CN103136310B - The method changed based on the historical query of individual Web sites to new inquiry - Google Patents
The method changed based on the historical query of individual Web sites to new inquiry Download PDFInfo
- Publication number
- CN103136310B CN103136310B CN201110413826.7A CN201110413826A CN103136310B CN 103136310 B CN103136310 B CN 103136310B CN 201110413826 A CN201110413826 A CN 201110413826A CN 103136310 B CN103136310 B CN 103136310B
- Authority
- CN
- China
- Prior art keywords
- query
- website
- original
- candidate
- term
- 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
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
Present invention is disclosed a kind of method changed based on the historical query of individual Web sites to new inquiry.Each specific website has the content-preference of oneself and the customer group of oneself.When user is inquired about in some websites, the inquiry is related with the web site contents to most cases.The present invention is changed using the historical search information in the website to the original query being potentially ambiguous.Scanned for afterwards based on the replacement query in advertisement engine, the advertisement searched is returned into user again, so that really the query term for advertisement search more matches with the target service of the website, while ad result is more matched with the search intention of user.
Description
Technical field
The present invention relates to advertisement search technology, more specifically at website monitoring (syndication network)
The technology that the middle information according to specific website is changed to original query.
Background technology
Commercial advertisement engine (such as bing, ***) generally provides advertisement to a lot of other websites and supported, these nets
Station possesses the respective page, search inputting interface and search engine, but when carrying out backstage search, uses same advertisement
Engine.When user scans in some websites, the website returns to two class results：One kind is search result, by website itself
There is provided or third party's search engine provides；Another kind of is ad result, is provided by the advertisement engine of website monitoring.
Different websites has different theme tendency and target service, therefore the hobby of the advertisement provided advertisement engine
Degree is also differed, and website can be wished the ads related to the target service that oneself pays close attention to the most in forward position,
So these advertisements are more likely clicked on by user, and the click of user can bring income to website.Click on oneself the most
The quantity of the related advertisement of the target service of concern is more, and the income that website obtains is also higher.
In existing website monitoring, the ad result that each website provides shared advertisement engine is according to oneself
Target service carried out certain processing.Such as website " jobs.Yahoo.com " shares advertisement engine, but two with " bing "
The target service of person is different.The target service of " jobs.yahoo.com " is working opportunity and job ad, and " bing "
Target service be universal search engine.Then, if a same keyword scans on the webpage of this website, obtain
Search result be different, accordingly, ad result also should be different.
For example equally scanned in " jobs.yahoo.com " and input " computer " in " bing ", obtain difference
Search result.
" jobs.yahoo.com " for arranged in the search result of " computer " forward search result all with work
Position is related, this be due to " jobs.yahoo.com " itself characteristic and caused by.Because " jobs.yahoo.com " is one
The search result and ad result related to job overall are seen in website of the family using job overall as target service, user's expectation,
They can only click on the webpage related to job overall or advertisement, and only advertisement can be that " jobs.yahoo.com " is brought
Income；User will not click on the advertisement unrelated with job overall, therefore can not be that " jobs.yahoo.com " brings income.
" bing " occupies search for being related to the webpage of computer product sale in the search result of " computer "
As a result forward position.The target service of " bing " is universal search engine, and its revenue source is with being similar to
Professional website as " jobs.yahoo.com " is different, and therefore " bing " can select the advertisement knot the most favourable to its income
Fruit, the result of these ad results and " jobs.yahoo.com " are very different.
, it is necessary to the advertisement fed back from advertisement engine for the website in the website monitoring using shared advertisement engine
As a result the middle advertisement for selecting to meet oneself target service, so could effectively improve clicking rate, improve income.Although at present
Through there is some to handle to ad result, but these processing not have very high efficiency and good applicability.In some passes
In the search of key word, these processing can reach certain effect, but in the processing of considerable keyword, current scheme
Effective processing is unable to reach, still can will be unrelated in own target business on its webpage for specific website
Ads are in forward position.
The content of the invention
The present invention is directed to propose it is a kind of can the historical query based on the specific website in website monitoring new inquiry is entered
The method of row conversion is so as to providing the method for more accurate advertisement search result.
In one embodiment, new inquiry is turned based on the historical query of individual Web sites present invention is disclosed a kind of
The method changed.Individual website of this method first in website monitoring receives the original query of user, the website monitoring bag
Containing several independent websites, the original query is related to the theme of the website.This method obtains all users from the website afterwards
Historical search information, the original query is changed based on the historical search information, obtains converted replacement query.
It is finally based on the replacement query to scan in advertisement engine, the advertisement searched is shown to user.
In one embodiment, new inquiry is turned based on the historical query of individual Web sites present invention is disclosed a kind of
The method changed.Individual website of this method first in website monitoring receives original query, and the website monitoring includes number
Individual independent website and the original query is related to the theme of the website.Then this method obtains all users in mistake from the website
The tendency query term related to original query gone in a period of time.The tendency query term is merged into original query, obtained
Converted replacement query.The replacement query is finally based on to scan in advertisement engine and be shown to the advertisement searched
User.
In one embodiment, new inquiry is turned based on the historical query of individual Web sites present invention is disclosed a kind of
The method changed.Individual website of this method first in website monitoring receives original query, and the website monitoring includes number
Individual independent website and the original query is related to the theme of the website.Then obtain candidate query from the website and candidate is looked into
Inquiry is screened, and the screening is that the attribute based on candidate query and original query is compared and made with similarity and candidate query
Carried out with frequency, screening obtains the candidate query matched with original query.The original is replaced using the candidate query of the matching
Begin to inquire about, obtain converted replacement query.The replacement query is finally based on to scan for and will search in advertisement engine
Advertisement be shown to user.
The present invention can be directed to the characteristic of specific website and historical data in website monitoring and query term is carried out pointedly
Conversion so that really the query term for advertisement search more matches with the target service of website, more has valency so as to obtain
The advertisement search result of value.
Brief description of the drawings
The above and other features of the present invention, property and advantage will pass through retouching with reference to the accompanying drawings and examples
State and become readily apparent from, in the accompanying drawings, identical reference represents identical feature all the time, wherein：
Fig. 1 discloses to be carried out according to the first embodiment of the present invention based on the historical query of individual Web sites to new inquiry
The flow chart of the method for conversion.
Fig. 2 discloses being carried out based on the historical query of individual Web sites to new inquiry according to the second embodiment of the present invention
The flow chart of the method for conversion.
Fig. 3 discloses being carried out based on the historical query of individual Web sites to new inquiry according to the third embodiment of the invention
The flow chart of the method for conversion.
Embodiment
The present invention proposes a kind of method changed based on the historical query of individual Web sites to new inquiry.Fig. 1 is disclosed
According to the flow chart of the method for the first embodiment of the present invention.Main idea is that：As one in website monitoring
After specific website receives an original query, for example user logs on the webpage of the website and have input an inquiry,
The website is analyzed, the information related to its target service is obtained at the website.These information will be used for original
Inquiry is changed, and the process of conversion is by the proportion increase related to the target service of the website, obtains one and website
The inquiry that target service more matches.When being scanned for using advertisement engine, by using the query term after changing, so obtain
Search result it is more related to the target service of website.
This method is applied to have any website for having limited occupation feature.With reference to shown in figure 1, this method 100 includes：
102. individual website in website monitoring receives the original query of user.The website monitoring includes several only
Vertical, the website with respective service feature, and the original query is related to the specific website.One reality of step 102
Now mode is：User's input inquiry on specific website scans for, and is taken as have input an original related to the website
Begin to inquire about.Give one example explanation, user logs on " jobs.yahoo.com " and have input " computer ", then it is assumed that be to use
Family have input an original query related to " jobs.yahoo.com ", and the original query is " computer ".
104. obtain the historical search information of all users from the website, original looked into described based on the historical search information
Inquiry is changed, and obtains converted replacement query.Historical search information is related to the specific website, typically from the spy
The historical data of fixed website, such as the website once used historical query, obtained historical search result, going through of browsing
History webpage etc..The historical search information has two kinds of forms used：A kind of mode is that historical search information is merged into original
Begin in inquiry, obtain converted replacement query, in this fashion, historical search information is a tendency query term.Second
Embodiment is related to this mode.Another way is that usage history search information replaces original query, as replacement query,
In this mode, historical search information is also an inquiry, and 3rd embodiment is related to this mode.
106. being scanned for based on the replacement query in advertisement engine, the advertisement searched is shown to user.
Fig. 2 discloses method according to the second embodiment of the present invention, and in this method 200, historical search information is tendency
The form of query term, tendency query term are integrated into original query.As shown in Fig. 2 this method 200 includes：
202. individual website in website monitoring receives original query.The website monitoring includes several independences
Website, the original query is related to the theme of the website.Step 202 is similar with step 102, is not repeated to describe herein.
204. obtain to original query related tendency query term of all users within the past period from the website.
The acquisition modes of tendency query term have following two kinds：
1) tendency query term is obtained from the historical query of the specific website.For example the history in the specific website is looked into
Frequency of occurrences highest query term is searched in inquiry as tendency query term.Referring again to above being lifted
The example of " computer " is inputted on " jobs.yahoo.com ", tendency query term is from website, i.e. " jobs.yahoo.com "
Middle analysis obtains, the historical query record of analysis " jobs.yahoo.com ", i.e., all to be inputted on " jobs.yahoo.com "
Inquiry record, find wherein frequency of occurrences highest query term, keyword is " job " in other words, then just will " job " choosing
It is selected as being inclined to query term.
2) tendency query term is obtained from the webpage of the specific site search.Such as the net in the specific site search
Frequency of occurrences highest query term is searched in page as tendency query term.Also with reference to being lifted at " jobs.yahoo.com "
Example, to the inquiry carried out by " jobs.yahoo.com " and the search result that searches, webpage is analyzed in other words, is looked into
The frequency highest query term occurred is looked on these webpages, is the discovery that " job ", then " job " is just selected as being inclined to
Query term.
206. are merged into the tendency query term in original query, obtain converted replacement query.Inquired about for tendency
For, it is considered to be a kind of extra querying condition for original query, to improve the accuracy rate of inquiry.Tendency inquiry
Item will merge in original query and obtain replacement query.For example in above institute's illustrated example, pass through " jobs.yahoo.com "
And the original query " computer " inputted merges the replacement query " computer+ after being changed with tendency query term " job "
job”。
208. are scanned for based on the replacement query in advertisement engine, and the advertisement searched is shown into user.
Fig. 3 discloses method according to the third embodiment of the invention, and in this method 300, historical search information is also one
It is individual to inquire about and be used to replace original query.As shown in figure 3, this method 300 includes：
302. individual website in website monitoring receives original query.The website monitoring includes several independent
Website, the original query are related to the theme of the website.Step 302 is similar with step 102 and step 202, no longer heavy herein
Multiple description.
304. obtain candidate query from the website.Candidate query is obtained from the historical query for the website that this is specified, and is
Ensure the relevance of candidate query and original query, candidate query has at least one identical query term with original query.
Again by by " jobs.yahoo.com " and input inquiry " computer " exemplified by, original query is " computer ",
In the historical query of " jobs.yahoo.com ", selection has at least one identical query term, i.e., all bags with original query
Inquiry containing query term " computer " is as candidate query.Here, query term refers to a byte in inquiry, or,
For text query, a query term refers to a word.More specifically, query term is a word or the English in Chinese
A word in language.
306. pairs of candidate queries screen, the screening be based on candidate query and the attribute of original query compare with it is similar
Degree and candidate query frequency of use and carry out, screening obtain the candidate query matched with original query.For candidate query
Screening include three aspect consideration：
1) degree that is consistent with original query on attribute；
2) with the similarity of original query；
3) frequency of use of candidate query.
For the consideration in terms of above three, it is proposed that following restrictive condition：
First restrictive condition is compared for the attribute of candidate query and original query, and candidate query and original query are all
It is to be made up of a group polling item, each query term is considered as a byte (term), and for text query, one is looked into
It is exactly a word to ask item, such as " for this inquiry of computer device ", it is believed that it has two query terms
(two bytes), it is respectively " computer " and " device " that the byte length of the inquiry is 2.First restrictive condition requirement
There is the difference no more than following one between candidate query and original query：
Candidate query has identical byte (term) length and a different query term from original query；
A candidate query byte fewer than original query；
Candidate query byte more than original query.
For Article 2 restrictive condition for the comparison of candidate query and the similarity of original query, the comparison is literary according to being inverted
This frequency (IDF) and carry out.Specifically, including：
Calculate the inversion text frequency (IDF) of candidate query and original query.For from the given of given website
The text frequency (DF) of query term, refers to the search rate in a period of time inner cap query term, and is inverted text frequency (IDF)
It can be calculated as follows：
Wherein maxDF is the search rate of the query term with highest search rate.
It is then based on being inverted the similarity that text frequency calculates candidate query and original query.If qiAnd qjInquired about for two,
Such as original query and candidate query, then the two inquiry qiAnd qjBetween similarity be calculated as：
Wherein S1And S2The respectively inversion text frequency of original query and candidate query.
Then the candidate query that similarity is more than predetermined threshold is screened, Article 2 restrictive condition requires original query and candidate
Similarity between inquiry is higher than the threshold value of a setting, and the threshold value can be adjusted according to the requirement of application.
Article 3 restrictive condition is the restrictive condition for the frequency of use of candidate query.It is expected that meet first
The candidate query of bar restrictive condition and Article 2 restrictive condition meeting more than one, then, Article 3 restrictive condition is to candidate query
Frequency of use limited.Typically it can select to have in the candidate query of restrictive condition of first and Article 2 is met
The candidate query of highest frequency of use.Realized as one kind, the candidate query with highest clicking rate can be selected.
First restrictive condition and Article 2 restrictive condition may insure the semantic phase between candidate query and original query
Like degree, the characteristic of website is embodied as far as possible in the case where meeting user's search request.And Article 3 restrictive condition ensures to wait
Choosing inquiry can be obtained more search results, can also be created preferably by frequency of use, the usually used high inquiry of frequency
Income.
308. replace original query using the candidate query of the matching, obtain converted replacement query.
310. are scanned for based on the replacement query in advertisement engine, and the advertisement searched is shown into user.
The present invention can be directed to the characteristic of specific website and historical data in website monitoring and query term is carried out pointedly
Conversion so that really the query term for advertisement search more matches with the target service of website, more has valency so as to obtain
The advertisement search result of value.
Claims (16)
- A kind of 1. method changed based on the historical query of individual Web sites to new inquiry, it is characterised in that this method bag Include：Individual website in website monitoring receives the original query of user, and the website monitoring includes several independent nets Stand, several independent websites each have a different service features, and the original query is related to the theme of the website；The historical search information of all users is obtained from the website, the original query is turned based on the historical search information Change, obtain converted replacement query, the replacement query compared to for the original query with the service feature of the website more Add matching；Scanned for based on the replacement query in advertisement engine, the advertisement searched is shown to user.
- 2. the method as described in claim 1, it is characterised in that the historical search information is related to the website.
- 3. method as claimed in claim 2, it is characterised in that the historical search information is integrated into the original query In, obtain converted replacement query.
- 4. method as claimed in claim 2, it is characterised in that the historical search information be used to replacing described original look into Ask, as converted replacement query.
- A kind of 5. method changed based on the historical query of individual Web sites to new inquiry, it is characterised in that this method bag Include：Individual website in website monitoring receives original query, and the website monitoring includes several independent websites, should Several independent websites each have a different service features, and the original query is related to the theme of the website；To original query related tendency query term of all users within the past period is obtained from the website；The tendency query term is merged into the original query, obtains converted replacement query, the replacement query compared to More matched with the service feature of the website for the original query；Scanned for based on the replacement query in advertisement engine, the advertisement searched is shown to user.
- 6. method as claimed in claim 5, it is characterised in that obtaining tendency query term is included from the historical query of the website Obtain tendency query term.
- 7. method as claimed in claim 6, it is characterised in that obtain tendency query term and be included in the historical query of the website Frequency of occurrences highest query term is searched for as tendency query term.
- 8. method as claimed in claim 5, it is characterised in that obtaining tendency query term is included from the webpage of the site search Obtain tendency query term.
- 9. method as claimed in claim 8, it is characterised in that obtain tendency query term and be included in the webpage of the site search Frequency of occurrences highest query term is searched for as tendency query term.
- A kind of 10. method changed based on the historical query of individual Web sites to new inquiry, it is characterised in that this method bag Include：Individual website in website monitoring receives original query, and the website monitoring includes several independent websites, should Several independent websites each have a different service features, and the original query is related to the theme of the website；Candidate query is obtained from the website；Candidate query is screened, the screening be the attribute based on candidate query and original query compare with similarity and The frequency of use of candidate query and carry out, screening obtain the candidate query matched with original query；The original query is replaced using the candidate query of the matching, obtains converted replacement query, the replacement query is compared More matched with the service feature of the website for the original query；Scanned for based on the replacement query in advertisement engine, the advertisement searched is shown to user.
- 11. method as claimed in claim 10, it is characterised in that the candidate query is obtained from the historical query of the website .
- 12. method as claimed in claim 11, it is characterised in that the candidate query has at least one phase with original query Same query term.
- 13. method as claimed in claim 12, it is characterised in that the query term is in a word or English in Chinese A word.
- 14. method as claimed in claim 11, it is characterised in that the attribute based on candidate query and original query compares progress Screening includes the candidate query that screening meets one of following conditions：Candidate query has identical byte (term) length and a different query term from original query；A candidate query byte fewer than original query；Candidate query byte more than original query.
- 15. method as claimed in claim 11, it is characterised in that the similarity based on candidate query and original query is sieved Choosing includes：Calculate the inversion text frequency (IDF) of candidate query and original query；Based on the similarity for being inverted text frequency calculating candidate query and original query；Screen the candidate query that similarity is more than predetermined threshold.
- 16. method as claimed in claim 11, it is characterised in that screening bag is carried out based on the frequency of use of candidate query Include：Candidate query of the screening with highest clicking rate.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110413826.7A CN103136310B (en) | 2011-12-02 | 2011-12-02 | The method changed based on the historical query of individual Web sites to new inquiry |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110413826.7A CN103136310B (en) | 2011-12-02 | 2011-12-02 | The method changed based on the historical query of individual Web sites to new inquiry |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103136310A CN103136310A (en) | 2013-06-05 |
| CN103136310B true CN103136310B (en) | 2017-11-28 |
Family
ID=48496141
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201110413826.7A Active CN103136310B (en) | 2011-12-02 | 2011-12-02 | The method changed based on the historical query of individual Web sites to new inquiry |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103136310B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170148107A1 (en) * | 2015-11-23 | 2017-05-25 | Linkedln Corporation | Prioritizing search terms representing locations |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1761972A (en) * | 2003-03-18 | 2006-04-19 | Nhn株式会社 | A method of determining an intention of internet user, and a method of advertising via internet by using the determining method and a system thereof |
| CN1877581A (en) * | 2006-07-12 | 2006-12-13 | 百度在线网络技术（北京）有限公司 | Advertisement display system and method used for Internet search engine |
| CN1961316A (en) * | 2004-05-29 | 2007-05-09 | Nhn株式会社 | Method and system for managing the impressing of the search listing based on advertisement group |
| CN102067105A (en) * | 2008-03-18 | 2011-05-18 | 雅虎公司 | Personalizing sponsored search advertising layout using user behavior history |
| CN102096882A (en) * | 2002-09-24 | 2011-06-15 | Google公司 | Methods and apparatus for serving relevant advertisements |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8943043B2 (en) * | 2010-01-24 | 2015-01-27 | Microsoft Corporation | Dynamic community-based cache for mobile search |
-
2011
- 2011-12-02 CN CN201110413826.7A patent/CN103136310B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102096882A (en) * | 2002-09-24 | 2011-06-15 | Google公司 | Methods and apparatus for serving relevant advertisements |
| CN1761972A (en) * | 2003-03-18 | 2006-04-19 | Nhn株式会社 | A method of determining an intention of internet user, and a method of advertising via internet by using the determining method and a system thereof |
| CN1961316A (en) * | 2004-05-29 | 2007-05-09 | Nhn株式会社 | Method and system for managing the impressing of the search listing based on advertisement group |
| CN1877581A (en) * | 2006-07-12 | 2006-12-13 | 百度在线网络技术（北京）有限公司 | Advertisement display system and method used for Internet search engine |
| CN102067105A (en) * | 2008-03-18 | 2011-05-18 | 雅虎公司 | Personalizing sponsored search advertising layout using user behavior history |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103136310A (en) | 2013-06-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11507551B2 (en) | Analytics based on scalable hierarchical categorization of web content | |
| Crescenzi et al. | Clustering web pages based on their structure | |
| CN105765573B (en) | Improvements in website traffic optimization | |
| JP5860456B2 (en) | Determination and use of search term weighting | |
| CN102687138B (en) | Search is advised cluster and is presented | |
| US9798820B1 (en) | Classification of keywords | |
| JP6343035B2 (en) | Generate ad campaign | |
| US20070214133A1 (en) | Methods for filtering data and filling in missing data using nonlinear inference | |
| US20060155751A1 (en) | System and method for document analysis, processing and information extraction | |
| US8577875B2 (en) | Presenting search results ordered using user preferences | |
| US20090276729A1 (en) | Adaptive user feedback window | |
| CN103400286A (en) | Recommendation system and method for user-behavior-based article characteristic marking | |
| US10296622B1 (en) | Item attribute generation using query and item data | |
| US20180218043A1 (en) | Information providing method and system | |
| US11941073B2 (en) | Generating and implementing keyword clusters | |
| CN103136213B (en) | Method and device for providing related words | |
| JP2002215659A (en) | Information retrieval support method and information retrieval support system | |
| EP2933734A1 (en) | Method and system for the structural analysis of websites | |
| US10095788B2 (en) | Context-sensitive deeplinks | |
| US20120179540A1 (en) | Method of finding commonalities within a database | |
| US20140108090A1 (en) | Price mining system and method for mining the price | |
| US11170062B2 (en) | Structured search via key-objects | |
| CN103136310B (en) | The method changed based on the historical query of individual Web sites to new inquiry | |
| TWI605351B (en) | Query method, system and device based on vertical search | |
| WO2019218151A1 (en) | Data searching method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| ASS | Succession or assignment of patent right |
Owner name: MICROSOFT TECHNOLOGY LICENSING LLC Free format text: FORMER OWNER: MICROSOFT CORP. Effective date: 20150727 |
|
| C41 | Transfer of patent application or patent right or utility model | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20150727 Address after: Washington State Applicant after: Micro soft technique license Co., Ltd Address before: Washington State Applicant before: Microsoft Corp. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |