CN108572971A - It is a kind of to be used to excavate and the method and apparatus of the relevant keyword of term - Google Patents
It is a kind of to be used to excavate and the method and apparatus of the relevant keyword of term Download PDFInfo
- Publication number
- CN108572971A CN108572971A CN201710138638.5A CN201710138638A CN108572971A CN 108572971 A CN108572971 A CN 108572971A CN 201710138638 A CN201710138638 A CN 201710138638A CN 108572971 A CN108572971 A CN 108572971A
- Authority
- CN
- China
- Prior art keywords
- result
- query
- history
- search
- polymerization
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Abstract
Include with the method and apparatus of the relevant keyword of term, this method for excavating the present invention provides a kind of:According to the search log information in search engine, obtaining with term query there is the history of high correlation to show result, wherein, the history show result include history show showed in the page for query historical search result and/or with query it is relevant auxiliary show result;Showed according to the history as a result, generating at least one polymerization result corresponding with query;Extraction and the relevant keywords of query from least one polymerization result.Scheme according to the present invention, introduce the guidance of user's history search behavior, and magnanimity historical search has been used for reference as a result, largely compensating for the problem of the information content deficiency of term itself, to be conducive to excavate the real keyword that can reflect user search intent.
Description
Technical field
The present invention relates to field of computer technology more particularly to a kind of for excavating and the side of the relevant keyword of term
Method and device.
Background technology
In the prior art, user mainly expresses its search intention in search by the term of its input, therefore, searches
Whether index holds up the intention understanding to term accurate very crucial.And for the keyword extraction techniques of term, it is search
The basic module that engine term understands.
The following two keyword extraction techniques of generally use at present:One, pass through various natural language processing tools, analysis inspection
The weight for each basic entry that rope word itself is included, to extract keyword therein from term;Two, it polymerize institute
There is term, is based on TF-IDF or various topic models (PLSA/LDA etc.), extracts keyword therein.Above two is crucial
Word extractive technique has the following disadvantages:On the one hand, term input by user usually compares random or even inclined colloquial style, and portion
There may be wrong word or even phonetics etc. for Divided Retrieving word, if only from term, then cannot preferably get
Match the real keyword of user search intent;On the other hand, the phase on existing magnanimity third party webpage is not used for reference fully
Knowledge is closed, it is difficult to excavate the real keyword of matching user search intent.
Invention content
The object of the present invention is to provide a kind of for excavating and the method and apparatus of the relevant keyword of term.
According to an aspect of the present invention, it provides a kind of for excavating and the method for the relevant keyword of term, wherein
This method includes:
According to the search log information in search engine, obtaining with term query there is the history of high correlation to show
As a result, wherein it includes that history shows the historical search result for query showed in the page that the history, which shows result,
And/or show result with the relevant auxiliary of query;
Showed according to the history as a result, generating at least one polymerization result corresponding with query;
Extraction and the relevant keywords of query from least one polymerization result.
According to another aspect of the present invention, it additionally provides a kind of for excavating and the dress of the relevant keyword of term
It sets, wherein the device includes:
First obtains device, for according to the search log information in search engine, obtain to have height with term query
The history of correlation shows result, wherein the history show result include history show showed in the page be directed to query
Historical search result and/or with query it is relevant auxiliary show result;
Generating means, for being showed according to the history as a result, generating at least one polymerization knot corresponding with query
Fruit;
First extraction element, for the extraction from least one polymerization result and the relevant keywords of query.
Compared with prior art, the present invention has the following advantages:It can be based on the history with term with high correlation
Show as a result, to polymerize to obtain at least one polymerization result corresponding with term, and from least one polymerization result
Extraction and the relevant keyword of term, the excavation and the scheme of the relevant keyword of term introduce user's history search row
For guidance, and use for reference magnanimity historical search as a result, largely compensating for the problem of the information content deficiency of term itself,
To be conducive to excavate the real keyword that can reflect user search intent.In addition, when user initiates actual search, it can be first
Find under line excavate obtain initiate to search for again with the relevant keyword of term input by user, so as to be carried for user
For higher-quality search service;If also, related to term to excavate based on the nearest historical search behavior of mass users
Keyword, then in the actual search of user based on it is such excavate obtain searched with what the relevant keyword of term was obtained
Rope is as a result, be more likely to meet the real-time search need of user.In addition, the excavation and the scheme of the relevant keyword of term are answered
When being triggered for the advertisement during actual search, search advertisements triggering ratio can be substantially improved, it is a large amount of to promote search flows
Realization efficiency.
Description of the drawings
By reading a detailed description of non-restrictive embodiments in the light of the attached drawings below, of the invention other
Feature, objects and advantages will become more apparent upon:
Fig. 1 is illustrating with the flow of the method for the relevant keyword of term for excavating for one embodiment of the invention
Figure;
Fig. 2 is illustrating with the flow of the method for the relevant keyword of term for excavating for another embodiment of the present invention
Figure;
Fig. 3 is one embodiment of the invention for excavating and the structural representation of the device of the relevant keyword of term
Figure;
Fig. 4 is another embodiment of the present invention for excavating and the structural representation of the device of the relevant keyword of term
Figure.
Same or analogous reference numeral represents same or analogous component in attached drawing.
Specific implementation mode
It should be mentioned that some exemplary embodiments are described as before exemplary embodiment is discussed in greater detail
The processing described as flow chart or method.Although operations are described as the processing of sequence by flow chart, therein to be permitted
Multioperation can be implemented concurrently, concomitantly or simultaneously.In addition, the sequence of operations can be rearranged.When it
The processing can be terminated when operation completion, it is also possible to the additional step being not included in attached drawing.The processing
It can correspond to method, function, regulation, subroutine, subprogram etc..
Alleged " computer equipment " within a context, also referred to as " computer ", referring to can be by running preset program or referring to
Enable and execute the intelligent electronic device of the predetermined process process such as numerical computations and/or logical calculated, may include processor with
Memory executes the program instruction that prestores in memory to execute predetermined process process by processor, or by ASIC,
The hardware such as FPGA, DSP execute predetermined process process, or are realized by said two devices combination.
The computer equipment is for example including user equipment and the network equipment.Wherein, the user equipment includes but unlimited
In PC machine, tablet computer, smart mobile phone, PDA etc.;The network equipment includes but not limited to single network server, multiple nets
The server group of network server composition or based on cloud computing (Cloud Computing) by a large amount of computers or network server
The cloud of composition, wherein cloud computing is one kind of Distributed Calculation, and one be made of the computer collection of a group loose couplings is super
Virtual machine.Wherein, the computer equipment can isolated operation realize the present invention, also can access network and by with network
In the interactive operations of other computer equipments realize the present invention.Wherein, the network residing for the computer equipment include but
It is not limited to internet, wide area network, Metropolitan Area Network (MAN), LAN, VPN network etc..
It should be noted that the user equipment, the network equipment and network etc. are only for example, other are existing or from now on may be used
The computer equipment that can occur such as is applicable to the present invention, should also be included within the scope of the present invention, and by reference
It is incorporated herein.
The method (some of them illustrated by flow) discussed herein below can by hardware, software, firmware, in
Between part, microcode, hardware description language or its arbitrary combination implement.When with software, firmware, middleware or microcode come real
Shi Shi, program code or code segment to implement necessary task can be stored in machine or computer-readable medium (such as
Storage medium) in.(one or more) processor can implement necessary task.
Specific structure and function details disclosed herein are only representative, and are for describing the present invention show
The purpose of example property embodiment.But the present invention can be implemented by many alternative forms, and be not interpreted as
It is limited only by the embodiments set forth herein.
Although it should be understood that may have been used term " first ", " second " etc. herein to describe each unit,
But these units should not be limited by these terms.The use of these items is only for by a unit and another unit
It distinguishes.For example, without departing substantially from the range of exemplary embodiment, it is single that first unit can be referred to as second
Member, and similarly second unit can be referred to as first unit.Term "and/or" used herein above include one of them or
The arbitrary and all combination of more listed associated items.
Term used herein above is not intended to limit exemplary embodiment just for the sake of description specific embodiment.Unless
Context clearly refers else, otherwise singulative used herein above "one", " one " also attempt to include plural number.Also answer
When understanding, term " include " and or " include " used herein above provide stated feature, integer, step, operation,
The presence of unit and/or component, and do not preclude the presence or addition of other one or more features, integer, step, operation, unit,
Component and/or a combination thereof.
It should further be mentioned that in some replace implementations, the function action being previously mentioned can be according to different from attached
The sequence indicated in figure occurs.For example, involved function action is depended on, the two width figures shown in succession actually may be used
Substantially simultaneously to execute or can execute in a reverse order sometimes.
Present invention is further described in detail below in conjunction with the accompanying drawings.
Fig. 1 is illustrating with the flow of the method for the relevant keyword of term for excavating for one embodiment of the invention
Figure.The method of the present embodiment is mainly implemented by the network equipment.
Method according to the present embodiment includes step S1, step S2 and step S3.
In step sl, the network equipment has according to the search log information in search engine, acquisition with term query
The history of high correlation shows result.
Wherein, described search log information includes that the historical search in search engine for user is operated and generated any
Log information is such as operated for the historical search of user, and log information includes term query input by user, examines
The rope time and shows the history for the search result of query to user and show the page;It should be noted that the history exhibition
Further include other information in the existing page in addition to the historical search result for query, such as shows with the relevant auxiliary of query
As a result input frame, classification for search information, website logo information etc., are searched for.
Wherein, the history show result include history show in the page with query is relevant any shows result items.It is excellent
Selection of land, the history show result include history show showed in the page for query historical search result and/or with
The relevant auxiliary of query shows result;Wherein, the auxiliary shows result expression history and shows in the page except historical search result
In addition related to query shows result items, is such as located at search result and shows promotion message on the right side of region, is located at result presentation
Relevant search recommendation information below region etc..
Wherein, various ways can be used according to the search log information in search engine, acquisition and term in the network equipment
There is query the history of high correlation to show result.
For example, the network equipment all history corresponding to query are showed be accessed by the user in result show result
Item as with query there is the history of high correlation to show result.
In another example the network equipment will show predetermined number accessed recently in result with all history corresponding to query
Amount, which shows result items, as with query there is the history of high correlation to show result.
As a preferred embodiment, the network equipment is according to the search log information in search engine, and combines desired indicator
Information, obtaining with query there is the history of high correlation to show result.
Wherein, the desired indicator information includes any and the relevant information of desired indicator, and the desired indicator is for sentencing
Disconnected query shows the correlation between result with history.Preferably, the desired indicator includes at least one of following:
The history amount of showing;
History shows position;
History click volume;
History clicks Annual distribution.
Specifically, the network equipment obtains all exhibitions corresponding with query according to the search log information in search engine
Existing result items, and combine desired indicator information, from it is described it is all show the going through with high correlation with query is determined in result items
History shows result.
As an example, desired indicator information includes multiple desired indicators for judging correlation, and each pre-
Determining the index condition to be met, (such as history amount of showing is more than the predetermined amount of showing, and it is that search result shows area that history, which shows position,
Column is promoted on domain or right side, and history click volume is more than predetermined click volume, and it is most that history clicks the Annual distribution instruction history click time
In nearest one week etc.), then when the value for showing the corresponding multiple desired indicator of result items when one is satisfied by condition, determine
It is with query there is the history of high correlation to show result that this, which shows result items,.
As another example, desired indicator information includes multiple desired indicators for judging correlation, and each
Weight corresponding to desired indicator, then according to the value for showing the corresponding multiple desired indicator of result items and each predetermined finger
The corresponding weight of mark, calculates the weights for showing result, when the weights for showing result are more than predetermined threshold, determines the exhibition
Existing result items are with query there is the history of high correlation to show result.
It should be noted that the above-mentioned examples are merely illustrative of the technical solutions of the present invention, rather than to the limit of the present invention
System, it should be appreciated by those skilled in the art that any search log information according in search engine, obtains and term query
History with high correlation shows the realization method of result, should be included in the scope of the present invention.
In step s 2, the network equipment shows according to the history as a result, generating corresponding with query at least one poly-
Close result.
Wherein, the polymerization result can behave as a variety of data modes, it is preferable that the polymerization result shows as document shape
Formula.
Specifically, the network equipment shows according to the history as a result, generating at least one polymerization knot corresponding with query
The realization method of fruit includes but not limited to:
1) history show result include history show the historical search for query showed in the page as a result,
For the network equipment according to the query and multiple urls corresponding with the historical search result, structure is multiple<Query, url
>Right, later, the network equipment is according to the multiple<Query, url>, it polymerize and obtains a polymerization result corresponding with query.
Wherein,<Query, url>It indicates term query and links pair of url compositions.For example, term " Conan " with
Url1 can build to obtain<Conan, url1>, also that is, the value of query is " Conan ", and the value of url is " url1 ".Its
In, include to show letter corresponding to each url in the multiple url in the polymerization result being polymerize based on this realization method
It ceases, title, abstract etc. in the page as corresponding to each url.
As an example, include 5 search results with the relevant historical search results of term query1 high, this 5
A search result corresponds to following 5 url respectively:Url1, url2, url3, url4, url5, then the network equipment according to query1 with
And above-mentioned 5 url, structure are 5 following<Query, url>It is right:<Query1, url1>、<Query1, ur2>、<Query1, ur3
>、<Query1, ur4>、<Query1, ur5>;Later, obtained according to above-mentioned structure 5 of the network equipment<Query, url>It is right,
Polymerization obtains a polymerization result corresponding with query1<Query, document>,<Query, document>Include upper
State the corresponding title of 5 url institutes.
2) history show result include history show showed in the page with query it is relevant auxiliary show knot
Fruit, the network equipment show all contents that show in result to the auxiliary and polymerize, and obtain one corresponding with query
Polymerization result.
For example, it includes that the auxiliary positioned at right side popularization column shows knot that the corresponding history of term " Conan ", which shows result,
Fruit, the auxiliary show result and show content comprising following:" the sub- elder sister of the people ", " rice draw princess ", " punching arrow is high " ..., " Conan's final result
Or dead very important person ", " Conan's animation new op by spit slot ", " Conan play worm continues surging ", " Conan, which sees more, to fall ill ", " Ke
Slot is spat by online friend in south " again.Then the network equipment by it is above-mentioned show it is content-aggregated be a polymerization result corresponding with " Conan ".
It should be noted that the above-mentioned examples are merely illustrative of the technical solutions of the present invention, rather than to the limit of the present invention
System, it should be appreciated by those skilled in the art that any show according to the history as a result, generating corresponding with query at least one
The realization method of a polymerization result, should be included in the scope of the present invention.
It should be noted that preferably, under different retrieval granularities, the network equipment executes step S2 respectively, to generate
Polymerization result corresponding with each retrieval granularity.
In step s3, the network equipment extracts and the relevant keywords of query from least one polymerization result.
Specifically, the network equipment can be used various ways extracted from least one polymerization result it is related to query
Keyword.
As a kind of realization method, for each word in query, the network equipment extracts at least one polymerization result
In include the word of the word, as with the relevant keywords of query.
As another realization method, the network equipment extracts multiple basic entries from least one polymerization result;
Then, the network equipment obtains the multinomial predetermined characteristic of each basic entry;Later, the network equipment is from the multiple basic entry
Extract the part basis entry that multinomial predetermined characteristic is satisfied by predetermined condition, as with the relevant keywords of query.
As another realization method, the network equipment extracts multiple basic entries from least one polymerization result;
Then, for the basic entry of each of the multiple basic entry, according to the feature calculation of the basis entry basis entry
Weights;Later, according to the corresponding weights of the multiple basis entry institute being calculated, from multiple basic entry
Extraction and the relevant keywords of query.This realization method will be described in detail in subsequent embodiment.
Wherein, it can be denoted as with the relevant keywords of query<Query, keywords>, wherein keywords indicates crucial
Word or keyword set.
It should be noted that the above-mentioned examples are merely illustrative of the technical solutions of the present invention, rather than to the limit of the present invention
System, it should be appreciated by those skilled in the art that any extraction from least one polymerization result and the relevant keywords of query
Realization method, should be included in the scope of the present invention.
It should be noted that operation of the network equipment by lower execution step S1, step S2 and step S3 online, to excavate
Or update and the relevant keywords of query, the network equipment after executing step S3 every time, by query and with query phases
The keyword of pass stores into database (database can be local data base, cloud database), in order in user's reality
It can be used in searching the term relevant keyword actually entered with user when search.
As a preferred embodiment, the method for the present embodiment further includes the step S4 executed in user's actual search and step
Rapid S5.
In step s 4, the network equipment is searched and the term phase input by user according to term input by user
The keyword of pass.
Specifically, the term that the network equipment is currently inputted according to the user received, in the database inquiry and inspection
The relevant keyword of rope word.
For example, lower the last step S1, step S2 and the step S3 of executing excavates to obtain and query1 the network equipment online
Relevant following keyword:Key1, key2, key3, and will be updated to the relevant keywords of query1 in database key1,
key2、key3;When user inputs term query1, the network equipment is according to the query1 input by user received, in data
In library search obtain be with the relevant keywords of query1:key1、key2、key3.
It should be noted that preferably, when the network equipment does not inquire and term input by user in the database
When relevant keyword, can by database with the relevant keyword of term matching degree input by user highest query institutes,
As with the relevant keyword of term input by user.
In step s 5, the network equipment be based on the term input by user and with the term input by user
Relevant keyword initiates search, and search result is supplied to the user.
For example, in step s 4, the network equipment is searched in the database according to the query1 input by user received
It is to the relevant keywords of query1:key1、key2、key3;In step s 5, the network equipment be based on query1, key1,
Key2, key3 initiate search, to obtain search result, and search result are supplied to user.
It, can be based on showing as a result, to polymerize with history of the term with high correlation according to the scheme of the present embodiment
At least one polymerization result corresponding with term is obtained, and extraction is related to term from least one polymerization result
Keyword, the scheme of the excavation and the relevant keyword of term introduces the guidance of user's history search behavior, and uses for reference
Magnanimity historical search is as a result, largely compensate for the problem of the information content deficiency of term itself, to be conducive to excavate
It can reflect the real keyword of user search intent.In addition, when user initiates actual search, it can first find under line and excavate
What is obtained initiates to search for again with the relevant keyword of term input by user, so as to provide higher-quality search to the user
Rope service;If also, based on the nearest historical search behavior of mass users come excavate with the relevant keyword of term, with
Based on such search result obtained with the relevant keyword of term for excavating and obtaining in the actual search at family, it is more likely to full
The real-time search need of sufficient user.In addition, the scheme of the excavation and the relevant keyword of term is applied to actual search mistake
When advertisement in journey triggers, search advertisements triggering ratio, a large amount of realization efficiency for promoting search flows can be substantially improved.
It should be noted that in the prior art usually during user's actual search based on keyword input by user come
It determines and needs the auxiliary showed for user to show as a result, and never considering that the auxiliary that history shows, which is oppositely showed result, used
It is searched in user, and customer service of the present invention technology prejudice, it was found that since the auxiliary showed for term institute history shows knot
Fruit is relevant with term, and such auxiliary shows the reality that result may meet user for term to a certain extent
Search need, therefore, can show result based on the auxiliary showed for term institute history to excavate may meet user's reality
The keyword of search need.
Fig. 2 is illustrating with the flow of the method for the relevant keyword of term for excavating for another embodiment of the present invention
Figure.Method according to the present embodiment includes step S1, step S2 and step S3, wherein the step S3 further comprises step
S31, step S32 and step S33.Wherein, the step S1 and step S2 has given in detail with reference to the embodiment shown in FIG. 1
It states, details are not described herein.
In the step S31, the network equipment extracts multiple basic entries from least one polymerization result.
Wherein, the network equipment can be used various ways and extract multiple basic entries from least one polymerization result.
For example, for each polymerization result at least one polymerization result, the network equipment to the polymerization result into
Row word segmentation processing obtains multiple basic entries corresponding with the polymerization result.
In another example for each polymerization result at least one polymerization result, the network equipment extracts the polymerization knot
Highest 3 words of the frequency of occurrences in fruit.
It should be noted that the above-mentioned examples are merely illustrative of the technical solutions of the present invention, rather than to the limit of the present invention
System, it should be appreciated by those skilled in the art that any realization for extracting multiple basic entries from least one polymerization result
Mode should be included in the scope of the present invention.
In step s 32, for the basic entry of each of the multiple basic entry, the network equipment is according to the basis word
The weights of the feature calculation of the item basis entry.
Wherein, the weights of basic entry are used to characterize the importance of basic entry.
Preferably, the feature of basic entry includes at least one of following:
1) part of speech/important level of basic entry.Different parts of speech can correspond to different important levels, such as verb is important
Grade highest, noun take second place, and preposition, adverbial word, the important level of modal particle are minimum.
2) TF-IDF (Term Frequency-InverseDocument of the basic entry in polymerization result
Frequency, the reverse document-frequency of word frequency -) feature.Wherein, frequency of occurrence of the basic entry in a polymerization result is TF,
Frequency of occurrence in all polymerization results is IDF, and the importance of the more high then basic entries of TF is higher, the more low then basic words of IDF
The importance of item is lower.
3) user behavior characteristics showed corresponding to result items where basic entry.Wherein, the user behavior characteristics
Include with user for the relevant any feature of the behavior showed performed by result items where basic entry, when being accessed such as user
Between, in access times, the page operation etc..
4) appearance situation of the basic entry in query.Wherein, described situation occur there are following three kinds of situations:Basis
Entry appears in query, and the partial content in basic entry appears in query, and basic entry and its partial content are not
It appears in query.The importance of basic entry is successively decreased successively according to the sequence of above-mentioned three kinds of situations.
It should be noted that the feature of above-mentioned basis entry is only for example, and the limitation of non-present invention, people in the art
Member is it will be understood that any of basic entry can be used in judging that the feature of its importance (shows result as where basic entry
The type of item, the type for showing result items includes but not limited to commercial results, natural result, Extended Results etc., for another example, basic word
The commercial attribute of item, the commercial attribute include business type, whether purchased, class of trade etc.), it should be included in the present invention
In the range of the feature of the basic entry.
Specifically, various ways root can be used for each of the multiple basic entry basic entry, the network equipment
According to the weights of the feature calculation of the basis entry basis entry.
For example, determining each of the basis entry for each of the multiple basic entry basic entry, the network equipment
The corresponding weights of feature institute, later, the network equipment by the various features of the basis entry corresponding weights phase
Add, obtains the weights for calculating the basis entry.
In another example can be further directed to different feature-set weighting coefficients, for every in the multiple basic entry
A basis entry, the network equipment determine the corresponding weights of various features institute of the basis entry, and later, the network equipment is based on
The various features of the basis entry corresponding weights and various features weighting coefficient, to calculate the power of basic entry
Value.
It should be noted that the above-mentioned examples are merely illustrative of the technical solutions of the present invention, rather than to the limit of the present invention
System, it should be appreciated by those skilled in the art that it is any for the basic entry of each of the multiple basic entry, according to the basis
The realization method of the weights of the feature calculation of the entry basis entry, should be included in the scope of the present invention.
In step S33, the network equipment according to the corresponding weights of the multiple basic entry institute being calculated, from
Extraction and the relevant keywords of query in multiple basis entry.
Specifically, the network equipment can be used various ways and be corresponded to respectively according to the multiple basic entry being calculated
Weights, extraction and the relevant keywords of query from multiple basic entry.
For example, the network equipment is ranked up the multiple basic entry according to corresponding weights from high to low, and cut
Take the preceding N number of basic entry of sequence as with the relevant N number of keywords of query.
In another example the network equipment extracts M of corresponding weights higher than predetermined weights from the multiple basic entry
Basic entry, as with the relevant M keyword of query.
It should be noted that the above-mentioned examples are merely illustrative of the technical solutions of the present invention, rather than to the limit of the present invention
System, it should be appreciated by those skilled in the art that the corresponding weights of the multiple basis entry institute that any basis is calculated,
The realization method of extraction and the relevant keywords of query, should be included in the scope of the present invention from multiple basic entry.
According to the scheme of the present embodiment, multiple basic entries are first extracted from least one polymerization result that polymerization obtains,
And according to its weights of the feature calculation of each basic entry, and then based on the weights of each basic entry come from multiple basic word
In item extraction with the relevant keyword of term so that excavate obtain be more likely to reflect with the relevant keyword of term
Or influence the real-time search need that user is directed to term.
Fig. 3 is one embodiment of the invention for excavating and the structural representation of the device of the relevant keyword of term
Figure.This includes the first acquisition device with the device of the relevant keyword of term (hereinafter referred to as " excavating gear ") for excavating
1, generating means 2 and the first extraction element 3.
The first of the network equipment obtains device 1 according to the search log information in search engine, obtains and term query
History with high correlation shows result.
Wherein, described search log information includes that the historical search in search engine for user is operated and generated any
Log information is such as operated for the historical search of user, and log information includes term query input by user, examines
The rope time and shows the history for the search result of query to user and show the page;It should be noted that the history exhibition
Further include other information in the existing page in addition to the historical search result for query, such as shows with the relevant auxiliary of query
As a result input frame, classification for search information, website logo information etc., are searched for.
Wherein, the history show result include history show in the page with query is relevant any shows result items.It is excellent
Selection of land, the history show result include history show showed in the page for query historical search result and/or with
The relevant auxiliary of query shows result;Wherein, the auxiliary shows result expression history and shows in the page except historical search result
In addition related to query shows result items, is such as located at search result and shows promotion message on the right side of region, is located at result presentation
Relevant search recommendation information below region etc..
Wherein, various ways can be used according to the search log information in search engine in the first acquisition device 1, obtain and examine
There is rope word query the history of high correlation to show result.
For example, all history corresponding to query are showed showing of being accessed by the user in result by the first acquisition device 1
Result items as with query there is the history of high correlation to show result.
In another example the first acquisition device 1 will show in result with all history corresponding to query it is accessed recently pre-
Fixed number amount, which shows result items, as with query there is the history of high correlation to show result.
As a preferred embodiment, the first acquisition device 1 further comprises the second acquisition device (not shown), and second obtains
Device combines desired indicator information according to the search log information in search engine, and obtain has high correlation with query
History shows result.
Wherein, the desired indicator information includes any and the relevant information of desired indicator, and the desired indicator is for sentencing
Disconnected query shows the correlation between result with history.Preferably, the desired indicator includes at least one of following:
The history amount of showing;
History shows position;
History click volume;
History clicks Annual distribution.
Specifically, second device is obtained according to the search log information in search engine, obtain institute corresponding with query
Have and show result items, and combine desired indicator information, from it is described it is all show in result items determine with query have high correlation
History show result.
As an example, desired indicator information includes multiple desired indicators for judging correlation, and each pre-
Determining the index condition to be met, (such as history amount of showing is more than the predetermined amount of showing, and it is that search result shows area that history, which shows position,
Column is promoted on domain or right side, and history click volume is more than predetermined click volume, and it is most that history clicks the Annual distribution instruction history click time
In nearest one week etc.), then when the value for showing the corresponding multiple desired indicator of result items when one is satisfied by condition, second
It is with query there is the history of high correlation to show result to obtain device and determine that this shows result items.
As another example, desired indicator information includes multiple desired indicators for judging correlation, and each
Weight corresponding to desired indicator, then according to the value for showing the corresponding multiple desired indicator of result items and each predetermined finger
The corresponding weight of mark, calculates the weights for showing result, and when the weights for showing result are more than predetermined threshold, second obtains
It is with query there is the history of high correlation to show result that device, which determines that this shows result items,.
It should be noted that the above-mentioned examples are merely illustrative of the technical solutions of the present invention, rather than to the limit of the present invention
System, it should be appreciated by those skilled in the art that any search log information according in search engine, obtains and term query
History with high correlation shows the realization method of result, should be included in the scope of the present invention.
The generating means 2 of the network equipment show according to the history as a result, generating corresponding with query at least one poly-
Close result.
Wherein, the polymerization result can behave as a variety of data modes, it is preferable that the polymerization result shows as document shape
Formula.
Specifically, generating means 2 show according to the history as a result, generating at least one polymerization corresponding with query
As a result realization method includes but not limited to:
1) generating means 2 further comprise construction device (not shown) and the first polyplant (not shown).The history
It includes that history shows the historical search for query showed in the page as a result, construction device is according to show result
Query and multiple urls corresponding with the historical search result, structure are multiple<Query, url>Right, later, first is poly-
It attaches together and sets according to the multiple<Query, url>, it polymerize and obtains a polymerization result corresponding with query.
Wherein,<Query, url>It indicates term query and links pair of url compositions.For example, term " Conan " with
Url1 can build to obtain<Conan, url1>, also that is, the value of query is " Conan ", and the value of url is " url1 ".Its
In, include to show letter corresponding to each url in the multiple url in the polymerization result being polymerize based on this realization method
It ceases, title, abstract etc. in the page as corresponding to each url.
As an example, include 5 search results with the relevant historical search results of term query1 high, this 5
A search result corresponds to following 5 url respectively:Url1, url2, url3, url4, url5, then construction device according to query1 with
And above-mentioned 5 url, structure are 5 following<Query, url>It is right:<Query1, url1>、<Query1, ur2>、<Query1, ur3
>、<Query1, ur4>、<Query1, ur5>;Later, obtained according to above-mentioned structure 5 of the first polyplant<Query, url>
Right, polymerization obtains a polymerization result corresponding with query1<Query, document>,<Query, document>Middle packet
Include the corresponding title of above-mentioned 5 url institutes.
2) generating means 2 further comprise the second polyplant (not shown).It includes history exhibition that the history, which shows result,
What is showed in the existing page shows with the relevant auxiliary of query as a result, the second polyplant shows in result the auxiliary
All contents that show are polymerize, and a polymerization result corresponding with query is obtained.
For example, it includes that the auxiliary positioned at right side popularization column shows knot that the corresponding history of term " Conan ", which shows result,
Fruit, the auxiliary show result and show content comprising following:" the sub- elder sister of the people ", " rice draw princess ", " punching arrow is high " ..., " Conan's final result
Or dead very important person ", " Conan's animation new op by spit slot ", " Conan play worm continues surging ", " Conan, which sees more, to fall ill ", " Ke
Slot is spat by online friend in south " again.Then the second polyplant by it is above-mentioned show it is content-aggregated for one corresponding with " Conan " polymerization tie
Fruit.
It should be noted that the above-mentioned examples are merely illustrative of the technical solutions of the present invention, rather than to the limit of the present invention
System, it should be appreciated by those skilled in the art that any show according to the history as a result, generating corresponding with query at least one
The realization method of a polymerization result, should be included in the scope of the present invention.
It should be noted that preferably, under different retrieval granularities, generating means 2 execute operation respectively, with generate with
The corresponding polymerization result of each retrieval granularity.
First extraction element 3 of the network equipment extracts and the relevant keys of query from least one polymerization result
Word.
Specifically, the first extraction element 3 can be used various ways and be extracted from least one polymerization result and query
Relevant keyword.
As a kind of realization method, for each word in query, the first extraction element 3 extracts described at least one poly-
Close and include the word of the word in result, as with the relevant keywords of query.
As another realization method, the first extraction element 3 extracts multiple bases from least one polymerization result
Entry;Then, the multinomial predetermined characteristic of each basic entry is obtained;Later, it is extracted from the multiple basic entry multinomial pre-
Determine the part basis entry that feature is satisfied by predetermined condition, as with the relevant keywords of query.
As another realization method, the first extraction element 3 extracts multiple bases from least one polymerization result
Entry;Then, for the basic entry of each of the multiple basic entry, according to the feature calculation of the basis entry basis
The weights of entry;Later, according to the corresponding weights of the multiple basis entry institute being calculated, from multiple basic word
Extraction and the relevant keywords of query in item.This realization method will be described in detail in subsequent embodiment.
Wherein, it can be denoted as with the relevant keywords of query<Query, keywords>, wherein keywords indicates crucial
Word or keyword set.
It should be noted that the above-mentioned examples are merely illustrative of the technical solutions of the present invention, rather than to the limit of the present invention
System, it should be appreciated by those skilled in the art that any extraction from least one polymerization result and the relevant keywords of query
Realization method, should be included in the scope of the present invention.
It should be noted that the first acquisition device 1, generating means 2 and the first extraction element 3 execute behaviour by lower online
Make, to excavate or update and the relevant keywords of query, after the first extraction element 3 executes operation every time, the network equipment
Stored into database by query and with the relevant keywords of query (database can be local data base, cloud database
Deng), in order to can be used in searching the relevant keyword of term actually entered with user in user's actual search.
As a preferred embodiment, the excavating gear of the present embodiment further includes the lookup dress executed in user's actual search
Set (not shown) and searcher (not shown).
Device is searched according to term input by user, is searched and the relevant keyword of term input by user.
Specifically, the term that device is currently inputted according to the user received is searched, in the database inquiry and the inspection
The relevant keyword of rope word.
For example, the first acquisition device 1, generating means 2 and the first extraction element 3 are dug under line after the last execution operation
Pick obtains and the relevant following keywords of query1:Key1, key2, key3, and by database with the relevant keys of query1
Word is updated to key1, key2, key3;When user inputs term query1, lookup device is input by user according to what is received
Query1, in the database search obtain be with the relevant keywords of query1:key1、key2、key3.
It should be noted that preferably, not inquired in the database and term input by user when searching device
When relevant keyword, can by database with the relevant keyword of term matching degree input by user highest query institutes,
As with the relevant keyword of term input by user.
Searcher be based on the term input by user and with the relevant key of term input by user
Word initiates search, and search result is supplied to the user.
For example, search device according to the query1 input by user received, lookup in the database obtains and query1
Relevant keyword is:key1、key2、key3;Searcher is based on query1, key1, key2, key3 and initiates search, to obtain
Search result is obtained, and search result is supplied to user.
It, can be based on showing as a result, to polymerize with history of the term with high correlation according to the scheme of the present embodiment
At least one polymerization result corresponding with term is obtained, and extraction is related to term from least one polymerization result
Keyword, the scheme of the excavation and the relevant keyword of term introduces the guidance of user's history search behavior, and uses for reference
Magnanimity historical search is as a result, largely compensate for the problem of the information content deficiency of term itself, to be conducive to excavate
It can reflect the real keyword of user search intent.In addition, when user initiates actual search, it can first find under line and excavate
What is obtained initiates to search for again with the relevant keyword of term input by user, so as to provide higher-quality search to the user
Rope service;If also, based on the nearest historical search behavior of mass users come excavate with the relevant keyword of term, with
Based on such search result obtained with the relevant keyword of term for excavating and obtaining in the actual search at family, it is more likely to full
The real-time search need of sufficient user.In addition, the scheme of the excavation and the relevant keyword of term is applied to actual search mistake
When advertisement in journey triggers, search advertisements triggering ratio, a large amount of realization efficiency for promoting search flows can be substantially improved.
It should be noted that in the prior art usually during user's actual search based on keyword input by user come
It determines and needs the auxiliary showed for user to show as a result, and never considering that the auxiliary that history shows, which is oppositely showed result, used
It is searched in user, and customer service of the present invention technology prejudice, it was found that since the auxiliary showed for term institute history shows knot
Fruit is relevant with term, and such auxiliary shows the reality that result may meet user for term to a certain extent
Search need, therefore, can show result based on the auxiliary showed for term institute history to excavate may meet user's reality
The keyword of search need.
Fig. 4 is another embodiment of the present invention for excavating and the structural representation of the device of the relevant keyword of term
Figure.Include the first acquisition device 1, generating means 2 and the first extraction element 3 according to the excavating gear of the present embodiment, wherein described
First extraction element 3 further comprises the second extraction element 31, computing device 32 and third extraction element 33.Wherein, described
One acquisition device 1 and generating means 2 are described in detail with reference to the embodiment shown in FIG. 3, and details are not described herein.
Second extraction element 31 extracts multiple basic entries from least one polymerization result.
Wherein, the second extraction element 31 can be used various ways and extract multiple bases from least one polymerization result
Entry.
For example, for each polymerization result at least one polymerization result, the second extraction element 31 is to the polymerization
As a result word segmentation processing is carried out, multiple basic entries corresponding with the polymerization result are obtained.
In another example for each polymerization result at least one polymerization result, the extraction of the second extraction element 31 should
Highest 3 words of the frequency of occurrences in polymerization result.
It should be noted that the above-mentioned examples are merely illustrative of the technical solutions of the present invention, rather than to the limit of the present invention
System, it should be appreciated by those skilled in the art that any realization for extracting multiple basic entries from least one polymerization result
Mode should be included in the scope of the present invention.
For the basic entry of each of the multiple basic entry, computing device 32 is according to the feature meter of the basis entry
Calculate the weights of the basis entry.
Wherein, the weights of basic entry are used to characterize the importance of basic entry.
Preferably, the feature of basic entry includes at least one of following:
1) part of speech/important level of basic entry.Different parts of speech can correspond to different important levels, such as verb is important
Grade highest, noun take second place, and preposition, adverbial word, the important level of modal particle are minimum.
2) TF-IDF feature of the basic entry in polymerization result.Wherein, basic entry going out in a polymerization result
The existing frequency is TF, and the frequency of occurrence in all polymerization results is IDF, and the importance of the more high then basic entries of TF is higher, and IDF is got over
The importance of low then basic entry is lower.
3) user behavior characteristics showed corresponding to result items where basic entry.Wherein, the user behavior characteristics
Include with user for the relevant any feature of the behavior showed performed by result items where basic entry, when being accessed such as user
Between, in access times, the page operation etc..
4) appearance situation of the basic entry in query.Wherein, described situation occur there are following three kinds of situations:Basis
Entry appears in query, and the partial content in basic entry appears in query, and basic entry and its partial content are not
It appears in query.The importance of basic entry is successively decreased successively according to the sequence of above-mentioned three kinds of situations.
It should be noted that the feature of above-mentioned basis entry is only for example, and the limitation of non-present invention, people in the art
Member is it will be understood that any of basic entry can be used in judging that the feature of its importance (shows result as where basic entry
The type of item, the type for showing result items includes but not limited to commercial results, natural result, Extended Results etc., for another example, basic word
The commercial attribute of item, the commercial attribute include business type, whether purchased, class of trade etc.), it should be included in the present invention
In the range of the feature of the basic entry.
Specifically, various ways can be used for each of the multiple basic entry basic entry, computing device 32
According to the weights of the feature calculation of the basis entry basis entry.
For example, determining the basis entry for each of the multiple basic entry basic entry, computing device 32
The corresponding weights of various features institute, later, computing device 32 by the various features of the basis entry corresponding power
Value is added, and obtains the weights for calculating the basis entry.
In another example can be further directed to different feature-set weighting coefficients, for every in the multiple basic entry
A basis entry, computing device 32 determine the corresponding weights of various features institute of the basis entry, later, computing device 32
Various features based on the basis entry corresponding weights and various features weighting coefficient, to calculate basic entry
Weights.
It should be noted that the above-mentioned examples are merely illustrative of the technical solutions of the present invention, rather than to the limit of the present invention
System, it should be appreciated by those skilled in the art that it is any for the basic entry of each of the multiple basic entry, according to the basis
The realization method of the weights of the feature calculation of the entry basis entry, should be included in the scope of the present invention.
Third extraction element 33 is according to the corresponding weights of the multiple basic entry institute being calculated, from multiple
Extraction and the relevant keywords of query in basic entry.
Specifically, third extraction element 33 can be used various ways and be divided according to the multiple basic entry being calculated
Not corresponding weights, extraction and the relevant keywords of query from multiple basic entry.
For example, third extraction element 33 arranges the multiple basic entry according to corresponding weights from high to low
Sequence, and intercept the preceding N number of basic entry of sequence as with the relevant N number of keywords of query.
In another example third extraction element 33 extracts corresponding weights from the multiple basic entry is higher than predetermined weights
M basic entry, as with the relevant M keyword of query.
It should be noted that the above-mentioned examples are merely illustrative of the technical solutions of the present invention, rather than to the limit of the present invention
System, it should be appreciated by those skilled in the art that the corresponding weights of the multiple basis entry institute that any basis is calculated,
The realization method of extraction and the relevant keywords of query, should be included in the scope of the present invention from multiple basic entry.
According to the scheme of the present embodiment, multiple basic entries are first extracted from least one polymerization result that polymerization obtains,
And according to its weights of the feature calculation of each basic entry, and then based on the weights of each basic entry come from multiple basic word
In item extraction with the relevant keyword of term so that excavate obtain be more likely to reflect with the relevant keyword of term
Or influence the real-time search need that user is directed to term.
It is obvious to a person skilled in the art that invention is not limited to the details of the above exemplary embodiments, Er Qie
In the case of without departing substantially from spirit or essential attributes of the invention, the present invention can be realized in other specific forms.Therefore, no matter
From the point of view of which point, the present embodiments are to be considered as illustrative and not restrictive, and the scope of the present invention is by appended power
Profit requires rather than above description limits, it is intended that all by what is fallen within the meaning and scope of the equivalent requirements of the claims
Variation includes within the present invention.Any reference signs in the claims should not be construed as limiting the involved claims.This
Outside, it is clear that one word of " comprising " is not excluded for other units or step, and odd number is not excluded for plural number.That is stated in system claims is multiple
Unit or device can also be realized by a unit or device by software or hardware.The first, the second equal words are used for table
Show title, and does not represent any particular order.
Claims (18)
1. a kind of for excavating and the method for the relevant keyword of term, wherein this method includes:
According to the search log information in search engine, obtain with term query there is the history of high correlation to show as a result,
Wherein, the history show result include history show showed in the page for query historical search result and/or with
The relevant auxiliary of query shows result;
Showed according to the history as a result, generating at least one polymerization result corresponding with query;
Extraction and the relevant keywords of query from least one polymerization result.
2. according to the method described in claim 1, wherein, it includes that history shows in the page and showed that the history, which shows result,
Historical search for query is as a result, described show according to the history as a result, generating corresponding with query at least one
The step of polymerization result includes:
According to the query and multiple urls corresponding with the historical search result, structure is multiple<Query, url>It is right;
According to the multiple<Query, url>, it polymerize and obtains a polymerization result corresponding with query.
3. according to the method described in claim 1, wherein, it includes that history shows in the page and showed that the history, which shows result,
Show as a result, described show according to the history as a result, generating corresponding with query at least one with the relevant auxiliary of query
The step of a polymerization result includes:
Show all contents that show in result to the auxiliary to polymerize, obtains polymerization knot corresponding with query
Fruit.
4. according to the method described in claim 1, wherein, extraction and the query phases from least one polymerization result
The step of keyword of pass includes:
Multiple basic entries are extracted from least one polymerization result;
For the basic entry of each of the multiple basic entry, according to the feature calculation of the basis entry basis entry
Weights;
According to the corresponding weights of the multiple basic entry institute being calculated, from multiple basic entry extraction with
The relevant keywords of query.
5. method according to claim 4, wherein described to extract multiple basic entries from least one polymerization result
Step includes:
For each polymerization result at least one polymerization result, word segmentation processing is carried out to the polymerization result, obtain with
The corresponding multiple basic entries of the polymerization result.
6. method according to claim 4 or 5, wherein the feature of basic entry includes at least one of following:
Part of speech/important level of basic entry;
TF-IDF feature of the basic entry in polymerization result;
The user behavior characteristics showed corresponding to result items where basic entry;
Appearance situation of the basic entry in query.
7. according to the method described in claim 1, wherein, the search log information according in search engine is obtained and is examined
There is rope word query history the step of showing result of high correlation to include:
According to the search log information in search engine, and desired indicator information is combined, obtain has high correlation with query
History shows result.
8. according to the method described in claim 7, wherein, the desired indicator indicated by the desired indicator information include with down toward
One item missing:
The history amount of showing;
History shows position;
History click volume;
History clicks Annual distribution.
9. according to the method described in claim 1, wherein, this method further includes:
According to term input by user, search and the relevant keyword of term input by user;
It initiates to search for based on the term input by user and with the relevant keyword of term input by user, and
Search result is supplied to the user.
10. a kind of for excavating and the device of the relevant keyword of term, wherein the device includes:
First obtains device, for according to the search log information in search engine, obtain to have height related to term query
Property history show result, wherein it includes that history shows being gone through for query of being showed in the page that the history, which shows result,
History search result and/or with query it is relevant auxiliary show result;
Generating means, for being showed according to the history as a result, generating at least one polymerization result corresponding with query;
First extraction element, for the extraction from least one polymerization result and the relevant keywords of query.
11. device according to claim 10, wherein it includes that history shows in the page and showed that the history, which shows result,
The historical search for query as a result, the generating means include:
Construction device, for according to the query and multiple urls corresponding with the historical search result, structure to be multiple<
Query, url>It is right;
First polyplant, for according to the multiple<Query, url>, it polymerize and obtains a polymerization corresponding with query
As a result.
12. device according to claim 10, wherein it includes that history shows in the page and showed that the history, which shows result,
With query it is relevant auxiliary show as a result, the generating means include:
Second polyplant polymerize for showing all contents that show in result to the auxiliary, obtains and query phases
A corresponding polymerization result.
13. device according to claim 10, wherein first extraction element includes:
Second extraction element, for extracting multiple basic entries from least one polymerization result;
Computing device is used for for each of the multiple basic entry basic entry, according to the feature meter of the basis entry
Calculate the weights of the basis entry;
Third extraction element, for according to the corresponding weights of the multiple basic entry institute being calculated, from multiple
Extraction and the relevant keywords of query in basic entry.
14. device according to claim 13, wherein second extraction element is used for:
For each polymerization result at least one polymerization result, word segmentation processing is carried out to the polymerization result, obtain with
The corresponding multiple basic entries of the polymerization result.
15. the device according to claim 13 or 14, wherein the feature of basic entry includes at least one of following:
Part of speech/important level of basic entry;
TF-IDF feature of the basic entry in polymerization result;
The user behavior characteristics showed corresponding to result items where basic entry;
Appearance situation of the basic entry in query.
16. device according to claim 10, wherein described first, which obtains device, includes:
Second obtains device, for according to the search log information in search engine, and combines desired indicator information, obtain with
There is query the history of high correlation to show result.
17. device according to claim 16, wherein the desired indicator indicated by the desired indicator information includes following
At least one of:
The history amount of showing;
History shows position;
History click volume;
History clicks Annual distribution.
18. device according to claim 10, wherein the device further includes:
Device is searched, for according to term input by user, searching and the relevant keyword of term input by user;
Searcher, for based on the term input by user and with the relevant key of term input by user
Word initiates search, and search result is supplied to the user.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710138638.5A CN108572971B (en) | 2017-03-09 | 2017-03-09 | Method and device for mining keywords related to search terms |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710138638.5A CN108572971B (en) | 2017-03-09 | 2017-03-09 | Method and device for mining keywords related to search terms |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108572971A true CN108572971A (en) | 2018-09-25 |
CN108572971B CN108572971B (en) | 2022-11-01 |
Family
ID=63578090
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710138638.5A Active CN108572971B (en) | 2017-03-09 | 2017-03-09 | Method and device for mining keywords related to search terms |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108572971B (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109299383A (en) * | 2018-11-02 | 2019-02-01 | 北京字节跳动网络技术有限公司 | Generate method, apparatus, electronic equipment and the storage medium for recommending word |
CN109740161A (en) * | 2019-01-08 | 2019-05-10 | 北京百度网讯科技有限公司 | Data generaliza-tion method, apparatus, equipment and medium |
CN109740075A (en) * | 2018-12-13 | 2019-05-10 | 北京百度网讯科技有限公司 | Event relatedness computation method, apparatus, equipment and storage medium |
CN111222918A (en) * | 2020-01-04 | 2020-06-02 | 厦门二五八网络科技集团股份有限公司 | Keyword mining method and device, electronic equipment and storage medium |
CN111782962A (en) * | 2020-09-04 | 2020-10-16 | 浙江口碑网络技术有限公司 | Pattern matching method and device and electronic equipment |
WO2021232292A1 (en) * | 2020-05-20 | 2021-11-25 | 深圳市欢太科技有限公司 | Log data processing method and related product |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103544186A (en) * | 2012-07-16 | 2014-01-29 | 富士通株式会社 | Method and equipment for discovering theme key words in picture |
CN104063454A (en) * | 2014-06-24 | 2014-09-24 | 北京奇虎科技有限公司 | Search push method and device for mining user demands |
CN104933100A (en) * | 2015-05-28 | 2015-09-23 | 北京奇艺世纪科技有限公司 | Keyword recommendation method and device |
US20150278376A1 (en) * | 2014-04-01 | 2015-10-01 | Baidu (China) Co., Ltd. | Method and apparatus for presenting search result |
-
2017
- 2017-03-09 CN CN201710138638.5A patent/CN108572971B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103544186A (en) * | 2012-07-16 | 2014-01-29 | 富士通株式会社 | Method and equipment for discovering theme key words in picture |
US20150278376A1 (en) * | 2014-04-01 | 2015-10-01 | Baidu (China) Co., Ltd. | Method and apparatus for presenting search result |
CN104063454A (en) * | 2014-06-24 | 2014-09-24 | 北京奇虎科技有限公司 | Search push method and device for mining user demands |
CN104933100A (en) * | 2015-05-28 | 2015-09-23 | 北京奇艺世纪科技有限公司 | Keyword recommendation method and device |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109299383A (en) * | 2018-11-02 | 2019-02-01 | 北京字节跳动网络技术有限公司 | Generate method, apparatus, electronic equipment and the storage medium for recommending word |
CN109299383B (en) * | 2018-11-02 | 2021-11-05 | 北京字节跳动网络技术有限公司 | Method and device for generating recommended word, electronic equipment and storage medium |
CN109740075A (en) * | 2018-12-13 | 2019-05-10 | 北京百度网讯科技有限公司 | Event relatedness computation method, apparatus, equipment and storage medium |
CN109740161A (en) * | 2019-01-08 | 2019-05-10 | 北京百度网讯科技有限公司 | Data generaliza-tion method, apparatus, equipment and medium |
CN111222918A (en) * | 2020-01-04 | 2020-06-02 | 厦门二五八网络科技集团股份有限公司 | Keyword mining method and device, electronic equipment and storage medium |
CN111222918B (en) * | 2020-01-04 | 2023-06-30 | 厦门二五八网络科技集团股份有限公司 | Keyword mining method and device, electronic equipment and storage medium |
WO2021232292A1 (en) * | 2020-05-20 | 2021-11-25 | 深圳市欢太科技有限公司 | Log data processing method and related product |
CN111782962A (en) * | 2020-09-04 | 2020-10-16 | 浙江口碑网络技术有限公司 | Pattern matching method and device and electronic equipment |
CN111782962B (en) * | 2020-09-04 | 2021-01-12 | 浙江口碑网络技术有限公司 | Pattern matching method and device and electronic equipment |
Also Published As
Publication number | Publication date |
---|---|
CN108572971B (en) | 2022-11-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108572971A (en) | It is a kind of to be used to excavate and the method and apparatus of the relevant keyword of term | |
CN103631929B (en) | A kind of method of intelligent prompt, module and system for search | |
US9934293B2 (en) | Generating search results | |
JP5311378B2 (en) | Feature word automatic learning system, content-linked advertisement distribution computer system, search-linked advertisement distribution computer system, text classification computer system, and computer programs and methods thereof | |
US10068022B2 (en) | Identifying topical entities | |
US20160026727A1 (en) | Generating additional content | |
CN103324665A (en) | Hot spot information extraction method and device based on micro-blog | |
CN106484764A (en) | User's similarity calculating method based on crowd portrayal technology | |
CN103577432A (en) | Method and system for searching commodity information | |
JP2013531289A (en) | Use of model information group in search | |
CN103631794A (en) | Method, device and equipment for sorting search results | |
CN105468790B (en) | A kind of comment information search method and device | |
CN103885971A (en) | Data pushing method and data pushing device | |
TW201401088A (en) | Search method and apparatus | |
US9009192B1 (en) | Identifying central entities | |
CN106293119A (en) | A kind of method and apparatus carrying out information recommendation in input method | |
CN111639255B (en) | Recommendation method and device for search keywords, storage medium and electronic equipment | |
CN105468649A (en) | Method and apparatus for determining matching of to-be-displayed object | |
Blanco et al. | Overview of NTCIR-13 Actionable Knowledge Graph (AKG) Task. | |
JP5427694B2 (en) | Related content presentation apparatus and program | |
KR101140724B1 (en) | Method and system of configuring user profile based on a concept network and personalized query expansion system using the same | |
US10255246B1 (en) | Systems and methods for providing a searchable concept network | |
CN107862620A (en) | A kind of similar users method for digging based on social data | |
KR20120038418A (en) | Searching methods and devices | |
JP2010123036A (en) | Document retrieval device, document retrieval method and document retrieval program |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |