CN105955993A - Method and device for sequencing search results - Google Patents
Method and device for sequencing search results Download PDFInfo
- Publication number
- CN105955993A CN105955993A CN201610245408.4A CN201610245408A CN105955993A CN 105955993 A CN105955993 A CN 105955993A CN 201610245408 A CN201610245408 A CN 201610245408A CN 105955993 A CN105955993 A CN 105955993A
- Authority
- CN
- China
- Prior art keywords
- word
- search results
- collocation
- query statement
- term vector
- 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
Classifications
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic 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)
Abstract
The invention discloses a method and device for sequencing search results. The method for sequencing the search results comprises the steps that an input query statement is received, and the query statement is segmented into multiple words; matching corresponding to the words is acquired; the multiple searching results corresponding to the query statement are acquired; and the multiple search results are sequenced based on a BOW model according to the words and the matching corresponding to the words. The method and device for sequencing the search results disclosed by the embodiment of the invention are characterized in that the input query statement is received; the query statement is segmented into the multiple words; the matching corresponding to the words is acquired; the multiple search results corresponding to the query statement are acquired; and the multiple search results are sequenced based on the BOW model according to the words and the matching corresponding to the words. In this way, sequencing of the search results is optimized; the search results which can better satisfy a user's intent can be provided preferentially; and the user's use experience is improved.
Description
Technical field
The present invention relates to Internet technical field, particularly relate to a kind of search result ordering method and device.
Background technology
Along with the high speed development of science and technology, the Internet has been deep in daily life.User can be looked into by search engine
Ask required information.And how Search Results is ranked up, the information that user needs is come Search Results before
Face, is the problem of current primary study.
At present, mainly Search Results has been assisted by the semantic similarity before the title of query statement and Search Results
Sequence work.Such as: calculate the term vector of each word of query statement, after addition, obtain the term vector of query statement.With
Reason, calculates the term vector of the title of Search Results.Finally, then calculate the semantic similarity between two term vectors, thus right
Search Results is ranked up.
But, said method does not takes into account the structural information of query statement, may cause final ranking results deviation,
Such as " Wu Song beats a dead tiger " is identical with the term vector of " tiger kills Wu Song " the two query statement, reduces sequence knot
The accuracy of fruit.
Summary of the invention
It is contemplated that one of technical problem solved the most to a certain extent in correlation technique.To this end, the one of the present invention
Purpose is to propose a kind of search result ordering method, it is possible to the sequence to Search Results is optimized, and preferential offer more accords with
Close the Search Results of user view, promote user's experience.
Second object of the present invention is to propose a kind of search results ranking device.
To achieve these goals, first aspect present invention embodiment proposes a kind of search result ordering method, including: connect
Receive the query statement of input, and be multiple word by query statement cutting;Obtain the collocation corresponding with word;Obtain inquiry language
Multiple Search Results that sentence is corresponding;And according to collocation corresponding to word and word, based on BOW model to multiple Search Results
It is ranked up.
The search result ordering method of the embodiment of the present invention, the query statement inputted by reception, and by query statement cutting be
Multiple words, obtain the collocation corresponding with word, and obtain multiple Search Results that query statement is corresponding, and according to word
And the collocation that word is corresponding, based on BOW model, multiple Search Results being ranked up, the sequence to Search Results is optimized,
The preferential Search Results more conforming to user view that provides, lifting user's experience.
Second aspect present invention embodiment proposes a kind of search results ranking device, including cutting module, is used for receiving defeated
The query statement entered, and be multiple word by query statement cutting;First acquisition module, for obtaining take corresponding with word
Join;Second acquisition module, for obtaining multiple Search Results that query statement is corresponding;And order module, for according to list
Multiple Search Results are ranked up by word and collocation corresponding to word based on BOW model.
The search results ranking device of the embodiment of the present invention, the query statement inputted by reception, and by query statement cutting be
Multiple words, obtain the collocation corresponding with word, and obtain multiple Search Results that query statement is corresponding, and according to word
And the collocation that word is corresponding, based on BOW model, multiple Search Results being ranked up, the sequence to Search Results is optimized,
The preferential Search Results more conforming to user view that provides, lifting user's experience.
Accompanying drawing explanation
Fig. 1 is the flow chart of search result ordering method according to an embodiment of the invention;
Fig. 2 is the effect schematic diagram obtaining candidate collocation;
Fig. 3 is the flow chart being ranked up multiple Search Results based on BOW model according to an embodiment of the invention;
Fig. 4 is the structural representation one of search results ranking device according to an embodiment of the invention;
Fig. 5 is the structural representation two of search results ranking device according to an embodiment of the invention.
Detailed description of the invention
Embodiments of the invention are described below in detail, and the example of described embodiment is shown in the drawings, the most identical
Or similar label represents same or similar element or has the element of same or like function.Retouch below with reference to accompanying drawing
The embodiment stated is exemplary, it is intended to is used for explaining the present invention, and is not considered as limiting the invention.
Below with reference to the accompanying drawings search result ordering method and the device of the embodiment of the present invention are described.
Fig. 1 is the flow chart of search result ordering method according to an embodiment of the invention.
As it is shown in figure 1, search result ordering method comprises the steps that
S1, the query statement of reception input, and be multiple word by query statement cutting.
For example, query statement is for " how on-vehicle Bluetooth is from being dynamically connected?", can be " car by this query statement cutting
Carry ", " bluetooth ", " how ", " automatically ", " connection " and "?”.
S2, obtain the collocation corresponding with word.
Specifically, can first obtain current word and with current word word in default window length, then by current word and with
Current word combinations of words in default window length is multiple candidate collocation, be finally based on default dictionary of collocations filter out with currently
The collocation that word is corresponding.
In continuation, example is described, it is assumed that current word is " vehicle-mounted ", presets window a length of 5.So it is in same with " vehicle-mounted "
Word in one window length can include " bluetooth ", " how ", " automatically " and " connection ", the then time that " vehicle-mounted " is corresponding
Choosing collocation is " on-vehicle Bluetooth ", " how vehicle-mounted ", " vehicle-mounted automatically ", " vehicle-mounted connection ".In like manner, such as Fig. 2 institute
Show, the word being in same window length with " bluetooth " can include " how ", " automatically ", " connection " and "?",
The candidate collocation that then " bluetooth " is corresponding is " bluetooth how ", " bluetooth is automatic ", " bluetooth connection ", " bluetooth?”.
By that analogy, each word in scan for inquiries statement, obtain the candidate collocation of its correspondence.
After this, can filter out based on default dictionary of collocations and reasonably arrange in pairs or groups.Wherein, reasonably collocation can be co-occurrence
The collocation that frequency is high, it is also possible to be the collocation that log-likelihood ratio score is high.Such as: " how vehicle-mounted ", " vehicle-mounted automatically ",
" vehicle-mounted connection ", " bluetooth how ", " bluetooth is automatic ", " bluetooth?" these candidate collocation, it is common appearance
The collocation that frequency is relatively low, or the collocation that log-likelihood ratio score is relatively low, therefore may filter that above-mentioned candidate collocation.Finally obtain
Obtain " on-vehicle Bluetooth ", " bluetooth connection " collocation that such reasonability is higher.
S3, multiple Search Results that acquisition query statement is corresponding.
Such as query statement is for " how on-vehicle Bluetooth is from being dynamically connected?", then can retrieve corresponding therewith based on this query statement
Multiple Search Results.
S4, according to collocation corresponding to word and word, based on BOW model, multiple Search Results are ranked up.
After obtaining word and the collocation of correspondence, and multiple Search Results, can based on BOW (Bag Of Words,
Word bag) multiple Search Results are ranked up by model.
Specifically, as it is shown on figure 3, the process of sequence can be divided into following step:
The term vector sum of the collocation that S41, the term vector calculating word and word are corresponding, as the term vector of query statement.
For example, query statement can include that word 1, word 2 and word 3, the collocation of word 1 correspondence are collocation 1, take
Joining 2, the collocation of word 2 correspondence is collocation 3, collocation 4 and collocation 5, and the collocation of word 3 correspondence is collocation 6, can distinguish
Calculate word 1, word 2, word 3, collocation 1, collocation 2, collocation 3, collocation 4, collocation 5 and the term vector of collocation 6,
Above-mentioned term vector is added, the final term vector obtaining query statement.When calculating the term vector of query statement, will take
The term vector joined adds to input feature vector, adds the feature of syntactic structure, it is possible to effectively make up only by the term vector of word
As the shortcoming of input feature vector, improve semantic space indexing.
S42, calculate the term vector of multiple Search Results.
In the present embodiment, the term vector of Search Results is actual is the term vector of the title that Search Results is corresponding.Search Results is corresponding
Title be all multiple word composition, therefore, the method identical with the term vector calculating query statement can be taked, calculate search
The term vector of result.
The semantic similarity of the term vector of S43, the term vector of calculating query statement and multiple Search Results.
After the term vector of the term vector and Search Results that calculate query statement, the term vector of query statement can be calculated respectively
Semantic similarity with the term vector of each Search Results.
S44, according to semantic similarity, multiple Search Results are ranked up.
In the present embodiment, according to semantic similarity order from high to low, Search Results can be ranked up so that with inquiry
The higher search result rank of statement semantics similarity is located further forward.
It is described in detail below, sets up the detailed process presetting dictionary of collocations.
First can obtain corpus, and extract the candidate word in corpus to language material, then can be based on co-occurrence frequency or right
Number likelihood ratio scores to filtering out word language material to language material from candidate word, are finally based on word and set up language material and preset dictionary of collocations.
As an example it is assumed that corpus is " automobile Bluetooth is from being dynamically connected ", each word can be regarded as a node, analyze every two
Grammer dependence whether is there is between individual node.Using there is grammer dependence two words as a candidate word to language material.
Wherein, corpus may exist such as " ", " it ", " " such function word node, may filter that void
Word node.Choose candidate word to language material after, language material can be screened by candidate word.Screening technique can be divided into two kinds.
First method is to be screened by co-occurrence frequency, the number of times that i.e. two words of language material are occurred by composition candidate word jointly.Sieve
Select the co-occurrence frequency candidate word more than preset times to language material.Second method is to be screened by log-likelihood ratio score,
I.e. filter out the candidate word of front N name of log-likelihood ratio highest scoring to language material.Wherein, log-likelihood ratio score is mainly passed through
Collocation frequency, word frequency, corpus size etc. calculate and obtain.Collocation frequency refers to form candidate word and takes two words of language material
The number of times joined;Word frequency is the number of times that two words of language material are represented respectively by composition candidate word for corpus;Corpus size is
The number of the word comprised.Finally, according to the word obtained after screening, dictionary of collocations is preset in language material foundation.
The search result ordering method of the embodiment of the present invention, the query statement inputted by reception, and by query statement cutting be
Multiple words, obtain the collocation corresponding with word, and obtain multiple Search Results that query statement is corresponding, and according to word
And the collocation that word is corresponding, based on BOW model, multiple Search Results being ranked up, the sequence to Search Results carries out excellent
Change, the preferential Search Results more conforming to user view that provides, lifting user's experience.
For achieving the above object, the present invention also proposes a kind of search results ranking device.
Fig. 4 is the structural representation one of search results ranking device according to an embodiment of the invention.
As shown in Figure 4, search results ranking device comprises the steps that cutting module the 110, first acquisition module 120, second obtains
Module 130 and order module 140.Wherein, the first acquisition module 120 can include acquiring unit 121, assembled unit 122 and
Screening unit 123.
Cutting module 110 is for receiving the query statement of input, and is multiple word by query statement cutting.For example,
Query statement is for " how on-vehicle Bluetooth is from being dynamically connected?", can by this query statement cutting be " vehicle-mounted ", " bluetooth ",
" how ", " automatically ", " connection " and "?”.
First acquisition module 120 is for obtaining the collocation corresponding with word.Specifically, acquiring unit 121 can first obtain currently
Word and with current word word in default window length, then assembled unit 122 by current word and with current word in advance
If the combinations of words in window length is multiple candidate collocation, last screening unit 123 based on default dictionary of collocations filter out with currently
The collocation that word is corresponding.In continuation, example is described, it is assumed that current word is " vehicle-mounted ", presets window a length of 5.So with
The word that " vehicle-mounted " is in same window length can include " bluetooth ", " how ", " automatically " and " connection ", then " car
Carrying " corresponding candidate collocation be " on-vehicle Bluetooth ", " how vehicle-mounted ", " vehicle-mounted automatic ", " vehicle-mounted connection ".With
Reason, as in figure 2 it is shown, the word being in same window length with " bluetooth " can include " how ", " automatically ", " connection "
"?", then the candidate collocation that " bluetooth " is corresponding be " bluetooth how ", " bluetooth is automatic ", " bluetooth connection ",
" bluetooth?”.By that analogy, each word in scan for inquiries statement, obtain the candidate collocation of its correspondence.This it
After, can filter out based on default dictionary of collocations and reasonably arrange in pairs or groups.Wherein, reasonably collocation can be high the taking of co-occurrence frequency
Join, it is also possible to be the collocation that log-likelihood ratio score is high.Such as: " how vehicle-mounted ", " vehicle-mounted automatically ", " vehicle-mounted company
Connect ", " bluetooth how ", " bluetooth is automatic ", " bluetooth?" these candidate collocation, it is the common frequency of occurrences relatively low
Collocation, or the collocation that log-likelihood ratio score is relatively low, therefore may filter that above-mentioned candidate collocation.Final acquisition is " vehicle-mounted
Bluetooth ", " bluetooth connection " collocation that such reasonability is higher.
Second acquisition module 130 is for obtaining multiple Search Results that query statement is corresponding.Such as query statement is " vehicle-mounted indigo plant
How tooth is from being dynamically connected?", then the second acquisition module 130 can retrieve corresponding multiple search based on this query statement
Result.
Multiple Search Results, for the collocation corresponding according to word and word, are carried out by order module 140 based on BOW model
Sequence.Specifically, order module 140 can first calculate the term vector of word and the term vector sum of collocation corresponding to word, will
It is as the term vector of query statement.For example, query statement can include word 1, word 2 and word 3, and word 1 is right
The collocation answered is collocation 1, collocation 2, and the collocation of word 2 correspondence is collocation 3, collocation 4 and collocation 5, word 3 correspondence
Collocation is collocation 6, can calculate word 1, word 2, word 3, collocation 1, collocation 2, collocation 3, collocation 4 respectively, take
Join the term vector of 5 and collocation 6, above-mentioned term vector is added, the final term vector obtaining query statement.Look in calculating
When asking the term vector of statement, the term vector of collocation is added to input feature vector, adds the feature of syntactic structure, it is possible to have
Effect make up only using the term vector of word as the shortcoming of input feature vector, improve semantic space and index.
Then, order module 140 calculates the term vector of multiple Search Results.In the present embodiment, the term vector of Search Results is real
Border is the term vector of the title that Search Results is corresponding.The title that Search Results is corresponding is all multiple word composition, therefore, can adopt
The method that the term vector that takes and calculate query statement is identical, calculates the term vector of Search Results.
Afterwards, order module 140 calculates the semantic similarity of the term vector of query statement and the term vector of multiple Search Results again.
Finally, multiple Search Results can be ranked up by order module 140 according to semantic similarity.Such as, can be according to semanteme
Search Results is ranked up by similarity order from high to low, so that search knot higher with query statement semantic similarity
Really ranking is located further forward.
Additionally, as it is shown in figure 5, the first acquisition module 120 may also include and sets up module 124.
Specifically, set up module 124 and can first obtain corpus, and extract the candidate word in corpus to language material, then
Can based on co-occurrence frequency or log-likelihood ratio score from candidate word to language material filtering out word to language material, be finally based on word to language material
Set up and preset dictionary of collocations.As an example it is assumed that corpus is " automobile Bluetooth is from being dynamically connected ", each word can be regarded as
One node, analyzes whether there is grammer dependence between each two node.Two words with grammer dependence are made
It is that a candidate word is to language material.Wherein, may exist in corpus as such in " ", " it ", " "
Function word node, may filter that function word node.Choose candidate word to language material after, language material can be screened by candidate word.
Screening technique can be divided into two kinds.First method is to be screened by co-occurrence frequency, i.e. forms two to language material of candidate word
The number of times that word occurs jointly.Filter out the co-occurrence frequency candidate word more than preset times to language material.Second method is by right
Number likelihood ratio score is screened, and i.e. filters out the candidate word of front N name of log-likelihood ratio highest scoring to language material.Wherein,
Log-likelihood ratio score is mainly calculated by collocation frequency, word frequency, corpus size etc. and obtains.Collocation frequency refers to composition
The candidate word number of times to two word collocation of language material;Word frequency is that two words of language material are opened up respectively by composition candidate word for corpus
Existing number of times;Corpus size is the number of the word comprised.Finally, take language material foundation is default according to the word obtained after screening
Join dictionary.
The search results ranking device of the embodiment of the present invention, the query statement inputted by reception, and by query statement cutting be
Multiple words, obtain the collocation corresponding with word, and obtain multiple Search Results that query statement is corresponding, and according to word
And the collocation that word is corresponding, based on BOW model, multiple Search Results being ranked up, the sequence to Search Results carries out excellent
Change, the preferential Search Results more conforming to user view that provides, lifting user's experience.
Additionally, term " first ", " second " are only used for describing purpose, and it is not intended that instruction or imply relatively important
Property or the implicit quantity indicating indicated technical characteristic.Thus, define " first ", " second " feature permissible
Express or implicitly include at least one this feature.In describing the invention, " multiple " are meant that at least two,
Such as two, three etc., unless otherwise expressly limited specifically.
In the description of this specification, reference term " embodiment ", " some embodiments ", " example ", " concrete example ",
Or specific features, structure, material or the feature bag that the description of " some examples " etc. means to combine this embodiment or example describes
It is contained at least one embodiment or the example of the present invention.In this manual, to the schematic representation of above-mentioned term necessarily
It is directed to identical embodiment or example.And, the specific features of description, structure, material or feature can be arbitrary
Individual or multiple embodiment or example combine in an appropriate manner.Additionally, in the case of the most conflicting, the skill of this area
The feature of the different embodiments described in this specification or example and different embodiment or example can be combined by art personnel
And combination.
Although above it has been shown and described that embodiments of the invention, it is to be understood that above-described embodiment is exemplary,
Being not considered as limiting the invention, those of ordinary skill in the art within the scope of the invention can be to above-described embodiment
It is changed, revises, replaces and modification.
Claims (10)
1. a search result ordering method, it is characterised in that comprise the following steps:
Receive the query statement of input, and be multiple word by described query statement cutting;
Obtain the collocation corresponding with described word;
Obtain multiple Search Results that described query statement is corresponding;And
According to the collocation that described word and described word are corresponding, based on word bag BOW model, the plurality of Search Results is carried out
Sequence.
2. the method for claim 1, it is characterised in that obtain the collocation corresponding with described word, including:
Obtain current word and with current word word in default window length;
It is multiple candidate collocation by described current word with current word combinations of words in default window length;
The collocation corresponding with described current word is filtered out based on default dictionary of collocations.
3. method as claimed in claim 2, it is characterised in that filtering out and described current list based on default dictionary of collocations
Before the collocation that word is corresponding, also include:
Set up and preset dictionary of collocations.
4. method as claimed in claim 3, it is characterised in that set up and preset dictionary of collocations, including:
Obtain corpus, and extract the candidate word in described corpus to language material;
Based on co-occurrence frequency or log-likelihood ratio score from described candidate word to language material filtering out word to language material, and based on described
Language material is set up and is preset dictionary of collocations by word.
5. the method for claim 1, it is characterised in that according to the collocation that described word and described word are corresponding, base
In BOW model, the plurality of Search Results is ranked up, including:
Calculate the term vector of described word and the term vector sum of collocation corresponding to described word, as the word of described query statement
Vector;
Calculate the term vector of the plurality of Search Results;
Calculate the semantic similarity of the term vector of described query statement and the term vector of the plurality of Search Results;
According to described semantic similarity, the plurality of Search Results is ranked up.
6. a search results ranking device, it is characterised in that including:
Cutting module, for receiving the query statement of input, and is multiple word by described query statement cutting;
First acquisition module, for obtaining the collocation corresponding with described word;
Second acquisition module, for obtaining multiple Search Results that described query statement is corresponding;And
Order module, for the collocation corresponding according to described word and described word, searches the plurality of based on BOW model
Fruit is ranked up hitch.
7. device as claimed in claim 6, it is characterised in that described first acquisition module, including:
Acquiring unit, for obtain current word and with current word word in default window length;
Assembled unit, for being that multiple candidate takes by described current word with current word combinations of words in default window length
Join;
Screening unit, for filtering out the collocation corresponding with described current word based on default dictionary of collocations.
8. device as claimed in claim 7, it is characterised in that described first acquisition module, also includes:
Set up unit, for before filtering out the collocation corresponding with described current word based on default dictionary of collocations, set up pre-
If dictionary of collocations.
9. device as claimed in claim 8, it is characterised in that described set up unit, is used for:
Obtain corpus, and extract the candidate word in described corpus to language material;
Based on co-occurrence frequency or log-likelihood ratio score from described candidate word to language material filtering out word to language material, and based on described
Language material is set up and is preset dictionary of collocations by word.
10. device as claimed in claim 6, it is characterised in that described order module, is used for:
Calculate the term vector of described word and the term vector sum of collocation corresponding to described word, as the word of described query statement
Vector;
Calculate the term vector of the plurality of Search Results;
Calculate the semantic similarity of the term vector of described query statement and the term vector of the plurality of Search Results;
According to described semantic similarity, the plurality of Search Results is ranked up.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610245408.4A CN105955993B (en) | 2016-04-19 | 2016-04-19 | Search result ordering method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610245408.4A CN105955993B (en) | 2016-04-19 | 2016-04-19 | Search result ordering method and device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105955993A true CN105955993A (en) | 2016-09-21 |
CN105955993B CN105955993B (en) | 2020-09-25 |
Family
ID=56917662
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610245408.4A Active CN105955993B (en) | 2016-04-19 | 2016-04-19 | Search result ordering method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105955993B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110019665A (en) * | 2017-09-30 | 2019-07-16 | 北京国双科技有限公司 | Text searching method and device |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100057725A1 (en) * | 2008-08-26 | 2010-03-04 | Norikazu Matsumura | Information retrieval device, information retrieval method, and program |
CN101840397A (en) * | 2009-03-20 | 2010-09-22 | 日电(中国)有限公司 | Word sense disambiguation method and system |
CN102214189A (en) * | 2010-04-09 | 2011-10-12 | 腾讯科技(深圳)有限公司 | Data mining-based word usage knowledge acquisition system and method |
CN103390004A (en) * | 2012-05-11 | 2013-11-13 | 北京百度网讯科技有限公司 | Determination method and determination device for semantic redundancy and corresponding search method and device |
CN105117487A (en) * | 2015-09-19 | 2015-12-02 | 杭州电子科技大学 | Book semantic retrieval method based on content structures |
CN105279252A (en) * | 2015-10-12 | 2016-01-27 | 广州神马移动信息科技有限公司 | Related word mining method, search method and search system |
-
2016
- 2016-04-19 CN CN201610245408.4A patent/CN105955993B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100057725A1 (en) * | 2008-08-26 | 2010-03-04 | Norikazu Matsumura | Information retrieval device, information retrieval method, and program |
CN101840397A (en) * | 2009-03-20 | 2010-09-22 | 日电(中国)有限公司 | Word sense disambiguation method and system |
CN102214189A (en) * | 2010-04-09 | 2011-10-12 | 腾讯科技(深圳)有限公司 | Data mining-based word usage knowledge acquisition system and method |
CN103390004A (en) * | 2012-05-11 | 2013-11-13 | 北京百度网讯科技有限公司 | Determination method and determination device for semantic redundancy and corresponding search method and device |
CN105117487A (en) * | 2015-09-19 | 2015-12-02 | 杭州电子科技大学 | Book semantic retrieval method based on content structures |
CN105279252A (en) * | 2015-10-12 | 2016-01-27 | 广州神马移动信息科技有限公司 | Related word mining method, search method and search system |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110019665A (en) * | 2017-09-30 | 2019-07-16 | 北京国双科技有限公司 | Text searching method and device |
Also Published As
Publication number | Publication date |
---|---|
CN105955993B (en) | 2020-09-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103365925B (en) | Obtain polyphone phonetic, method based on phonetic retrieval and related device thereof | |
CN103678576B (en) | The text retrieval system analyzed based on dynamic semantics | |
CN103886063A (en) | Text retrieval method and device | |
US20180246955A1 (en) | Method and device for searching legal provision | |
US20150269175A1 (en) | Query Interpretation and Suggestion Generation under Various Constraints | |
CN106776564B (en) | Semantic recognition method and system based on knowledge graph | |
CN105843850A (en) | Searching optimization method and device | |
US20090112903A1 (en) | Ontology data import/export method and apparatus | |
CN111949758A (en) | Medical question and answer recommendation method, recommendation system and computer readable storage medium | |
CN102456054B (en) | A kind of searching method and system | |
WO2013170587A1 (en) | Multimedia question and answer system and method | |
JPH07219969A (en) | Device and method for retrieving picture parts | |
CN102999625A (en) | Method for realizing semantic extension on retrieval request | |
CN106202294A (en) | The related news computational methods merged based on key word and topic model and device | |
CN105930362A (en) | Search target recognition method and device, and terminal | |
CN111104488B (en) | Method, device and storage medium for integrating retrieval and similarity analysis | |
CN105989058A (en) | Chinese news brief generating system and method | |
CN106294845A (en) | The many emotions sorting technique extracted based on weight study and multiple features and device | |
CN110970112B (en) | Knowledge graph construction method and system for nutrition and health | |
CN102314464B (en) | Lyrics searching method and lyrics searching engine | |
CN110659310A (en) | Intelligent search method for vehicle information | |
CN102999489A (en) | Method and system for image search of community website page | |
US10114891B2 (en) | Method and system of audio retrieval and source separation | |
KR20130022075A (en) | Method for building emotional lexical information and apparatus for the same | |
CN105955993A (en) | Method and device for sequencing search results |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |