JP6448606B2 - Apparatus, method, program, and system for providing purpose-specific application in search engine result page - Google Patents

Description

Copyright Notice The disclosure part of this patent document contains the subject matter of copyright protection. The copyright owner does not object to any third party copying the patent document or patent disclosure as it appears in the US Patent and Trademark Office patent file or record, but otherwise Reserving copyright.

  The invention described herein generally relates to providing a purpose-specific context application in response to a search query. More particularly, the present invention analyzes a user query, identifies a user context, associates the user query and actions based on the user context with a stored application, and presents one or more applications on the user device. The present invention relates to a system and method.

  As the Internet has expanded rapidly, the role of search engines has grown dramatically. It has rapidly become impossible for human users to manually use vast amounts of Internet content. Therefore, the use of search engine technology has become an indispensable tool for increasing the usefulness of the Internet.

  While increasing the relevance of search results has made great strides, less progress has been made in considering search engine user contextual data and historical data. Also, there has been little progress in incorporating applications or widgets within the search results page interface. In contrast, current search engines typically provide a simple list of hyperlinks in response to a user query. Therefore, the user needs to identify and select the link most relevant to his query. In addition, the user needs to take additional steps to obtain the results he is seeking.

  An example where a user tries to plan a vacation is particularly remarkable. In the current environment, the user can enter a query such as “Airplane from Boston to Los Angeles”. In response, the user can receive multiple sites for ticket purchases. This paradigm is disadvantageous for a number of reasons. First, the user still needs to search a myriad of search results that include many websites that may not be what they want. Second, current search engines only provide information about instant search queries. That is, with current state-of-the-art technology, a user planning a vacation from Boston to Los Angeles can book a hotel (in addition to a search for “Airplanes from Boston to Los Angeles”), rent a car, Need to do multiple searches to find and do other sightseeing.

  Therefore, the current state-of-the-art technology cannot consider the “overall picture” of user search behavior. That is, current search engine technology simply responds directly to the user's query and does not attempt to confirm the nature of the query by using the user's context and past search history. Therefore, current state-of-the-art technology is required to enhance the user search experience by providing contextual applications that are provided according to the anticipated user objectives.

  The present invention relates to a system and method for providing purpose-specific applications in search engine results pages. The method receives a search query from a user and interprets the search query. In one embodiment, interpreting the user search query includes generating a tokenized form of the search query. In an alternative embodiment, interpreting the user search query further includes adding to the search query data representing a user context that includes data stored in the user profile. In an alternative embodiment, interpreting the user search query further includes adding to the search query data representing a user context that includes the user's past search query. The method then first associates the interpreted query with one or more action templates, and associating the interpreted query with one or more action templates includes one or more associated with the interpreted query. The step of selecting the operation.

  The method then associates the selected one or more actions associated with the interpreted query with a plurality of applications and selects one or more applications associated with the one or more actions. In one embodiment, associating the selected one or more actions associated with the interpreted query with the plurality of applications includes using an edit decision. In an alternative embodiment, the step of associating the selected one or more actions associated with the interpreted query with a plurality of applications is performed algorithmically. In an alternative embodiment, associating the selected one or more actions associated with the interpreted query with a plurality of applications and selecting the one or more applications associated with the one or more actions comprises: The method further includes pre-filling the application based on the search query and data extracted from the user context and the search history.

  In one embodiment, the method ranks one or more selected applications. In one embodiment, ranking one or more selected applications includes selecting a top application based on the popularity of the application in the user's community. In an alternative embodiment, ranking one or more selected applications includes selecting a top application based on the application's historical revenue. Finally, the method displays these one or more applications in the search results page.

  The invention further relates to a system for providing purpose-specific applications in search engine results pages. The system includes a plurality of client devices coupled to a search engine via a network, and these client devices send a search query to the search engine. The system further comprises a query interpreter that interprets the search query. In one embodiment, the query interpreter further generates a tokenized form of the search query. In an alternative embodiment, the query interpreter further adds data representing the user context to the search query. In one embodiment, the user context includes data stored in the user profile. In an alternative embodiment, the user context includes the user's past search query. The system includes a behavior reasoning module that associates the interpreted query with one or more behavior templates, and associating the interpreted query with one or more behavior templates includes one or more associated with the interpreted query. Selecting an action.

  The system associates one or more selected actions associated with the interpreted query with multiple applications and selects one or more applications associated with one or more actions stored in the application database. An application scorer is further provided. In one embodiment, associating the selected one or more actions associated with the interpreted query with the plurality of applications includes using an edit decision. In another embodiment, the step of associating the selected one or more actions associated with the interpreted query with a plurality of applications is performed algorithmically. In another embodiment, the application scorer further ranks one or more selected applications.

  In a further embodiment, an application scorer can further rank one or more selected applications. In one embodiment, ranking one or more selected applications includes selecting a top application based on the popularity of the application in the user's community. In an alternative embodiment, ranking one or more selected applications includes selecting a top application based on the application's historical revenue. The system further comprises a front-end server that displays these one or more applications in a search results page.

  The invention is illustrated in the drawings of the accompanying drawings, which are intended to be illustrative rather than limiting, wherein the same reference numerals indicate the same or corresponding parts.

FIG. 2 is a block diagram illustrating a system for generating a context action based on a user context, according to one embodiment of the invention. FIG. 2 is a flow diagram illustrating a system for generating contextual behavior based on a user query, according to one embodiment of the invention. FIG. 6 is a flow diagram illustrating a method for generating a context action based on a user context, according to one embodiment of the invention. FIG. 6 is a screen diagram illustrating a context application on a search results page according to one embodiment of the invention. FIG. 6 is a screen diagram illustrating a context application on a search results page according to one embodiment of the invention. FIG. 6 is a screen diagram illustrating a context application on a search results page according to one embodiment of the invention.

  In the following description, reference is made to the accompanying drawings that form a part hereof, and in which is shown by way of illustration specific embodiments in which the invention may be practiced. It should be understood that other embodiments may be utilized and structural changes may be made without departing from the scope of the present invention.

  FIG. 1 is a block diagram illustrating a system for generating a context action based on a user context, according to one embodiment of the present invention. As shown in the embodiment of FIG. 1, a plurality of client devices 102 a, 102 b, 102 c are connected to a search engine 106 via a network 104. The search engine 106 includes a plurality of components including a front-end server 108, a behavior inference module 110, a query interpreter 112, an application scorer 114, an application parameterization module 116, an application index 118 and an application database 120.

  The client devices 102a, 102b and 102c are general-purpose computer devices (personal computers, TV set top boxes, mobile devices, etc.) having a central processing unit, a memory unit, a permanent storage device, an audio / video output device, a network interface, and the like. Can do. Client devices 102a, 102b and 102c communicate via a network 104, which can be a local network or a wide area network such as the Internet. In this embodiment, client devices 102a, 102b, and 102c send requests to search engine 106 via HTTP, WAP, or similar protocols for exchanging text, images, and other data between the client / server.

  The client devices 102a, 102b, and 102c send requests for search results to the search provider 110, and these search requests include search terms. This search term may be a search term generated by the user, or may be a user-supported, ie, proposed search term. In one embodiment, client devices 102a, 102b, and 102c can send an HTTP request that includes a text query string. For example, a user operating client devices 102a, 102b, and 102c may enter a query string that includes a search term such as “New York, ny sushi” and submit this query via an HTML-format element. .

  The search engine 106 receives a query from the client device via the front end server 108. In one embodiment, the front-end server 108 may include various hardware components that run web server software, such as APACHE HTTP SERVER, MICROSOFT IIS, or any other suitable web server software. In the illustrated embodiment, the front end server 108 receives a search request entered by a given user. The front-end server 108 can first retrieve multiple search results from a search index (not shown) in response to a given query. In one embodiment, the plurality of search results includes a plurality of web pages that include at least one or more of the submitted search terms. In the illustrated embodiment, various methods known in the art can be used to identify multiple search results within the search index. The front-end server 108 further transmits the search result and application to the client devices 102a, 102b, and 102c.

  The front-end server 108 communicates with the behavior reasoning module 110. In the illustrated embodiment, the front-end server 108 sends the received query to the behavior reasoning module 110. In an alternative embodiment, the front-end server 108 further sends metadata about this query to the motion reasoning module 110. For example, the front-end server 108 can send information in a cookie associated with the query submitting user, including data such as a user ID. In the illustrated embodiment, the motion reasoning module 110 indexes the motion for fast retrieval, stores metadata for retrieving and weighting the motion, and selecting the motion based on context.

  When the query interpreter 112 receives the query, it analyzes the given query and provides an interpretation of the query that represents the user's context. For example, the query interpreter 112 may receive a query “a plane from bos (Boston) to lax (Los Angeles)”. In response, the query interpreter 112 can translate the plaintext query into a structured format such as “<token: plane> <city: Boston> <token: bound> <city: Los Angeles>”. As shown, the query interpreter 112 converts the plaintext query into a tokenized version based on the characteristics of the received query. In the previous example, the query interpreter 112 extracts the terms “airplane” and “bound” as tokens, and the terms “bos (Boston)” and “lax (Los Angeles)” as cities. In an alternative embodiment, the query interpreter can utilize multiple external metrics in generating the interpretation. For example, the query interpreter 112 can utilize user profiles (including long-term and short-term activity) to match concepts and entities or objects with the original user query. In addition, the query interpreter 112 can use the collected index regarding user interaction with the application. In one embodiment, the metrics represent the quality of the user experience using the application and are quickly improved based on user behavior.

  The query interpreter 112 returns the extracted context to the behavior reasoning module 110. The behavior inference module 110 selects a behavior based on this identified context. In one embodiment, the motion reasoning module 110 compares the interpreted context with multiple templates. For example, the motion reasoning module 110 can match the above interpreted context to two template actions “<purchase air ticket>” and “<check flight arrival / departure time>”.

  The behavior reasoning module 110 sends the identified behavior and the interpreted context to the application scorer 114. In the illustrated embodiment, the application scorer 114 selects an application from the application index 118 based on the received context and behavior. The application scorer 114 can also score and rank these identified applications, as described herein. In addition to the application scorer module 114, an application parameterization module 116 extracts relevant parameters from the interpreted context. For example, in the above example, the application parameterization module 116 can extract the terms “bos (Boston)” and “lax (Los Angeles)” as related parameters. In an alternative embodiment, the application parameterization module 116 may further convert “box (Boston)” and “lax (Los Angeles)” to “Boston” and “Los Angeles”, or perform other conversion operations. it can. In addition to identifying parameters, the application parameterization module 116 can also pre-fill application fields with the identified parameters. In the given example, application parameterization module 116 may insert the cities Boston and Los Angeles into an identified application, such as an application that provides arrival and departure times.

  The behavior reasoning module 110 and the application parameterization module 116 can access the application via the application index 118. Application index 118 indexes applications for fast searching. In the illustrated embodiment, the application index 118 stores metadata to assist in fast indexing and ranking of applications. In addition to providing a fast search, the application index 118 can further process additional inputs to the application database 120. That is, the application index 118 can receive new applications and data from the developer, index the applications, and store them in the application database 120. Search engine 106 may further provide an API that allows developers to create applications and maintain applications stored in application database 120.

  FIG. 2 is a flow diagram illustrating a system for generating a context action based on a user query, according to one embodiment of the invention. In the illustrated embodiment, the method 200 first receives a search query at step 202. In the illustrated embodiment, the method 200 can receive a query via a search form such as an HTML format or other input mechanism such as a browser toolbar. To illustrate the previous example, the method 200 may receive the query as a plain text string such as “Bos (Boston) to Lax (Los Angeles)”.

  After receiving the search query, method 200 parses and interprets the query at step 204. In the illustrated embodiment, the method 200 converts a plain text string into a tokenized form. For example, the method 200 uses a plaintext string “airplane from box (Boston) to lax (Los Angeles)” and a token “<token: plane> <city: Boston> <token: bound> <city: Los Angeles>”. Can be converted into a generalized interpretation. As can be seen, the method 200 translates “bos” into a token “<city: Boston>” and “lax” into a token “<city: Los Angeles>”. In one embodiment, the method 200 performs various text processing operations to reformat the query in the event that various terms are misspelled or misplaced.

  Next, the method 200 associates the identified and interpreted query with an action template at step 206. In the illustrated embodiment, the method 200 can identify multiple action templates that match the tokenized interpretation. For example, the method 200 can identify that the tokens “Airplane” and “To” indicate that the user is moving between two locations. Based on this interpretation, the method 200 can identify a plurality of templates, such as a “buy air ticket” operation and a “check flight arrival / departure time” operation. In one embodiment, each identified action has a given weight. For example, the “buy air ticket” operation may be given higher weight than the “check flight arrival / departure time” operation. This behavioral weight determination may be based on historical user metrics (eg, users are often interested in purchasing tickets), revenue metrics (eg, revenue per application), or a combination thereof. it can.

  The method 200 then associates these operations with multiple applications at step 208. For example, the method 200 can map “buy air tickets” to an ORBITZ application that allows a user to purchase air tickets between desired destinations, and “check flight arrival / departure times”. Can also be associated with an aircraft scheduling application such as that provided by a FLIGHTSSTATS or similar application developer. In one embodiment, the method 200 can pre-fill the application with the data tokenized in step 204. For example, the method 200 can pre-fill the ORBITZ application with the departure city (Boston) and the arrival city (Los Angeles). In one embodiment, the association between actions and applications is based on edit decisions. In an alternative embodiment, the association between actions and applications is based on algorithmic decisions.

  The method 200 then ranks the identified applications at step 210. Application ranking can utilize various indicators such as the popularity of the application based on the number of users using the application after entering a particular type of query. For example, a user who inputs a query of the form “airplane from CITY 1 (city 1) to CITY 2 (city 2)” (in this case, city 1 represents a departure city and city 2 represents an arrival city) can be obtained from the FLAGTSTATS application. May use the ORBITZ application frequently. Various user index collection devices embedded in the application or application container page can collect this data. Based on this data, the method 200 may determine that the ORBITZ application should be ranked higher than the FLIGHTSTATS application. In alternative embodiments, the method 200 may use other rankings, such as ranking by the amount of advertising revenue earned by the application. Alternatively or in combination, the method 200 may rank applications based on predetermined application quality, reliability, or various other metadata parameters associated with the application.

  After ranking the applications, method 200 displays the applications with the desired search results at step 212. The method 200 can display an application with search results in various ways. In one embodiment, the method 200 displays the application alongside the search results as shown in FIGS. In another embodiment, the method 200 displays the application as a hyperlink below the search results as shown in FIG.

  FIG. 3 is a flow diagram illustrating a method for generating a context action based on a user context, according to one embodiment of the invention. As shown in the embodiment of FIG. 3, the method 300 first receives a search query at step 302. In the illustrated embodiment, the method 300 can receive a query via a search form such as an HTML format or other input mechanism such as a browser toolbar. To illustrate the previous example, the method 300 may receive the query as a clear text string such as “Bos (Boston) to Lax (Los Angeles)”.

  After receiving the search query, method 300 determines in step 304 whether a user context exists. In one embodiment, determining whether a user context exists involves determining whether a user profile or similar structure can be associated with the query submitting user. Further, determining whether a context exists can involve determining whether there are additional past queries associated with a given user. In one embodiment, a search engine can store context data. In an alternative embodiment, the context data can be stored on the client device in the form of a cookie or similar mechanism. If the method 300 determines that the user context is not found, the method 300 continues based on the query as described in connection with FIG.

  On the other hand, if the method 300 determines that a context exists, it first analyzes and interprets the query at step 306. The steps for analyzing and interpreting the query are virtually similar to the method described in connection with FIG. 2 and will not be repeated for clarity. The method 300 adds context data to the query at step 308 after parsing and interpreting the query.

  Continuing with the example of FIG. 2, the method 300 begins with the query “<Token: Airplane> <City: Boston> <Token: Toward> for the query“ Airplane from bos (Boston) to lax (Los Angeles) ”. An interpretation of <city: Los Angeles> can be generated, and the method 300 adds such contextual data, such as various features of the user and past queries, after generating such an interpretation. By using user profiles or estimating user trends, users are geographically located in Boston, frequently visit the AMERIKA AIRLINES website, ORBITZ more than TRAVELOCITY Prefer to visit the restaurant web page often Various aspects of the user's personality can be added to the query, etc. Examples of these user behaviors can be based on a user profile, historical user data, or a combination thereof, and the method 300 can be The past query can be utilized to adapt the identified action, for example, the method 300 can identify that the user has recently performed a search for “Sightseeing Los Angeles” or “Hotel Los Angeles”. Therefore, it is possible to select an operation relating to tourism in spite of an instant search relating to an air ticket.

  After adding the context to the interpreted user query, the method 300 associates the context with the action at step 310. In the illustrated embodiment, the association of contexts and actions can be performed in a manner similar to that described in connection with FIG. However, the method 300 can utilize the user's context in addition to simply using the user query. For example, if the user has recently searched for sightseeing in Los Angeles, the method 300 may deduce that the user is interested in traveling rather than observing someone else's trip. Only the “do” action can be returned and the “check flight arrival / departure time” action can be ignored. As another example, the method 300 may return a “rental car” operation and / or a “check weather in destination city” operation based on a user search for tourism. As yet another example, the method 300 may “find a restaurant in the destination city” based on a combination of the user's interest in tourism and the user's frequent visits to the restaurant website. The action can be returned.

  In an alternative embodiment, method 300 can chain applications, in which subsequent applications are “chained” based on user interaction with the initial application. For example, the method 300 can determine that the user first interacted with the restaurant application, and then chained additional applications, such as a movie application, based on the interaction with the first application. In this embodiment, the method 300 chains the applications based on a recognized purpose, such as the goal of “going out at night” that combines a restaurant application and a movie application. In one embodiment, chained applications can be provided to users asynchronously. For example, the method 300 can display only one application (such as a restaurant application) on a search results page, and if the user interacts with this application, a second chained application (such as a movie application) is displayed. The search result page can be updated to include. In contrast, if the user does not interact with the restaurant application, the search result page remains unchanged.

  After determining a plurality of actions, method 300 associates actions with applications based on historical data at step 312 and ranks the actions at step 314. In the illustrated embodiment, the operation based on the history data can be associated with the application in the same manner as described with reference to FIG. However, the method 300 can return applications that are highly relevant to individual users based on the user's context, preferences, and past queries. In one embodiment, the method 300 can organize related applications into an application hierarchy. Continuing the above example, the method 300 may select multiple applications (such as ORBITZ, TRAVELOCITY, and EXPEDIA applications) related to air ticket purchases. The method 300 can further indicate that these applications have the highest priority and should place these applications within the highest priority level. However, the method 300 uses context information that the user is more likely to use ORBITZ than TRAVELOCITY to rank the application so that ORBITZ is the highest priority application within this priority level. can do. The method 300 can determine as many priority levels as necessary. For example, the method 300 may further create a second priority level that includes tourism, such as a HERTZ car rental application, a YAHOO WEATHER application, and a YELP restaurant application. As described in connection with FIG. 2, application ranking can utilize various indicators such as the popularity of the application based on the number of users using the application when entering a particular type of query. This data can be collected by various user indicator collection devices embedded within the application or application container page. In another embodiment, the method 300 may utilize a past user search to identify a given search goal. For example, the method 300 may identify that the user has searched for a meal suggestion and movie theater at a given location. Based on this search pattern, the method 300 can bundle related applications to match the search history, such as bundling meal and theater applications.

  The method 300 displays the application along with the desired search results in step 316 after ranking the application. Search results and applications can be displayed in various ways. In one embodiment, the application is displayed alongside the search results as shown in FIGS. In another embodiment, the application can be displayed as a hyperlink below the search results as shown in FIG.

  FIG. 4 is a screen diagram illustrating a context application on a search results page, according to one embodiment of the invention. As shown in the embodiment of FIG. 4, a search engine results page is displayed in response to a user query 402 ("bos (Boston) to la (Los Angeles)"). The search engine result page first displays the search result 404. In addition to the search results, the search engine results page includes the application 406 selected for the query submitting user. Specifically, the search engine results page displays the ORBITZ application returned in response to a user query of an airplane from Boston to Los Angeles. Further, the application 406 fills in the “From” field and the “To” field based on this user query.

  FIG. 5 is a screen diagram illustrating a context application on a search results page, according to one embodiment of the invention. As shown in the embodiment of FIG. 5, the user inputs a search query 502 of “mortgage interest rate” and receives a plurality of search results 504. The embodiment of FIG. 5 shows an application 508 selected without using user profile data or history data and an application 506 selected based on the user profile data.

  Application 508 is presented to the user based primarily on the query “mortgage interest rate”. Thus, the search engine determines that the mortgage interest rate query may be related to an application that shows the average mortgage interest rate across the country. In contrast, application 506 is presented to the user based on a user profile or similar mechanism. As shown, application 506 represents an application that allows a user to search for a local mortgage rate. The search engine pre-fills the application with user locations that may be stored in the user profile or similar device based on the user profile. The search engine also pre-fills loan amount, loan purpose and loan type. In one embodiment, the present invention can determine which fields to pre-fill based on privacy considerations. In one embodiment, such data can be collected from past user data. For example, the search engine may identify that the user has recently entered the queries “$ 200,000 refinancing” and “interest-only loan wikipedia” within one or two hours. Based on these queries, the search engine may determine that the user is interested in an interest-only $ 200,000 refinancing loan and pre-fill application 506 with this contextual data.

  FIG. 6 is a screen diagram illustrating a context application on a search results page, according to one embodiment of the invention. According to the embodiment shown in FIG. 6, the user inputs a search query 602 “Napa's Restaurant” and receives a plurality of search results 604. The user also receives a plurality of application hyperlinks 606. Among other things, these application links are presented to the “guest” user 608. Accordingly, these application links are generated based on the interpretation of the query “Napa's Restaurant”. As shown, these interpretations correspond to table reservations, restaurant ratings, and sharing a restaurant with others.

  FIG. 1 to FIG. 6 are conceptual explanatory diagrams that allow the description of the present invention. It should be understood that various aspects of embodiments of the present invention can be implemented in hardware, firmware, software, or a combination thereof. In such embodiments, various components and / or steps can be implemented in hardware, firmware, and / or software to implement the functionality of the present invention. That is, the same part of a hardware, firmware, or software module can execute one or more of the illustrated blocks (such as components or steps).

  In a software implementation, computer software and / or data (such as programs or other instructions) are stored on a machine-readable medium as part of a computer program product, and the computer is connected via a removable storage drive, hard drive, or communication interface. Loaded into a system or other device or machine. A computer program (also referred to as computer control logic or computer readable program code) is stored in main memory and / or secondary memory and executed by one or more processors (such as a controller) as described herein. Causes one or more processors to perform the functions of the present invention. As used herein, the terms “machine-readable medium”, “computer program medium” and “computer-usable medium” are generally referred to as random access memory (RAM), read-only memory (ROM), (magnetic or optical disk, or flash Used to mean media such as removable storage units (such as memory devices), hard disks, electrical, electromagnetic, optical, acoustic, or other forms of propagation signals (such as carrier waves, infrared signals, digital signals, etc.) .

  It should be noted that the above figures and examples are intended to limit the scope of the invention to a single embodiment, as other embodiments are possible by replacing some or all of the elements described or illustrated. It is not a thing. Further, where known components may be used to partially or fully implement some of the elements of the present invention, such known component parts are not subject to the present invention in order to avoid obscuring the present invention. Only those necessary to understand are described, and detailed descriptions of other parts of these known components are omitted. In this specification, unless expressly stated otherwise, an embodiment showing a singular component should not necessarily be limited to other embodiments comprising a plurality of identical components, and vice versa. Moreover, Applicants intend that any terms in the specification or claims have no unusual or special meaning unless explicitly stated otherwise. In addition, the present invention includes presently known equivalents and future known equivalents of known components cited herein for illustrative purposes.

  The above description of specific embodiments is based on the application of knowledge within the pertinent art (s) (including the content of the documents cited herein and incorporated herein by reference). Such specific embodiments can be modified and / or easily adapted to various applications without undue experimentation by others without departing from the general concept of the present invention. Thus, the general properties of the present invention are fully clarified. Accordingly, such adaptations and modifications are intended to be included within the meaning and scope of equivalents of the disclosed embodiments based on the teachings and guidance presented herein.

  Although various embodiments of the present invention have been described above, it should be understood that these have been presented for purposes of illustration and not limitation. It will be apparent to those skilled in the relevant art (s) that various changes in form and detail may be made without departing from the spirit and scope of the invention. Accordingly, the present invention should not be limited by any of the above-described exemplary embodiments, but should be defined only in accordance with the following claims and their equivalents.

In addition, this invention can claim a patent for every item described below as an example.
Item 1.
A method for providing purpose-specific applications within search engine results pages,
Receiving a search query from a user;
Interpreting the search query, including adding data representing a user context to the search query;
Associating the interpreted query with one or more action templates comprising selecting one or more actions associated with the interpreted query;
Associating the selected one or more actions associated with the interpreted query with a plurality of applications and selecting one or more applications associated with the one or more actions;
Displaying the one or more applications in a search results page;
A method comprising the steps of:
Item 2.
Interpreting the user search query includes generating a tokenized form of the search query;
The method according to item 1, characterized in that:
Item 3.
The user context includes data stored in a user profile;
The method according to item 1, characterized in that:
Item 4.
The user context includes the user's past search queries,
The method according to item 1, characterized in that:
Item 5.
Displaying the one or more applications in a search results page includes displaying at least one application based on the search query and at least one application based on the search query and the user context.
The method according to item 1, characterized in that:
Item 6.
Associating the selected one or more actions associated with the interpreted query with a plurality of applications is performed algorithmically;
The method according to item 1, characterized in that:
Item 7.
Further comprising ranking the one or more selected applications;
The method according to item 1, characterized in that:
Item 8.
Ranking the one or more selected applications includes selecting a top application based on the popularity of the application in a user community.
8. The method according to item 7, wherein
Item 9.
Ranking the one or more selected applications includes selecting the top application based on past revenue of the application.
8. The method according to item 7, wherein
Item 10.
Associating the selected one or more actions associated with the interpreted query with a plurality of applications, and selecting one or more applications associated with the one or more actions from the search query Further comprising pre-filling said application based on the extracted data;
The method according to item 1, characterized in that:
Item 11.
A system for providing purpose-specific applications in search engine results pages,
A plurality of client devices coupled to a search engine via a network for sending a search query to the search engine;
A query interpreter that interprets the search query and adds data representing a user context to the search query;
An action inference module that maps the interpreted query to one or more action templates;
And associating the interpreted query with one or more action templates includes selecting one or more actions associated with the interpreted query;
Associating the selected one or more actions associated with the interpreted query with a plurality of applications and selecting one or more applications associated with the one or more actions stored in an application database Application scorer,
A front-end server that displays the one or more applications in a search results page;
A system further comprising:
Item 12.
The query interpreter further generates a tokenized form of the search query;
Item 12. The system according to Item 11, wherein
Item 13.
The user context includes data stored in a user profile;
Item 12. The system according to Item 11, wherein
Item 14.
The user context includes the user's past search queries,
Item 12. The system according to Item 11, wherein
Item 15.
The front end server further displays at least one application based on the search query and at least one application based on the search query and the user context;
Item 12. The system according to Item 11, wherein
Item 16.
Associating the selected one or more actions associated with the interpreted query with a plurality of applications is performed algorithmically;
Item 12. The system according to Item 11, wherein
Item 17.
The application scorer further ranks the one or more selected applications;
Item 12. The system according to Item 11, wherein
Item 18.
Ranking the one or more selected applications includes selecting a top application based on the popularity of the application in a user community.
Item 12. The system according to Item 11, wherein
Item 19.
Ranking the one or more selected applications includes selecting the top application based on past revenue of the application.
Item 19. The system according to Item 18, wherein
Item 20.
An application parameterization module that pre-fills the application based on data extracted from the search query;
Item 12. The system according to Item 11, wherein

Claims (5)

A device for providing purpose-specific applications in web content,
A means of receiving user input;
Means for analyzing the input and selecting an application based on the analysis result;
Means for displaying in the web content the result of executing the selected application using the input subject to analysis;
A device comprising: A device for providing purpose-specific applications in web content,
A means of receiving user input;
Means for adding data representing a user context to the input, interpreting the input, and selecting an application associated with the interpreted input;
Means for displaying the selected application in the web content;
A device comprising: A method for providing purpose-specific applications in web content,
Receiving user input; and
Analyzing the input and selecting an application based on the analysis results;
Displaying the result of executing the selected application using the input subject to analysis in the web content;
A method comprising the steps of: A program for providing purpose-specific applications in web content,
On the computer,
Receiving user input; and
Analyzing the input and selecting an application based on the analysis results;
Displaying the result of executing the selected application using the input subject to analysis in the web content;
A program characterized by having executed. A system for providing purpose-specific applications in web content,
Multiple client devices that receive user input; and
Means for receiving the input; means for analyzing the input and selecting an application based on the analysis result; and executing the selected application using the input subject to analysis in the web content. A server device comprising: means for displaying;
A system comprising:

Images