WO2017001944A1 - Procédé, système et mémoire lisible par ordinateur de génération de résultats de recherche classés incorporant des suggestions - Google Patents

Procédé, système et mémoire lisible par ordinateur de génération de résultats de recherche classés incorporant des suggestions Download PDF

Info

Publication number
WO2017001944A1
WO2017001944A1 PCT/IB2016/050563 IB2016050563W WO2017001944A1 WO 2017001944 A1 WO2017001944 A1 WO 2017001944A1 IB 2016050563 W IB2016050563 W IB 2016050563W WO 2017001944 A1 WO2017001944 A1 WO 2017001944A1
Authority
WO
WIPO (PCT)
Prior art keywords
search
search result
query
general
server
Prior art date
Application number
PCT/IB2016/050563
Other languages
English (en)
Inventor
Andrey Sergeevich YAROSHEVSKY
Pavel Alekseevich SHISHKIN
Aydar Ilshatovich BIKTIMIROV
Original Assignee
Yandex Europe Ag
Yandex Llc
Yandex Inc.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Yandex Europe Ag, Yandex Llc, Yandex Inc. filed Critical Yandex Europe Ag
Publication of WO2017001944A1 publication Critical patent/WO2017001944A1/fr

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/332—Query formulation
    • G06F16/3322—Query formulation using system suggestions

Definitions

  • a method, system, and computer readable memory provide, inter alia, for a display to a user including both general search results and suggests. More specifically, the method, system, and computer readable memory rank the suggests together with the general search results to generate the combination of the suggests and the general search results to the user in response to the user's search query.
  • a search engine may organize the general search results on the SERP so that a listing of web resources (i. e. , uniform resource locators ("URLs")) is presented to the user in the center of the page.
  • URLs uniform resource locators
  • the search engine i.e., as executed by the search server
  • the search engine may output one or more widgets, which are blocks of results.
  • the widgets may present information such as links to shopping sites, news, images, videos, etc. that may be associated with one or more terms in the user's search query.
  • “suggests” that are generated by a search engine while a user is typing a search query into a search query field. Suggests are generated using several variables, not the least of which is an assessment of the common phrases entered by other users, for previous searches, when searching for the same information.
  • U.S. Patent No. 8,631,030 (“the '030 Patent”) describes query suggestions with high diversity. Query suggestions are suggests, as discussed above.
  • a search engine results page 300 may include a search results block 301.
  • the search results provided include those that a search engine has identified as being relevant to the initial query 302 (in the figure, "childztoyz").
  • the search engine results page also includes references 303A to 303C to each additional query (in the figure, "childztoyz trading cards,” “childztoyz pets,” and “childztoyz charms,” respectively).
  • the search engine results page 300 also includes an advertising block 304 that includes an advertisement 306A that is targeted to the initial query 302, an advertising block 305A that includes reference 303A and advertisements 306B-D that are targeted to the additional query 303 A, an advertising block 305B that includes reference 306B and advertisements 306E-F that are targeted to the additional query 303B, and an advertising block 305C that includes reference 303C and advertisements 306G-H that are targeted to the additional query 303C.
  • the search engine results page 300 does not include a reference that refers to the initial query 302 directly above the advertising block 304.
  • the additional advertising blocks may be displayed in order of, for example, relevance of each additional query to the initial query, historical click through rate for the ads in each ad block, number of ads available for each ad block, random placement, and the like.
  • the '750 Patent discusses query suggestions, which may be ranked according to an order.
  • the query suggestions may be ranked based on the probability that the user will select the respective query suggestion.
  • the reference describes a search engine that can re -rank the query suggestions so natural query extensions are presented together in a group. For example, according to an identified re -ranking, "Italy” has the highest rank, "Italy map” (with the same first term) has the second rank, “Italian” has the third rank, and "Italian History” has the fourth rank.
  • one or more quality features are extracted based on the first and second sets of documents resulting from the searches performed at blocks 610 and 612. Such quality features generally indicate the relevance of the two sets of search results and provide an indication of the quality of each reformulation candidate as compared to the original query. These quality features may include ranking features and topic drift features.
  • U.S. Patent No. 8,375,049 (“the '049 Patent”) describes query revision using known highly-ranked queries.
  • the '049 Patent discusses a reviser confidence estimator 112 operating on the assumption that certain behaviors, e.g., a long click by a user on a revised query link 302, indicates that the user is satisfied with the revision as being an accurate representation of the user's original information need.
  • a long click can be deemed to occur when the user stays on the clicked through page for some minimum period of time, for example a minimum of 60 seconds.
  • the reviser confidence estimator 112 can train the predictive model to predict the likelihood of a long click given the various features of the revised query and the original query. Revised queries having high predicted likelihoods of a long click are considered to be better (i.e., more successful) revisions for their associated original queries.
  • Embodiments of the method described herein address one or more of the deficiencies with respect to the prior art.
  • One contemplated embodiment relies on a separate machine learning formula to calculate the relevance of a suggest. It may be that the formula determines the probability or other relevant value associated with a "long" click.
  • a long click may be a sequence of a transition due to a related query from a suggest to a SERP and a subsequent transition from this SERP to a web- resource. The obtained value is used to rank the suggest among other elements of search results.
  • the suggest(s) may be embodied in a widget.
  • Another embodiment is contemplated to allow the integration of suggests into the search results.
  • Interesting, related queries may be presented as links. They may be consolidated into a widget that appears as one of the elements of the search results. It is contemplated that clicking on a link initiates a search associated to a related query and brings to the user the relevant search results.
  • one or more embodiments provide a method of generating a search engine results page in a system comprising a search server, a user device, and a network connecting the search server to the user device.
  • the method includes receiving, by the search server, a search query from the user device. Based on the search query, the method determines, by the search server, a first search result set, where the first search result set encompasses at least a first general search result and a second general search result. In the method, the first general search result and the second general search result are ranked using at least a first ranking algorithm.
  • the method includes retrieving, by the server, at least one related query, where the at least one related query is related to the search query.
  • the method generates, by the server, a first suggest object comprising at least one clickable element, where the at least one clickable element is associated with the at least one related query. Then, the method applies, by the server, a second ranking algorithm to rank the first suggest object relative to the first general search result and the second general search result. The method subsequently generates, by the server, a second search result set that comprises, in a ranked order, the first suggest object, the first general search result, and the second general search result. In addition, the method generates, by the server, the search result page based on the second search result set.
  • the method includes sending, via the network, the search result page for display on the user device and receiving, from the user device, a selection of the at least one clickable element.
  • the method also includes applying, by the server, a third raking algorithm to a plurality of related queries including the at least one related query and generating, by the server, a third search result set that comprises the plurality of related queries in ranked order.
  • the third ranking algorithm may employ a variable dependent on a long click to generate a related long click.
  • the second ranking algorithm may employ the related long click to generate the second search result set.
  • the method may include generating, by the server, a related queries search result page based on the third search result set and sending, via the network, the related queries search result page for display on the user device.
  • the method includes, in response to receipt of the selection of the at least one clickable element at the user device, causing a display, on the user device, of the related queries search result page. [0034] Still further, it is contemplated that the method includes causing a display of the related queries search result page on the user device.
  • selection of the at least one clickable element at the user device causes a display of a related queries search result page comprising the at least one related query.
  • the at least one related query may be a suggest.
  • the first search result set may include at least a vertical search result ranked according to a fourth ranking algorithm.
  • the vertical search result may be ranked, as a first search result, together with other first search results via the first ranking algorithm.
  • the first ranking algorithm may employ a variable dependent on a long click.
  • the second ranking algorithm may employ a variable dependent on a related long click, where the related long click is a function of a long click associated with a related search result of the at least one related query.
  • the related search result may encompass a web resource.
  • the long click may be calculated based on an average time users dwell on at least one of the first general search result and the second general search result.
  • the first general search result may include a plurality of web resources.
  • the first general search result may include a first plurality of web resources and the second general search result comprises a second plurality of web resources.
  • the second general search result may include a plurality of web resources.
  • the vertical search result may be embodied in a widget.
  • Still another contemplated embodiment provides a method of generating search engine results page.
  • the method is contemplated to be executed at a server and include acquiring a search query from a user device. Based on the search query, the method may include determining a search result set, the search result set including a first general search result and a second general search result, the first general search result and the second general search result having been ranked using a first ranking algorithm.
  • the method is contemplated to include retrieving a list of related queries, the list of related queries including at least one related query, each query in the list of related queries being related with the search query and generating a first suggest object including at least one clickable element, the at least one clickable element being associated with at least one related query from the list of related queries, the click on the least one clickable element causing a displaying of a page, on the user device, with search results for the related query associated with the at least one clickable element.
  • the method further includes applying a second ranking algorithm to rank the first suggest object relative to the first general search result and the second general search result and generating a search result page (SERP) that includes, in a ranked order, the first suggest object, the first general search result, and the second general search result.
  • SERP search result page
  • the method contemplates that the selection of the at least one clickable element at the user device may cause a display of a related queries search result page comprising the at least one related query.
  • the at least one related query is contemplated to be a suggest.
  • One other embodiment also provides a computer readable storage medium having executable instructions for operation of a method of generating a search engine results page in a system with a search server, a user device, and a network connecting the search server to the user device.
  • the executable instructions when executed, result in operations including receiving, by the search server, a search query from the user device and, based on the search query, determining, by the search server, a first search result set, where the first search result set comprises at least a first general search result and a second general search result, and where the first general search result and the second general search result are ranked using at least a first ranking algorithm.
  • the executable instructions also are contemplated to operate such that, based on previous search sessions, the server retrieves at least one related query, where the at least one related query is related to the search query.
  • the instructions are contemplated to operate such that the server generates a first suggest object encompassing at least one clickable element, where the at least one clickable element is associated with the at least one related query.
  • the instructions are contemplated to operate such that the server applies a second ranking algorithm to rank the first suggest object relative to the first general search result and the second general search result, such that the server generates a second search result set that includes, in a ranked order, the first suggest object, the first general search result, and the second general search result, and such that the server generates the search result page based on the second search result set.
  • the executable instructions are contemplated to operate so that selection of the at least one clickable element at the user device causes a display of a related queries search result page comprising the at least one related query.
  • the at least one related query is contemplated to be a suggest.
  • a system is also provided for generating a search engine results page.
  • the system is contemplated to include a search server, a user device, and a network connecting the search server to the user device.
  • the system is contemplated to operate by receiving, by the search server, a search query from a user device, based on the search query, determining, by the search server, a first search result set, where the first search result set includes at least a first general search result and a second general search result, and where the first general search result and the second general search result are ranked using at least a first ranking algorithm, based on previous search sessions retrieving, by the server, at least one related query, where the at least one related query is related to the search query, generating, by the server, a first suggest object encompassing at least one clickable element, where the at least one clickable element is associated with the at least one related query, applying, by the server, a second ranking algorithm to rank the first suggest object relative to the first general search result and the second general search result, generating, by the
  • selection of the at least one clickable element at the user device causes a display of a related queries search result page comprising the at least one related query.
  • the at least one related query is contemplated to be a suggest.
  • FIG. 1 is a graphical overview of an example of a system operating according to the method of the present technology
  • FIG. 2 is an example of a screenshot from a prior art search engine, illustrating various aspects of a prior art SERP;
  • Fig. 3 is a modified version of the screenshot of the SERP from Fig. 2, showing a drop down menu from the search query input bar with suggests;
  • FIG. 4 is a screenshot of a SERP according to a first embodiment of the present technology
  • Fig. 5 is a screenshot of a SERP according to a second embodiment of the present technology.
  • Fig. 6 is a flow chart illustrating one method according to the present technology.
  • a “server” is a computer program that operates on selected, appropriate hardware and is capable of receiving requests (e.g., from one or more user devices (also referred to as “client devices")) over a network, carrying out those requests, or causing those requests to be carried out.
  • the hardware may be one physical computer or one physical computer system, but neither is required explicitly therefor.
  • a “server” may be embodied in software, hardware, or a combination of software and hardware.
  • server is not intended to mean that every task (e.g., received instructions or requests) or any particular task will have been received, carried out, or caused to be carried out by the same server (i.e. , the same software and/or hardware).
  • server is intended to mean that any number of software elements and/or hardware devices may be involved in receiving, sending, carrying out, and/or causing to be carried out any task or request, or the consequences of any task or request.
  • the software and/or hardware may encompass one server or multiple servers.
  • a user device is contemplated to encompass any computer hardware that is capable of running software appropriate to the relevant task at hand.
  • a user device may be embodied in a personal computer (i.e., a desktop, laptop, netbook, etc.), a smartphone, and/or a tablet. It is noted that a user device in the present context should not preclude multiple devices from being used in concert with one another.
  • a user device may encompass multiple devices that are used for sending/receiving, carrying out, or causing to be carried out any task or request, the consequences of any task or request, or the steps of any method described herein.
  • database is intended to encompass any collection of data irrespective of the structure of that data, the database management software, or the computer hardware on which the data is stored, implemented, or otherwise rendered available for use.
  • a database may reside on the same hardware as the process that stores or makes use of the information stored in the database or it may reside on separate hardware, such as a dedicated server or plurality of servers.
  • component is meant to include both software (appropriate to a particular hardware context) and/or hardware that is both necessary and sufficient to achieve the specific function(s) being performed.
  • component therefore, is intended to encompass software, hardware, and the combination of software and hardware, as appropriate.
  • computer usable information storage medium (or any variant thereon) is intended to include media of any nature and kind whatsoever including random access memory (“RAM”), read only memory (“ROM”), disks (such as CD-ROMs, DVDs, floppy disks, hard drivers, etc.), USB keys, solid state drives, tape drives, etc.
  • RAM random access memory
  • ROM read only memory
  • disks such as CD-ROMs, DVDs, floppy disks, hard drivers, etc.
  • USB keys solid state drives, tape drives, etc.
  • solid state drives tape drives, etc.
  • Fig. 1 is an exemplary schematic of one non-limiting diagram of a system 10.
  • the system 10 is constructed to receive input, in the form of a user's search query, and provide output, in the form of a SERP, among other output formats.
  • the system 10 includes a user device 12 connected via a network 14 to a search server 16.
  • the user device 12 is contemplated to be connected to the network 14 via a first communication link 18.
  • the server 16 is connected to the network 14 via a second communication link 20.
  • the first and second communication links 18, 20 are contemplated to be wireless, two-way communication links.
  • one or both of the first and second communication links 18, 20 may be wired two-way communication links.
  • a plurality of one way communication links may be employed, as required or as desired.
  • the user device 12 is contemplated to be any of a number of electronic devices that are capable of receiving input from a user and providing output to the user. While the user device 12 is contemplated to embody both an input element and an output element in the same device (i.e., via a touch screen), the user device 12 may have separate input and output elements. For example, the user device 12 may combine a keyboard with a monitor, as would be expected for a desktop computer.
  • the user device 12 may be a personal computer, cellular telephone, smart phone, personal data assistant, or any other type of electronic device capable of interfacing with the search server 16.
  • User input may be made via keystrokes, a touch screen interface, by voice recognition, or via any other methodology suitable for receiving and transmitting a search query.
  • output to the user will be provided via a suitable visual display.
  • the method, system, and/or computer readable memory of the instant description should not be understood to be limited solely to visual displays. Audio displays, among others, also are contemplated to fall within the scope of the present disclosure.
  • the search server 16 is contemplated to be any suitable device capable of executing the instructions for conducting a search of one or more resources available on the Internet, for example. While the search engine 16 is contemplated to be embodied in s single electronic device with resident software, the present technology is not limited solely to such a construction.
  • the search server 16 may be constructed from any number of electronic devices (embedding any number of different software instructions) that are connected together to execute the search instructions. Still further, the electronic devices that make up the search server 16 may not be in the same physical location. To the contrary, the electronic devices may be connected to one another via the Internet (or other suitable network) without departing from the scope of the present discussion.
  • the network 14 is contemplated to be the Internet. However, the embodiments described herein are not limited to a system 10 that relies on the Internet. To the contrary, the system 10 is contemplated to encompass any suitable network 14 including, but not limited to a local area network (“LAN”), wide area network (“ ' WAN”), personal area network (“PAN”), or the like.
  • LAN local area network
  • ' WAN wide area network
  • PAN personal area network
  • Figs. 2 and 3 are graphical presentations of a SERP 22 generated using the search engine available from Google Inc.
  • the SERP 22 is an example of the type of results generated by search engines in the prior art.
  • the SERP 22 is described in connection with Figs. 2 and 3 to identify the various elements of that are common to a SERP 22 that is generated to present results responsive to a search query from a user. While the fields in the SERP 22 are described using selected terms, it is noted that the terms are employed merely for purposes of description. To the extent that the same terms are used to describe aspects of various embodiments described herein, the terms should not be considered to be limiting thereof.
  • the SERP 22 includes a query field 24, a search tool field 26, a general search results field 28, a first vertical search results field 30, a second vertical search results field 32, and an object card 34. Each of these areas presents different information relevant to a user's search query 36.
  • the query field 24 is presented as a blank box into which a user may type the search query 36.
  • the search query 36 comprises the words "new york.”
  • the search tool field 26 presents a number of search icons 38, 40, 42, 44, 46, 48.
  • the search icons 38, 40, 42, 44, 46, 48 may be selected, by clicking, to present results relevant to that category of search.
  • the default search icon 38 is listed as "web," and provides a listing of web resources 50, such as URLs.
  • the URLs are clickable and redirect the user to the website associated with the URL.
  • the web resources 50 also are referred to as general search results 50. As indicated, the general search results 50 are listed in the general search results field 28.
  • search icons 40, 42, 44, 46, 48 may be referred to as vertical search icons 40,
  • search icons 42, 44, 48 because these search icons direct the user to a separate SERP that presents the selected search results. For example, if the user were to select the "Images" search icon 42, the user would be directed to a SERP that presents a number of images responsive to the search query 36. If the user were to select the "Videos" search icon 46, the user would be redirected to a SERP that presents a number of videos that are responsive to the search query 36.
  • a first vertical search result field 30 is presented.
  • the first vertical search result field 30 displays a number of images thumbnails 52 that are associated with the "Images" search icon 42. Clicking on one of the images thumbnails 52 causes the user to be redirected to a SERP listing various images, as discussed above. In other words, the first vertical search result field 30 presents a more visual avenue to images than the "Images" search icon 42.
  • a second vertical search field 32 also is presented.
  • the second vertical search result field 32 presents one or more results associated with the "News" icon 40.
  • the object card 34 is contemplated to present any of a number of results that are associated with the search query 36.
  • the object card 34 may include object card thumbnails 54 that call out points of interest, upcoming events, shopping opportunities, any or all of which may be of interest to the user who entered the search query 36 into the query field 24.
  • the content of the object card 34 is pre -populated with information about an "object" which is deemed to be a target object of the user's search query 36. For example, if the user's search query 36 comprises "New York" the target object may be determined to be "New York City”.
  • the object card 34 may present the results of related objects. Related objects include those objects processed by the search engine (and/or other search engines, as appropriate) on other occasions.
  • the object card 34 also is referred to as a "widget,” as should be apparent to those skilled in the art.
  • An object card 34 or widget 34 need not be presented solely on the right-hand side of the SERP 22.
  • the object card 34 or widget 34 may be intermixed with the general search results 50 in the general search results field 28.
  • another implementation of the widget 34 can display some or all results from a particular vertical, such as images and the like.
  • the SERP 22 generated by a search engines ranks the general search results 50 of the search query 36 and presents the general search results 50 in a results list 56.
  • the ranking may be established using any number of different criteria, including, for example, "long click" data.
  • the term "long click” refers to a web resource 50 that a user dwells on (or visits) for a long period of time after clicking on that web resource 50 (or general search result 50).
  • a "long” period of time typically refers to a period of time of 60 seconds or more. The understanding is that a person who spends 60 seconds or more on a web resource 50 typically has found information of interest on that web resource 50.
  • Fig. 3 replicates the same information presented in Fig. 2, except that a drop down box 58 has been exposed.
  • the drop down box 58 is provided in connection with the query field 24 and provides a number of suggests 60 that are associated with the search query 36. As indicated above, suggests 60 are presented to complete a search query 36 based on other, related queries entered by users in the query field 24 on other, prior occasions.
  • a suggest 60 redirects the user to a new SERP. Therefore, when a user clicks on the suggest 60, the amount of time the user spends on the new SERP cannot be used to rank the suggest 60. Long clicks are associated only with web resources and not with suggests 60. Accordingly, a search engine would not consider a click on a suggest 60 as a "long" one, even if the user were to spend 60 seconds or more on the SERP to which the suggest 60 directed the user. In this environment, the assessment of "long" clicks for suggests 60 leads to the value assigned being zero. With a value of "zero" any results associated with a suggest would be placed at the bottom of any general search result field 28, rendering the entry largely inconsequential.
  • Figs. 4 and 5 are provided to illustrate two possible SERPs 62, 64 that may be generated as a result of that method.
  • Fig. 4 is an illustration of a first embodiment of a SERP 62 generated according to the method described below.
  • suggests 60 are presented in a first widget 66.
  • the first widget 66 includes four suggest icons 68, 70, 72, 74, each of which is clickable and will redirect the user to additional web resources 50 that are associated with the search query 36 entered by the user into the query field 24.
  • Fig. 5 is an illustration of a second embodiment of a SERP 64 consistent with embodiments described herein.
  • suggests 60 are presented in the first widget 66 as in Fig. 4.
  • a second widget 76 is provided, below the first widget 66.
  • the second widget also includes four suggest icons 78, 80, 82, 84.
  • the suggest icons 78, 80, 82, 84 are clickable and will redirect the user to additional web resources 50 that are associated with the search query 36.
  • the suggest icons 68, 70, 72, 74, 78, 80, 82, 84 in the widgets 66, 76 direct the user to new SERPs, each of which present at least general search results 50 that are responsive to the search query 36 associated with each suggest 60.
  • the suggests 60 lack any long click data because they do not refer to specific web resources 50 on which a user might dwell for any period of measurable time.
  • the related long click is a variable that associates long click data with an individual suggest 60 by analyzing, among other things, a dwell time associated with the web resources 50 presented on the SERP associated with the suggest 60.
  • the calculation of related long click data for a suggest 60 may be made by any number of methodologies. For example, it may be prudent to use an average of the long click data for the top web resources 50 (perhaps the top 5, 10, 20, 50, 100, etc.) associated with the SERP for the selected suggest 60. As should be apparent, there are innumerable methodologies that may be employed. The method, system, and computer readable memory described herein are not intended to be limited to any one of them. To the contrary, any number of methodologies may be relied upon to determine the related long click data for a particular suggest 60. Web resources 50 and suggests 60 are contemplated to be ranked using the long click data and related long click data, respectively.
  • a separate machine learning formula may calculate the position of the widget 66, 76.
  • the formula is contemplated to allow calculation of the probability of a "long" click - in a sense of a sequence of transitions from the suggest 60, to the associated SERP, and, finally, to the web resource 50.
  • the related long click data may be obtained from a data with regard to the transition sequence (at first, to the SERP, then, to the subsequent web resource 50), which data may be obtained from logs of prior search sessions.
  • the search engine is contemplated to receive information that the user clicked on a related query in the widget 66, 76 and save the information relating not only to the separate transitions but to the relationship between them.
  • this information can be obtained from a search engine log (not depicted), which search engine log stores information about past user searches, as well as past user interactions with the past SERPs generated in response to the past user searches. This data may be employed in future searches, as expected. Regardless of the manner in which the related long click data is obtained and/or calculated, the obtained value for the related long click may then be used to rank the widget 66, 76 among other elements of search results, including the web resources 50.
  • the suggests 60 when the suggests 60 are presented to the user, they may be presented in any of a number of potential formats. In Figs. 4 and 5, the suggests 60 are presented in the widgets 66, 76. The suggests 60 may be presented, alternatively, as alphanumeric descriptions, similar to the web resources 50. It is preferred that the suggest 60 be presented in the form of pictures or visual images, but this is not required.
  • Fig. 6 is a flow chart that provides a graphical summary of one exemplary, non- limiting method 90 consistent with the instant disclosure.
  • the method 90 is merely exemplary of the types of methods that are contemplated to fall within the scope of the present discussion. In other words, the method 90 is not intended to be limiting of the any of the described embodiments.
  • the method 90 starts at block 92. [00104] From block 92, the method 90 proceeds to block 94 where the search server 16 receives a search query 36 from the user device 12.
  • the method proceeds to block 96.
  • the search server 16 determines a first search result set or results list 56.
  • the first search result set is contemplated to include at least a first general search result 50 and a second general search result 51.
  • the first general search result 50 and the second general search result 51 are ranked using at least a first ranking algorithm.
  • the 16 based on previous search sessions, retrieves at least one related query.
  • the at least one related query is related to the search query 36.
  • the method 90 proceeds to block 100.
  • the search server 16 generates a first suggest object, i.e., a suggest 60, comprising at least one clickable element.
  • the at least one clickable element is associated with the at least one related query.
  • the method then proceeds to block 102, where the search server 16 applies a second ranking algorithm to rank the first suggest object 60 relative to the first general search result 50 and the second general search result 51.
  • the search server 16 generates a second search result set that includes, in a ranked order, the first suggest object 60, the first general search result 50, and the second general search result 51.
  • the search server 16 generates the search result page (i.e., SERP 62 or SERP 64) based on the second search result set.
  • the method 90 ends at block 108.
  • the method 90 also may include additional steps.
  • the method 90 may include sending, via the network 14, the search result page for display on the user device 12 and receiving, from the user device 12, a selection of the at least one clickable element.
  • the method 90 may include applying, by the search server 16, a third raking algorithm to a plurality of related queries including the at least one related query and generating, by the search server 16, a third search result set that comprises the plurality of related queries in ranked order.
  • the third ranking algorithm may employ a variable dependent on a long click to generate a related long click.
  • the second ranking algorithm may employ the related long click to generate the second search result set.
  • the method 90 may include generating, by the search server 16, a related queries search result page based on the third search result set and sending, via the network 14, the related queries search result page for display on the user device 12.
  • the method 90 may cause a display, on the user device 12, of the related queries search result page.
  • the at least one related query is contemplated to be a suggest 60.
  • the first search result set 50 may include a vertical search result ranked according to a fourth ranking algorithm.
  • the fourth ranking algorithm may be a ranking algorithm specifically dedicated to ranking search results from the associated vertical search result. If so, the vertical search result may be ranked, as a first search result, together with other first search results via the first ranking algorithm.
  • Vertical search results may be embodied in a widget.
  • Suggests 60 may be embodied in widgets 66, 76, as discussed above.
  • the first ranking algorithm is contemplated to employ a variable dependent on a long click.
  • the second ranking algorithm is contemplated to employ a variable dependent on a related long click, where the related long click is a function of a long click associated with a related search result of the at least one related query.
  • the long click may be calculated based on an average time users dwell on at least one of the first general search result 50 and the second general search result 51.
  • the search server (16) determines, by the search server (16), a first search result set, [00124] wherein the first search result set comprises at least a first general search result (50) and a second general search result (51), and
  • first general search result (50) and the second general search result (51) are ranked using at least a first ranking algorithm
  • the at least one related query is related to the search query (36);
  • the at least one clickable element is associated with the at least one related query
  • (51) comprises a plurality of web resources.
  • search result set including a first general search result (50) and a second general search result (51), the first general search result (50) and the second general search result (51) having been ranked using a first ranking algorithm;
  • a computer readable storage medium comprising executable instructions for operation of a method (90) of generating a search engine results page in a system (10) comprising a search server (16), a user device (12), and a network (14) connecting the search server (16) to the user device (12), where the executable instructions, when executed, result in operations comprising:
  • the first search result set comprises at least a first general search result (50) and a second general search result (51), and
  • the at least one clickable element is associated with the at least one related query
  • the first search result set comprises at least a first general search result (50) and a second general search result (51), and
  • first general search result (50) and the second general search result (51) are ranked using at least a first ranking algorithm

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • User Interface Of Digital Computer (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

L'invention concerne un procédé de génération d'une page de résultats de moteur de recherche consistant à recevoir, par un serveur de recherche, une interrogation de recherche d'un dispositif d'utilisateur. En se basant sur l'interrogation de recherche, le serveur de recherche détermine un premier ensemble de résultats de recherche, des premier et second résultats de recherche globaux étant classés à l'aide d'un premier algorithme de classement. En se basant sur des sessions de recherche antérieures, le serveur de recherche récupère une interrogation apparentée relative à l'interrogation de recherche et génère un premier objet de suggestion qui peut être un élément cliquable. Le serveur de recherche applique ensuite un second algorithme de classement pour classer le premier objet de suggestion par rapport aux premier et second résultats de recherche globaux afin de générer un second ensemble de résultats de recherche qui comprend, dans un ordre classé, le premier objet de suggestion, le premier résultat de recherche global et le second résultat de recherche global. Le serveur de recherche génère la page de résultats de recherche sur la base du second ensemble de résultats de recherche.
PCT/IB2016/050563 2015-06-30 2016-02-04 Procédé, système et mémoire lisible par ordinateur de génération de résultats de recherche classés incorporant des suggestions WO2017001944A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
RU2015125961A RU2015125961A (ru) 2015-06-30 2015-06-30 Способ (варианты), система и машиночитаемый носитель для создания ранжированных результатов поиска
RU2015125961 2015-06-30

Publications (1)

Publication Number Publication Date
WO2017001944A1 true WO2017001944A1 (fr) 2017-01-05

Family

ID=57607940

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/IB2016/050563 WO2017001944A1 (fr) 2015-06-30 2016-02-04 Procédé, système et mémoire lisible par ordinateur de génération de résultats de recherche classés incorporant des suggestions

Country Status (2)

Country Link
RU (1) RU2015125961A (fr)
WO (1) WO2017001944A1 (fr)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
RU2677380C2 (ru) * 2017-04-05 2019-01-16 Общество С Ограниченной Ответственностью "Яндекс" Способ и система ранжирования множества документов на странице результатов поиска
WO2023142938A1 (fr) * 2022-01-27 2023-08-03 华为技术有限公司 Procédé d'interrogation de carte et dispositif électronique

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050149504A1 (en) * 2004-01-07 2005-07-07 Microsoft Corporation System and method for blending the results of a classifier and a search engine
US20070124301A1 (en) * 2004-09-30 2007-05-31 Elbaz Gilad I Methods and systems for improving text segmentation
US20100082604A1 (en) * 2008-09-22 2010-04-01 Microsoft Corporation Automatic search query suggestions with search result suggestions from user history
US20100228710A1 (en) * 2009-02-24 2010-09-09 Microsoft Corporation Contextual Query Suggestion in Result Pages
US20130091116A1 (en) * 2006-12-12 2013-04-11 Yahoo! Inc. Selecting and presenting search results based on distinct taxonomies
US20130226913A1 (en) * 2012-02-29 2013-08-29 International Business Machines Corporation Use of statistical language modeling for generating exploratory search results
US20140379689A1 (en) * 2009-12-07 2014-12-25 Google Inc. Framework for Suggesting Search Terms
WO2015028898A1 (fr) * 2013-08-29 2015-03-05 Yandex Europe Ag Système et procédé pour l'affichage de résultats de recherche verticale les plus pertinents
US20150161255A1 (en) * 2004-06-28 2015-06-11 Google Inc. Systems and Methods for Deriving and Using an Interaction Profile

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050149504A1 (en) * 2004-01-07 2005-07-07 Microsoft Corporation System and method for blending the results of a classifier and a search engine
US20150161255A1 (en) * 2004-06-28 2015-06-11 Google Inc. Systems and Methods for Deriving and Using an Interaction Profile
US20070124301A1 (en) * 2004-09-30 2007-05-31 Elbaz Gilad I Methods and systems for improving text segmentation
US20130091116A1 (en) * 2006-12-12 2013-04-11 Yahoo! Inc. Selecting and presenting search results based on distinct taxonomies
US20100082604A1 (en) * 2008-09-22 2010-04-01 Microsoft Corporation Automatic search query suggestions with search result suggestions from user history
US20100228710A1 (en) * 2009-02-24 2010-09-09 Microsoft Corporation Contextual Query Suggestion in Result Pages
US20140379689A1 (en) * 2009-12-07 2014-12-25 Google Inc. Framework for Suggesting Search Terms
US20130226913A1 (en) * 2012-02-29 2013-08-29 International Business Machines Corporation Use of statistical language modeling for generating exploratory search results
WO2015028898A1 (fr) * 2013-08-29 2015-03-05 Yandex Europe Ag Système et procédé pour l'affichage de résultats de recherche verticale les plus pertinents

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
RU2677380C2 (ru) * 2017-04-05 2019-01-16 Общество С Ограниченной Ответственностью "Яндекс" Способ и система ранжирования множества документов на странице результатов поиска
US10754863B2 (en) 2017-04-05 2020-08-25 Yandex Europe Ag Method and system for ranking a plurality of documents on a search engine results page
WO2023142938A1 (fr) * 2022-01-27 2023-08-03 华为技术有限公司 Procédé d'interrogation de carte et dispositif électronique

Also Published As

Publication number Publication date
RU2015125961A (ru) 2017-01-11

Similar Documents

Publication Publication Date Title
KR101303488B1 (ko) 인라인 문맥 질의들을 사용하는 검색 시스템들 및 방법들
US9760541B2 (en) Systems and methods for delivery techniques of contextualized services on mobile devices
KR101667344B1 (ko) 검색 결과들을 제공하는 방법 및 시스템
US7899818B2 (en) Method and system for providing focused search results by excluding categories
US10061820B2 (en) Generating a user-specific ranking model on a user electronic device
US9348935B2 (en) Systems and methods for augmenting a keyword of a web page with video content
US8725725B2 (en) Method and system for assessing relevant properties of work contexts for use by information services
JP5572596B2 (ja) 検索結果内におけるプレーストコンテンツの順序付けのパーソナライズ
KR101872547B1 (ko) 엔티티와 연관된 액션 및 제공자의 제시 기법
US9053158B1 (en) Method for human ranking of search results
US20130054371A1 (en) Systems and methods for contextualizing services for inline mobile banner advertising
US20150178350A1 (en) Automatic method and system for formulating and transforming representations of context used by information services
US20130054672A1 (en) Systems and methods for contextualizing a toolbar
US10503803B2 (en) Animated snippets for search results
KR102148968B1 (ko) 컨텍스트 정보 제공 시스템 및 방법
CA2571867A1 (fr) Personnalisation d'annonces publicitaires basee sur des resultats dans un moteur de recherche
EP2689349A1 (fr) Systèmes et procédés de récolte de contenu étendue pour une contextualisation
US10339191B2 (en) Method of and a system for processing a search query
WO2017001944A1 (fr) Procédé, système et mémoire lisible par ordinateur de génération de résultats de recherche classés incorporant des suggestions
US10235679B2 (en) Learning a ranker to rank entities with automatically derived domain-specific preferences
WO2013033445A2 (fr) Systèmes et procédés permettant de contextualiser une barre d'outils, une image et un bandeau publicitaire mobile en ligne
WO2023028596A1 (fr) Systèmes et procédés d'hyperliens dynamiques

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16817324

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16817324

Country of ref document: EP

Kind code of ref document: A1