CN101986293B - For presenting the method and apparatus of search answer information in search interface - Google Patents

For presenting the method and apparatus of search answer information in search interface Download PDF

Info

Publication number
CN101986293B
CN101986293B CN201010271796.6A CN201010271796A CN101986293B CN 101986293 B CN101986293 B CN 101986293B CN 201010271796 A CN201010271796 A CN 201010271796A CN 101986293 B CN101986293 B CN 101986293B
Authority
CN
China
Prior art keywords
answer
question
information
search
entity
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201010271796.6A
Other languages
Chinese (zh)
Other versions
CN101986293A (en
Inventor
戴帅湘
徐犇
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Baidu Netcom Science and Technology Co Ltd
Original Assignee
Beijing Baidu Netcom Science and Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Beijing Baidu Netcom Science and Technology Co Ltd filed Critical Beijing Baidu Netcom Science and Technology Co Ltd
Priority to CN201010271796.6A priority Critical patent/CN101986293B/en
Publication of CN101986293A publication Critical patent/CN101986293A/en
Application granted granted Critical
Publication of CN101986293B publication Critical patent/CN101986293B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Landscapes

  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The invention provides a kind of method and apparatus for presenting search answer information in search interface, the present invention is by obtaining the search information from user, by described search information with each question and answer prestored to mating, obtain one or more question and answer pair that all or part of content with described search information matches, and according to the one or more question and answer pair, provide a user with the answer information corresponding with this search information.The invention have the advantages that 1) directly can provide a user with definitiveness answer based on this search information in search interface according to the search information of user's input, and Search Results based on this search information can be provided a user with simultaneously;2) by the analysis to web page contents and encyclopaedic knowledge, generation question and answer, to storehouse, provide a user with comprehensive, real-time answer;3) by the relevant information of user and/or the most mutual with user, improve answer accuracy, give user more preferable individualized experience.

Description

For presenting the method and apparatus of search answer information in search interface
Technical field
The present invention relates to computer realm, answer particularly for presenting search in search interface The method, apparatus and system of case information.
Background technology
In prior art, if user intentionally gets some information, often at search engine or special The information bank of door scans for.Wherein, the information of definitiveness answer is likely to be obtained for some, Search engine provides a user with link according only to the search information of user, and user also needs to from link Middle selection also searches answer;And special information bank mostly is the special storehouse in terms of certain, it is difficult to User provides comprehensive information, and while providing a user with answer information, it is impossible to user Other search content is provided.
Therefore, how the answer of certain problem can be provided to user simultaneously all sidedly, again can Search information according to user provides Search Results, it has also become those skilled in the art need to solve Problem.
Summary of the invention
It is an object of the invention to provide a kind of for presenting search answer information in search interface Method, apparatus and system.
According to an aspect of the present invention, it is provided that one is answered for presenting search in search interface The method of case information, the method comprises the following steps:
A obtains the search information from user;
B by described search information with each question and answer prestored to mating, it is thus achieved that search with described One or more question and answer pair that all or part of content of rope information matches;
C, according to the one or more question and answer pair, provides a user with corresponding with this search information Answer information.
According to a further aspect in the invention, additionally provide one to search for presenting in search interface The equipment of rope answer information, wherein, this equipment includes:
First acquisition device, for obtaining from the search information of user;
Coalignment, for by described search information with each question and answer prestored to mating, Obtain one or more question and answer pair that all or part of content with described search information matches;
Device is provided, is used for according to the one or more question and answer pair, provide a user with and search with this The answer information that rope information is corresponding.
Compared with prior art, the invention have the advantages that 1) can input according to user Search information, in search interface, directly provide a user with definitiveness based on this search information Answer, and, additionally it is possible to provide a user with Search Results based on this search information simultaneously;2) By the analysis to web page contents and encyclopaedic knowledge, generation question and answer, to storehouse, provide a user with complete Face, real-time answer;3) by the relevant information of user and/or hand over the further of user Mutually, improve the accuracy of answer, give user more preferable individualized experience.
Accompanying drawing explanation
The detailed description that non-limiting example is made made with reference to the following drawings by reading, The other features, objects and advantages of the present invention will become more apparent upon:
Fig. 1 is the side for presenting search answer information in search interface of one aspect of the invention Method flow chart;
Fig. 2 be a preferred embodiment of the invention for present in search interface search answer letter The method flow diagram of breath;
Fig. 3 be another preferred embodiment of the present invention for present in search interface search answer letter The method flow diagram of breath;
Fig. 4 is that the network equipment generates the method flow of question and answer pair according to the question and answer content from webpage Figure;
Fig. 5 is the network equipment method flow according to the data genaration question and answer pair from encyclopaedia webpage Figure.
Fig. 6 is the system presenting search answer information in search interface of one aspect of the invention Structural representation;
Fig. 7 is the structural representation of the coalignment of a preferred embodiment of the invention;
Fig. 8 is the structural representation of the coalignment of another preferred embodiment of the present invention;
Fig. 9 is the network equipment according to from the question and answer content of webpage and the data genaration of encyclopaedia webpage The apparatus structure schematic diagram of question and answer pair.
In accompanying drawing, same or analogous reference represents same or analogous parts.
Detailed description of the invention
Below in conjunction with the accompanying drawings the present invention is described in further detail.
Fig. 1 illustrates and presents search answer information in search interface according to one aspect of the invention Method flow diagram.It illustrates that the network equipment 2 passes through subscriber equipment via Network Capture from user The search information of 1 input, by search information with question and answer to mating, obtains answer information and is searching Rope interface is presented to the process of user.
Wherein, network includes but not limited to the Internet, wide area network, Metropolitan Area Network (MAN), LAN, VPN Network, wireless self-organization network (Ad Hoc network) etc..Subscriber equipment 1 includes but does not limits Carry out man-machine with user by keyboard, remote controller, touch pad or voice-operated device in any one Mutual electronic product, such as computer, smart mobile phone, PDA, game machine or IPTV Deng.The network equipment 2 includes but not limited to that single network server, multiple webserver form Server group or based on cloud computing (Cloud Computing) by a large amount of computers or network The cloud that server is constituted, wherein, cloud computing is the one of Distributed Calculation, by the loose coupling of a group One super virtual machine of the computer collection composition closed.Wherein, the network equipment 2 is preserved Question and answer are to storehouse and lexical types storehouse.Question and answer centering comprises problem and the corresponding answer of this problem, and Question and answer are the set comprising a large amount of question and answer pair to storehouse.Lexical types storehouse have recorded vocabulary or vocabulary The corresponding relation of type that may describe of the combination of combination and this vocabulary or vocabulary.
When user is desired with search, the input mode input search provided by subscriber equipment Information.Wherein, this input mode includes but not limited to: 1) word input;2) phonetic entry; 3) handwriting input.Wherein, in aforesaid way, the position of input search information includes but does not limits In: 1) search column of the page that provides of search engine;2) searched page that client provides; 3) search column etc. in embedded web page or client.
Specifically, in step s101, subscriber equipment 1 is carried out with user by any one The interactive device of man-machine interaction obtains the search information of user's input.This interactive device can be Keyboard, remote controller, touch pad or voice-operated device etc..Then, in step s102, user The search information that described user is inputted by equipment 1 sends to the network equipment 2.In step s103, The network equipment 2 obtains above-mentioned search information.
Then, in step s104, the network equipment 2 is by the search information received and this locality Question and answer each question and answer to prestoring in storehouse, to mating, are i.e. searched in each question and answer centering and are searched Same or analogous lexical information in rope information, to obtain the whole or portion with described search information Point one or more question and answer pair that content matches.
Specifically, the problem of search information with question and answer centering is mated by the network equipment 2, root According to search information, may occur in which the problem of one or more question and answer pair and the feelings of this search information matches Condition.
When search information only implies single problem, the network equipment 2 can obtain this search information In the result that can match with the problem of one or more question and answer centerings of full content.
Such as, when search information is " birthplace of Mao Zedong is at which ", and above-mentioned search information is hidden Containing single problem.This search information is mated in question and answer are to storehouse by the network equipment 2, obtains The problem of one question and answer pair is " Mao Zedong birthplace ", the problem of these question and answer pair and search information Full content match, then the network equipment 2 obtains question and answer to " Mao Zedong birthplace, lake Shaoshan, south ".
When searching for the multiple problem of embodying information, the network equipment 2 can mate and obtains multiple question and answer pair, The problem of each question and answer pair matches with the partial content in this search information.
Such as, when search information is " date of birth of Mao Zedong and place ", and above-mentioned search is believed Implicit two problems of breath.This search information is mated in question and answer are to storehouse by the network equipment 2, Obtain question and answer to " Mao Zedong birthplace, Shaoshan, Hunan " and question and answer to " Mao Zedong's date of birth, On December 26th, 1893 ".Wherein, the problem of previous question and answer pair is " Mao Zedong birthplace ", This problem matches with " Mao Zedong place of birth " in search information, and later question and answer are to asking Entitled " Mao Zedong's date of birth ", this problem and " Mao Zedong's date of birth " in search information Match.
When there are the question and answer pair that relative search information more refines during question and answer are to storehouse, the network equipment 2 can match multiple question and answer pair.
Such as, when search information is " No. two line last bus moment of subway ", due to each city No. two line final vehicle hours of subway are different, and question and answer are to often there being each different cities in storehouse The question and answer pair of No. two line last buses of subway.The network equipment 2 is by above-mentioned search information and each question and answer After mating, it is thus achieved that multiple question and answer pair: " No. two line last bus moment of Shanghai Underground, 23:00 ", " No. two line last bus moment of Beijing Metro, 23:15 ", " No. two line ends of Guangzhou Underground The regular bus moment, 23:30 " etc., the problem of above-mentioned question and answer pair all can be with the full content of search information Coupling.It should be noted that above-mentioned example is only better described the solution of the present invention, but this Invention is not limited thereto, and it should be appreciated by those skilled in the art that any according to search information, Obtain the scheme of one or more question and answer pair, should be included in the scope of the present invention.
In step s105, the network equipment 2 according to the question and answer obtained in step s104 to generation Answer information.
When the question and answer with search information matches are to only one, the network equipment 2 obtains above-mentioned asking Answer in answering questions, in conjunction with the problem of these question and answer pair as answer information.
When the question and answer with search information matches are to when having multiple, it may be judged whether can be by each question and answer To answer integrate.If each question and answer above-mentioned are to semantically related, can integrate, Each question and answer above-mentioned are integrated by the network equipment 2 to by the mode meeting natural language custom, and raw Become answer information.Such as, for search information " date of birth of Mao Zedong and place " phase To " Mao Zedong's date of birth, on December 26th, 1893 " and " Mao Zedong goes out the question and answer of coupling The Radix Rehmanniae, Shaoshan, Hunan ", the main body of its problem is identical, and the network equipment 2 can be integrated generation Article one, " Mao Zedong is born in December in 1893 26, lake for semantically coherent answer information Shaoshan, south ";If each question and answer are to integrating, the network equipment 2 is the most respectively according to question and answer pair Answer and the problem of these question and answer pair, generate multiple answer unit as answer information, such as, and will The question and answer that search information " No. two line last bus moment of subway " matches are to " Shanghai Underground two Line last bus moment, 23:00 ", " No. two line last bus moment of Beijing Metro, 23:15 ", " wide No. two line last bus moment of state subway, 23:30 " generate such as " No. two line last buses of Shanghai Underground Moment is 23:00;No. two line last bus moment of Beijing Metro are 23:15;No. two lines of Guangzhou Underground The last bus moment is 23:30 " answer information.
Need it is further noted that the network equipment 2 can wrap during generating answer information Multiple can ask obtain according to search information containing the generation of compound question and answer pair, the i.e. network equipment 2 Answer questions and merge to generate new compound question and answer pair, such as, can be according to the question and answer obtained to " hair pool East birthplace, Shaoshan, Hunan " and question and answer to " Mao Zedong's date of birth, December 26 in 1893 Day " generate question and answer to " date and place of birth of Mao Zedong, December in 1893 Hunan on the 26th Shaoshan ".
Subsequently, in step s107, described answer information is sent to user by the network equipment 2 Equipment 1.Finally, in step s108, subscriber equipment 1 is believed according to the described answer received Breath, updates the page, presents to user with the page after described answer information is incorporated renewal.
Preferably, between step s105 and step s107, (figure is not also to include step s106 Show).In step s106, the network equipment 2 updates searched page, will generate in step s105 Answer information be dissolved in the searched page of renewal.
Wherein, the position that answer information presents in the page includes but not limited to following at least one :
-Search Results Article 1, such as, be presented on the link according to search information acquisition by result In Article 1;
-search suggestion, such as, be presented on the search key providing a user with suggestion by result In position;
-input method candidate bar, such as, is presented on result in the option of user's input method;
-search column candidate item hurdle, such as, is presented on the offer search that search column is listed by result In the drop-down hurdle of candidate item;
Candidate item hurdle under-WEB input field, such as, is presented on WEB input field by result In for listing in the drop-down hurdle of candidate item.
It should be noted that above-mentioned example is only better described the solution of the present invention, but this Bright be not limited thereto, it should be appreciated by those skilled in the art that any to searched page at Reason, with scheme answer information being included in searched page, should be included in the model of the present invention In enclosing.
Accordingly, in step s107, the network equipment 2 is by the described answer information that contains The page after renewal is sent to subscriber equipment 1.In step s108, subscriber equipment 1 receives After the page comprising answer information after described renewal, present to user.
Preferably, in step s104, each question and answer are to being with one in the network equipment 2 The frame mode storage of four-tuple.One question and answer to be expressed as comprising question and answer classification, entity, Substance feature describes, the tetrameric four-tuple of answer.
Here, " question and answer classification " represents the classification of problem contained by this question and answer centering, including but not It is limited to: the time, place, product attribute etc.;It is right that " entity " represents that question and answer centering is asked As, include but not limited to: name, place name, product, event, proper noun etc.;" entity Feature description " represent the content of described object wanted to know about;" answer " is these question and answer pair Answer.Such as, question and answer are to " Shaoshan, Hunan, Mao Zedong birthplace, place ", wherein " Point " it is the question and answer classification of these question and answer pair, " Mao Zedong " is the entity of these question and answer pair, " birthplace " Substance feature for these question and answer pair describes, and " Shaoshan, Hunan " is the answer of these question and answer pair.With four Tuple storage question and answer are to can be by question and answer to data structured, it is simple to deposit with relational database etc. Storage.Therefore, search information is turned to search to mating to have further by the network equipment 2 with question and answer Entity to being comprised of rope information and described each question and answer prestored and substance feature describe and carry out Coupling, obtains described entity and the description of described substance feature all can be complete with described search information Portion or one or more question and answer pair of partial content coupling.
Such as, when search information is " birthplace of Mao Zedong is at which ", and the network equipment 2 is by upper State search information and described each question and answer prestored entity to being comprised and substance feature describes Coupling, obtains question and answer to " Shaoshan, Hunan, Mao Zedong birthplace, place ", its entity " hair East, pool ", feature description " birthplace " mates with the full content of described search information.So Just obtain the question and answer pair matched with the information full content of search.
And for example, when search information is " date of birth of Mao Zedong and place ", and above-mentioned search is believed Implicit two problems of breath, the network equipment 2 by above-mentioned search information in question and answer are to storehouse with question and answer pair Entity and substance feature describe mate, obtain question and answer to " time Mao Zedong be born 1893 dates December 26 days " and question and answer to " Hunan, Mao Zedong birthplace, place Shaoshan ", wherein, above-mentioned two question and answer are to entity " Mao Zedong ", feature description " date of birth " And " birthplace " respectively with the partial content " date of birth of Mao Zedong " of described search information Birthplace with Mao Zedong " match.
Preferably, in step s104, the network equipment 2 is by search information and each question and answer pair Before mating, also will determine that the question and answer classification that described search information comprises.Question and answer classification is Section 1 in above-mentioned four-tuple, question and answer classification represents the classification of question and answer centering problem.Such as " time Between 1893 Mao Zedong's dates of birth December 26 days " belong to the question and answer class of time character Not;" Shaoshan, Hunan, Mao Zedong birthplace, place " belongs to the question and answer classification of place character. Such as, when search information is " date of birth of Mao Zedong and place ", and the network equipment 2 is by upper State search information to mate in lexical types storehouse, it is thus achieved that search message part content " date of birth " The type " time " of this word combination description and search message part content " birthplace " The type " place " that this word combination describes.Then judge that above two type is the most right Answering the question and answer classification of timeliness matter and the question and answer classification of place character, then the network equipment 2 is by institute State search information to interrogate in the question and answer classification of its above-mentioned time character comprised and location respectively Answer the question and answer centering that two question and answer classifications of classification are comprised to mate, obtain location and interrogate Answer the question and answer in classification to " Shaoshan, Hunan, Mao Zedong birthplace, place " and time character Question and answer in question and answer classification are to " 1893 Mao Zedong's dates of birth time December 26 Day ".Wherein, the problem of previous question and answer pair is " Mao Zedong birthplace ", this problem and search In information, " Mao Zedong place of birth " matches, and the problem of later question and answer pair is " Mao Zedong Date of birth ", this problem matched with " date of birth of Mao Zedong " in search information.
Preferably, in step s104, the network equipment 2 first judges whether to search from described The entity and the substance feature that extract comprised problem in rope information describe, if can search from described Rope information extracts entity and substance feature describes, by the described entity that can extract and reality Body characteristics describes to describe with each the question and answer described entity to being comprised and substance feature and carries out Join inquiry.Such as, when search information is " birthplace of Mao Zedong is at which ", the network equipment 2 Judging according to entity recognition techniques and proper noun recognition technology can be from described search information The entity and the substance feature that extract comprised problem describe and identify in described search information and ask The entity of topic is " Mao Zedong ", judges that the substance feature of this entity is described as " birthplace ".Its In, entity recognition techniques is object or theme described by a kind of identification content of text, and preferably, The technology that described description object or theme are sorted out;Proper noun recognition technology is a kind of knowledge The proprietary name occurred in other text and significant numeral classifier phrase, and preferably, to described specially There is the technology that title and significant numeral classifier phrase are sorted out.Then the network equipment 2 will extract The entity " Mao Zedong " gone out and substance feature describe " birthplace " with each question and answer to being comprised Entity and substance feature describe mate, it is thus achieved that the question and answer matched are to " place Mao Ze Shaoshan, Hunan, birthplace, east ".
Preferably, in step s104, the network equipment 2 first judges whether to search from described Whether the entity and the substance feature that extract comprised problem in rope information describe, and can interpolate that Go out the classification of this comprised problem of search information.If reality can be extracted from described search information Body and substance feature describe, and can interpolate that out the classification of this comprised problem of search information, net The classification of this comprised problem of search information, entity and substance feature are described with each by network equipment 2 The classification of individual question and answer centering, entity and substance feature describe and mate respectively.Such as, when search letter Breath is " date of birth of Mao Zedong ", and the network equipment 2 is according to entity recognition techniques and proprietary name Word identification technology judges to extract the classification of comprised problem, reality from described search information Body and substance feature describe and identify that in described search information, the classification of problem is " time ", real Body is " Mao Zedong ", and judges that the substance feature of this entity is described as " date of birth ".Then The network equipment 2 is special by the problem category " time " extracted, entity " Mao Zedong " and entity Levy description " date of birth " special with each question and answer problem category, entity and entity to being comprised Levy description to mate, it is thus achieved that the question and answer matched are to " Mao Zedong time, lake date of birth Shaoshan, south ".
Fig. 2 is answering for presenting search in search interface according to one embodiment of the present invention The method flow diagram of case information.
First the network equipment 2 obtains the search information inputted from user by subscriber equipment 1, When there are the question and answer pair that relative search information more refines during question and answer are to storehouse, can match multiple Question and answer pair, as question and answer to candidate item.Then the network equipment 2 will exist according to user related information Above-mentioned multiple question and answer are to choosing one or more question and answer pair in candidate item further, the most further Obtain answer information and in search interface, present to user.Wherein user related information include but It being not limited to: 1) individual subscriber attribute (includes but not limited to: IP address, subscriber equipment Classification, user's Sex, Age etc.);2) user preference is arranged;3) user's search history record Deng.
Specifically, step s201 to s204 and above reference step s101 described by Fig. 1 Same or similar to s104, comprise by reference at this, repeat no more.
In step s205, the network equipment 2 obtains the relevant letter of user from subscriber equipment 1 Breath.Wherein, the acquisition of described relevant information includes but not limited in the following manner:
1) user related information that subscriber equipment 1 sends directly is obtained;
2) identity or the information of identification, the network equipment 2 of the user that subscriber equipment 1 sends are obtained According to this identity or the relevant letter of the identification acquisition of information record this user in the network equipment 2 Breath;
3) network equipment 2 sets according to the user obtained when setting up communicate with described subscriber equipment 1 The identification information of standby 1, the cell-phone number of the subscriber equipment such as obtained or hardware sequence number etc., come Judge the identity of user, and obtain the relevant information of this user according to this identity.
In step s206, the network equipment 2 is according to the relevant letter of the user obtained in step s205 Cease obtained multiple question and answer to candidate item is chosen one or more question and answer further Right.Wherein, choose described in the method for one or more question and answer pair include following at least one:
1) specifically, each problem of candidate item is carried out by the network equipment 2 by the question and answer of acquisition Comparison, extracts incoherent vocabulary, and by described incoherent vocabulary in described lexical types storehouse Middle lookup, it is thus achieved that the type of described uncorrelated vocabulary, and extract according to the type or obtain corresponding User related information, choose in question and answer candidate item.
Such as, if search information be " No. two line last bus moment of subway ", it is thus achieved that question and answer pair For " No. two line last bus moment of Shanghai Underground, 23:00 ", " during No. two line last buses of Beijing Metro Carve, 23:00 ", " No. two line last bus moment of Guangzhou Underground, 23:00 " etc., by above-mentioned question and answer The problem of candidate item is compared, extract incoherent vocabulary be " Shanghai ", " Beijing ", " wide State " etc., search in lexical types storehouse and obtain above-mentioned uncorrelated lexical types and be " place ", And the user related information corresponding to " place " is IP address, the network equipment 2 basis IP address, it is judged that user location is Shanghai, and then select question and answer to " Shanghai Underground No. two line last bus moment, 23:00 ".
Wherein, if question and answer are to being to store in the way of four-tuple, the most described comparison can be further It is limited to the comparison that entity and substance feature are described.
Such as, search information is " No. two line last bus moment of subway ", and the network equipment 2 is by this Entity to being comprised of search information and each question and answer and substance feature describe and mate, and obtain " No. two line last bus moment 23:00 of time Shanghai Underground ", " time Beijing Metro two Number line last bus moment 23:15 ", " during No. two line last buses of time Guangzhou Underground Carve 23:30 " etc. multiple question and answer pair, then the network equipment 2 analyzes these question and answer pair, extracts this A little question and answer are to the difference " Shanghai " of " entity " item, " Beijing ", " Guangzhou " etc., at word In remittance typelib, coupling obtains its type is " place ", the relevant letter of the user corresponding to " place " Breath is IP address, and then according to user equipment (UE) IP position, the network equipment 2 is asked above-mentioned Question and answer are chosen further to " during No. two line last buses of time Shanghai Underground in answering questions Carve 23:00 ".
2) arrange according to the user preference obtained, it is judged that how to choose question and answer to candidate item.
Such as, in user preference is arranged, it is set in search interface and only presents a number of answer During information, if the network equipment 2 is by entity to being comprised of search information and each question and answer and reality Body characteristics description carries out mating the number of the coupling question and answer pair obtained and exceedes this setting quantity, then net Network equipment 2 leaves out unnecessary question and answer pair, and only retains user preference and arrange asking of middle setting quantity Answer questions.
And for example, user preference can set in arranging when obtaining multiple question and answer to candidate item, chooses The question and answer pair that answer is most.
For another example, if question and answer store with quadruple form, user preference also can arrange each and ask Answer the priority of classification, as preferentially chosen the question and answer of which classification to candidate item etc..
3) according to the user's search history record obtained, it is judged that how to choose question and answer to candidate item.
Specifically, each problem of the question and answer pair of acquisition is compared by the network equipment 2, extracts Incoherent vocabulary, and described incoherent vocabulary is mated with the search information of user, In selection user's search information, the question and answer at the vocabulary place that matching degree is the highest are to candidate item.Wherein, The factor of judgment of described matching degree includes but not limited to: the quantity of the vocabulary matched, match The search time of vocabulary, the frequency etc. that matches.
Such as, if the search information of user is " the niciest river system ", the question and answer pair that coupling obtains For " the niciest river system dish, XXX " and " the niciest river system restaurant, YYY " etc., network sets Standby 2 extract incoherent vocabulary " dish ", " restaurant " etc..Then the network equipment 2 analyzes user Search history record, obtain " dining room ", " restaurant ", " hotel " etc. with " restaurant " more The most more vocabulary of coupling, then the network equipment 2 selects to comprise asking of " restaurant " Answer questions " the niciest river system restaurant, YYY ".
Step s207 to s210 is identical with step s105 to s108 as described in Figure 1 or phase Seemingly, comprise by reference at this, repeat no more.
Fig. 3 be according to another preferred embodiment of the present invention for presenting search in search interface The method flow diagram of answer information.It illustrates that first the network equipment 2 obtains from user by using , in question and answer be to storehouse, there is relative search information more refine in the search information of family equipment 1 input Question and answer pair time, can match multiple question and answer to as question and answer to candidate item.Then network sets Standby 2 will choose one in above-mentioned multiple question and answer centerings alternately further according to the further of user Or multiple question and answer pair, finally obtain answer information and in search interface, present to user.
Step s301 to s304 is identical with step s101 to s104 as described in Figure 1 or phase Seemingly, comprise by reference at this, repeat no more.
In step s305, the network equipment 2 to candidate item, obtains phase according to the question and answer obtained The options answered.
Specifically, described options to the problem of candidate item or can be used for describing according to each question and answer The entity of problem and substance feature describe and generate, or the direct problem by each question and answer to candidate item As corresponding options.
Such as, when input search information is " Mao Zedong's birth ", and the network equipment 2 is searched described Entity to being comprised of rope information and each question and answer and substance feature describe after mating will To question and answer to candidate item " 1893 Mao Zedong's dates of birth time December 26 days ", " Shaoshan, Hunan, Mao Zedong birthplace, place ", the network equipment 2 extracts above-mentioned asking respectively Entity and the substance feature of answering questions candidate item describe and are combined into the options " date of birth of Mao Zedong Phase ", " place of birth of Mao Zedong ".
In step s306, above-mentioned two options is sent to subscriber equipment 1 by the network equipment 2. Then, in step s307, above-mentioned options is presented to user by subscriber equipment 1.In step In rapid s308, subscriber equipment 1 obtains the options selected by user.Subsequently, in step s309 In, the options selected by user is sent to the network equipment 2 by subscriber equipment 1.Finally, exist In step s310, the network equipment 2 selects its corresponding question and answer pair according to this options.
Step s207 to s210 described by step s311 to s314 and Fig. 1 is same or similar, Comprise by reference at this, repeat no more.
Preferably, in step s304, the network equipment 2 first judges whether to search from described The entity and the substance feature that extract comprised problem in rope information describe, if can search from described Rope information extracts entity and substance feature describes, by the described entity that can extract and reality Body characteristics describes to describe with each the question and answer described entity to being comprised and substance feature and carries out Join;If the network equipment 2 can not extract the entity comprising problem and substance feature describes these two, And one of above-mentioned two can only be extracted, the network equipment 2 first will describe two at entity or substance feature The question and answer centering coupling that in Xiang, one of which determines, obtains multiple coupling question and answer to as question and answer pair Candidate item.
Such as, when search information is " birth of Mao Zedong ", and the network equipment 2 is known according to entity Other technology and proper noun recognition technology can only extract to be comprised from described search information asks The entity of topic is " Mao Zedong ", and can not extract its substance feature and describe.The network equipment 2 will " it is born " and the substance feature profile matching of question and answer centering that entity is " Mao Zedong ", obtains Substance feature describes two the coupling question and answer comprising " birth " to " time Mao Zedong is born 1893 dates December 26 days " and " Shaoshan, Hunan, Mao Zedong birthplace, place ".
Fig. 4 illustrates that the above-mentioned network equipment 2 generates question and answer pair according to the question and answer content from webpage Method flow diagram.
Specifically, in step s401, the network equipment 2 obtains from default storehouse, website can Can be containing the website of question and answer content information, as Baidu knows, searches and ask.Then, in step In rapid s402, the network equipment 2 can use the mode such as Web Spider, web crawlers, and capturing should The web page contents that may contain question and answer content information in website, and question and answer may be contained to above-mentioned The web page contents of content information is analyzed, and according to the position of web page code decision problem, and carries Taking-up problem.The html format sources code of such as webpage, there will be on its question text one hurdle The such as labelling of code " title ", the network equipment 2 obtains the problem literary composition on above-mentioned relevant position This information, the problem i.e. obtaining the question and answer content of this webpage.
Then, the network equipment 2 obtains the answer of this question and answer content.The method obtaining answer includes But be not limited to following at least one:
1) whether this question and answer content there is the optimum answer being identified, have then with this optimum answer Answer as this question and answer content;
Such as, analyze webpage, it is judged that the Baidu got know in question and answer content in whether have " optimum answer " this hurdle." optimum answer " represents the answer corresponding to this problem.
2) using the highest answer of clicking rate in all answers of this question and answer content or positive rating as The answer of this question and answer content.
Such as, analyze obtain certain answer " push up " by online friend and " favorable comment " at most, then sentence This answer disconnected is the answer of this problem.
Subsequently, in step s403, the network equipment 2 judges the question and answer obtained in step s402 Whether content is definitiveness question and answer content, and "Yes" then carries out step s405;"No" carries out step S404, i.e. gives up this question and answer content.
Judge that whether question and answer content is that the method for definitiveness question and answer content includes but not limited to:
1) classification searching this question and answer content of matching judgment in lexical types storehouse is first passed through, When prestore in lexical types storehouse classification can matched time, further according to entity recognition techniques and The entity of problem in question and answer content described in proper noun recognition technology identification, judge the reality of this entity Body characteristics describes.When classification, entity and the substance feature of described question and answer content describe and answer After being determined by above-mentioned steps, i.e. show that described question and answer content is definitiveness question and answer content;
2) by proper noun recognition technology and entity recognition techniques, it may be judged whether be capable of identify that The substance feature going out entity and this entity describes, if can, then judge that this question and answer content is as determining Property question and answer content, and determine whether the classification of this question and answer content.
It should be noted that it should be appreciated by those skilled in the art that and determine whether that definitiveness is asked The method answered is not exemplified as limit with above-mentioned, it is true that any according to whether entity can be extracted And substance feature describes, or whether can extract entity and substance feature describes and judges question and answer class , do not judge that whether question and answer content is the method for definitiveness question and answer content, should be included in this In bright scope.
Need it is further noted that it should be appreciated by those skilled in the art that and judge question and answer content When whether being definitiveness question and answer content, can only be judged by the problem in question and answer content, because of This, the order of step s403 can be before step s402, accordingly, in step s402, Only in the case of judging that this question and answer content is definitiveness question and answer content, just extract answer.
Finally, in step s405, the network equipment 2 is above-mentioned according to obtain in above-mentioned steps The classification of problem, entity and substance feature describe, in conjunction with the answer in above-mentioned question and answer content information Information, generates asking of four-tuple structure (" class instance substance feature describing answer ") Answer questions.
Fig. 5 illustrates that the above-mentioned network equipment 2 is according to the data genaration question and answer pair from encyclopaedia webpage Method flow diagram.
In step s501, the network equipment 2 can obtain network encyclopaedia character by the Internet Web page address, such as Baidupedia, wikipedia etc., thus obtains the encyclopaedia number in such website According to.The network equipment 2 uploads also by this locality or the mode of network obtains encyclopaedia data.
Then, in step s502, the network equipment 2 can be by obtaining in step s501 Network encyclopaedia webpage is analyzed, and is judged entry and the entry solution of encyclopaedia data by web page code Release.In the relevant position of webpage html source code, such as obtain the descriptor of these encyclopaedia data The entry explanation etc. of bar and described entry;Also by the parsing (mould to the encyclopaedia data uploaded Plate analysis judgment etc.) obtain the theme entry of encyclopaedia data and the entry explanation of described entry Deng.
Then, in step s504, the network equipment 2 using entry as entity, and according to word Bar and entry explain that generating substance feature describes.Then describe according to described entity and substance feature Generation problem.Such as " Mao Zedong " this entry, it is corresponding that the network equipment 2 will generate it Substance feature be described as " life ", and generate problem " life of Mao Zedong ";For " trade Easily favorable balance " this entry, generation substance feature is described as " implication " etc. by the network equipment 2, Generation problem " what favourable trade balance is meant that " etc..
Finally, in step s505, obtained by the network equipment 2 is according to step s503 and s504 Problem, and it is construed to answer with the entry obtained in step s502, generate this encyclopaedia data Corresponding question and answer pair.
Preferably, between described step s502 and step s504, also include step s503 (figure Do not show).In step s503, the network equipment 2 is according to above-mentioned entry and the entry of described entry Explain in lexical types storehouse, search coupling, to determine the entry solution of above-mentioned entry and described entry Release the classification of its correspondence.Such as " Mao Zedong " these encyclopaedia data, the network equipment 2 will " Mao Zedong " this entry is searched in lexical types storehouse, and it is right to judge that this entry and entry are explained The classification answered should be " personage ".Accordingly, in step s504, the network equipment 2 is by entry As entity, and explain that corresponding classification generates substance feature and describes according to entry and entry.Example As for " Mao Zedong " this entry, the substance feature generating its correspondence is retouched by the network equipment 2 State as " life ";For " favourable trade balance " this entry, the network equipment 2 will generate entity Feature description is " implication " etc..In step s505, the network equipment 2 is according to step s503 And entry classification, entity and the substance feature obtained by s504 describes, in integrating step s502 The entry obtained is construed to answer, generates the question and answer pair that these encyclopaedia data are corresponding.
It should be noted that as one of the preferred version of the present invention, question and answer to generate or During joining, entity and substance feature can be described and be normalized.
Wherein, described normalized includes but not limited to:
1) the multiple entities belonged in same synonym phrase or substance feature are described with wherein One entity or substance feature describe states, and wherein, described synonym phrase is stored in synonym In storehouse;
Such as, during question and answer are to generating, the entity obtaining these question and answer pair is " Chairman Mao ", Substance feature is " date of birth ", and during question and answer are to generation or after this process, The network equipment 2 is searched in thesaurus and is obtained " Chairman Mao " and be included in a synonym phrase, This synonym phrase is all using " Mao Zedong " as unified description, then in the generation process of question and answer pair In, entity " Chairman Mao " is normalized to entity " Mao Zedong ";In the matching process or After coupling, the network equipment 2 is searched in thesaurus and is obtained " date of birth " and be included in one In synonym phrase, and this synonym phrase is all using " birthday " as unified description, then in question and answer To generation during, substance feature description " date of birth " is normalized to substance feature is retouched State " birthday ".
The most such as, during search information is mated, obtain the reality of this search information Body is " Chairman Mao ", and substance feature is " date of birth ", and maybe this mistake during coupling After journey, the network equipment 2 search in thesaurus obtain " Chairman Mao " be included in one with In justice phrase, this synonym phrase all using " Mao Zedong " as unified description, then will searched for During information and question and answer are to mating, entity " Chairman Mao " is normalized to entity " hair East, pool " mate;In the matching process or coupling after, the network equipment 2 is at synonym Storehouse is searched and obtains " date of birth " and be included in a synonym phrase, and this synonym phrase is equal Using " birthday " as unified description, then by mistake to mating of search information and question and answer Cheng Zhong, is normalized to substance feature by substance feature description " date of birth " and describes " birthday " Mate.
2) the entity unification that similarity exceedes predetermined threshold is identical entity, by similarity The substance feature exceeding predetermined threshold describes unified for identical substance feature description.
It will be understood by those skilled in the art that the similarity that entity or substance feature describe can be by many Kind of mode calculates, such as, by similar portion proportion, or preset similar to aforementioned Numerical value corresponding to the different range of part proportion determines.It addition, people in the art Member should rule of thumb or actual demand determines aforementioned predetermined threshold not repeat at this.
It should be noted that as one of the preferred version of the present invention, also include the network equipment 2 The answer information comprised is analyzed by the question and answer obtaining coupling, it is judged that described answer information Whether comprising answer and obtain information, if described answer information comprises answer and obtains information, network sets Standby 2 obtain information according to this answer, are called the step obtaining corresponding answer by api interface. Wherein, above-mentioned answer acquisition information includes but not limited to 1) webpage url links and answer is at this Particular location in webpage;2) customizing messages obtained from special interface.Above-mentioned answer obtains Information can be by manually presetting.
Such as, certain user inputs search information " Shanghai weather condition " on August 31st, 2010, The network equipment 2 obtains corresponding question and answer pair according to described search information matches, and these question and answer are to answering The network address comprising certain webpage in case information and the content wishing acquisition position in the web page Information, this positional information includes but not limited to described content position range in the web page or institute Place's module, then the network equipment 2 judges that these question and answer comprise answer and obtain the answer information comprised Information, then, the network equipment 2 calls corresponding webpage by api interface, and at this webpage On according to described positional information, capturing the content wishing to obtain is " Shanghai, in August, 2010 31 days, 25~29 DEG C, drizzle to moderate rain, southeaster 4-5 level ", and present as answer To user, the appearance form of this answer include but not limited to comprise the textual form of above-mentioned answer or Graphic form.
According to a further aspect in the invention, if subscriber equipment 1 by the network equipment 2 according to Fig. 4 and The question and answer that embodiment illustrated in fig. 5 generates are downloaded to this locality to storehouse and default lexical types storehouse, use Family equipment 1 can the function of complete independently embodiment as shown in Figure 1, Figure 2 and Figure 3.
Specifically, it is with the difference of the embodiment shown in Fig. 1, in the present embodiment, uses After family equipment 1 obtains the search information of user's input, it is not necessary to be sent to the network equipment 2, but Directly perform to set with network in described step s104, described step s105 and described step s106 Standby same or analogous step performed by 2, it is thus achieved that after containing the searched page of answer information, It is presented directly to user.Wherein, step s104 described in Fig. 1, described step s105 and institute State in step s106 that all operations performed by the network equipment 2 all can be same by subscriber equipment Complete, at this by with in the way of comprise, repeat no more.
It is with the difference of the embodiment shown in Fig. 2, in the present embodiment, subscriber equipment 1 After obtaining the search information of user's input, it is not necessary to be sent to the network equipment 2, but directly perform Same or analogous step with performed by the network equipment 2 in described step s204, obtains Question and answer are to candidate item, and subsequently, subscriber equipment 1 directly obtains user related information from this locality, then Perform and the network equipment 2 in described step s206, described step s207 and described step s208 The same or analogous step done, it is thus achieved that after containing the searched page of answer information, directly Present to user.Wherein, step s204 described in Fig. 2, described step s206, described step In s207 and described step s208, all operations performed by the network equipment 2 all can be same by Subscriber equipment 1 completes, at this by with in the way of comprise, repeat no more.
It is with the difference of the embodiment shown in Fig. 3, in the present embodiment, subscriber equipment 1 After obtaining the search information of user's input, it is not necessary to be sent to the network equipment 2, but directly perform Same or similar with performed by the network equipment 2 in described step s304 and described step s305 Step, obtain options, subsequently, described options is presented directly to by subscriber equipment 1 User, and after obtaining the selection of user, directly perform and described step s310, described step s311 And the same or analogous step that in described step s312, the network equipment 2 is done, obtain bag After having contained the searched page of answer information, it is presented directly to user.Wherein, walk described in Fig. 3 Rapid s304, described step s305, described step s310, described step 311 and described step What in s312, all operations performed by the network equipment 2 all can be same is completed by subscriber equipment, This by with in the way of comprise, repeat no more.
Fig. 6 illustrates and presents search answer information in search interface according to one aspect of the invention System construction drawing.
Wherein, subscriber equipment 1 is connected with the network equipment 2 via network, described network include but It is not limited to the Internet, wide area network, Metropolitan Area Network (MAN), LAN, VPN, mobile Ad hoc network Network (Ad Hoc network) etc..
Subscriber equipment 1 includes but not limited to that any and user passes through keyboard, remote controller, touches Template or voice-operated device carry out the electronic product of man-machine interaction, such as computer, smart mobile phone, PDA, game machine or IPTV etc..Subscriber equipment 1 includes the first dispensing device 11, first Receive device 12 and the 4th acquisition device 13.
The network equipment 2 includes but not limited to that single network server, multiple webserver form Server group or based on cloud computing (Cloud Computing) by a large amount of computers or network The cloud that server is constituted.Wherein, cloud computing is the one of Distributed Calculation, by the loose coupling of a group One super virtual machine of the computer collection composition closed.Wherein, the network equipment 2 includes One acquisition device 21, coalignment 22, offer device 23.Described question and answer are to storehouse 26 and vocabulary Typelib 25 can be included in the network equipment 2, it is possible to the network equipment 2 physical separation but logical Letter connects.Wherein, question and answer centering comprises problem and the corresponding answer of this problem, and question and answer are to storehouse 26 is the set comprising a large amount of question and answer pair.Lexical types storehouse 25 have recorded vocabulary or vocabulary Combination is the corresponding relation of the possible type described with the combination of this vocabulary or vocabulary.
When user is desired with search, the input mode input search provided by subscriber equipment Information.Wherein, this input mode includes but not limited to: 1) word input;2) phonetic entry; 3) handwriting input.Wherein, in aforesaid way, the position of input search information includes but does not limits In: 1) search column of the page that provides of search engine;2) searched page that client provides; 3) search column etc. in embedded web page or client.
Specifically, the 4th acquisition device 13 in subscriber equipment 1 is by any and user Carry out the interactive device of man-machine interaction to obtain the search information of user's input.This interactive device can To be keyboard, remote controller, touch pad or voice-operated device etc..Then, the first dispensing device 11 The search information described user inputted is sent to the network equipment 2 by the Internet.
Then, the first acquisition device 21 in the network equipment 2 obtains above-mentioned user and searches for information, Coalignment 22 is by the search information got and question and answer each question and answer pair to prestoring in storehouse 26 Mate, i.e. same or analogous vocabulary letter in each question and answer centering lookup with search information Breath, to obtain and all or part of content of described search information matches one or more asks Answer questions.
Specifically, the problem of search information with question and answer centering is mated by coalignment 22. According to search information, may occur in which problem and this search information matches of one or more question and answer pair Situation.
When search information only implies single problem, coalignment 22 can obtain this search letter The result that full content in breath can match with the problem of one or more question and answer centerings.
Such as, when search information is " birthplace of Mao Zedong is at which ", and above-mentioned search information is hidden Containing single problem.This search information is mated in question and answer are to storehouse 26 by coalignment 22, Obtaining question and answer is that " Mao Zedong is born to the problem in " Mao Zedong birthplace, Shaoshan, Hunan " Ground ", the problem of these question and answer pair matches with the full content of search information.
When searching for the multiple problem of embodying information, coalignment 22 can mate and obtains multiple question and answer Right, the problem of each question and answer pair matches with the partial content in this search information.
Such as, when search information is " date of birth of Mao Zedong and place ", and above-mentioned search is believed Implicit two problems of breath.This search information is carried out in question and answer are to storehouse 26 by coalignment 22 Join, obtain question and answer to " Mao Zedong birthplace, Shaoshan, Hunan " and " Mao Zedong's date of birth, On December 26th, 1893 ".Wherein, the problem of previous question and answer pair is " Mao Zedong birthplace ", This problem matches with " Mao Zedong place of birth " in search information, and later question and answer are to asking Entitled " Mao Zedong's date of birth ", this problem and " Mao Zedong's date of birth " in search information Match.
When there are the question and answer pair that relative search information more refines during question and answer are to storehouse 26, coupling Device 22 can match multiple question and answer pair.
Such as, when search information is " No. two line last bus moment of subway ", due to each city No. two line final vehicle hours of subway are different, and question and answer are to often there being each different cities in storehouse 26 The question and answer pair of No. two line last buses of subway in city.Specifically, coalignment 22 is by above-mentioned search After information and each question and answer are to mating, it is thus achieved that multiple question and answer pair: " No. two lines of Shanghai Underground Last bus moment, 23:00 ", " No. two line last bus moment of Beijing Metro, 23:15 ", " Guangzhou No. two line last bus moment of subway, 23:30 " etc., the problem of above-mentioned question and answer pair all can be with search letter The full content coupling of breath.
It should be noted that above-mentioned example is only better described the solution of the present invention, but this Bright it is not limited thereto, it should be appreciated by those skilled in the art that any according to search information, obtain Obtain the scheme of one or more question and answer pair, should be included in the scope of the present invention.
Then, it is provided that device 23 is believed extracting answer according to said one or multiple coupling question and answer Breath.
When the question and answer with search information matches are to only one, it is provided that device 23 obtains above-mentioned The answer of question and answer centering, in conjunction with the problem of these question and answer pair as answer information.
When the question and answer with search information matches are to when having multiple, it may be judged whether can be by each question and answer To answer integrate.Device 23 is provided also to include integrating apparatus (not shown), if above-mentioned Each question and answer are to semantically related and can integrate, and above-mentioned each is asked by described integrating apparatus The mode with meeting natural language custom of answering questions is integrated, and generates answer information.Such as, for With search information " date of birth of Mao Zedong and the place " question and answer that match to " Mao Zedong goes out Raw days, on December 26th, 1893 " and " Mao Zedong birthplace, Shaoshan, Hunan ", it is asked The main body of topic is identical, and integrating apparatus can be integrated generates a semantically coherent answer information " Mao Zedong is born in December in 1893 26, Shaoshan, Hunan ";If each question and answer are to cannot Integrating, integrating apparatus is the most respectively according to answer and the problem of these question and answer pair of question and answer pair, and generation is many Individual answer unit is as answer information, such as, by search information " during No. two line last buses of subway Carve " question and answer that match are to " No. two line last bus moment of Shanghai Underground, 23:00 ", " ground, Beijing No. two line last bus moment of ferrum, 23:15 ", " No. two line last bus moment of Guangzhou Underground, 23:30 " Such as " No. two line last bus moment of Shanghai Underground are 23:00 in generation;No. two line ends of Beijing Metro The regular bus moment is 23:15;No. two line last bus moment of Guangzhou Underground are 23:30 " answer information.
Need it is further noted that the network equipment 2 can wrap during generating answer information Containing the generation of compound question and answer pair, i.e. provide in device 23 and also comprise integrating apparatus (not shown), The multiple question and answer obtained according to search information can be combined to generate new by described integrating apparatus Compound question and answer pair, such as, can be according to the question and answer obtained to " Mao Zedong birthplace, Shaoshan, Hunan " And question and answer generate question and answer to " hair pool to " Mao Zedong's date of birth, on December 26th, 1893 " The date and place of birth in east, Shaoshan, December in 1893 Hunan on the 26th ".
Then, it is provided that described answer information is sent to subscriber equipment 1 by device 23.First receives Device 12, according to the described answer information received, updates the page, with by described answer information Incorporate the page after renewal and present to user.
Preferably, the network equipment 2 also includes webpage updating device (not shown), described webpage The answer information that updating device provides according to described offer device 23, updates searched page, with Answer information is dissolved in the searched page of renewal.
The position that wherein answer information presents in the page include but not limited to following at least one:
-Search Results Article 1, such as, be presented on the link according to search information acquisition by result In Article 1;
-search suggestion, such as, be presented on the search key providing a user with suggestion by result In position;
-input method candidate bar, such as, is presented on result in the option of user's input method;
-search column candidate item hurdle, such as, is presented on the offer search that search column is listed by result In the drop-down hurdle of candidate item;
Candidate item hurdle under-WEB input field, such as, is presented on WEB input field by result In for listing in the drop-down hurdle of candidate item.
It should be noted that above-mentioned example is only better described the solution of the present invention, but this Bright be not limited thereto, it should be appreciated by those skilled in the art that any to searched page at Reason, with scheme answer information being included in searched page, should be included in the model of the present invention In enclosing.
Accordingly, the page after the described renewal containing answer information is sent out by webpage updating device Give subscriber equipment 1.First receiving device 12 receive described renewal after comprise answer information The page after, present to user.
Preferably, each question and answer in question and answer to storehouse 26 in be with the structure side of a four-tuple Formula storage.One question and answer to being expressed as comprising question and answer classification, entity, substance feature describe, The tetrameric four-tuple of answer.Wherein, question and answer classification represents the classification (bag of this question and answer centering problem Include but be not limited to: the time, place, product attribute etc.) entity represents what question and answer centering was asked Object (includes but not limited to: name, place name, product, event, proper noun etc.);Real Body characteristics describes the content representing the described object wanted to know about;Answer is these question and answer to answering Case.Such as question and answer are to " Shaoshan, Hunan, Mao Zedong birthplace, place ", wherein " place " For the question and answer classification of these question and answer pair, " Mao Zedong " is the entity of these question and answer pair, and " birthplace " is The substance feature of these question and answer pair describes, and " Shaoshan, Hunan " is the answer of these question and answer pair.With quaternary Group storage question and answer are to can be by question and answer to data structured, it is simple to deposit with relational database etc. Storage.Therefore, search information can be entered one with each question and answer prestored to mating by coalignment 22 Step tool turns to entity to being comprised of search information and described each question and answer prestored and entity Feature description is mated, and obtains described entity and the description of described substance feature all can be with institute State one or more question and answer pair of all or part of content matching of search information.
Such as, when search information is " birthplace of Mao Zedong is at which ", and coalignment 22 is by upper State search information and described each question and answer prestored entity to being comprised and substance feature describes Coupling, obtains question and answer to " Shaoshan, Hunan, Mao Zedong birthplace, place ", its entity " hair East, pool ", feature description " birthplace " mates with the full content of described search information.So Just obtain the question and answer pair matched with the information full content of search.
And for example, when search information is " date of birth of Mao Zedong and place ", and above-mentioned search is believed Implicit two problems of breath, coalignment 22 by above-mentioned search information in question and answer are to storehouse 26 with ask The entity answered questions and substance feature describe and mate, and obtain question and answer to " time, Mao Zedong went out 1893 phases birthday December 26 days " and question and answer to " lake, Mao Zedong birthplace, place Shaoshan, south ", wherein, above-mentioned two question and answer are to entity " Mao Zedong ", feature description " date of birth Phase " and " birthplace " respectively with the partial content " year of birth of Mao Zedong of described search information Month " and the birthplace of Mao Zedong " match.
Fig. 7 illustrates the coalignment of a preferred embodiment of the network according to the invention equipment 2 The schematic diagram of 22.Wherein coalignment 22 also includes first judgment means the 221, first son It is equipped with and puts 222, second judgment means the 227, second sub-coalignment the 223, the 3rd judgment means 228, the 3rd sub-coalignment 225.Wherein the first judgment means 221 and the first sub-coalignment 222, the second judgment means 227 and the second sub-coalignment the 223, the 3rd judgment means 228 Being respectively combined with the 3rd sub-coalignment 225 is three covering devices.
Preferably, by search information and each question and answer to mating before, in coalignment 22 First judgment means 221 also will determine that the question and answer classification that described search information comprises.Question and answer classification Being the Section 1 in above-mentioned four-tuple, question and answer classification represents the classification of question and answer centering problem.Example As " 1893 Mao Zedong's dates of birth time December 26 days " belongs to time character Question and answer classification;" Shaoshan, Hunan, Mao Zedong birthplace, place " belongs to asking of place character Answer classification.Specifically, when search information is " date of birth of Mao Zedong and place ", first Above-mentioned search information is mated in lexical types storehouse 25 by judgment means 221, it is thus achieved that search letter Type " time " and search that breath partial content " date of birth " this word combination describes are believed The type " place " that breath partial content " birthplace " this word combination describes.Then One judgment means 221 judges question and answer classification and the ground of the most corresponding time character of above two type The question and answer classification of some character.Then the first sub-coalignment 222 utilize question and answer to storehouse 26, by institute State search information to interrogate in the question and answer classification of its above-mentioned time character comprised and location respectively Answer the question and answer centering that two question and answer classifications of classification are comprised to mate, obtain location and interrogate Answer the question and answer in classification to " Shaoshan, Hunan, Mao Zedong birthplace, place " and time character Question and answer in question and answer classification are to " 1893 Mao Zedong's dates of birth time December 26 Day ".Wherein, the problem of previous question and answer pair is " Mao Zedong birthplace ", this problem and search In information, " Mao Zedong place of birth " matches, and the problem of later question and answer pair is " Mao Zedong Date of birth ", this problem matched with " date of birth of Mao Zedong " in search information.
Preferably, by search information and each question and answer to mating before, the second judgment means 227 First judge whether to extract from described search information entity and the entity of comprised problem Feature description, if entity and substance feature description can be extracted from described search information, the Two judgment means 227 the described entity that can extract and substance feature are described with described each The question and answer entity to being comprised and substance feature describe and mate.Such as, when search information it is " birthplace of Mao Zedong is at which ", the second judgment means 227 is according to entity recognition techniques and specially Noun identification technology is had to judge to extract from described search information the reality of comprised problem Body and substance feature describe, and identify that in described search information, the entity of problem is " Mao Zedong ", Substance feature is described as " birthplace ".Then, the second sub-coalignment 223 will extract Entity " Mao Zedong " and substance feature describe " birthplace " and each question and answer reality to being comprised Body and substance feature describe and mate, it is thus achieved that the question and answer matched are to " place Mao Zedong go out Shaoshan, Radix Rehmanniae Hunan ".
Preferably, by search information and each question and answer to mating before, in coalignment 22 3rd judgment means 228 first judges whether to extract to be comprised from described search information to ask The entity of topic and substance feature describe, and whether can interpolate that out this comprised problem of search information Classification.If entity can be extracted from described search information and substance feature describes, and energy Enough judge that this is searched by the classification of this comprised problem of search information, the 3rd sub-coalignment 225 The classification of the comprised problem of rope information, entity and substance feature describe the class with each question and answer centering Not, entity and substance feature describe and mate respectively.Such as, when search information be " Mao Zedong's Date of birth ", the 3rd judgment means 228 is according to entity recognition techniques and proper noun recognition skill Art judges to extract the classification of comprised problem, entity and entity from described search information Feature description also identifies that in described search information, the classification of problem is " time ", and entity is " hair East, pool ", and judge that the substance feature of this entity is described as " date of birth ".Then, the 3rd son Coalignment 225 is by the problem category " time " extracted, entity " Mao Zedong " and entity Feature description " date of birth " and each question and answer problem category, entity and entity to being comprised Feature description is mated, it is thus achieved that the question and answer matched are to " Mao Zedong time, lake date of birth Shaoshan, south ".
Fig. 8 illustrates the coupling dress of another preferred embodiment of the network according to the invention equipment 2 Put the schematic diagram of 22.Wherein coalignment 22 also includes the 4th sub-coalignment 223, selects Device the 224, the 5th sub-coalignment 225 and interactive device 226.
Preferably, in coalignment 22 the 4th sub-coalignment 229 by described search information with To mating, in question and answer are to storehouse 26, there is relative search information more in each question and answer prestored For refinement question and answer pair time, can match multiple question and answer to as question and answer to candidate item.Then select Select device 224 to enter in above-mentioned multiple question and answer are to candidate item according to the user related information got One step chooses one or more question and answer pair.Wherein user related information includes but not limited to: 1) Individual subscriber attribute (includes but not limited to: IP address, subscriber equipment classification, user Sex, Age etc.);2) user preference is arranged;3) user's search history record etc..
Wherein, the acquisition of described relevant information includes but not limited in the following manner:
1) network equipment 2 directly obtains the user related information of subscriber equipment 1 transmission and provides Give and select device 224;
2) network equipment 2 obtains the identity of user that subscriber equipment 1 sends or identification information also It is supplied to select device 224, selects device 224 according to this identity or to identify acquisition of information record The relevant information of this user in the network equipment 2;
3) when selecting device 224 to communicate with the foundation of described subscriber equipment 1 according to the network equipment 2 The information of the subscriber equipment 1 obtained is (such as the cell-phone number of subscriber equipment obtained or hardware sequence number Deng), it is judged that the identity of user, and the relevant information of this user is obtained according to this identity.
Then, select device 224 according to the user related information obtained at the 4th sub-coalignment The multiple question and answer obtained in 229 are to choosing one or more question and answer pair in candidate item further.Its In, described in choose the method for one or more question and answer pair include following at least one:
1) specifically, device 224 is selected the question and answer of acquisition to be entered by each problem of candidate item Row comparison, extracts incoherent vocabulary, and by described incoherent vocabulary at described lexical types Storehouse is searched, it is thus achieved that the type of described uncorrelated vocabulary, and extract according to the type or obtain phase The user related information answered, chooses in question and answer candidate item.
Such as, if search information be " No. two line last bus moment of subway ", it is thus achieved that question and answer pair Candidate item is " No. two line last bus moment of Shanghai Underground, 23:00 ", " No. two line ends of Beijing Metro Regular bus moment, 23:00 ", " No. two line last bus moment of Guangzhou Underground, 23:00 " etc., select The problem of candidate item is compared by device 224 by above-mentioned question and answer, extracts incoherent vocabulary and is " Shanghai ", " Beijing ", " Guangzhou " etc., then search on obtaining in lexical types storehouse 25 State uncorrelated lexical types to be in " place ", and the user related information corresponding to " place " For IP address, select device 224 according to IP address, it is judged that user location is Shanghai, and then select question and answer to " No. two line last bus moment of Shanghai Underground, 23:00 ".
Wherein, if question and answer are to being to store in the way of four-tuple, the most described comparison can be further It is limited to the comparison that entity and substance feature are described.
Such as, search information is " No. two line last bus moment of subway ", the 4th sub-coalignment Entity to being comprised of this search information and each question and answer and substance feature are described and to carry out by 229 Join, obtain " No. two line last bus moment 23:00 of time Shanghai Underground ", " time north No. two line last bus moment 23:15 of capital subway ", " No. two lines of time Guangzhou Underground are last Car moment 23:30 " etc. multiple question and answer pair, then select device 224 analyze these question and answer pair, Extract these question and answer to the difference " Shanghai " of " entity " item, " Beijing ", " Guangzhou " etc., And in lexical types storehouse 25, coupling obtains its type for " place ", corresponding to " place " User related information is IP address, then according to user equipment (UE) IP position, selects device 224 choose question and answer to " No. two line ends of time Shanghai Underground further in above-mentioned question and answer centering Regular bus moment 23:00 ".
2) arrange according to the user preference obtained, it is judged that how to choose question and answer to candidate item.
Such as, in user preference is arranged, it is set in search interface and only presents a number of answer During information, if the 4th sub-coalignment 229 by search information and each question and answer to being comprised Entity and substance feature description carry out mating the number of the coupling question and answer pair obtained and exceed this setting Quantity, then select device 224 to leave out unnecessary question and answer pair, and only retains during user preference arranges Set the question and answer pair of quantity.
And for example, user preference can set in arranging when obtaining multiple question and answer to candidate item, chooses The question and answer pair that answer is most.
For another example, if question and answer store with quadruple form, user preference also can arrange each and ask Answer the priority of classification, as preferentially chosen the question and answer of which classification to candidate item etc..
3) according to the user's search history record obtained, it is judged that how to choose question and answer to candidate item.
Specifically, select device 224 each problem of the question and answer pair of acquisition to be compared, carry Take incoherent vocabulary, and the search information of described incoherent vocabulary Yu user is carried out Joining, in selection user's search information, the question and answer at the vocabulary place that matching degree is the highest are to candidate item.Its In, the factor of judgment of described matching degree includes but not limited to: the quantity of the vocabulary matched, phase The search time of the vocabulary of coupling, the frequency etc. that matches.
Such as, if the search information of user is " the niciest river system ", the question and answer pair that coupling obtains For " the niciest river system dish, XXX " and " the niciest river system restaurant, YYY " etc., select dress Put the incoherent vocabulary of 224 extraction " dish ", " restaurant " etc..Then device 224 basis is selected The search history record of user obtains " dining room ", " restaurant ", " hotel " etc. with " restaurant " relatively For the most more vocabulary of coupling, device 224 is therefore selected to select to comprise vocabulary " meal Shop " question and answer to " the niciest river system restaurant, YYY ".
Preferably, in coalignment 22 the 5th sub-coalignment 230 by described search information with To mating, in question and answer are to storehouse 26, there is relative search information more in each question and answer prestored For refinement question and answer pair time, can match multiple question and answer to as question and answer to candidate item.Then Interactive device 226 by with user further and according to the feedback information of user in above-mentioned multiple question and answer One or more question and answer pair are chosen in centering further.
Specifically, interactive device 226 is first according to asking of obtaining in the 5th sub-coalignment 230 Answer questions candidate item, obtain corresponding options.Wherein, described options can be according to each question and answer The problem of candidate item or the entity and substance feature for describing problem are described and generate, or directly Using each question and answer to the problem of candidate item as corresponding options.
Such as, when input search information is " Mao Zedong's birth ", the 5th sub-coalignment 230 By described search information in question and answer are to storehouse 26 with each question and answer entity to being comprised and entity Feature description will obtain question and answer to candidate item " time Mao Zedong's date of birth after mating 1893 phases December 26 days ", " Shaoshan, Hunan, Mao Zedong birthplace, place ", hand over Mutual device 226 extracts above-mentioned question and answer respectively and entity and the substance feature of candidate item is described and be combined into Options " date of birth of Mao Zedong ", " place of birth of Mao Zedong ".
Then, above-mentioned two options is sent to subscriber equipment 1, user by interactive device 226 Options is presented to user and is selected for it by equipment 1, have selected one of them options user After, this options is sent to interactive device 226 by the Internet by subscriber equipment 1, then, Interactive device 226 obtains the options selected by above-mentioned user, and selects its phase according to this options The question and answer pair answered.
It should be noted that coalignment 22 can be by three covering devices shown in Fig. 7 at least one Covering device and Fig. 8 select at least one in device 224, both interactive devices 226 to carry out appointing Meaning combination, to realize further function.Such as, coalignment 22 is by the second judgment means 227 and second sub-coalignment 223 constitute with interactive device 226 combination;When being judged by second Device 227 judges that entity and substance feature describe, and is mated by the second sub-coalignment 223 When obtaining corresponding multiple question and answer pair, interactive device 226 obtains with the plurality of question and answer corresponding Options is supplied to user, and obtains the question and answer pair corresponding with the options that user selects.This Skilled person should be appreciated that the present invention is not exemplified as limit with above-mentioned, it is true that any By in described three covering devices the most a set of with select in device 224 and interactive device 226 at least A set of combination, to select the scheme of optimum question and answer pair, is all contained in the scope of the present invention.
Fig. 9 illustrates that the network equipment 2 is according to from the question and answer content of webpage and the data of encyclopaedia webpage Generate the apparatus structure schematic diagram of question and answer pair.Wherein, the second acquisition device 27 also includes the 4th Judgment means (not shown), the first generating means 28 also includes the first sub-generating means, and (figure is not Show).
Specifically, the network equipment 2 can generate question and answer pair, this generation according to the content from webpage The process of question and answer pair relates to the second acquisition device 27 and the first generating means 28.Second obtains dress Put 27 and from default storehouse, website, obtain the website that may contain question and answer content information, such as Baidu Know, search and ask.Then, the second acquisition device 27 can use Web Spider, network The modes such as reptile, capture the web page contents that may contain question and answer content information in this website, and May be analyzed containing the web page contents of question and answer content information above-mentioned, sentence according to web page code The position of disconnected problem, and extract problem.The html format sources code of such as webpage, at it Question text one hurdle there will be the labelling of such as code " title ", and the second acquisition device 27 obtains Question text information on above-mentioned relevant position, i.e. obtains the asking of question and answer content of this webpage Topic.
Then the second acquisition device 27 obtains answer from described question and answer content.Obtain answer Method include but not limited to following at least one:
1) whether this question and answer content there is the optimum answer being identified, have then with this optimum answer Answer as this question and answer content;
Such as, analyze webpage, it is judged that the Baidu got know in question and answer content in whether have " optimum answer " this hurdle." optimum answer " represents the answer corresponding to this problem.
2) in all answers of this question and answer content, clicking rate or the highest answer of positive rating.
Such as, analyze obtain certain answer " push up " by online friend and " favorable comment " at most, then sentence This answer disconnected is the answer of this problem.
Subsequently, in the 4th judgment means judges whether the question and answer content obtained is definitiveness question and answer Holding, the first son that this question and answer content is then supplied in the first generating means 28 by "Yes" generates dress Put to generate question and answer pair;"No" then gives up this question and answer content.
4th judgment means judges that whether question and answer content is that the method for definitiveness question and answer content includes But it is not limited to:
1) classification searching this question and answer content of matching judgment in lexical types storehouse is first passed through, When prestore in lexical types storehouse classification can matched time, further according to entity recognition techniques and The entity of problem in question and answer content described in proper noun recognition technology identification, judge the reality of this entity Body characteristics describes.When classification, entity and the substance feature of described question and answer content describe and answer After being determined by above-mentioned steps, i.e. show that described question and answer content is definitiveness question and answer content;
2) by proper noun recognition technology and entity recognition techniques, it may be judged whether be capable of identify that The substance feature going out entity and this entity describes, if can, then judge that this question and answer content is as determining Property question and answer content, and determine whether the classification of this question and answer content.
It should be noted that it should be appreciated by those skilled in the art that and determine whether that definitiveness is asked The method answered is not exemplified as limit with above-mentioned, it is true that any according to whether entity can be extracted And substance feature describes, or whether can extract entity and substance feature describes and judges question and answer class , do not judge that whether question and answer content is the method for definitiveness question and answer content, should be included in this In bright scope.
Need it is further noted that it should be appreciated by those skilled in the art that and judge question and answer content When whether being definitiveness question and answer content, can only be judged by the problem in question and answer content, because of This, the 4th judgment means can i.e. judge that before the second acquisition device 27 obtains answer question and answer are No for definitiveness question and answer, accordingly, the second acquisition device 27 only judges in the 4th judgment means In the case of this question and answer content is definitiveness question and answer content, just extract answer.
Finally, the first generating means 28 according to the classification of the problems referred to above obtained in above-mentioned steps, Entity and substance feature describe, and in conjunction with the answer information in above-mentioned question and answer content information, generate four The question and answer pair of tuple structure (" class instance substance feature describes answer ").
Specifically, the network equipment 2 can be according to the data genaration question and answer pair from encyclopaedia webpage, should The process generating question and answer pair relates to the 3rd acquisition device 29 and the second generating means 30.3rd obtains Fetching put 29 can by the Internet obtain network encyclopaedia character web page addresses, as Baidupedia, Wikipedia etc., thus obtain the encyclopaedia data in such website.3rd acquisition device 29 is also Can be uploaded by this locality or by the way of network, obtain encyclopaedia data.
Second generating means 30 can by network encyclopaedia webpage obtained above is analyzed, Judge that the entry of encyclopaedia data and entry are explained by web page code.Such as in webpage html source The relevant position of code obtains the theme entry of these encyclopaedia data and the entry solution of described entry Release;Obtain also by the parsing (template analysis judgment etc.) of the encyclopaedia data uploaded The theme entry of encyclopaedia data and the entry explanation etc. of described entry.
Second generating means 30 also includes that the second judgment means (not shown), the 3rd son generate dress Put (not shown) and the 4th sub-generating means (not shown).3rd sub-generating means is by entry As entity, and explain that according to entry and entry generating substance feature describes.Then according to described Entity and substance feature describe generation problem.Such as " Mao Zedong " this entry, the 3rd The substance feature generating its correspondence is described as " life " by sub-generating means, and generates problem " hair The life in east, pool ";For " favourable trade balance " this entry, the 3rd sub-generating means will generate Substance feature is described as " implication " etc., and generates problem " what favourable trade balance is meant that " Deng.
Finally, the 4th sub-generating means is according to the problem obtained by above-mentioned, and to obtain the 3rd The entry obtained in device 29 is construed to answer, generates the question and answer pair that these encyclopaedia data are corresponding.
Preferably, before the 3rd sub-generating means generation problem, the second judgment means can basis The entry obtained in the 3rd acquisition device 29 and the entry of described entry are explained at lexical types Storehouse 25 is searched coupling, to determine that its correspondence explained in the entry of above-mentioned entry and described entry Classification.Such as " Mao Zedong " these encyclopaedia data, the second judgment means is by " Mao Zedong " This entry is searched in lexical types storehouse, and judges that this entry and entry explain that corresponding classification should For " personage ".Accordingly, the 3rd sub-generating means using entry as entity, and according to entry And the classification generation substance feature description that entry explanation is corresponding.Such as " Mao Zedong " this Entry, the substance feature generating its correspondence is described as " life " by the 3rd sub-generating means;Right In " favourable trade balance " this entry, generation substance feature is described as " containing by the 3rd sub-generating means Justice " etc..Finally, the 4th sub-generating means is according to step entry obtained above classification, entity And substance feature describes, it is combined in the entry obtained in the 3rd acquisition device 29 and is construed to answer Case, generates the question and answer pair that these encyclopaedia data are corresponding.
For brevity, Fig. 6 to Fig. 9 eliminates the dress unrelated with the embodiment shown in this figure Put, it should be appreciated by those skilled in the art that the network equipment 2 can comprise described in Fig. 6 to Fig. 9 All devices or the combination in any of all devices.
It should be noted that as one of the preferred version of the present invention, in the question and answer mistake to generating Cheng Zhong, can describe entity and substance feature and be normalized, in conjunction with in corresponding question and answer Problem category and answer in appearance generate corresponding question and answer pair.
Wherein, described normalized includes but not limited to:
1) the multiple entities belonged in same synonym phrase or substance feature are described with wherein One entity or substance feature describe states, and wherein, described synonym phrase is stored in synonym In storehouse;
Such as, during question and answer are to generating, the entity obtaining these question and answer pair is " Chairman Mao ", Substance feature is " date of birth ", and during question and answer are to generation or after this process, The network equipment 2 is searched in thesaurus and is obtained " Chairman Mao " and be included in a synonym phrase, This synonym phrase is all using " Mao Zedong " as unified description, then in the generation process of question and answer pair In, entity " Chairman Mao " is normalized to entity " Mao Zedong ";In the matching process or After coupling, the network equipment 2 is searched in thesaurus and is obtained " date of birth " and be included in one In synonym phrase, and this synonym phrase is all using " birthday " as unified description, then in question and answer To generation during, substance feature description " date of birth " is normalized to substance feature is retouched State " birthday ".
The most such as, during search information is mated, obtain the reality of this search information Body is " Chairman Mao ", and substance feature is " date of birth ", and maybe this mistake during coupling After journey, the network equipment 2 search in thesaurus obtain " Chairman Mao " be included in one with In justice phrase, this synonym phrase all using " Mao Zedong " as unified description, then will searched for During information and question and answer are to mating, entity " Chairman Mao " is normalized to entity " hair East, pool " mate;In the matching process or coupling after, the network equipment 2 is at synonym Storehouse is searched and obtains " date of birth " and be included in a synonym phrase, and this synonym phrase is equal Using " birthday " as unified description, then by mistake to mating of search information and question and answer Cheng Zhong, is normalized to substance feature by substance feature description " date of birth " and describes " birthday " Mate.
2) the entity unification that similarity exceedes predetermined threshold is identical entity, by similarity The substance feature exceeding predetermined threshold describes unified for identical substance feature description.
It will be understood by those skilled in the art that the similarity that entity or substance feature describe can be by many Kind of mode calculates, such as, by similar portion proportion, or preset similar to aforementioned Numerical value corresponding to the different range of part proportion determines.It addition, people in the art Member should rule of thumb or actual demand determines aforementioned corresponding predetermined threshold not do superfluous at this State.
It should be noted that as one of the preferred version of the present invention, the network equipment 2 also includes 6th judgment means (not shown) and calling device (not shown).6th judgment means is to basis Join question and answer the answer information obtained is analyzed, it is judged that whether described answer information comprises answer Acquisition information, if described answer information comprises answer and obtains information, calling device is according to this answer Acquisition information, is called by api interface and obtains corresponding answer.Wherein, above-mentioned answer obtains Information includes but not limited to: 1) webpage url link and answer particular location in the web page; 2) customizing messages obtained from special interface.Above-mentioned answer obtains information can be pre-by manually carrying out If.
Such as, certain user inputs search information " Shanghai weather condition " on August 31st, 2010, The network equipment 2 obtains corresponding question and answer pair according to described search information matches, and these question and answer are to answering The network address comprising certain webpage in case information and the content wishing acquisition position in the web page Information, this positional information includes but not limited to described content position range in the web page or institute Place's module, then the 6th judgment means judges that these question and answer comprise answer to the answer information comprised and obtain Winning the confidence breath, then, calling device calls corresponding webpage by api interface, and at this webpage On according to described positional information, capturing the content wishing to obtain is " Shanghai, in August, 2010 31 days, 25~29 DEG C, drizzle to moderate rain, southeaster 4-5 level ", and present as answer To user, the appearance form of this answer include but not limited to comprise the textual form of above-mentioned answer or Graphic form.
According to a further aspect in the invention, if subscriber equipment 1 by the network equipment 2 according to Fig. 9 institute Show that the question and answer that embodiment generates are downloaded to this locality to storehouse 26 and default lexical types storehouse 25, use Family equipment 1 can the function of complete independently embodiment as shown in Fig. 6, Fig. 7 and Fig. 8.
Specifically, it is with the embodiment difference shown in Fig. 6, in the present embodiment, user Equipment 1 comprises the 4th acquisition device 13, and comprises further and embodiment as shown in Figure 1 The first acquisition device 21 that the middle network equipment 2 is comprised, coalignment 22, provide device 23 The same or analogous device of function, but do not comprise the first dispensing device 11 and first receiving device 12.After subscriber equipment 1 obtains the search information of user's input by the 4th acquisition device 13, Without being sent to the network equipment 2, but directly by described with the first acquisition device 21, mate Device 22, the offer device 23 same or analogous device of function, it is thus achieved that contain answer information Searched page after, be presented directly to user.
In the present embodiment, subscriber equipment 1 can comprise each dress included in coalignment 22 Put, be so with the embodiment difference shown in Fig. 8, owing to selecting device 224 and mutual Device 226 is arranged in subscriber equipment 1, therefore selects device 224 directly from subscriber equipment 1 Middle acquisition user related information, and interactive device 226 directly possessed by subscriber equipment 1 Interactive device (such as: display, touch screen, mouse, keyboard, felt pen etc.) enters with user Row is mutual, but selects how device 224 selects one or to individual question and answer according to user related information To and interactive device 226 how to obtain question and answer pair according to the selection of user, shown in Fig. 8 Embodiment same or similar, comprise by reference at this, repeat no more;First judges Device 221, first sub-coalignment the 222, second judgment means the 227, second sub-coalignment 223, the 3rd judgment means 228 and the 3rd sub-coalignment 225 institute in subscriber equipment 1 is real Existing function and to realize the mode of function same or similar with the embodiment shown in Fig. 7, at this with The mode quoted comprises, and repeats no more.
It is obvious to a person skilled in the art that the invention is not restricted to above-mentioned one exemplary embodiment Details, and without departing from the spirit or essential characteristics of the present invention, it is possible to it His concrete form realizes the present invention.Therefore, no matter from the point of view of which point, all should be by embodiment Regarding exemplary as, and be nonrestrictive, the scope of the present invention is by claims Rather than described above limit, it is intended that by fall claim equivalency implication and In the range of all changes be included in the present invention.Should be by any accompanying drawing mark in claim Note is considered as limiting involved claim.Furthermore, it is to be understood that " an including " word is not excluded for other lists Unit or step, odd number is not excluded for plural number.The multiple unit stated in system claims or device Can also be realized by software or hardware by a unit or device.The first, the second word such as grade Pragmatic represents title, and is not offered as any specific order.

Claims (30)

1., for the method presenting search answer information in search interface, the method includes Following steps:
A obtains the search information from user;
B judges whether to extract entity from described search information and substance feature describes;
If c can extract entity from described search information and substance feature describes, by described energy The entity enough extracted and substance feature describe special with each question and answer entity to being comprised and entity Levy description to mate, it is thus achieved that the one or more question and answer pair matched;
D, according to the one or more question and answer pair, provides a user with and this corresponding answering of search information Case information;
Wherein, the method is further comprising the steps of:
-while described answer information is provided, provide based on described search information to described user Search Results.
Method the most according to claim 1, wherein, also includes:
-user is inputted the page of described search information be updated processing, to be updated to bag The renewal page containing described answer information.
Method the most according to claim 1, wherein, described step b comprises the following steps:
-judge whether to extract from described search information entity and substance feature description, and Whether can interpolate that out the classification that this search information comprises;
Wherein, described step c comprises the following steps:
-if entity and substance feature description can be extracted from described search information, and can sentence Breaking and the classification that this search information comprises, entity, the substance feature that can extract described describe The classification judged described in and and each question and answer described question and answer classification, entity and entity to being comprised Feature description is mated, it is thus achieved that the one or more question and answer pair matched.
Method the most according to claim 1, wherein, described step d also includes following step Rapid:
-the question and answer that if desired present to for multiple, then integrate the plurality of question and answer to generate answer letter Breath.
Method the most according to claim 1, wherein, described answer information can be presented on Below at least one position:
-Search Results Article 1;
-search suggestion;
-input method candidate bar;
-search column candidate item hurdle;
Candidate item hurdle under-WEB input field.
Method the most according to claim 1, wherein, the method is further comprising the steps of:
E obtains the question and answer content from webpage, it is judged that whether described question and answer content is definitiveness question and answer Content;
F generates question and answer pair according to the content being judged as definitiveness question and answer.
Method the most according to claim 6, wherein, described step e also includes following step Rapid:
-entity and substance feature description can be extracted by the problem judging described question and answer content, Judge whether described question and answer content is definitiveness question and answer content;
Described step f is further comprising the steps of:
-describe according to described entity and substance feature, in conjunction with the answer of described question and answer content, generate Question and answer pair.
Method the most according to claim 1, wherein, the method is further comprising the steps of:
G obtains encyclopaedia data;
H explains according to entry corresponding in described encyclopaedia data and entry and generates question and answer pair.
Method the most according to claim 8, wherein, described step h also includes following step Rapid:
-classification judging question and answer pair to be generated is explained according to described entry and entry;
-using described entry as entity, and according to described entry explain generate substance feature describe;
-combine the classification of described question and answer pair, described entity, the description of described substance feature and institute's predicate Bar is explained, generates question and answer pair.
Method the most according to claim 1, wherein, described step b also includes following step Rapid:
-by described search information with each question and answer prestored to mating, it is thus achieved that multiple question and answer pair Candidate item;
-according to user related information, one or more to candidate item is chosen from the plurality of question and answer Question and answer pair.
11. methods according to claim 10, wherein, described user related information include with Descend at least one:
-individual subscriber attribute;
-user preference is arranged;
-user search history record.
12. methods according to claim 1, wherein, described step b also includes following step Rapid:
-by described search information with each question and answer prestored to mating, it is thus achieved that multiple question and answer pair Candidate item;
-basis is the most mutual with user's, obtains one from the plurality of question and answer to candidate item Or multiple question and answer pair.
13. methods according to claim 1, wherein, the method is further comprising the steps of:
-judge whether the answer information of the question and answer pair obtained comprises answer and obtain information;
-obtain information when the answer information of the question and answer pair obtained comprises answer, then obtain according to this answer Win the confidence breath, by the corresponding answer of API Calls.
14. according to the method according to any one of claim 1 to 13, and wherein, the method is by net Network equipment completes.
15. according to the method according to any one of claim 1 to 5 and claim 10 to 12, Wherein, the method is completed by subscriber equipment.
16. 1 kinds of equipment searching for answer information for presenting in search interface, wherein, this sets For including:
First acquisition device, for obtaining the search information from user;
Second judgment means, be used for judging whether can extracting from described search information entity and Substance feature describes;
Second sub-coalignment, if for extracting entity and entity from described search information Feature description, describes the described entity that can extract and substance feature with each question and answer being wrapped The entity contained and substance feature describe and mate, it is thus achieved that the one or more question and answer pair matched;
Device is provided, for according to the one or more question and answer pair, provides a user with and this search The answer information that information is corresponding;
Search Results provides device, for while providing described answer information, to described user Search Results based on described search information is provided.
17. equipment according to claim 16, wherein, this equipment also includes:
Webpage updating device, is updated processing for user inputs the page of described search information, To be updated to comprise the renewal page of described answer information.
18. equipment according to claim 16, wherein, this equipment also includes:
3rd judgment means, be used for judging whether can extracting from described search information entity and Substance feature describes, and whether can interpolate that out the classification that this search information comprises;
Wherein, described second sub-coalignment includes:
3rd sub-coalignment, if for extracting entity and entity from described search information Feature description, and can interpolate that out the classification that this search information comprises, can extract described Entity, substance feature describe and described in the classification judged with each question and answer described to asking of being comprised Answer classification, entity and substance feature description to mate, it is thus achieved that the one or more question and answer matched Right.
19. equipment according to claim 16, wherein, described offer device also includes:
Integrating apparatus, the question and answer being used for if desired presenting to for multiple, then integrate the plurality of question and answer pair To generate answer information.
20. equipment according to claim 16, wherein, described answer information can be presented on Below at least one position:
-Search Results Article 1;
-search suggestion;
-input method candidate bar;
-search column candidate item hurdle;
Candidate item hurdle under-WEB input field.
21. equipment according to claim 16, wherein, this equipment also includes:
Second acquisition device, for obtaining the question and answer content from webpage, it is judged that described question and answer content Whether it is definitiveness question and answer content;
First generating means, for generating question and answer pair according to the content being judged as definitiveness question and answer.
22. equipment according to claim 21, wherein, described second acquisition device also includes:
Can the 4th judgment means, for extract entity by the problem judging described question and answer content And substance feature describes, judge whether described question and answer content is definitiveness question and answer content;
Described first generating means also includes:
First sub-generating means, for describing according to described entity and substance feature, in conjunction with described in ask Answer the answer of content, generate question and answer pair.
23. equipment according to claim 16, wherein, this equipment also includes:
3rd acquisition device, is used for obtaining encyclopaedia data;
Second generating means, for explaining raw according to entry corresponding in described encyclopaedia data and entry Become question and answer pair.
24. equipment according to claim 23, wherein, described second generating means includes:
5th judgment means, judges question and answer pair to be generated for explaining according to described entry and entry Classification;
3rd sub-generating means, is used for described entry as entity, and explains according to described entry Generation substance feature describes;
4th sub-generating means, for combining the classification of described question and answer pair, described entity, described reality Body characteristics describes and described entry is explained, generates question and answer pair.
25. equipment according to claim 16, wherein, described coalignment also includes:
4th sub-coalignment, for by described search information and each question and answer prestored to carrying out Join, it is thus achieved that multiple question and answer are to candidate item;
Select device, for according to user related information, selecting candidate item from the plurality of question and answer Take one or more question and answer pair.
26. equipment according to claim 25, wherein, described user related information include with Descend at least one:
-individual subscriber attribute;
-user preference is arranged;
-user search history record.
27. equipment according to claim 16, wherein, described coalignment also includes:
5th sub-coalignment, for by described search information and each question and answer prestored to carrying out Join, it is thus achieved that multiple question and answer are to candidate item;
Interactive device, further mutual, from the plurality of question and answer to candidate for according to user One or more question and answer pair are obtained in Xiang.
28. equipment according to claim 16, wherein, this equipment also includes:
-the six judgment means, for judging whether the answer information of the question and answer pair obtained comprises answer Acquisition information;
-calling device, for obtaining information when the answer information of the question and answer pair obtained comprises answer, Then obtain information according to this answer, by the corresponding answer of API Calls.
29. according to the equipment according to any one of claim 16 to 28, and wherein, this equipment is The network equipment.
30. according to setting according to any one of claim 16 to 18 and claim 25 to 28 Standby, wherein, this equipment is subscriber equipment.
CN201010271796.6A 2010-09-03 2010-09-03 For presenting the method and apparatus of search answer information in search interface Active CN101986293B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201010271796.6A CN101986293B (en) 2010-09-03 2010-09-03 For presenting the method and apparatus of search answer information in search interface

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201010271796.6A CN101986293B (en) 2010-09-03 2010-09-03 For presenting the method and apparatus of search answer information in search interface

Publications (2)

Publication Number Publication Date
CN101986293A CN101986293A (en) 2011-03-16
CN101986293B true CN101986293B (en) 2016-08-24

Family

ID=43710640

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201010271796.6A Active CN101986293B (en) 2010-09-03 2010-09-03 For presenting the method and apparatus of search answer information in search interface

Country Status (1)

Country Link
CN (1) CN101986293B (en)

Families Citing this family (35)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102214209A (en) * 2011-04-27 2011-10-12 百度在线网络技术(北京)有限公司 Method and equipment for identifying homonymous information entities
CN103186643A (en) * 2011-12-30 2013-07-03 安凯(广州)微电子技术有限公司 Autonomous learning method for realizing association of teaching contents, terminal and system
CN103838554B (en) 2012-11-21 2017-12-12 腾讯科技(北京)有限公司 The generation method and device of a kind of interactive event
CN104375845A (en) * 2013-08-14 2015-02-25 中兴通讯股份有限公司 Application startup method and device and terminal
WO2015058604A1 (en) * 2013-10-21 2015-04-30 北京奇虎科技有限公司 Apparatus and method for obtaining degree of association of question and answer pair and for search ranking optimization
CN103699590B (en) * 2013-12-09 2017-04-05 北京奇立软件技术有限公司 The method and server of graphic tutorial problem solution are provided
CN103760991B (en) * 2014-01-13 2017-02-15 北京搜狗科技发展有限公司 Physical input method and physical input device
CN103853842B (en) * 2014-03-20 2017-07-18 百度在线网络技术(北京)有限公司 A kind of automatic question-answering method and system
CN104376046A (en) * 2014-10-24 2015-02-25 北京奇虎科技有限公司 Browsing method based on query result provided by search engine and browser client-side
CN104331441A (en) * 2014-10-24 2015-02-04 北京奇虎科技有限公司 Method and device for providing answers to questions based on search engine
CN104331440A (en) * 2014-10-24 2015-02-04 北京奇虎科技有限公司 Instant messaging method and client for providing query results based on search engine
CN105786874A (en) * 2014-12-23 2016-07-20 北京奇虎科技有限公司 Method and device for constructing question-answer knowledge base data items based on encyclopedic entries
CN105786851A (en) * 2014-12-23 2016-07-20 北京奇虎科技有限公司 Question and answer knowledge base construction method as well as search provision method and apparatus
CN105786871B (en) * 2014-12-23 2019-03-19 北京奇虎科技有限公司 Question and answer class search result rendering method and device based on search term
CN105786869B (en) * 2014-12-23 2020-05-29 北京奇虎科技有限公司 Method and device for obtaining question and answer special topic data based on search
CN105786872A (en) * 2014-12-23 2016-07-20 北京奇虎科技有限公司 Method and device for providing question-answer onebox based on user searches
CN104933084B (en) * 2015-05-04 2018-11-09 上海智臻智能网络科技股份有限公司 A kind of method, apparatus and equipment for obtaining answer information
CN105117398B (en) * 2015-06-25 2018-10-26 扬州大学 A kind of software development problem auto-answer method based on crowdsourcing
CN106407198A (en) * 2015-07-28 2017-02-15 百度在线网络技术(北京)有限公司 Question and answer information processing method and device
CN106919589A (en) * 2015-12-24 2017-07-04 北京奇虎科技有限公司 Customer problem analysis method and device
CN105740362A (en) * 2016-01-26 2016-07-06 百度在线网络技术(北京)有限公司 Information display method and display apparatus
CN106168962B (en) * 2016-06-30 2020-02-21 北京奇虎科技有限公司 Search method and device for providing accurate viewpoint based on natural search result
CN106776797A (en) * 2016-11-22 2017-05-31 中国人名解放军理工大学 A kind of knowledge Q-A system and its method of work based on ontology inference
CN107590252A (en) * 2017-09-19 2018-01-16 百度在线网络技术(北京)有限公司 Method and device for information exchange
CN107798126B (en) * 2017-11-13 2021-11-02 北京邮电大学 Question-answer processing method based on knowledge base
CN108959559B (en) * 2018-06-29 2021-02-26 北京百度网讯科技有限公司 Question and answer pair generation method and device
CN109191940B (en) * 2018-08-31 2021-09-24 广东小天才科技有限公司 Interaction method based on intelligent equipment and intelligent equipment
CN109800286B (en) * 2018-12-17 2021-05-11 北京百度网讯科技有限公司 Dialog generation method and device
CN109635214B (en) * 2018-12-20 2020-11-06 广东小天才科技有限公司 Learning resource pushing method and electronic equipment
CN109710747B (en) * 2019-01-16 2021-04-06 北京猎户星空科技有限公司 Information processing method and device and electronic equipment
CN110246493A (en) * 2019-05-06 2019-09-17 百度在线网络技术(北京)有限公司 Address book contact lookup method, device and storage medium
CN112214692A (en) * 2019-07-11 2021-01-12 北京搜狗科技发展有限公司 Data processing method and device based on input method and machine readable medium
CN110502689A (en) * 2019-08-28 2019-11-26 上海智臻智能网络科技股份有限公司 The crawling method and device of knowledge point, storage medium, terminal
CN112579642A (en) * 2019-09-30 2021-03-30 北京国双科技有限公司 Data processing method, data processing device, storage medium and electronic equipment
CN113377934B (en) * 2021-05-21 2022-07-05 海南师范大学 System and method for realizing intelligent customer service

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1821991A (en) * 2005-02-18 2006-08-23 上海赢思软件技术有限公司 Knowledge question-and-answer quick processing system based on artificial intelligence
CN1928864A (en) * 2006-09-22 2007-03-14 浙江大学 FAQ based Chinese natural language ask and answer method
CN101118554A (en) * 2007-09-14 2008-02-06 中兴通讯股份有限公司 Intelligent interactive request-answering system and processing method thereof

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1821991A (en) * 2005-02-18 2006-08-23 上海赢思软件技术有限公司 Knowledge question-and-answer quick processing system based on artificial intelligence
CN1928864A (en) * 2006-09-22 2007-03-14 浙江大学 FAQ based Chinese natural language ask and answer method
CN101118554A (en) * 2007-09-14 2008-02-06 中兴通讯股份有限公司 Intelligent interactive request-answering system and processing method thereof

Also Published As

Publication number Publication date
CN101986293A (en) 2011-03-16

Similar Documents

Publication Publication Date Title
CN101986293B (en) For presenting the method and apparatus of search answer information in search interface
Li et al. Media representation of digital-free tourism: A critical discourse analysis
CN105068661B (en) Man-machine interaction method based on artificial intelligence and system
CN106227815B (en) Multi-modal clue personalized application program function recommendation method and system
CN103544176B (en) Method and apparatus for generating the page structure template corresponding to multiple pages
Gungwu et al. The Chinese Overseas
CN103544178B (en) It is a kind of for providing the method and apparatus of reconstruction page corresponding with target pages
CN102163198B (en) A method and a system for providing new or popular terms
CN107636648A (en) Response is constructed based on mood mark
CN106570106A (en) Method and device for converting voice information into expression in input process
CN102117317A (en) Blind person Internet system based on voice technology
CN107145496A (en) The method for being matched image with content item based on keyword
CN102970326B (en) A kind of method and apparatus of the mood indication information for sharing users
CN104809142A (en) Trademark inquiring system and method
CN102483756A (en) An assistant-adviser using the semantic analysis of community exchanges
CN102314461B (en) Navigation prompt method and system
CN102314440B (en) Utilize the method and system in network operation language model storehouse
CN106326452A (en) Method for human-machine dialogue based on contexts
CN103649953A (en) Method and system for processing a search request
CN111046272A (en) Intelligent question-answering system based on medical knowledge map
CN111553138B (en) Auxiliary writing method and device for standardizing content structure document
Hlava The taxobook: Principles and practices of building taxonomies, part 2 of a 3-part series
CN101770291B (en) Semantic analysis data hashing storage and analysis methods for input system
WO2015198112A1 (en) Processing search queries and generating a search result page including search object related information
CN108874789A (en) Generation method, device, storage medium and the electronic device of sentence

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant