Patent Application
20130086509
Search systems often provide query suggestions as a user inputs a portion of a user query into a search box. In many cases, query suggestions expand or auto-complete a query being entered thereby providing a user with additional characters and/or terms to suggest a more complete user query. Such query suggestions enhance a user's search experience by providing query suggestions that may be relevant or related to the intent of the user and, as such, result in search results relevant to the user. In some instances, however, a user might input a query, or portion thereof, for which very few, if any, query suggestions are available. For example, a user might input a very specific query that has rarely been previously searched. As such, a limited amount of query suggestions are provided to the user. As a result, search results provided to a user may not be content desired by the user resulting in a poor user experience.
This Summary is provided to introduce a selection of concepts in a simplified form that are further described below in the Detailed Description. This Summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used in isolation as an aid in determining the scope of the claimed subject matter.
According to embodiments of the invention, alternative query suggestions are provided to a user. In this regard, as a user inputs a search query, the user is provided with query suggestions other than auto-complete query suggestions. Such alternative query suggestions are based on an initial query, such as a query input by a user into a search box, that is modified by removing insignificant or frequently used terms. The modified query can be utilized to identify alternative query suggestions, for example, by matching terms within the modified query to queries within a query log.
Embodiments of the invention are described in detail below with reference to the attached drawing figures, wherein:
The subject matter of embodiments of the invention disclosed herein is described with specificity to meet statutory requirements. However, the description itself is not intended to limit the scope of this patent. Rather, the inventors have contemplated that the claimed subject matter might also be embodied in other ways, to include different steps or combinations of steps similar to the ones described in this document, in conjunction with other present or future technologies. Moreover, although the terms “step” and/or “block” may be used herein to connote different elements of methods employed, the terms should not be interpreted as implying any particular order among or between various steps herein disclosed unless and except when the order of individual steps is explicitly described.
Embodiments of the invention described herein include computer-readable media having computer-executable instructions for performing a method of facilitating presentation of alternative query suggestions. The method includes referencing an initial query having a plurality of terms. A term(s) to remove from the initial query is identified based at least on a number of times the term(s) occurs in previous queries. The initial query is modified by removing the term(s) from the initial query. The modified query is utilized to identify an alternative query suggestion(s), wherein the alternative query suggestion(s) comprises at least one term that matches at least one term of the modified query.
In a second illustrative embodiment, a graphical user interface embodied on one or more computer-readable media and executable on a computer is provided. The graphical user interface includes a first display area configured for displaying at least a portion of a search query. The graphical interface also includes a second display area adjacent to the first display area. The second display area is configured for displaying an auto-complete query suggestion and an alternative query suggestion. The alternative query suggestion is generated based on one or more terms removed from the at least the portion of the search query input into the first display area.
In a third illustrative embodiment, computer-executable instructions cause a computing device to perform a method of facilitating presentation of alternative query suggestions. The method includes referencing a search query input by a user into a search box, the search query comprising a plurality of terms. A specificity score for each of the plurality of terms of the search query is referenced. A first term and a second term to remove from the search query are identified based on a comparison of the specificity scores with a specificity predetermined threshold. The second term is identified to retain based on a comparison of a ngram score associated with the second term with a ngram predetermined threshold. The search query is modified by removing the first term from the search query and retaining the second term. The modified search query and a query log of previously submitted search queries are used to identify an alternative query suggestion that has one or more terms that match one or more terms of the modified query.
Various aspects of embodiments of the invention may be described in the general context of computer program products that include computer code or machine-useable instructions, including computer-executable instructions such as program modules, being executed by a computer or other machine, such as a personal data assistant or other handheld device. Generally, program modules including routines, programs, objects, components, data structures, etc., refer to code that perform particular tasks or implement particular abstract data types. Embodiments of the invention may be practiced in a variety of system configurations, including dedicated servers, general-purpose computers, laptops, more specialty computing devices, and the like. The invention may also be practiced in distributed computing environments where tasks are performed by remote-processing devices that are linked through a communications network.
Computer-readable media include both volatile and nonvolatile media, removable and nonremovable media, and contemplate media readable by a database, a processor, and various other networked computing devices. By way of example, and not limitation, computer-readable media include media implemented in any method or technology for storing information. Examples of stored information include computer-executable instructions, data structures, program modules, and other data representations. Media examples include, but are not limited to RAM, ROM, EEPROM, flash memory and other memory technology, CD-ROM, digital versatile discs (DVD), holographic media and other optical disc storage, magnetic cassettes, magnetic tape, magnetic disk storage, and other magnetic storage devices. These technologies can store data momentarily, temporarily, or permanently.
An exemplary operating environment in which various aspects of the present invention may be implemented is described below in order to provide a general context for various aspects of the present invention. Referring initially to
The computing device 100 includes a bus 110 that directly or indirectly couples the following devices: a memory 112, one or more processors 114, one or more presentation components 116, input/output (I/O) ports 118, input/output components 120, and an illustrative power supply 122. The bus 110 represents what may be one or more busses (such as an address bus, data bus, or combination thereof). Although the various blocks of
The memory 112 includes computer-executable instructions (not shown) stored in volatile and/or nonvolatile memory. The memory may be removable, nonremovable, or a combination thereof. Exemplary hardware devices include solid-state memory, hard drives, optical-disc drives, etc. The computing device 100 includes one or more processors 114 coupled with a system bus 110 that read data from various entities such as the memory 112 or I/O components 120. In an embodiment, the one or more processors 114 execute the computer-executable instructions to perform various tasks and methods defined by the computer-executable instructions. The presentation component(s) 116 are coupled to the system bus 110 and present data indications to a user or other device. Exemplary presentation components 116 include a display device, speaker, printing component, and the like.
The I/O ports 118 allow computing device 100 to be logically coupled to other devices including the I/O components 120, some of which may be built in. Illustrative components include a microphone, joystick, game pad, satellite dish, scanner, printer, wireless device, keyboard, pen, voice input device, touch-input device, touch-screen device, interactive display device, or a mouse. The I/O components 120 can also include communication connections that can facilitate communicatively connecting the computing device 100 to remote devices such as, for example, other computing devices, servers, routers, and the like.
Embodiments of the present invention are directed to providing alternative query suggestions. A query suggestion, as used herein, refers to a query that is suggested to a user based on the query input, or partially input, by a user. Generally, a query suggestion is presented to a user as a user inputs a search query, for example, into a search box on a webpage (e.g., a search page). In embodiments, one or more query suggestions can be displayed to a user via a drop box descending from the search box or in any other position or placement within the webpage. By way of example and with reference to
In some embodiments, a query suggestion is an auto-suggest query (also referred to as a auto-complete suggestion or auto-complete query). An auto-suggest query may be based on previously-submitted user queries identified from a query log, which correspond to the character sequence submitted as part of a user query. As such, an auto-complete query suggestion may provide additional characters and/or terms that complete at least a portion of a user query. For example, a query suggestion for the user query “aut” may be “autotrader,” adding the character “otrader” in the query suggestion. In another example, a query suggestion for the user query “auto s” may be “auto salavage,” adding the characters/terms “alavage” to the submitted query.
An alternative query suggestion refers to a query suggestion that is alternative or additional to other query suggestions provided to a user, such as auto-suggest queries. In this regard, embodiments of the present invention describe aspects of generating and providing alternative query suggestions. Such alternative query suggestions can be presented to a user in response to a query, or partial query, input to a search box. In one embodiment, one or more alternative query suggestions are solely provided to a user. That is, other query suggestions, such as auto-suggest queries, are not provided to a user. Such might be the case when auto-suggest queries do not exist for a particular query or partial query entered by a user. In other embodiments, one or more alternative query suggestions are provided to a user in combination with other query suggestions, such as auto-complete suggestions. For instance, alternative query suggestions can be presented in a separate region from the other query suggestions, juxtaposed or adjacent to other query suggestions, integrated with other query suggestions, etc.
The term “query” refers to any type of request containing one or more search terms (e.g., words, acronyms, etc.) that can be submitted to a search engine (or multiple search engines) for identifying search results based on the search term(s) contained in the query. An input query refers to any portion of a query that has been input into a search box by a user. For example, a portion of a user query may include any number of characters and/or terms, such as the query “aut” being a portion for the query for “autotrader.” A query or input query may include any number of characters and/or terms, whether complete or incomplete, such as a user query including two terms and a portion of a third term.
Turning now to
The user device 210 can be any kind of computing device capable of allowing a user to submit a search query, for example, to the query suggestion service 214 and to receive, in response to the input query, one or more query suggestions. For example, in an embodiment, the user device 210 can be a computing device such as computing device 100, as described above with reference to
In an embodiment, the user device 210 is separate and distinct from the database 212 and/or the query suggestion service 214 illustrated in
The database 212 of
With continued reference to
The query suggestion service 214 provides alternative query suggestions. In embodiments, the query suggestion service 214 includes a query referencing component 220, a query modifying component 222, a query suggesting component 224, and a presenting component 226. According to embodiments of the invention, the query suggestion service 214 can include any number of other components not illustrated. In some embodiments, one or more of the illustrated components 220, 222, 224, and 226 can be integrated into a single component or can be divided into a number of different components. Components 220, 222, 224, and 226 can be implemented on any number of machines and can be integrated, as desired, with any number of other functionalities or services.
The query referencing component 220 is generally configured to reference initial queries. An initial query can be used to generate or identify an alternative query suggestion(s). In this regard, an initial query is utilized as a basis or initial point for providing alternative query suggestions.
As can be appreciated, an initial query can be any query or query suggestion that can be used to dynamically identify an alternative query suggestion. By way of example only, an initial query can be an input query. As previously mentioned, an input query refers to any portion of a query that has been entered or input into a search box. Such an input query can be obtained, received, retrieved, or otherwise referenced by the query suggestion service 214, or a portion thereof, upon input by a user (e.g., a character input by a user via a search box). An input query (i.e., search request) can be input or entered by a user via the user device 210. Input queries received from a user can include search queries that were manually or verbally input by the user or other search queries received by the query suggestion service 214 that were somehow approved or entered by the user. As can be appreciated, an input query can be the portion entered into a search box prior to a user selecting to search for related content (i.e., the user has not yet selected the search icon).
Alternatively or additionally, an initial query may be an auto-suggest query. Such an auto-suggest query can be obtained, received, retrieved, or otherwise referenced by the query suggestion service 214, or a portion thereof, for example, from query suggestion service 214 or a service(s). In this regard, the query referencing component 220 can communicate with another service(s) to obtain auto-suggest queries associated with a query, or portion thereof, input by a user. For instance, auto-suggest queries might be retrieved or received from a service (not shown) to reference an auto-suggest query(s) to be provided as a query suggestion in response to a query, or portion thereof, input by a user.
Although initial queries are described herein as input queries and/or auto-suggest queries, other forms of queries or query suggestions or contemplated to be within the scope of embodiments of the present invention. In this regard, any form of query suggestion can be used as an initial query to derive an alternative query suggestion.
In embodiments, the query referencing component 220 can dynamically reference input queries. Accordingly, an input query can be referenced upon entering or modifying any portion of the query (e.g., an addition or deletion of a character). Similarly, an auto-suggest query can be referenced upon an auto-suggest service (not shown) identifying or determining an auto-suggest query (e.g., prior to, simultaneous with, or after presenting the auto-suggest query to the user, for example, via an query suggestion display).
The query modifying component 222 is generally configured to modify initial queries. Initial queries are modified or reformulated so that alternative query suggestions can be identified. In some cases, prior to modifying initial queries, a determination may be made as to whether to provide alternative query suggestions based on the initial query. In this way, alternative query suggestions may not be provided for each input query and/or auto-complete query. In one example, alternative query suggestions may be provided for input queries that are tail queries. A tail query, as used herein, refers to a query for which there is a limited number of auto-suggest queries. In this regard, tail queries are queries that occur very infrequently in query logs or histograms (e.g., generally long queries, detailed queries, etc.). For example, assume a user enters a detailed or specific query into a search box. Further assume that the number of auto-suggest queries that exist for that particular query is below a predetermined threshold (e.g., five, three, one, etc.). In such a case, a determination can be made that alternative query suggestions are to be provided to the user. In another example, a determination of whether to provide an alternative query suggestion(s) can be based on, for instance, a type of initial query (e.g., input query, auto-select query), a user preference, or the like.
In other cases, a determination as to whether to provide alternative query suggestions is not required. For example, each input query and/or auto-suggest query associated with an input query can be used to identify alternative query suggestions.
The query modifying component 222 can modify an initial query by removing various terms within the initial query. That is, insignificant or non-specific terms within an initial query can be removed to result in a more succinct and directed modified query. In one embodiment, specificity scores associated with terms are utilized to identify terms to remove from a query. A specificity score indicates an extent of specificity of a particular term. Specificity scores can be calculated or determined using any method, calculation, formula, lookup index, etc. In embodiments, a specificity score uses or is based at least in part on a number of times or frequency that a particular term occurs in previous queries. In some cases, a specificity score is a value (e.g., number) that indicates a frequency of a particular term within previous queries relative to other terms. In this regard, the frequency of a particular term within previous queries can be compared to frequency of other terms within previous queries or frequency of the particular term occurring with other terms. In one implementation, a specificity score S(t) for a term t within an initial query is calculated using the following calculation:
wherein f(i) is the number of queries in which term i occurs, p(i) is the frequency with which term i appears across all queries (i.e., f(i)/N), and N is the total number of queries in the corpus. g(i) is the number of queries in which both terms i and t occur. q(i) is the probability of i's occurrence conditioned on t (i.e., g(i)/N_t), wherein N_t is the number of queries containing t in the corpus. The specificity of t is then defined as the Kullback-Leibler divergence of q from p (i.e., DKL(q∥P)).
As such, previously received queries are analyzed to identify which terms frequently occur with a large number of other terms. For instance, the term “free” may occur with “free camera,” “free milkshake,” “free cd,” etc. In the above calculation, “free” is the term t for which a specificity score is calculated while i is every other term in the corpus other than t. Because “free” occurs with many other terms, the term “free” may result in a low specificity score and deemed unimportant or insignificant.
In embodiments, a specificity score is calculated for each term within an initial query. Specificity scores may be dynamically calculated or calculated offline and stored in a database, such as database 212 of
More specific, or less-common, terms can be provided with a higher or greater specificity score that indicates a stronger term or otherwise more specific term. For example, a specificity score for a term with great specificity that is determined using equation 1 discussed above generally results in a higher specificity score (e.g., value). In other cases, a more specific, or less-common, term can be provided with a lower specificity score that indicates a stronger term or otherwise more specific term. For example, a specificity score based solely on frequency of a term within previous queries generally results in terms used less frequently having a lower specificity score.
Specificity scores can be analyzed to determine which, if any, terms to remove from an initial query. In some cases, specificity scores below or above a predetermined threshold are removed or deleted from the initial query. A predetermined threshold can be any threshold that indicates a maximum or minimum value. In other cases, specificity scores can be compared to one another to determine which, if any, terms to remove from an initial query. For instance, two terms with the lowest or greatest specificity score may be removed.
To prevent removing terms that are part of a phrase that may be significant to the initial query, the query modifying component 222 can detect and retain phrases recognized as such. In embodiments, phrases to be retained can be identified using ngram scores. An ngram refers to a subsequence of n items from a given sequence (e.g., bigram (two sequential terms), trigram (three sequential terms), etc.). A ngram score can be calculated or identified in any number of ways. In some cases, a ngram score can be obtained by mining query logs, such as query logs stored in database 212, to identify frequencies or scores for various ngrams. For instance, a ngram score might be a frequency or a number of times that a particular ngram occurred in previously submitted queries. In embodiments, a ngram score can be identified for each ngram within an initial query or a particular set of ngrams within an initial query (e.g., bigrams, bigrams and trigrams, a specific ngram, etc.). Ngram scores may be dynamically calculated or calculated offline and stored in a database, such as database 212 of
Ngram scores can be analyzed to determine which, if any, phrases to retain within the initial query. In some cases, ngram scores below or above a predetermined threshold are maintained within the initial query. In other cases, ngram scores can be compared to one another to determine which, if any, phrases to retain. For instance, an ngram(s) with the lowest or greatest ngram score may be retained. In some cases, more common ngrams (phrases) are associated with a higher or greater ngram score. For example, an ngram score based on frequency of the ngram within previous queries generally results ngrams used more frequently having a greater ngram score. In other cases, more common ngrams may be associated with a lower ngram score.
By way of example only, assume that a user inputs a query “information about tom green.” The phrase “tom green” would be detected as a bigram with a high ngram score (e.g., high frequency) and thereby retained in the modified query even though the word “green” may have a low specificity score resulting in a term that would otherwise be removed from the initial query.
The query modifying component 222 can reformulate the initial query to generate a modified query in accordance with the specificity scores and ngram scores. In this regard, any terms that are not part of a phrase to be retained are removed from the initial query to result in a modified or reformulated query. In some cases, terms suggested for removal may be evaluated in light of the ngram data to determine whether such terms should be retained as part of a phrase. In other cases, phrases or ngrams to be retained can be initially recognized and, thereafter, specificity analysis can be performed to determine if any of the remaining terms are to be removed from the initial query.
By way of example only and with reference to
The query suggesting component 224 is generally configured to provide alternative query suggestions. The query suggesting component 224 can use the modified query to attain alternative query suggestions. In embodiments, a modified query is mapped to existing queries in a query log, query histogram, existing query suggestions log, or the like. In this regard, one or more of the terms within the modified query can be utilized (e.g., looked up) within a query log or log of previous query suggestions to identify if such a term(s)/phrase exists in the log. As can be appreciated, in some embodiments, any number of the remaining terms can be searched for in such query logs. For example, a modified query may be “mattress stores las vegas” but alternative query suggestions may additionally include “mattress las vegas” and “mattress vegas.” Additionally or alternatively, in some embodiments, an alternative query suggestion can include terms within a modified query that are in any order. In this regard, assume a modified query is “mattress stores las vegas.” In such a case, alternative query suggestions may include “las vegas mattress” and “vegas mattress.”
By way of example and with reference to
Returning to
By way of example only,
In some cases, an alternative suggestion indicator may be displayed in association with a corresponding alternative query suggestion to provide an indication that the particular query suggestion is an alternative query suggestion. In this regard, an alternative suggestion indicator may provide an indication that a specific query suggestion is based on modifications made to the initial query (e.g., query input by user). An example of an alternative suggestion indicator 514 is provided in
To recapitulate, embodiments of the invention include systems, machines, media, methods, techniques, processes and options for providing alternative query suggestions. Such alternative query suggestions can be displayed in proximity to a corresponding query input by a user. Turning to
Initially, at block 610, an initial query is referenced. Such an initial query might be input by a user using a user device, such as user device 210 of
Turning now to
With reference to
At block 824, it is determined if there is another term within the search query to analyze. If there is another term within the search query to analyze, the method returns to block 812 to analyze that term. If there is not another term within the search query to analyze, at block 826, a modified query is identified or generated. Such a modified query retains terms identified at block 816 and removes terms identified at block 822. At block 828, the terms within the modified query are compared to previous queries to identify one or more alternative query suggestions. In this regard, the modified search query and a query log of previously submitted search queries are utilized to identify an alternative query suggestion that has one or more terms that match one or more terms of the modified query. Although
Various embodiments of the invention have been described to be illustrative rather than restrictive. Alternative embodiments will become apparent from time to time without departing from the scope of embodiments of the inventions. It will be understood that certain features and sub-combinations are of utility and may be employed without reference to other features and sub-combinations. This is contemplated by and is within the scope of the claims.
