2, searching for the character resource index and being adopted by the value length of M
With different fault-tolerant strategies, initial search result is obtained；And initial search result is ranked up by the temperature of resource, it obtains most
Whole phonetic search result.
It should be noted that the device in the embodiment of the present invention can be used to implement whole skills in above method embodiment
The function of art scheme, each functional module can be implemented according to the method in above method embodiment, specific implementation
Process can refer to the associated description in examples detailed above, and details are not described herein again.
In conclusion the embodiment of the present invention is by pre-establishing the inverted index of resource at calculating center, while updating money
The temperature in source；User voice data is acquired using equipment；The user voice data is identified first, is converted into textual data
According to；And semantic analysis is carried out to the text data, obtain core keyword；Further according to the length of the core keyword, adopt
With different phonetic search strategies, obtain initial search result；And initial search result is ranked up by the temperature of resource, it obtains
To final phonetic search result；To propose a kind of new phonetic search mode, the energy in centainly misrecognition range is realized
Search result is fast and accurately exported, and can greatly improve user speech search experience effect.
Those of ordinary skill in the art may be aware that list described in conjunction with the examples disclosed in the embodiments of the present disclosure
Member and algorithm steps can be realized with the combination of electronic hardware or computer software and electronic hardware.These functions are actually
It is implemented in hardware or software, the specific application and design constraint depending on technical solution.Professional technician
Each specific application can be used different methods to achieve the described function, but this realization is it is not considered that exceed
The scope of the present invention.
It is apparent to those skilled in the art that for convenience and simplicity of description, the device of foregoing description
It with the specific work process of unit, can refer to corresponding processes in the foregoing method embodiment, details are not described herein.
In several embodiments provided herein, it should be understood that disclosed method and device can pass through it
Its mode is realized.For example, the apparatus embodiments described above are merely exemplary, for example, the module, unit are drawn
Point, only a kind of logical function partition, there may be another division manner in actual implementation, such as multiple units or components can
To combine or be desirably integrated into another system, or some features can be ignored or not executed.Another point, it is shown or beg for
The mutual coupling, direct-coupling or communication connection of opinion can be through some interfaces, the INDIRECT COUPLING of device or unit
Or communication connection, it can be electrical property, mechanical or other forms.
The unit as illustrated by the separation member may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple
In network unit.It can select some or all of unit therein according to the actual needs to realize the mesh of this embodiment scheme
's.
In addition, each functional unit in each embodiment of the present invention, module can integrate in one processing unit,
Can be each unit, module physically exists alone, can also be integrated in one unit with two or more units, module
In.
It, can be with if the function is realized in the form of SFU software functional unit and when sold or used as an independent product
It is stored in a computer readable storage medium.Based on this understanding, technical solution of the present invention is substantially in other words
The part of the part that contributes to existing technology or the technical solution can be embodied in the form of software products, the meter
Calculation machine software product is stored in a storage medium, including some instructions are used so that a computer equipment (can be a
People's computer, server or network equipment etc.) it performs all or part of the steps of the method described in the various embodiments of the present invention.
And storage medium above-mentioned includes: that USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), arbitrary access are deposited
The various media that can store program code such as reservoir (RAM, Random Access Memory), magnetic or disk.
The above description is merely a specific embodiment, but scope of protection of the present invention is not limited thereto, any
Those familiar with the art in the technical scope disclosed by the present invention, can easily think of the change or the replacement, and should all contain
Lid is within protection scope of the present invention.Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (8)
1. a kind of method of phonetic search, which is characterized in that the described method includes:
Pre-establish the inverted index of resource, while the temperature of more new resources；
Acquire user voice data；
The user voice data is identified, is converted into text data；
Semantic analysis is carried out to the text data, obtains core keyword；
According to the length of the core keyword, using different phonetic search strategies, initial search result is obtained；And press resource
Temperature initial search result is ranked up, obtain final phonetic search result.
2. the method for phonetic search as described in claim 1, which is characterized in that the inverted index tool for pre-establishing resource
Body includes:
To all resource numbers to be searched, unique ID of generation resource index；
According to resource content, extract keyword and sentence for retrieving the resource；
It is segmented, is individual Chinese character by the keyword and sentence cutting, obtains set of keywords；
According to the size of the set of keywords, inverted index is established by relative strategy.
3. the method for phonetic search as claimed in claim 2, which is characterized in that described according to the big of the set of keywords
It is small, inverted index, which is established, by relative strategy specifically includes:
Judge the value of the size K of the set of keywords；
As K≤2, the pronunciation of each word in the set of keywords is obtained, obtains pronunciation set { a1, a2 ..., an }, and give
Each pronunciation in pronunciation set establishes corresponding phonetic resource index；
As K > 2, the coding of each word in the set of keywords is obtained, obtains character code set { c1, c2 ..., cn },
And corresponding character resource is established to each character in character code set and is indexed.
4. the method for phonetic search as described in claim 1, which is characterized in that the length according to the core keyword
Degree, using different phonetic search strategies, obtains initial search result；And initial search result is arranged by the temperature of resource
Sequence obtains final phonetic search as a result, specifically including:
The core keyword obtained described in cutting obtains user's search key set；
Judge the value of the size M of user's search key set；
As M≤2, the phonetic resource index is searched for, initial search result is obtained；And by the temperature of resource to initial search knot
Fruit is ranked up, and obtains final phonetic search result；
As M > 2, searches for the character resource index and by the value length of M using different fault-tolerant strategies, obtain initial search
As a result；And initial search result is ranked up by the temperature of resource, obtain final phonetic search result.
5. a kind of device of phonetic search, which is characterized in that described device includes:
Preprocessing module, for pre-establishing the inverted index of resource, while the temperature of more new resources；
Speech data collection module, for acquiring user voice data；
Speech recognition module, for being identified to the user voice data, being converted into text data；
Semantic module obtains core keyword for carrying out semantic analysis to the text data；
Phonetic search and result output module, for the length according to the core keyword, using different phonetic search plans
Slightly, initial search result is obtained；And initial search result is ranked up by the temperature of resource, obtain final phonetic search knot
Fruit.
6. the device of phonetic search as claimed in claim 5, which is characterized in that the preprocessing module further include:
ID generation module is indexed, for unique ID to all resource numbers to be searched, generation resource index；
Crucial words and phrases abstraction module, for according to resource content, extract keyword and sentence for retrieving the resource；
The keyword and sentence cutting are individual Chinese character, obtain set of keywords by word segmentation module for being segmented；
Inverted index establishes module, for the size according to the set of keywords, establishes inverted index by relative strategy.
7. the device of phonetic search as claimed in claim 6, which is characterized in that the inverted index is established module and specifically wrapped
It includes:
The size judgment module of set of keywords, the value of the size K for judging the set of keywords；
Phonetic resource index establishes module, for obtaining the pronunciation of each word in the set of keywords, being read as K≤2
Sound set { a1, a2 ..., an }, and corresponding phonetic resource index is established to each pronunciation in pronunciation set；
Character resource index establishes module, for obtaining the coding of each word in the set of keywords, obtaining word as K > 2
It accords with code set { c1, c2 ..., cn }, and establishes corresponding character resource to each character in character code set and index.
8. the device of phonetic search as claimed in claim 7, which is characterized in that the phonetic search and result output module tool
Body includes:
Cutting core key word modules, the core keyword for obtaining described in cutting obtain user's search key set；
The size judgment module of user's search key set, for judging the size M of user's search key set
Value；
First phonetic search object module, for searching for the phonetic resource index, obtaining initial search result as M≤2；
And initial search result is ranked up by the temperature of resource, obtain final phonetic search result；
Second phonetic search object module, for as M > 2, searching for the character resource index and by the value length of M using not
Same fault-tolerant strategy, obtains initial search result；And initial search result is ranked up by the temperature of resource, obtain final language
Sound search result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811645648.9A CN109727591B (en) | 2018-12-29 | 2018-12-29 | Voice search method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811645648.9A CN109727591B (en) | 2018-12-29 | 2018-12-29 | Voice search method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109727591A true CN109727591A (en) | 2019-05-07 |
| CN109727591B CN109727591B (en) | 2020-08-11 |
Family
ID=66299402
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811645648.9A Active CN109727591B (en) | 2018-12-29 | 2018-12-29 | Voice search method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109727591B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113421565A (en) * | 2021-07-19 | 2021-09-21 | 北京百度网讯科技有限公司 | Search method, search device, electronic equipment and storage medium |
| CN120956833A (en) * | 2025-07-30 | 2025-11-14 | 杭州易小智科技有限公司 | An AI voice customer service platform |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105590626A (en) * | 2015-12-29 | 2016-05-18 | 百度在线网络技术（北京）有限公司 | Continuous speech man-machine interaction method and system |
| JP2017107472A (en) * | 2015-12-11 | 2017-06-15 | 日本電信電話株式会社 | Character string search method and apparatus |
-
2018
- 2018-12-29 CN CN201811645648.9A patent/CN109727591B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2017107472A (en) * | 2015-12-11 | 2017-06-15 | 日本電信電話株式会社 | Character string search method and apparatus |
| CN105590626A (en) * | 2015-12-29 | 2016-05-18 | 百度在线网络技术（北京）有限公司 | Continuous speech man-machine interaction method and system |
Non-Patent Citations (1)
| Title |
|---|
| 苟廷熹: "云环境下个性化推送搜索引擎的设计", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113421565A (en) * | 2021-07-19 | 2021-09-21 | 北京百度网讯科技有限公司 | Search method, search device, electronic equipment and storage medium |
| CN120956833A (en) * | 2025-07-30 | 2025-11-14 | 杭州易小智科技有限公司 | An AI voice customer service platform |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109727591B (en) | 2020-08-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109918680B (en) | Entity identification method and device and computer equipment | |
| CN111708869B (en) | Method and device for man-machine dialogue processing | |
| CN107291783B (en) | Semantic matching method and intelligent equipment | |
| Tur et al. | What is left to be understood in ATIS? | |
| CN108170859B (en) | Voice query method, device, storage medium and terminal equipment | |
| CN106649825B (en) | Voice interaction system and creation method and device thereof | |
| CN105718586B (en) | The method and device of participle | |
| CN111046656B (en) | Text processing method, text processing device, electronic equipment and readable storage medium | |
| CN111209363B (en) | Corpus data processing method, corpus data processing device, server and storage medium | |
| CN110147445A (en) | Intension recognizing method, device, equipment and storage medium based on text classification | |
| CN108304372A (en) | Entity extraction method and apparatus, computer equipment and storage medium | |
| CN109165386A (en) | A kind of Chinese empty anaphora resolution method and system | |
| CN101566998A (en) | Chinese question-answering system based on neural network | |
| CN107704453A (en) | A kind of word semantic analysis, word semantic analysis terminal and storage medium | |
| CN111858888B (en) | Multi-round dialogue system of check-in scene | |
| WO2022141875A1 (en) | User intention recognition method and apparatus, device, and computer-readable storage medium | |
| CN108388553B (en) | Method for eliminating ambiguity in conversation, electronic equipment and kitchen-oriented conversation system | |
| CN113157887B (en) | Knowledge question and answer intention recognition method and device and computer equipment | |
| CN113326702A (en) | Semantic recognition method and device, electronic equipment and storage medium | |
| CN111046653A (en) | Sentence recognition method, sentence recognition device and intelligent equipment | |
| CN106294460A (en) | A kind of Chinese speech keyword retrieval method based on word and word Hybrid language model | |
| CN114116975A (en) | Multi-intention identification method and system | |
| JP2006244262A (en) | Question answer search system, method and program | |
| CN111680514B (en) | Information processing and model training method, device, equipment and storage medium | |
| TW202034207A (en) | Dialogue system using intention detection ensemble learning and method thereof |
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 |