WO2021185326A1 - 用于搜索的用户交互 - Google Patents

用于搜索的用户交互 Download PDF

Info

Publication number
WO2021185326A1
WO2021185326A1 PCT/CN2021/081612 CN2021081612W WO2021185326A1 WO 2021185326 A1 WO2021185326 A1 WO 2021185326A1 CN 2021081612 W CN2021081612 W CN 2021081612W WO 2021185326 A1 WO2021185326 A1 WO 2021185326A1
Authority
WO
WIPO (PCT)
Prior art keywords
search
search result
clusters
display
navigation
Prior art date
Application number
PCT/CN2021/081612
Other languages
English (en)
French (fr)
Inventor
刘铭
魏波
仙云森
王仲远
王可
许鑫
王珺
陈恩地
汪祖海
吕梅
Original Assignee
北京三快在线科技有限公司
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 北京三快在线科技有限公司 filed Critical 北京三快在线科技有限公司
Publication of WO2021185326A1 publication Critical patent/WO2021185326A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/954—Navigation, e.g. using categorised browsing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/951—Indexing; Web crawling techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9538—Presentation of query results

Definitions

  • This application relates to the field of search engines, in particular to user interaction methods, devices, mobile terminals, and storage media for searching.
  • This application provides a user interaction method, device, mobile terminal, and storage medium for searching.
  • a user interaction method for search including: in response to a search request, determining one or more search result clusters and navigation items of each of the search result clusters; The display priority of the search result clusters, displaying the navigation items of each of the search result clusters and at least part of the search result clusters; in response to an interactive request, the display focus of at least part of the search result clusters and The display focus of the navigation items in each of the search result clusters is adjusted in linkage.
  • displaying the navigation items of each of the search result clusters and at least part of the search result clusters includes: displaying the navigation items of each of the search result clusters in the navigation area of the search result navigation page;
  • the content area of the search result navigation page displays at least part of the search preview results of the search result cluster.
  • the linkage adjustment of the display focus of at least part of the search result clusters and the display focus of the navigation items of each of the search result clusters includes: responding to the clustering of the target search results
  • the selection request of the navigation item makes the navigation item of the target search result cluster become the display focus in the navigation area, and causes the target search result cluster to become the display focus in the content area.
  • the linkage adjustment of the display focus of at least part of the search result clusters and the display focus of the navigation items of each of the search result clusters includes: responding to browsing of the content area It is requested that the navigation items of the target search result cluster become the display focus in the navigation area, and the target search result clusters become the display focus in the content area.
  • the method further includes: in response to a selection request for the target search result cluster, displaying the search result set of the target search result cluster in the form of a waterfall in the search result details page.
  • displaying the clustered search result set of the target search result in the form of a waterfall flow on the search result details page includes: displaying filter items in the search result details page; The selection request is used to filter the search result set of the target search result cluster according to the target filter item, and the filtered search result set is displayed in the form of a waterfall flow on the search result details page.
  • each of the one or more search result clusters corresponds to one search intent
  • each search intent corresponds to one or more business categories.
  • the number of preview results in the search result cluster is determined according to the display priority of the search result cluster and the maximum number of displays corresponding to the search result cluster;
  • the maximum number of impressions is determined based on the number of search results viewed by at least some users who placed the order before placing the order in the search log.
  • determining one or more of the search result clusters and navigation items of each of the search result clusters includes: determining one or more search intents corresponding to the search request, and generating the one or more search intents.
  • generating the search result cluster corresponding to the search intention includes: determining the search result set of the search intention; determining the result weight of the search result set according to the search relevance; determining the result weight of the search result set according to the result weight of the search result set The intent weight of the search intent; the display priority of the search result cluster corresponding to the search intent is determined according to the intent weight of the search intent.
  • determining the search result set of the search intention includes: obtaining a search result set corresponding to the search intention from one or more indexing services through distributed search.
  • the search relevance includes at least one of the following: text relevance of the search request and search result description information, relevance of user preferences and search results, and relevance of search scene information and search results.
  • determining the intention weight of the search intention according to the result weight of the search result set includes: determining the intention weight of the search intention according to the result weight of the search result set and the initial weight of the search intention.
  • a user interaction device for searching including: a clustering unit for determining one or more search result clusters and each of the search result clusters in response to a search request
  • a display unit for displaying the navigation items of each of the search result clusters and at least part of the search result clusters according to the display priority of each of the search result clusters; and for responding to an interactive request , Performing linkage adjustment on the display focus of at least part of the search result clusters and the display focus of the navigation items of each search result cluster.
  • the display unit is configured to: display the navigation items of each of the search result clusters in the navigation area of the search result navigation page; display at least part of the search result clusters in the content area of the search result navigation page At least part of the search preview results.
  • the display unit is further configured to: in response to a selection request for a target search result clustered navigation item and/or a browse request for the content area, make the target search result clustered navigation item become The display focus in the navigation area, and clustering the target search result into the display focus in the content area.
  • the display unit is further configured to: in response to a selection request for the target search result cluster, display the search result set of the target search result cluster in the form of a waterfall in the search result details page.
  • the display unit may be used to: display filter items in the search result details page; in response to a selection request for target filter items, filter the target search result clustered search result set according to the target filter items, The filtered search result set is displayed in the form of a waterfall on the search result detail page.
  • each of the one or more search result clusters corresponds to one search intent
  • each search intent corresponds to one or more business categories.
  • the number of preview results in the search result cluster is determined according to the display priority of the search result cluster and the maximum number of displays corresponding to the search result cluster; the maximum number of displays is based on the search
  • the number of search results viewed by at least some users who placed the order before placing the order is determined.
  • the clumping unit is configured to: determine one or more search intents corresponding to the search request, generate search result clumps corresponding to each of the one or more search intents, and generate each of the search The navigation item of the result cluster; wherein each of the search result clusters includes a preview of the search result and a display priority.
  • the blocking unit is configured to, for each search intent: determine the search result set of the search intent; determine the result weight of the search result set according to the search relevance; determine the result weight of the search result set The intent weight of the search intent; the display priority of the search result cluster corresponding to the search intent is determined according to the intent weight of the search intent.
  • the agglomeration unit may obtain a search result set corresponding to the search intention from one or more indexing services through distributed search.
  • the search relevance may include at least one of the following: text relevance of the search request and search result description information, relevance of user preferences and search results, and relevance of search scene information and search results.
  • the blocking unit is specifically configured to determine the intention weight of the search intention according to the result weight of the search result set of the search intention and the initial weight of the search intention.
  • a mobile terminal including a processor and a memory arranged to store computer-executable instructions, which when executed, cause the processor to execute any of the above The described user interaction method for searching.
  • a computer-readable storage medium stores one or more programs, and the one or more programs, when executed by a processor, implement any of the foregoing One of the user interaction methods used in search.
  • the technical solution of the present application can determine one or more search result clusters and the navigation items of each search result cluster according to the user's search request, and according to the display priority of each search result cluster Display the navigation items of each of the search result clusters and at least part of the search result clusters, including the display focus of at least part of the search result clusters and the display focus of the navigation items of each search result cluster in response to an interactive request Make linkage adjustments.
  • This technical solution displays search results in clusters of search results (including clustered navigation items and at least part of the content), which increases the amount of information contained in the search result page; and clusters the search results through interaction with the user.
  • the display focus is adjusted in conjunction with the display focus of the navigation item clustered in search results, which improves the user experience.
  • the search result clusters can correspond to the user's search intent according to the needs, thereby realizing a new multi-intent interactive organization form of the search result display.
  • Fig. 1 shows a schematic flowchart of a user interaction method for searching according to an embodiment of the present application
  • Fig. 2 shows a schematic diagram of a search result navigation page according to an embodiment of the present application
  • Fig. 3 shows a schematic diagram of a change state of a search result navigation page according to an embodiment of the present application
  • FIG. 4 shows a schematic diagram of interactive display of search results according to an embodiment of the present application
  • FIG. 5 shows a schematic flowchart of a method for generating user interaction data for searching according to an embodiment of the present application
  • Fig. 6 shows a schematic structural diagram of a user interaction device for searching according to an embodiment of the present application
  • FIG. 7 shows a schematic structural diagram of an apparatus for generating user interaction data for searching according to an embodiment of the present application
  • Fig. 8 shows a schematic structural diagram of a mobile terminal according to an embodiment of the present application.
  • Fig. 9 shows a schematic structural diagram of a computer-readable storage medium according to an embodiment of the present application.
  • Fig. 1 shows a schematic flowchart of a user interaction method for searching according to an embodiment of the present application.
  • the user interaction method for searching includes the following steps S110, S120, and S130.
  • Step S110 in response to the search request, determine one or more search result clusters and navigation items of each of the search result clusters.
  • the embodiments of this application can be applied to various scenarios using search engine technology, including but not limited to general search engines such as Baidu and Google (the business names here are only illustrative), and special search engines in the fields of patents, trademarks, etc. , And the search engine in the app.
  • the search request may come from a user, including the user performing a specific operation on a terminal device, such as a mobile terminal.
  • the user can generate a search request in various ways such as text, image, voice, etc.
  • the text can be a search keyword or a search sentence.
  • Search result clumping is an aggregation of search results.
  • the search results can be clustered according to categories to obtain search result clusters.
  • the search result cluster covers search results belonging to that category.
  • the set of search results covered by the search result cluster is referred to as the search result set of the search result cluster.
  • clustering methods For example, clustering based on semantic relevance, clustering based on geographic location corresponding to search results, and so on.
  • One solution is to cluster according to the user's search intent.
  • Each of the one or more search result clusters corresponds to a search intent, and the search intent corresponds to one or more business categories.
  • each search intent can correspond to a local life service category, and the local life service category can include takeaway, flash sale, recipes, and so on.
  • the local life service category can include takeaway, flash sale, recipes, and so on.
  • the search intention is determined based on the search request and the search scene information related to the search request.
  • a user's search request can obtain the distribution of multiple search intents, and the display priority of the search result cluster is determined according to the intensity distribution of these search intents.
  • the search scene information can include weather, location, user behavior, and so on.
  • the search scene information and the search request can be pre-compounded modeling, and the search intention recognition model can be trained. Therefore, in actual application scenarios, the search intent can be determined by inputting the search request and the search scenario information related to the search request into the pre-trained search intent recognition model.
  • the robustness of search intent recognition can be ensured by the way in which the business party recommends the search intent.
  • Step S120 according to the display priority of each of the search result clusters, display the navigation items of each of the search result clusters and at least part of the search result clusters.
  • the display priority may only be expressed in the order of first and second, or may be specifically expressed as a display weight. For example, assuming that the display weights of the five search result clusters A, B, C, D, and E are 0.33, 0.1, 0.2, 0.3, and 0.07, respectively, then the display order of the navigation items of the search result clusters and the search result clusters The order of display is A, D, C, B, E.
  • each search result block is an aggregation of many search results, it cannot display all the search results in the screen area, but can display preview search results.
  • the search result block has a high weight among the many search results aggregated Search results for. Specifically, the weight of search results can be determined through semantic relevance or business recommendation strategies.
  • Step S130 In response to the interaction request, the display focus of the one or more search result clusters and the display focus of the navigation items of each search result cluster are coordinated and adjusted.
  • the interaction request may come from the user, including the user performing a specific operation on a terminal device, such as a mobile terminal.
  • the navigation items of the search result cluster can be used to quickly locate the search result cluster, including the index or quick entry of the search result cluster.
  • the user actively browses the search result clumps he can also familiarize himself with the corresponding relationship between the navigation items and the search result clumps by changing the display focus of the navigation items. And, the whole process can be completed in one screen page.
  • the display focus adjustment here refers to displaying the target search results or their navigation items in a more eye-catching manner, including but not limited to highlighting, pop-up display, or moving to a prominent position such as the first position of the display area for display, etc. .
  • the user interaction method for search shown in Figure 1 shows the preview search results in clusters of search results, which increases the amount of information contained in the search result page; and through interaction with the user, the focus of the display of the search results is clustered Linked adjustments with the display focus of the navigation items in the search results block, which improves the user experience.
  • search result clusters can correspond to search intents according to needs, thereby realizing a new multi-intent interactive organization form.
  • Displaying the navigation items of each search result cluster and at least part of the search result clusters may include: sequentially displaying the navigation items of each search result cluster in the navigation area of the search result navigation page; displaying at least part of the navigation items in the content area of the search result navigation page At least part of the search preview results of the search result cluster.
  • the navigation area is on the top and the content area is on the bottom; the navigation area is on the side and can be hidden as required, and the content area is permanently displayed in the center.
  • the layout of the navigation area and the content area can also be different.
  • a dual-screen device dual screens are not necessarily two independent display hardware, for example, two screen areas formed by a folding screen according to a folding mechanism) can display a navigation area on one screen and a content area on the other screen.
  • Fig. 2 shows a schematic diagram of a search result navigation page according to an embodiment of the present application.
  • a search bar for the user to input voice, search keywords, images, etc. to generate a search request based on the user's input.
  • the six controls under the search bar "Order Takeaway”, “Eat at a Store”, “Discount Group Purchase”, “Buy Ingredients”, “Recipes” and “Reviews” respectively correspond to a navigation item in the search result cluster.
  • the area where the navigation items of these search results are clumped is the navigation area.
  • the content area which shows the two preview search results of the "Takeaway" search result cluster.
  • the linkage adjustment of the display focus of the search result clumps and the display focus of the navigation items of the search result clumps may include: responding to a request for selecting the navigation items of the target search result clumps to make the target search result
  • the clustered navigation items become the display focus in the navigation area
  • the target search result clusters become the display focus in the content area.
  • the display position above the content area is the display focus position.
  • the first preview search result contained in the search result block will be moved to the display focus position for display, which makes the search result block become content The display focus of the area.
  • the linkage adjustment of the display focus of the search result clusters and the display focus of the navigation items of the search result clusters may include: in response to a request for browsing the content area, making the navigation items of the target search result cluster become The display focus in the navigation area, and clustering the target search results into the display focus in the content area.
  • the user can also browse through multiple search result clusters through operations such as sliding.
  • the display focus of the navigation items in the navigation area will also change accordingly. For example, if the user finishes browsing the "buy ingredients” cluster and starts to browse the "recipe", the display focus in the navigation area will change from “buy ingredients” to "recipe", specifically by moving or highlighting the navigation items. Way to achieve.
  • Figure 3 shows a schematic diagram of the changing state of the search result navigation page. Specifically, FIG. 3 shows the two search result clusters "to eat at the store” and "preferential group purchase”. Among them, limited by the screen size, the "Preferential Group Buying" search result block only shows part of the preview search results.
  • the blank areas are themed content, such as themed wallpapers, poster pictures, video covers, business avatars, control icons, etc.
  • the crosses represent text and/or numbers.
  • the user interaction method for searching may further include: in response to a selection request for a target search result cluster, displaying the search result of the target search result cluster in the form of a waterfall in the search result details page set.
  • the left side is the search result navigation page
  • the right side is the search result details page.
  • the search result navigation page jumps to the search result details page, and each search in the search result set of the search result block is displayed in turn on the search result details page result.
  • the waterfall flow is a multi-column layout. As the page scroll bar scrolls down, data blocks are continuously loaded and appended to the current tail, which is suitable for display scenarios of multiple search results.
  • the user interaction method for search may further include: displaying filter items in the search result details page;
  • the filtered search result set is displayed in the form of waterfall on the search result details page.
  • Fig. 4 also shows filter items, and the user can filter the search result set by checking the filter items such as "only look at brand merchants" and "only look at business". The user can also return to the search result navigation page by swiping to the left from the search result details page (as shown by the "back" arrow in Figure 4).
  • the number of preview results in the search result cluster may be determined according to the display priority of the search result cluster and the maximum number of display results of the search result cluster.
  • the display priority can be expressed in the form of display weight.
  • the number of preview results of each search result cluster may be calculated in the manner of display weight ⁇ maximum display number. If the calculation result is not an integer, then round down or up. For example, assuming that the display weight of the search result cluster A is 0.33, and the maximum display number is 10, the number of preview results of the search result cluster A is 4 according to rounding up.
  • the maximum number of impressions may be determined based on the number of search results viewed by at least some users who placed the order before placing the order in the search log related to the search result cluster.
  • the maximum number of impressions reflects the tendency of all or at least most users to select specific application scenarios or search intentions. For example, in shopping scenarios, placing an order is a factor that can strongly reflect a positive tendency. If most users will browse 4-5 products and choose one to place an order, then given 4-5 products as the preview search result, the user is likely to make the order after viewing these 4-5 products There is no need to jump to the search result details page, which further saves the time cost of searching.
  • statistics can be made on the number of search results viewed by 95% of the ordering users before placing the order, and the maximum number of impressions can be determined by finding the median.
  • This method has high versatility and is suitable for cold start scenarios.
  • the proportion of users placing orders and the statistical method can be modified according to requirements, or the maximum display quantity can be directly specified manually, which is not limited in this application.
  • Determining search result clusters and navigation items of search result clusters may include: determining one or more search intents corresponding to the search request, generating search result clusters corresponding to the search intents, and generating each of the search results Clustered navigation items.
  • the search result clusters each include a preview search result and a display priority, and the one or more search result clusters each correspond to a search intent.
  • search intent usually refers to the real needs of users behind the search behavior. For example, searching for "badminton” may be because the user wants to buy badminton equipment, it may be looking for badminton courts, or it may be learning the rules of badminton, and so on. "Buying equipment”, “finding venues” and “learning rules” are three different search intents related to the search keyword "badminton".
  • the search intent may include takeaway, dine-in, recipes, reviews, offers, and so on. These search intents can reflect user needs, and the specific names and categories of search intents can be determined by business parties or domain experts. In other words, it can be understood as a generalized user demand.
  • the search intent can correspond to the category of goods or services.
  • the categories of goods and services can be defined according to business needs.
  • the take-out and dine-in services given above are the classification of service provision methods.
  • One search result can correspond to one or more search intents. For example, if a restaurant provides both dine-in sales and take-out services, the corresponding search intent of the restaurant may include take-out and dine-in; while another restaurant only provides take-out services, the corresponding search intent of the restaurant only includes take-out. Conversely, a search intent can also correspond to one or more search results, and generally multiple search results, for example, there are many restaurants that provide take-out services. The more the search intent matches the user's real needs, the correspondingly, the easier it is for the search results displayed to the user to achieve the user's search purpose.
  • clustering search results according to search intentions and generating search result clusters can make users more quickly pay attention to the search results they need.
  • Generating the search result cluster corresponding to the search intent may include: determining the search result set of the search intent; determining the result weight of each search result in the search result set according to the search relevance; according to the result weight of each search result in the search result set Determine the intent weight of the search intent; determine the display priority of the corresponding search result cluster according to the intent weight of the search intent.
  • determining the search result set of each search intention includes: obtaining a search result set corresponding to the search intention from one or more indexing services through distributed search.
  • the search result set is obtained based on the search intent, and the result weight of the search result set is further determined based on the search relevance, and the intent weight is determined by the result weight, and then the display priority of the corresponding search result cluster is determined by the intent weight.
  • the solution enhances the robustness of search results that fits the user’s intentions.
  • one search intent corresponds to a search result set including one or more search results
  • one search result cluster corresponds to one search intent.
  • the result weight of the search result set can be considered as the result weight of the search result in the search result set;
  • the result weight of the multiple search results can be calculated by the mean operation or the weighted mean operation to obtain the search result set Outcome weight.
  • Search relevance includes at least one of the following: text relevance of search request and search result description information, relevance of user preferences and search results, and relevance of search scene information and search results.
  • the search request can be regarded as a direct expression of the user's search intent.
  • User preferences can be obtained through user pre-input information or user portraits.
  • the search scene information can be regarded as an indirect expression of the user's search intent, and it can supplement the potential search intent not reflected by the search keywords.
  • the search scene information can cover multiple scene dimensions, such as time dimension, location dimension, weather dimension, and so on.
  • a user searches for "Kung Pao Chicken", it may be because they want to learn the practice of Kung Pao Chicken, or they want to order Kung Pao Chicken for takeout, or they may want to go to a restaurant that sells Kung Pao Chicken.
  • searching users do not necessarily use search keywords to clearly express their search intent at one time. This requires users to further search in the search results or perform a second search, which reduces the user experience.
  • this problem can be appropriately improved. For example, if the user is searching for Kung Pao Chicken in a shopping mall, it is more likely that they want to go to a restaurant that sells Kung Pao Chicken instead of looking for recipes or ordering takeaway.
  • the role of the environment is reflected. And if the search log shows that the user has skipped multiple physical restaurants that sell Kung Pao chicken, clicked into the page of multiple take-out restaurants, and placed an order at a take-out restaurant, it can be determined that the user wants to order take-out rather than Other intentions. This reflects the role of user behavior.
  • the search relevance includes multiple categories
  • different relevance weights can be assigned to each category.
  • the relevance weight can be obtained through experience, for example, the relevance weight of the search keyword can be assigned higher.
  • Relevance weights can also be obtained according to statistical modeling, for example, by searching logs, the exposure of the search result clusters with clicks generates positive examples, and the exposures without clicks generates negative examples, and the relevance weights are determined through supervised learning.
  • Determining the intention weight of the search intention according to the result weight of each search result in the search result set may include: determining the intention weight of the search intention according to the result weight of each search result in the search result set and the initial weight of the search intention.
  • One search request may correspond to multiple search intents, and these search intents have different initial weights.
  • the initial weight of the search intent for example, can be determined manually, or determined by statistical modeling, and so on.
  • the result weight of the search result set corresponding to the search intent and the initial weight of the search intent are linearly weighted and summed to obtain the intent weight of the search intent.
  • the sum of weights of the same type should be 1, which can be implemented by using Softmax normalization calculation.
  • the user interaction method used for search can be applied to the mobile terminal side, and specifically can be applied to the application program APP. However, this does not mean that the entire flow of the user interaction method for searching needs to be completely implemented on the mobile terminal side. In fact, part of the processing that has a large amount of calculation or does not depend on private information can be implemented by the server side.
  • the determination of the search intent can be implemented by the server side.
  • the server side may be responsible for giving the distribution of search intention strength and returning it to the mobile terminal side, and then the mobile terminal side will perform search result clustering and navigation item generation, and so on.
  • this application also provides an example in which some steps are executed on the server side, which is specifically used as a user interaction data generation method used in search, which is introduced through the following embodiments.
  • Fig. 5 shows a schematic flowchart of a method for generating user interaction data for searching according to an embodiment of the present application.
  • the user interaction data generation method for searching includes S510, S520, and S530.
  • Step S510 Obtain the search request sent by the mobile terminal.
  • search engine technology including but not limited to general search engines such as Baidu and Google (the business names here are only illustrative), and special search engines in the fields of patents, trademarks, etc. , And the search engine in the app.
  • the user can generate a search request in various ways such as text, image, voice, etc.
  • the text can be a search keyword or a search sentence.
  • Step S520 Determine one or more search intents corresponding to the search request, generate search result clusters corresponding to the one or more search intents, and generate navigation items for each of the search result clusters.
  • the search result clusters each include preview search results and display priority.
  • Step S530 Return the navigation items of each of the search result clusters and at least part of the search result clusters to the mobile terminal.
  • step S530 the mobile terminal is enabled to display the navigation items in each of the search result clusters and at least part of the search result clusters according to the display priority, and the mobile terminal is enabled to display according to the interaction request from the user
  • the display focus of at least part of the search result clusters and the display focus of the navigation items of each search result cluster are coordinated and adjusted.
  • the user interaction data generation method for search shown in FIG. 5 can be implemented in conjunction with the aforementioned user interaction method for search, and provides data support for the display of search results based on user interaction.
  • the various technical details introduced in the aforementioned user interaction method for search can be appropriately applied to the user interaction data generation method for search, and will not be repeated here.
  • the search intent is determined by matching the search keywords with the text of the rules; the search intent prediction is based on text classification or clustering; the keywords are mapped to high-dimensional semantic vectors through topic models and other methods Space to express search intent, etc.
  • these schemes all have the problem of focusing only on the text and lacking attention to other factors.
  • the search scene information associated with the search request can be obtained, the composite feature used to identify the search intent can be generated according to the search scene information and the search request, and the composite feature can be input into the search intent recognition model to obtain the output of the search intent recognition model.
  • the search intent recognition results In this way, not only the search request is concerned, but also the search scene information such as weather, location, and user behavior.
  • the search intent recognition model based on composite feature modeling is used to predict the real needs of users with reference to many factors, which improves the problem of low accuracy in identifying search intents only based on search requests. It is especially suitable for life services, LBS ( Location-Based Services, location-based services) search scenarios.
  • LBS Location-Based Services, location-based services
  • the displayed search results only correspond to a single search intent and are displayed using a fixed template.
  • This type of method needs to accurately identify the user’s first search intention and organize it with the corresponding template.
  • the drawbacks are obvious: first, the search intention is completely predicted and accurate, and the elimination of ambiguity is almost impossible; the second is the fixed template display even in certain situations. The situation meets the current main user needs and also kills potential growth space.
  • search results corresponding to multiple search intents can be mixed in the search result page.
  • This type of method can perform mixed sorting on search result sets of multiple different businesses and satisfying different needs. Since it is difficult for users to predict the next page after clicking on the list, it will cause a confused understanding of the search result set and affect the user experience.
  • a supplementary intention card in the idea map for display.
  • This type of method generally needs to identify the user's first search intention as the idea map, and use the other identified search intentions as the auxiliary intention.
  • the main intent is directly expanded into the result list, while the auxiliary intent only displays the summary card and requires the user to click to expand the result list. Since the display of auxiliary intentions may be incomplete due to the ordering is too low, the user cannot perceive it. Moreover, the insertion of multiple cards will interfere with the user's browsing of the result list, which is not conducive to the user experience.
  • this application combines the search result display solution of search intent, which can predict multiple search intents of the current search request for each search request of the user, and obtain the search result set corresponding to the multiple search intents through distributed retrieval recall. And generate corresponding search result clusters, and then automatically generate a navigation area through the final judgment of the quality of the search result set to display the navigation items and content areas of the search result clusters to display at least part of the search result clusters, and can be linked according to interactive requests Adjust the display corners of the navigation area and content area to form an interactive display form of search results.
  • This kind of cause-and-effect solution has greatly improved the problem of insufficient display of multi-intent search results caused by screen size limitations on mobile terminals, and can effectively solve the problem of single search intent dominating the screen and potential search intents unable to be displayed. .
  • Fig. 6 shows a schematic structural diagram of a user interaction device for searching according to an embodiment of the present application.
  • the user interaction device 600 used for searching includes a block unit 610 and a display unit 620.
  • the clumping unit 610 is configured to obtain one or more search result clumps and navigation items of each of the search result clumps in response to a search request.
  • the display unit 620 is configured to display the navigation items and at least part of the search result clusters according to the display priority of each of the search result clusters; and to respond to at least part of the search result clusters in response to an interactive request.
  • the display focus of the search result cluster and the display focus of the navigation item of each search result cluster are adjusted in linkage.
  • the user interaction device for search shown in FIG. 6 displays search results in clusters of search results (including clustered navigation items and at least part of the content), which increases the amount of information contained in the search result page ; And through the interaction with the user, the display focus of the search result clumps and the display focus of the navigation items of the search result clumps are linked and adjusted to improve the user experience.
  • the search result clusters can correspond to the user's search intent according to the needs, thereby realizing a new multi-intent interactive organization form of the search result display.
  • the display unit 620 may be used to display the navigation items of each search result cluster in the navigation area of the search result navigation page; display at least part of the search preview results of at least part of the search result cluster in the content area of the search result navigation page.
  • the display unit 620 may be used to: in response to a request for selection of navigation items in the target search result block, make the navigation item in the target search result block become the focus of display in the navigation area, and make the target search result block become the focus in the content area. Display focus; or, in response to a browsing request for the content area, make the navigation items of the target search result block become the display focus in the navigation area, and make the target search result block the display focus in the content area.
  • the display unit 620 can also be configured to display the search result set of the target search result cluster in the form of a waterfall in the search result detail page in response to the selection request for the target search result cluster.
  • the display unit 620 can be used to display filter items in the search result details page, and in response to the selection request for the target filter items, filter the search result set of the target search results clustered according to the target filter items, and display the filter items in the search result details page Display the filtered search result set in the form of waterfall.
  • the clustering unit 610 may be used to determine one or more search intents corresponding to the search request, generate search result clusters corresponding to the one or more search intents, and generate navigation items for each of the search result clusters.
  • each search result cluster includes preview search results and display priority.
  • the blocking unit 610 can also be used for each search intent: determine the search result set of the search intent; determine the result weight of the search result set according to the search relevance; determine the intent of the search intent according to the result weight of the search result set Weight: Determine the display priority of the corresponding search result cluster according to the intent weight of the search intent.
  • the agglomeration unit 610 may obtain a search result set corresponding to the search intention from one or more indexing services through distributed search.
  • search relevance includes at least one of the following: text relevance of search request and search result description information, relevance of user preferences and search results, and relevance of search scene information and search results.
  • the blocking unit 610 may also be used to determine the intention weight of the search intention according to the result weight of the search result set of the search intention and the initial weight of the search intention.
  • the aforementioned user interaction device for searching can be used to execute the aforementioned user interaction method for searching, and the various technical details introduced for the aforementioned method are applicable to the corresponding device, and therefore will not be repeated.
  • Fig. 7 shows a schematic structural diagram of an apparatus for generating user interaction data for searching according to an embodiment of the present application.
  • the user interaction data generating device 700 used for searching includes a communication unit 710 and a generating unit 720.
  • the communication unit 710 is configured to obtain a search request sent by the mobile terminal.
  • the generating unit 720 is configured to determine one or more search intents corresponding to the search request, generate search result clusters corresponding to each of the one or more search intents, and generate navigation items for each of the search result clusters.
  • the search result clusters each include preview search results and display priority.
  • the communication unit 710 may also be configured to return the navigation items of each of the search result clusters and at least part of the search result clusters to the mobile terminal, so that the mobile terminal can cluster and aggregate the search results according to the display priority.
  • the navigation items of the search result clusters are displayed, and the mobile terminal can coordinate the display focus of the search result clusters and the navigation items of each search result cluster to adjust the display focus of the search result clusters according to the interaction request.
  • the user interaction data generating device for searching shown in FIG. 7 can be implemented in cooperation with the aforementioned user interaction device for searching, providing data support for the display of search results based on user interaction.
  • the generating unit 720 can be used for each search intent: determine the search result set of the search intent; determine the result weight of the search result set according to the search relevance; determine the intent weight of the search intent according to the result weight; according to the intent weight Determine the display priority of the search result cluster corresponding to the search intent.
  • the generating unit 720 may obtain the search result set corresponding to the search intention from one or more indexing services through distributed search.
  • Search relevance includes at least one of the following: text relevance of search request and search result description information, relevance of user preferences and search results, and relevance of search scene information and search results.
  • the generating unit 720 may be used to determine the intention weight of the search intention according to the result weight of the search result set of the search intention and the initial weight of the search intention.
  • the foregoing user interaction data generating device for searching can be used to execute the foregoing user interaction data generating method for searching, and the various technical details introduced for the foregoing method are all applicable to the device, and therefore will not be repeated.
  • the technical solution of the present application can determine one or more search result clusters and the navigation items of each of the search result clusters according to the user's search request, and display priority according to the search result clusters
  • the navigation items of each of the search result clusters and at least part of the search result clusters are displayed in stages, including the display focus of at least part of the search result clusters and the display of the navigation items of each of the search result clusters in response to an interactive request
  • the focus is adjusted in linkage.
  • This technical solution displays search results in clusters of search results (including clustered navigation items and at least part of the content), which increases the amount of information contained in the search result page; and clusters the search results through interaction with the user.
  • the display focus is adjusted in conjunction with the display focus of the navigation item clustered in search results, which improves the user experience.
  • the search result clusters can correspond to the user's search intent according to the needs, thereby realizing a new multi-intent interactive organization form of the search result display.
  • modules or units or components in the embodiments can be combined into one module or unit or component, and in addition, they can be divided into multiple sub-modules or sub-units or sub-components. Except that at least some of such features and/or processes or units are mutually exclusive, any combination can be used to compare all the features disclosed in this specification (including the accompanying claims, abstract and drawings) and any method or methods disclosed in this manner or All the processes or units of the equipment are combined. Unless expressly stated otherwise, each feature disclosed in this specification (including the accompanying claims, abstract and drawings) may be replaced by an alternative feature providing the same, equivalent or similar purpose.
  • the various component embodiments of the present application may be implemented by hardware, or by software modules running on one or more processors, or by a combination of them.
  • a microprocessor or a digital signal processor (DSP) can be used in practice to implement the user interaction device for searching and the user interaction data generation for searching according to the embodiments of the present application.
  • DSP digital signal processor
  • This application can also be implemented as a device or device program (for example, a computer program and a computer program product) for executing part or all of the methods described herein.
  • Such a program for implementing the present application may be stored on a computer-readable medium, or may have the form of one or more signals.
  • Such a signal can be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.
  • FIG. 8 shows a schematic structural diagram of a mobile terminal according to an embodiment of the present application.
  • the mobile terminal 800 includes a processor 810 and a memory 820 arranged to store computer-executable instructions (computer-readable program code).
  • the memory 820 may be an electronic memory such as flash memory, EEPROM (Electrically Erasable Programmable Read Only Memory), EPROM, hard disk, or ROM.
  • the memory 820 has a storage space 830 for storing computer-readable program codes 831 for executing any method steps in the above methods.
  • the storage space 830 for storing computer-readable program codes may include various computer-readable program codes 831 respectively used to implement various steps in the above method.
  • the computer-readable program code 831 can be read from or written into one or more computer program products.
  • FIG. 9 shows a schematic structural diagram of a computer-readable storage medium according to an embodiment of the present application.
  • the computer-readable storage medium 900 stores computer-readable program codes 831 for executing the method steps according to the present application, which can be read by the processor 810 of the mobile terminal 800.
  • the computer-readable program codes 831 are run by the mobile terminal 800 , Causing the mobile terminal 800 to execute each step in the method described above.
  • the computer readable program code 831 stored in the computer readable storage medium can execute the method shown in any of the above embodiments.
  • the computer-readable program code 831 can be compressed in an appropriate form.

Landscapes

  • Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • User Interface Of Digital Computer (AREA)

Abstract

本申请公开了用于搜索的用户交互方法、装置、移动终端和存储介质。根据用于搜索的用户交互方法的一个示例，响应于来自用户的搜索请求，确定一个或多个搜索结果聚块和各所述搜索结果聚块的导航项，并根据各所述搜索结果聚块的展示优先级展示各所述搜索结果聚块的导航项以及至少部分所述搜索结果聚块。然后，可响应于来自用户的交互请求，对所述搜索结果聚块的展示焦点与各所述搜索结果聚块的导航项的展示焦点进行联动调整。

Description

用于搜索的用户交互 技术领域
本申请涉及搜索引擎领域，具体涉及用于搜索的用户交互方法、装置、移动终端和存储介质。
背景技术
如何在搜索场景下提供更好的用户交互方式一直是搜索领域，特别是移动搜索领域所关注的问题。尤其是移动终端上的可展示空间相对较小，就更需要在搜索结果页面中展示符合用户需求的搜索结果，并根据用户的操作进行展示效果的调整。
发明内容
本申请提供一种用于搜索的用户交互方法、装置、移动终端和存储介质。
依据本申请的第一方面，提供了一种用于搜索的用户交互方法，包括：响应于搜索请求，确定一个或多个搜索结果聚块和各所述搜索结果聚块的导航项；根据各所述搜索结果聚块的展示优先级，展示各所述搜索结果聚块的导航项以及至少部分所述搜索结果聚块；响应于交互请求，对至少部分所述搜索结果聚块的展示焦点与各所述搜索结果聚块的导航项的展示焦点进行联动调整。
可选地，展示各所述搜索结果聚块的导航项以及至少部分所述搜索结果聚块，包括：在搜索结果导航页的导航区域，展示各所述搜索结果聚块的导航项；在所述搜索结果导航页的内容区域，展示至少部分所述搜索结果聚块的至少部分搜索预览结果。
可选地，响应于交互请求，对至少部分所述搜索结果聚块的展示焦点与各所述搜索结果聚块的导航项的展示焦点进行联动调整，包括：响应于对目标搜索结果聚块的导航项的选取请求，使所述目标搜索结果聚块的导航项成为所述导航区域中的展示焦点，以及使所述目标搜索结果聚块成为所述内容区域中的展示焦点。
可选地，响应于交互请求，对至少部分所述搜索结果聚块的展示焦点与各所述搜索结果聚块的导航项的展示焦点进行联动调整，包括：响应于对所述内容区域的浏览请求，使目标搜索结果聚块的导航项成为所述导航区域中的展示焦点，以及使所述目标搜索结果聚块成为所述内容区域中的展示焦点。
可选地，所述方法还包括：响应于对目标搜索结果聚块的选取请求，在搜索结果详情页中以瀑布流形式展示所述目标搜索结果聚块的搜索结果集。
可选地，在所述搜索结果详情页中以瀑布流形式展示所述目标搜索结果聚块的搜索结果集，包括：在所述搜索结果详情页中展示筛选项；响应于对目标筛选项的选取请求，根据目标筛选项对所述目标搜索结果聚块的搜索结果集进行筛选，在所述搜索结果详情页中以瀑布流形式展示所述筛选后的搜索结果集。
可选地，所述一个或多个搜索结果聚块各自对应一个搜索意图，各所述搜索意图对应一个或多个业务类别。
可选地，对于每个所述搜索结果聚块，该搜索结果聚块中的预览结果数量是根据该搜索结果聚块的展示优先级以及该搜索结果聚块对应的最大展示数量确定的；所述最大展示数量是根据搜索日志中，至少部分下单用户在下单前浏览过的搜索结果数量确定的。
可选地，确定一个或多个所述搜索结果聚块和各所述搜索结果聚块的导航项，包括：确定与所述搜索请求对应的一个或多个搜索意图，生成所述一个或多个搜索意图各自对应的搜索结果聚块，以及生成各所述搜索结果聚块的导航项；其中，各所述搜索结果聚块包括预览搜索结果和展示优先级。
可选地，生成所述搜索意图对应的搜索结果聚块，包括：确定该搜索意图的搜索结果集；根据搜索相关性确定该搜索结果集的结果权重；根据该搜索结果集的结果权重确定该搜索意图的意图权重；根据该搜索意图的意图权重确定该搜索意图对应的搜索结果聚块的展示优先级。
可选地，确定该搜索意图的搜索结果集，包括：通过分布式搜索，从一个或多个索引服务中获取与该搜索意图对应的搜索结果集。
可选地，所述搜索相关性包括如下的至少一种：搜索请求与搜索结果描述信息的文本相关性，用户偏好与搜索结果的相关性，搜索场景信息与搜索结果的相关性。
可选地，根据该搜索结果集的结果权重确定该搜索意图的意图权重，包括：根据该搜索结果集的结果权重和该搜索意图的初始权重，确定该搜索意图的意图权重。
依据本申请的第二方面，提供了一种用于搜索的用户交互装置，包括：聚块单元，用于响应于搜索请求，确定一个或多个搜索结果聚块和各所述搜索结果聚块的导航项；展示单元，用于根据各所述搜索结果聚块的展示优先级，展示各所述搜索结果聚块的导航项以及至少部分所述搜索结果聚块；以及用于响应于交互请求，对至少部分所述搜索 结果聚块的展示焦点与各所述搜索结果聚块的导航项的展示焦点进行联动调整。
可选地，所述展示单元用于：在搜索结果导航页的导航区域展示各所述搜索结果聚块的导航项；在所述搜索结果导航页的内容区域展示至少部分所述搜索结果聚块的至少部分搜索预览结果。
可选地，所述展示单元还用于：响应于对目标搜索结果聚块的导航项的选取请求和/或对所述内容区域的浏览请求，使所述目标搜索结果聚块的导航项成为所述导航区域中的展示焦点，以及使所述目标搜索结果聚块成为所述内容区域中的展示焦点。
可选地，所述展示单元还用于：响应于对目标搜索结果聚块的选取请求，在搜索结果详情页中以瀑布流形式展示所述目标搜索结果聚块的搜索结果集。例如，所述展示单元可用于：在所述搜索结果详情页中展示筛选项；响应于对目标筛选项的选取请求，根据目标筛选项对所述目标搜索结果聚块的搜索结果集进行筛选，在所述搜索结果详情页中以瀑布流形式展示所述筛选后的搜索结果集。
可选地，所述一个或多个搜索结果聚块各自对应一个搜索意图，各所述搜索意图对应一个或多个业务类别。
可选地，所述搜索结果聚块中的预览结果数量是根据所述搜索结果聚块的展示优先级以及所述搜索结果聚块对应的最大展示数量确定的；所述最大展示数量是根据搜索日志中，至少部分下单用户在下单前浏览过的搜索结果数量确定的。
可选地，所述聚块单元用于：确定与所述搜索请求对应的一个或多个搜索意图，生成所述一个或多个搜索意图各自对应的搜索结果聚块，以及生成各所述搜索结果聚块的导航项；其中，各所述搜索结果聚块包括预览搜索结果和展示优先级。
可选地，所述聚块单元用于，针对各所述搜索意图：确定该搜索意图的搜索结果集；根据搜索相关性确定该搜索结果集的结果权重；根据该搜索结果集的结果权重确定该搜索意图的意图权重；根据该搜索意图的意图权重确定该搜索意图对应的搜索结果聚块的展示优先级。其中，所述聚块单元可通过分布式搜索，从一个或多个索引服务中获取与搜索意图对应的搜索结果集。所述搜索相关性可包括如下的至少一种：搜索请求与搜索结果描述信息的文本相关性，用户偏好与搜索结果的相关性，搜索场景信息与搜索结果的相关性。
可选地，所述聚块单元具体用于根据该搜索意图的搜索结果集的结果权重和该搜索意图的初始权重确定该搜索意图的意图权重。
依据本申请的第三方面，提供了一种移动终端，包括处理器以及被安排成存储计算机可执行指令的存储器，所述可执行指令在被执行时使所述处理器执行如上述任一项所述的用于搜索的用户交互方法。
依据本申请的第四方面，提供了一种计算机可读存储介质，所述计算机可读存储介质存储一个或多个程序，所述一个或多个程序当被处理器执行时，实现如上述任一项所述的用于搜索中的用户交互方法。
由上述可知，本申请的技术方案，能够根据用户的搜索请求确定一个或多个搜索结果聚块和各所述搜索结果聚块的导航项，并根据各所述搜索结果聚块的展示优先级展示各所述搜索结果聚块的导航项以及至少部分所述搜索结果聚块，包括响应于交互请求对至少部分搜索结果聚块的展示焦点与各所述搜索结果聚块的导航项的展示焦点进行联动调整。该技术方案以搜索结果聚块展示搜索结果(包括聚块的导航项和至少部分聚块的内容)，提升了搜索结果页面所包含的信息量；并通过与用户的交互对搜索结果聚块的展示焦点与搜索结果聚块的导航项的展示焦点进行联动调整，提升了用户体验。其中，搜索结果聚块可以根据需求与用户的搜索意图对应起来，从而实现了搜索结果展示的全新的多意图交互组织形态。
上述说明仅是本申请技术方案的概述，为了能够更清楚了解本申请的技术手段，而可依照说明书的内容予以实施，并且为了让本申请的上述和其它目的、特征和优点能够更明显易懂，以下特举本申请的具体实施方式。
附图说明
通过阅读下文优选实施方式的详细描述，本申请的各种其他优点和益处对于本领域普通技术人员将变得清楚明了。附图仅用于示出优选实施方式，而并不认为是对本申请的限制。而且在整个附图中，用相同的参考符号表示相同的部件。在附图中：
图1示出了根据本申请一个实施例的一种用于搜索的用户交互方法的流程示意图；
图2示出了根据本申请一个实施例的搜索结果导航页示意图；
图3示出了根据本申请一个实施例的搜索结果导航页变化状态示意图；
图4示出了根据本申请一个实施例的搜索结果的交互展示示意图；
图5示出了根据本申请一个实施例的一种用于搜索的用户交互数据生成方法的流程示意图；
图6示出了根据本申请一个实施例的一种用于搜索的用户交互装置的结构示意图；
图7示出了根据本申请一个实施例的一种用于搜索的用户交互数据生成装置的结构示意图；
图8示出了根据本申请一个实施例的移动终端的结构示意图；
图9示出了根据本申请一个实施例的计算机可读存储介质的结构示意图。
具体实施方式
下面将参照附图更详细地描述本申请的示例性实施例。虽然附图中显示了本申请的示例性实施例，然而应当理解，可以以各种形式实现本申请而不应被这里阐述的实施例所限制。实际上，提供这些实施例是为了能够更透彻地理解本申请，并且能够将本申请的范围完整的传达给本领域的技术人员。
图1示出了根据本申请一个实施例的一种用于搜索的用户交互方法的流程示意图。如图1所示，用于搜索的用户交互方法包括以下步骤S110、S120和S130。
步骤S110，响应于搜索请求，确定一个或多个搜索结果聚块和各所述搜索结果聚块的导航项。
本申请的实施例可以应用于各类使用搜索引擎技术的场景，包括但不限于百度、谷歌(此处的商业名称仅作示例性说明)等通用搜索引擎，专利、商标等领域的专用搜索引擎，以及应用APP内的搜索引擎等。搜索请求可来自于用户，包括用户在终端设备、例如移动终端上执行特定的操作。
用户可以通过文本、图像、语音等各类方式生成搜索请求，例如文本可以是搜索关键词或者搜索语句。
搜索结果聚块是搜索结果的聚合。具体来说，可以将搜索结果按照类别进行聚类，得到搜索结果聚块。换言之，搜索结果聚块涵盖属于该类别的搜索结果。下文中，将搜索结果聚块所涵盖的搜索结果的集合称为该搜索结果聚块的搜索结果集。聚类方法有很多。例如按照语义相关性进行聚类、按照搜索结果对应的地理位置进行聚类，等等。
一种方案是按照用户的搜索意图进行聚类。所述一个或多个搜索结果聚块各自对应一个搜索意图，所述搜索意图对应一个或多个业务类别。
例如，每一个搜索意图可对应一种本地生活服务类别，本地生活服务类别可以包括 外卖、闪购、菜谱等。参看下面这个具体场景的例子：用户搜索麻婆豆腐，与麻婆豆腐相关的搜索结果包括将麻婆豆腐作为外卖商品的商家、售卖豆腐和调料的超市、麻婆豆腐的做法等等。那么就可以产生至少三个搜索结果聚块：外卖聚块、闪购聚块以及菜谱聚块。以外卖聚块为例，其涵盖的搜索结果就是将麻婆豆腐作为外卖商品的商家。
搜索意图是根据搜索请求和搜索请求相关的搜索场景信息确定的。具体地，用户的一个搜索请求可以得到多个搜索意图的分布，根据这些搜索意图的强度分布来确定搜索结果聚块的展示优先级。搜索场景信息可以包括天气、位置、用户行为等。可以预先对搜索场景信息和搜索请求进行复合建模，训练得到搜索意图识别模型。从而在实际应用场景中，可通过将搜索请求和搜索请求相关的搜索场景信息输入到预训练好的搜索意图识别模型，来确定搜索意图。并且，针对应用冷启动(初始安装后的首次启动，或者在未被使用达到一段时间后的首次启动)场景，可以通过业务方推荐搜索意图的方式来保证搜索意图识别的鲁棒性。
步骤S120，根据各所述搜索结果聚块的展示优先级，展示各所述搜索结果聚块的导航项以及至少部分搜索结果聚块。
展示优先级可以仅以第一、第二这样的次序表示，也可以具体表现为一个展示权重。例如，假设五个搜索结果聚块A、B、C、D、E的展示权重分别为0.33，0.1，0.2，0.3以及0.07，那么搜索结果聚块的导航项的展示次序，以及搜索结果聚块的展示次序均为A、D、C、B、E。
在移动终端等设备上，由于屏幕大小限制，能够展示的信息有限。因此，每个搜索结果聚块虽然是众多搜索结果的聚合，但不能在屏幕区域内展示全部搜索结果，而可以展示预览搜索结果，例如搜索结果聚块所聚合的众多搜索结果中具有较高权重的搜索结果。具体地，可以通过语义相关性或是业务推荐策略等，来确定搜索结果的权重大小。
步骤S130，响应于交互请求，对所述一个或多个搜索结果聚块的展示焦点与各所述搜索结果聚块的导航项的展示焦点进行联动调整。
交互请求可来自于用户，包括用户在终端设备、例如移动终端上执行特定的操作。搜索结果聚块的导航项可以用来快速定位到搜索结果聚块，包括搜索结果聚块的索引或是快速入口。而用户主动浏览搜索结果聚块时，也可以通过导航项展示焦点的变动来熟悉导航项与搜索结果聚块的对应关系。并且，整个过程可以在一屏页面内完成。
这里的展示焦点调整，是指将目标搜索结果聚块或其导航项以较醒目的方式进行展 示，包括但不限于高亮展示、弹窗展示或者移动到醒目位置如展示区域的首位进行展示等。
可见，图1所示的用于搜索的用户交互方法，以搜索结果聚块展示预览搜索结果，提升了搜索结果页面所包含的信息量；并通过与用户的交互对搜索结果聚块的展示焦点与搜索结果聚块的导航项的展示焦点进行联动调整，提升了用户体验。其中，搜索结果聚块可以根据需求与搜索意图对应起来，从而实现了全新的多意图交互组织形态。
展示各搜索结果聚块的导航项以及至少部分搜索结果聚块，可包括：在搜索结果导航页的导航区域依次展示各搜索结果聚块的导航项；在搜索结果导航页的内容区域展示至少部分搜索结果聚块的至少部分搜索预览结果。
在具体实现时，导航区域与内容区域可以存在多种不同的布局。例如，导航区域在上，内容区域在下；导航区域在侧边并可以根据需求隐藏，内容区域在中心部分持久展示。针对不同的设备，导航区域与内容区域的布局也可以不同。例如双屏设备(双屏不一定是两个独立的显示屏硬件，例如可以是折叠屏根据折叠机构形成的两个屏幕区域)可以在一个屏幕显示导航区域，在另一个屏幕显示内容区域。
图2示出了根据本申请一个实施例的搜索结果导航页示意图。在图2中，最上方是搜索栏，用于供用户输入语音、搜索关键词、图像等以基于用户的该输入来生成搜索请求。搜索栏下的六个控件“点外卖”、“到店吃”、“优惠团购”、“买食材”、“菜谱”以及“点评”分别对应一个搜索结果聚块的导航项。这些搜索结果聚块的导航项所在的区域就是导航区域。在导航区域下面是内容区域，其中示出的是“点外卖”这个搜索结果聚块的两个预览搜索结果。
响应于交互请求，对搜索结果聚块的展示焦点与搜索结果聚块的导航项的展示焦点进行联动调整，可包括：响应于对目标搜索结果聚块的导航项的选取请求，使目标搜索结果聚块的导航项成为导航区域中的展示焦点，以及使目标搜索结果聚块成为内容区域中的展示焦点。
以图2为例，位于内容区域上方的展示位即为展示焦点位。例如，当一个搜索结果聚块的导航项被点选，该搜索结果聚块包含的首个预览搜索结果就会相应移动至该展示焦点位进行展示，也就使得该搜索结果聚块成为了内容区域的展示焦点。
响应于交互请求，对搜索结果聚块的展示焦点与搜索结果聚块的导航项的展示焦点进行联动调整，可包括：响应于对内容区域的浏览请求，使目标搜索结果聚块的导航项 成为导航区域中的展示焦点，以及使目标搜索结果聚块成为内容区域中的展示焦点。
除了点选目标搜索结果聚块的导航项外，用户还可以通过滑动等操作，实现对多个搜索结果聚块的浏览。在这种情况下，随着用户的滑动，导航区域中的导航项的展示焦点也会随之变动。例如，用户浏览完了“买食材”聚块，开始浏览“菜谱”，则导航区域中的展示焦点就从“买食材”变为“菜谱”，具体地可以是通过导航项的移动或高亮等方式来实现。
可以看出，用户可点击任一搜索结果聚块的导航项，使界面快速滑动并使相应搜索结果聚块成为内容区域的展示焦点。用户也可点击搜索结果导航页中任意控件执行更多操作。当用户执行上滑操作时，搜索结果聚块也随之移动，可以查看到图2中未被展示出的其他搜索结果聚块，如“到店吃”、“优惠团购”、“买食材”、“菜谱”以及“点评”。图3示出了搜索结果导航页的变化状态示意图。具体地，在图3中示出了“到店吃”和“优惠团购”两个搜索结果聚块。其中，受限于屏幕尺寸，“优惠团购”搜索结果聚块只示出了部分预览搜索结果。
图2和图3中，空白区域为主题内容，例如主题壁纸、海报图片、视频封面、商户头像、控件图标等。叉号代表文字和/或数字。
根据本申请实施例的用于搜索的用户交互方法，还可包括：响应于对目标搜索结果聚块的选取请求，在搜索结果详情页中以瀑布流形式展示该目标搜索结果聚块的搜索结果集。
例如，如图4所示，左侧为搜索结果导航页，右侧为搜索结果详情页。搜索结果导航页中的任一搜索结果聚块被点击后，从搜索结果导航页跳转至搜索结果详情页，并在搜索结果详情页中依次展示该搜索结果聚块的搜索结果集中的各搜索结果。瀑布流是多栏布局，随着页面滚动条向下滚动，不断加载数据块并附加至当前尾部，适合于多搜索结果的展示场景。
根据本申请实施例的用于搜索的用户交互方法，还可包括：在搜索结果详情页中展示筛选项；响应于对目标筛选项的选取请求，根据目标筛选项对搜索结果集进行筛选，在搜索结果详情页中以瀑布流形式展示筛选后的搜索结果集。
图4中还示出了筛选项，用户可通过勾选“只看品牌商家”、“只看营业中”等筛选项对搜索结果集进行筛选。用户还可以从搜索结果详情页通过向左滑动(如图4的“回退”箭头所示)等方式回到搜索结果导航页。
搜索结果聚块中的预览结果数量可以是根据搜索结果聚块的展示优先级以及搜索结果聚块的最大展示数量确定的。
如前面提到，展示优先级可以以展示权重的方式表示。在一个具体实施例中，可以以展示权重×最大展示数量的方式计算各搜索结果聚块的预览结果数量。如果计算结果不为整数，则向下取整或向上取整。例如，假设搜索结果聚块A的展示权重为0.33，最大展示数量为10，则按照向上取整，搜索结果聚块A的预览结果数量为4。
最大展示数量可以是根据与该搜索结果聚块相关的搜索日志中，至少部分下单用户在下单前浏览过的搜索结果数量确定的。
最大展示数量反映了全体或至少大部分用户对特定应用场景或搜索意图的选择倾向性。例如，在购物等场景下，下单是能够强烈体现正向倾向的一个因素。如果大部分用户都会浏览4-5个商品后选择一个进行下单，那么给出4-5个商品作为预览搜索结果，用户就有很大可能在浏览了这4-5个商品后即进行下单，也不需要跳转至搜索结果详情页，进一步节约了搜索的时间成本。
在一个具体的实施例中，可以对95％的下单用户在下单前浏览过的搜索结果数量做统计，通过求中位数的方式确定最大展示数量。该方式具有较高的泛用性，适合冷启动场景。在其他实施例中，下单用户的比例以及统计方式可以根据需求修改，或者直接以人工方式指定最大展示数量，本申请对此不做限制。
确定搜索结果聚块和搜索结果聚块的导航项，可包括：确定与搜索请求对应的一个或多个搜索意图，生成所述搜索意图各自对应的搜索结果聚块，以及生成各所述搜索结果聚块的导航项。其中，所述搜索结果聚块各自包括预览搜索结果和展示优先级，所述一个或多个搜索结果聚块各自对应一个搜索意图。
准确预测用户的搜索意图，是搜索引擎至关重要的能力。其中，搜索意图通常指搜索行为背后承载的用户真实需求。例如，搜索“羽毛球”，可能是因为用户想购买羽毛球器械，也可能是寻找羽毛球场馆，亦可能是学习羽毛球规则，等等。“购买器械”、“寻找场馆”以及“学习规则”就是与“羽毛球”这个搜索关键词相关的三类不同搜索意图。
在更具体的例子中，搜索意图可以包括外卖、堂食、菜谱、点评、优惠等等。这些搜索意图能够反映出用户需求，具体可以由业务方或是领域专家等进行搜索意图的名称确定以及类别划分。换句话说，可以理解为是概括出的用户需求。
具体到业务场景，搜索意图可以是和商品或者服务的类别相对应的。而商品和服务的类别可以根据业务需求进行定义，例如上面给出的外卖、堂食就是对服务提供方式的分类。
一个搜索结果可以对应一个或多个搜索意图。例如某餐馆既提供堂食售卖，也提供外卖服务，则该餐馆对应的搜索意图可以包括外卖和堂食；而另一餐馆只提供外卖服务，则该餐馆对应的搜索意图仅包括外卖。反过来，一个搜索意图也能够对应一个或多个搜索结果，并且一般是多个搜索结果，比如提供外卖服务的餐馆很多。搜索意图与用户的真实需求越匹配，对应地，展示给用户的搜索结果也就更容易达到用户的搜索目的。
可见，按照搜索意图进行对搜索结果进行聚类并生成搜索结果聚块，能够使用户更快速地关注到自己需要的搜索结果。
生成与搜索意图对应的搜索结果聚块，可包括：确定该搜索意图的搜索结果集；根据搜索相关性确定该搜索结果集中各搜索结果的结果权重；根据该搜索结果集中各搜索结果的结果权重确定该搜索意图的意图权重；根据该搜索意图的意图权重确定对应的搜索结果聚块的展示优先级。
具体地，确定各搜索意图的搜索结果集包括：通过分布式搜索，从一个或多个索引服务中获取与搜索意图对应的搜索结果集。
通过基于搜索意图获取搜索结果集，再基于搜索相关性进一步确定搜索结果集的结果权重，以及通过该结果权重确定意图权重，再通过该意图权重确定对应的搜索结果聚块的展示优先级，这样的方案增强了搜索结果展示贴合用户意图的鲁棒性。具体地，一个搜索意图对应一个包括一个或多个搜索结果的搜索结果集，一个搜索结果聚块对应一个搜索意图。在搜索结果集仅包括一个搜索结果的情况下，也即在搜索意图与搜索结果为一对一关系的情况下，可以认为搜索结果集的结果权重＝搜索结果集中该搜索结果的结果权重；在搜索结果集包括多个搜索结果的情况下，也即在搜索意图与搜索结果为一对多关系的情况下，可以通过多个搜索结果的结果权重的均值运算或加权均值运算得到搜索结果集的结果权重。上述方式可以根据需求组合实现，本申请对此不作限制。
搜索相关性包括如下的至少一种：搜索请求与搜索结果描述信息的文本相关性，用户偏好与搜索结果的相关性，搜索场景信息与搜索结果的相关性。
搜索请求可以看做用户对其搜索意图给出的直接表达。用户偏好可以通过用户预先输入的资料或是用户画像得到。搜索场景信息可以看作是用户对其搜索意图给出的间接 表达，并且能够补充搜索关键词所没有体现出的潜在搜索意图。具体来说，搜索场景信息可以覆盖多个场景维度，例如时间维度、位置维度、天气维度等等。
例如，用户搜索“宫保鸡丁”，可能是因为想学习宫保鸡丁的做法，也可能是因为想点宫保鸡丁的外卖，也可能是希望前往售卖宫保鸡丁的餐馆就餐。但用户在搜索时，并不一定会利用搜索关键词一次性表达清楚自己的搜索意图，这就需要用户在搜索结果中进一步查找，或是进行二次检索，降低了用户体验。但是，从搜索场景信息入手，就能够适当改善这一问题。例如，用户是在商场内搜索宫保鸡丁，那么就更有可能是希望前往售卖宫保鸡丁的餐馆就餐，而非查找菜谱或点外卖。此时，环境的作用就体现了出来。而如果搜索日志显示用户略过了多个售卖宫保鸡丁的实体餐馆，点击进入了多个外卖餐馆的页面，并在一家外卖餐馆下过单，就能够确定用户是希望点外卖，而非其他意图。这就体现了用户行为的作用。
在搜索相关性包括多个类时，可以为每类分配不同的相关性权重。相关性权重可以通过经验得到，如搜索关键词的相关性权重可以分配得较高。还可以根据统计建模得到相关性权重，例如通过搜索日志，对搜索结果聚块的曝光且有点击生成正例，曝光且无点击生成负例，并通过监督式学习确定相关性权重。
根据该搜索结果集中各搜索结果的结果权重确定该搜索意图的意图权重，可包括：根据该搜索结果集中各搜索结果的结果权重和该搜索意图的初始权重确定该搜索意图的意图权重。
一个搜索请求可能对应多个搜索意图，这些搜索意图具有不同的初始权重。搜索意图的初始权重，例如可通过人工方式确定，或者通过统计建模方式确定等等。将搜索意图对应的搜索结果集的结果权重与该搜索意图的初始权重通过线性加权求和等方式，即可得到该搜索意图的意图权重。
上述实施例中涉及权重的方案，优选地，同一类型的权重总和应当为1，可以利用Softmax归一化计算来实现。
这里需要说明的是，用于搜索的用户交互方法可以应用于移动终端侧，具体可以应用于应用程序APP中。但是，这并不意味用于搜索的用户交互方法的全部流程都需要完全在移动终端侧实现。实际上，可以将部分具有较大计算量或是不依赖于隐私信息的处理交由服务器侧来实现。
例如，搜索意图的确定可以由服务器侧来实现。具体地，服务器侧可负责给出搜索 意图强度的分布，并返回给移动终端侧，再由移动终端侧进行搜索结果聚块及导航项的生成，等等。
对此，本申请还给出了将部分步骤放在服务器侧执行的示例，具体将其作为用于搜索中的用户交互数据生成方法，通过下面的实施例进行介绍。
图5示出了根据本申请一个实施例的一种用于搜索的用户交互数据生成方法的流程示意图。如图5所示，用于搜索的用户交互数据生成方法包括S510、S520和S530。
步骤S510，获取移动终端发送的搜索请求。
本申请的实施例可以应用于各类使用搜索引擎技术的场景，包括但不限于百度、谷歌(此处的商业名称仅作示例性说明)等通用搜索引擎，专利、商标等领域的专用搜索引擎，以及应用APP内的搜索引擎等。
用户可以通过文本、图像、语音等各类方式生成搜索请求，例如文本可以是搜索关键词或者搜索语句。
步骤S520，确定与搜索请求对应的一个或多个搜索意图，生成所述一个或多个搜索意图各自对应的搜索结果聚块，以及生成各所述搜索结果聚块的导航项。其中，所述搜索结果聚块各自包括预览搜索结果和展示优先级。
步骤S530，将各所述搜索结果聚块的导航项和至少部分所述搜索结果聚块返回给所述移动终端。
通过步骤S530，使得所述移动终端能够根据展示优先级对各所述搜索结果聚块的导航项和至少部分所述搜索结果聚块进行展示，以及使得所述移动终端能够根据来自用户的交互请求对至少部分所述搜索结果聚块的展示焦点和各所述搜索结果聚块的导航项的展示焦点进行联动调整。
图5所示的用于搜索的用户交互数据生成方法，可以与前述用于搜索的用户交互方法配合实施，为基于用户交互的搜索结果展示提供了数据支撑。并且，针对前述用于搜索的用户交互方法所介绍的各种技术细节，均可适当应用于该用于搜索的用户交互数据生成方法，并在此不再赘述。
确定搜索请求对应的搜索意图可以有多种方式。例如，基于业务专家制定的规则，利用搜索关键词与规则的文本匹配确定搜索意图；基于文本分类或聚类进行搜索意图的预测；通过主题模型等方式，将关键词映射到高维度的语义向量空间，以表达搜索 意图，等等。不过这些方案均存在着仅关注文本，缺少对其他因素关注的问题。
在一个方案中，可以获取与搜索请求关联的搜索场景信息，根据搜索场景信息以及搜索请求生成用于识别搜索意图的复合特征，将复合特征输入到搜索意图识别模型中，获取搜索意图识别模型输出的搜索意图识别结果。这样不仅关注搜索请求，还关注天气、位置、用户行为等搜索场景信息。利用基于复合特征建模实现的搜索意图识别模型，参考多方面因素对用户真实需求进行预测，改善了仅根据搜索请求识别出搜索意图准确率较低的问题，特别适合于生活服务类、LBS(Location Based Services，基于位置的服务)类搜索场景。并且，针对应用冷启动(初始安装后的首次启动，或者在未被使用达到一段时间后的首次启动)场景，可以通过业务方推荐搜索意图的方式来保证搜索意图识别的鲁棒性。
除了本申请给出的上述方案外，还有一些方案也可以基于搜索意图展示搜索结果，但是展示方式都存在一些缺陷性。
例如，所展示的搜索结果仅对应单一搜索意图，并使用固定模板进行展示。该类方法需要准确识别用户第一搜索意图，并以对应的模板组织展现，弊端是显而易见的：一是搜索意图想要完全预测准确、排除歧义几乎无法实现；二是固定模板展示即便在一定的情况下符合了当前主要的用户需求，也扼杀了潜在的增长空间。
又例如，可以将多搜索意图对应的搜索结果，在搜索结果页面中混插。该类方法可以对多个不同业务、满足不同需求的搜索结果集进行混合排序。由于用户难以预知点击列表后的下一个页面，会造成对搜索结果集混乱的理解，影响使用体验。
再例如，可以在主意图中插入辅意图卡片进行展示。该类方法一般需要识别出用户的第一搜索意图作为主意图、并将识别出的其他搜索意图作为辅意图。主意图直接展开为结果列表，而辅意图仅展示摘要卡片，需要用户点击才展开结果列表。由于辅意图的展现可能因排序过于靠下而不完整，用户无法感知。并且，多个卡片的插入会干扰用户对结果列表的浏览，不利于使用体验。
而本申请结合了搜索意图的搜索结果展示方案，能够针对用户的每次搜索请求，预测出当前搜索请求的多个搜索意图，经过分布式的检索召回获取多个搜索意图各自对应的搜索结果集并生成相应的搜索结果聚块，再通过对搜索结果集质量的终判自动生成导航区域以展示搜索结果聚块的导航项和内容区域以展示至少部分搜索结果聚块，并能够根据交互请求联动调整导航区域和内容区域的展示角点，从而形成搜索结果的交互展 示形态。这类由因至果的解决方式，使移动终端因为屏幕大小限制而造成的多意图搜索结果展示不足的问题有了极大改观，能够有效解决单搜索意图霸屏、潜在搜索意图无法展示的问题。
图6示出了根据本申请一个实施例的一种用于搜索的用户交互装置的结构示意图。如图6所示，用于搜索中的用户交互装置600包括聚块单元610和展示单元620。
聚块单元610用于响应于搜索请求，获取一个或多个搜索结果聚块和各所述搜索结果聚块的导航项。
展示单元620用于根据各所述搜索结果聚块的展示优先级，展示各所述搜索结果聚块的导航项以及至少部分搜索结果聚块；以及用于响应于交互请求，对至少部分所述搜索结果聚块的展示焦点与各所述搜索结果聚块的导航项的展示焦点进行联动调整。
可见，图6所示的用于搜索的用户交互装置，以以搜索结果聚块展示搜索结果(包括聚块的导航项和至少部分聚块的内容)，提升了搜索结果页面所包含的信息量；并通过与用户的交互对搜索结果聚块的展示焦点与搜索结果聚块的导航项的展示焦点进行联动调整，提升了用户体验。其中，搜索结果聚块可以根据需求与用户的搜索意图对应起来，从而实现了搜索结果展示的全新的多意图交互组织形态。
展示单元620可用于在搜索结果导航页的导航区域展示各搜索结果聚块的导航项；在搜索结果导航页的内容区域展示至少部分搜索结果聚块的至少部分搜索预览结果。
展示单元620可用于：响应于对目标搜索结果聚块的导航项的选取请求，使目标搜索结果聚块的导航项成为导航区域中的展示焦点，以及使目标搜索结果聚块成为内容区域中的展示焦点；或者，响应于对内容区域的浏览请求，使目标搜索结果聚块的导航项成为导航区域中的展示焦点，以及使目标搜索结果聚块成为内容区域中的展示焦点。
展示单元620还可用于响应于对目标搜索结果聚块的选取请求，在搜索结果详情页中以瀑布流形式展示目标搜索结果聚块的搜索结果集。例如，展示单元620可用于在搜索结果详情页中展示筛选项，响应于对目标筛选项的选取请求根据目标筛选项对目标搜索结果聚块的搜索结果集进行筛选，并在搜索结果详情页中以瀑布流形式展示筛选后的搜索结果集。
聚块单元610可用于：确定与搜索请求对应的一个或多个搜索意图，生成所述一个或多个搜索意图各自对应的搜索结果聚块，以及生成各所述搜索结果聚块的导航项。其中，各搜索结果聚块包括预览搜索结果和展示优先级。
聚块单元610还可用于，针对每个搜索意图：确定该搜索意图的搜索结果集；根据搜索相关性确定该搜索结果集的结果权重；根据该搜索结果集的结果权重确定该搜索意图的意图权重；根据该搜索意图的意图权重确定相应的搜索结果聚块的展示优先级。例如，聚块单元610可通过分布式搜索，从一个或多个索引服务中获取与搜索意图对应的搜索结果集。其中，搜索相关性包括如下的至少一种：搜索请求与搜索结果描述信息的文本相关性，用户偏好与搜索结果的相关性，搜索场景信息与搜索结果的相关性。
聚块单元610还可用于根据搜索意图的搜索结果集的结果权重和该搜索意图的初始权重确定该搜索意图的意图权重。
需要说明的是，上述用于搜索的用户交互装置可用于执行前述用于搜索的用户交互方法，且针对前述方法介绍的各种技术细节均适用于相应的装置，因此不再赘述。
图7示出了根据本申请一个实施例的一种用于搜索的用户交互数据生成装置的结构示意图。如图7所示，用于搜索中的用户交互数据生成装置700包括通信单元710和生成单元720。
通信单元710用于获取移动终端发送的搜索请求。
生成单元720用于确定与搜索请求对应的一个或多个搜索意图，生成所述一个或多个搜索意图各自对应的搜索结果聚块，以及生成各所述搜索结果聚块的导航项。其中，所述搜索结果聚块各自包括预览搜索结果和展示优先级。
通信单元710还可用于将各所述搜索结果聚块的导航项和至少部分所述搜索结果聚块返回给移动终端，以使得所述移动终端能够根据展示优先级对所述搜索结果聚块和搜索结果聚块的导航项进行展示，以及使得所述移动终端能够根据交互请求而对所述搜索结果聚块的展示焦点和各所述搜索结果聚块的导航项的展示焦点进行联动调整。
图7所示的用于搜索的用户交互数据生成装置，可以与前述用于搜索的用户交互装置配合实施，为基于用户交互的搜索结果展示提供了数据支撑。
生成单元720可用于，针对每个搜索意图：确定该搜索意图的搜索结果集；根据搜索相关性确定该搜索结果集的结果权重；根据该结果权重确定该搜索意图的意图权重；根据该意图权重确定该搜索意图对应的搜索结果聚块的展示优先级。例如，生成单元720可通过分布式搜索，从一个或多个索引服务中获取与搜索意图对应的搜索结果集。搜索相关性包括如下的至少一种：搜索请求与搜索结果描述信息的文本相关性，用户偏好与搜索结果的相关性，搜索场景信息与搜索结果的相关性。
生成单元720可用于根据搜索意图的搜索结果集的结果权重和该搜索意图的初始权重确定该搜索意图的意图权重。
需要说明的是，上述用于搜索的用户交互数据生成装置可用于执行前述用于搜索的用户交互数据生成方法，并且针对前述方法介绍的各种技术细节均适用于该装置，因此不再赘述。
综上所述，本申请的技术方案，能够根据用户的搜索请求确定一个或多个搜索结果聚块和各所述搜索结果聚块的导航项，并根据各所述搜索结果聚块的展示优先级展示各所述搜索结果聚块的导航项以及至少部分所述搜索结果聚块，包括响应于交互请求对至少部分搜索结果聚块的展示焦点与各所述搜索结果聚块的导航项的展示焦点进行联动调整。该技术方案以搜索结果聚块展示搜索结果(包括聚块的导航项和至少部分聚块的内容)，提升了搜索结果页面所包含的信息量；并通过与用户的交互对搜索结果聚块的展示焦点与搜索结果聚块的导航项的展示焦点进行联动调整，提升了用户体验。其中，搜索结果聚块可以根据需求与用户的搜索意图对应起来，从而实现了搜索结果展示的全新的多意图交互组织形态。
需要说明的是：在此提供的算法和显示不与任何特定计算机、虚拟装置或者其它设备固有相关。各种通用装置也可以与基于在此的示教一起使用。根据上面的描述，构造这类装置所要求的结构是显而易见的。此外，本申请也不针对任何特定编程语言。应当明白，可以利用各种编程语言实现在此描述的本申请的内容，并且上面对特定语言所做的描述是为了披露本申请的最佳实施方式。
在此处所提供的说明书中，说明了大量具体细节。然而，能够理解，本申请的实施例可以在没有这些具体细节的情况下实践。在一些实例中，并未详细示出公知的方法、结构和技术，以便不模糊对本说明书的理解。
类似地，应当理解，为了精简本申请并帮助理解各个发明方面中的一个或多个，在上面对本申请的示例性实施例的描述中，本申请的各个特征有时被一起分组到单个实施例、图、或者对其的描述中。然而，并不应将该公开的方法解释成反映如下意图：即所要求保护的本申请要求比在每个权利要求中所明确记载的特征更多的特征。更确切地说，如下面的权利要求书所反映的那样，发明方面在于少于前面公开的单个实施例的所有特征。因此，遵循具体实施方式的权利要求书由此明确地并入该具体实施方式，其中每个权利要求本身都作为本申请的单独实施例。
本领域那些技术人员可以理解，可以对实施例中的设备中的模块进行自适应性地改变并且把它们设置在与该实施例不同的一个或多个设备中。可以把实施例中的模块或单元或组件组合成一个模块或单元或组件，以及此外可以把它们分成多个子模块或子单元或子组件。除了这样的特征和/或过程或者单元中的至少一些是相互排斥之外，可以采用任何组合对本说明书(包括伴随的权利要求、摘要和附图)中公开的所有特征以及如此公开的任何方法或者设备的所有过程或单元进行组合。除非另外明确陈述，本说明书(包括伴随的权利要求、摘要和附图)中公开的每个特征可以由提供相同、等同或相似目的的替代特征来代替。
此外，本领域的技术人员能够理解，尽管在此所述的一些实施例包括其它实施例中所包括的某些特征而不是其它特征，但是不同实施例的特征的组合意味着处于本申请的范围之内并且形成不同的实施例。例如，在下面的权利要求书中，所要求保护的实施例的任意之一都可以以任意的组合方式来使用。
本申请的各个部件实施例可以以硬件实现，或者以在一个或者多个处理器上运行的软件模块实现，或者以它们的组合实现。本领域的技术人员应当理解，可以在实践中使用微处理器或者数字信号处理器(DSP)来实现根据本申请实施例的用于搜索中的用户交互装置以及用于搜索中的用户交互数据生成装置中的一些或者全部部件的一些或者全部功能。本申请还可以实现为用于执行这里所描述的方法的一部分或者全部的设备或者装置程序(例如，计算机程序和计算机程序产品)。这样的实现本申请的程序可以存储在计算机可读介质上，或者可以具有一个或者多个信号的形式。这样的信号可以从因特网网站上下载得到，或者在载体信号上提供，或者以任何其他形式提供。
例如，图8示出了根据本申请一个实施例的移动终端的结构示意图。该移动终端800包括处理器810和被安排成存储计算机可执行指令(计算机可读程序代码)的存储器820。存储器820可以是诸如闪存、EEPROM(电可擦除可编程只读存储器)、EPROM、硬盘或者ROM之类的电子存储器。存储器820具有存储用于执行上述方法中的任何方法步骤的计算机可读程序代码831的存储空间830。例如，用于存储计算机可读程序代码的存储空间830可以包括分别用于实现上面的方法中的各种步骤的各个计算机可读程序代码831。计算机可读程序代码831可以从一个或者多个计算机程序产品中读出或者写入到这一个或者多个计算机程序产品中。这些计算机程序产品包括诸如硬盘，紧致盘(CD)、存储卡或者软盘之类的程序代码载体。这样的计算机程序产品通常为例如图9所述的计算机可读存储介质。图9示出了根据本申请一个实施例的一种计算机可读存储 介质的结构示意图。该计算机可读存储介质900存储有用于执行根据本申请的方法步骤的计算机可读程序代码831，可以被移动终端800的处理器810读取，当计算机可读程序代码831由移动终端800运行时，导致该移动终端800执行上面所描述的方法中的各个步骤，具体来说，该计算机可读存储介质存储的计算机可读程序代码831可以执行上述任一实施例中示出的方法。计算机可读程序代码831可以以适当形式进行压缩。
应该注意的是上述实施例对本申请进行说明而不是对本申请进行限制，并且本领域技术人员在不脱离所附权利要求的范围的情况下可设计出替换实施例。在权利要求中，不应将位于括号之间的任何参考符号构造成对权利要求的限制。单词“包含”不排除存在未列在权利要求中的元件或步骤。位于元件之前的单词“一”或“一个”不排除存在多个这样的元件。本申请可以借助于包括有若干不同元件的硬件以及借助于适当编程的计算机来实现。在列举了若干装置的单元权利要求中，这些装置中的若干个可以是通过同一个硬件项来具体体现。单词第一、第二、以及第三等的使用不表示任何顺序。可将这些单词解释为名称。

Claims (19)

  1. 一种用于搜索的用户交互方法，包括：
    响应于搜索请求，确定一个或多个搜索结果聚块和各所述搜索结果聚块的导航项；
    根据各所述搜索结果聚块的展示优先级，展示各所述搜索结果聚块的导航项以及至少部分所述搜索结果聚块；
    响应于交互请求，对至少部分所述搜索结果聚块的展示焦点与各所述搜索结果聚块的导航项的展示焦点进行联动调整。
  2. 如权利要求1所述的方法，其特征在于，展示各所述搜索结果聚块的导航项以及至少部分所述搜索结果聚块，包括：
    在搜索结果导航页的导航区域，展示各所述搜索结果聚块的导航项；
    在所述搜索结果导航页的内容区域，展示至少部分所述搜索结果聚块的至少部分搜索预览结果。
  3. 如权利要求2所述的方法，其特征在于，响应于交互请求，对至少部分所述搜索结果聚块的展示焦点与各所述搜索结果聚块的导航项的展示焦点进行联动调整，包括：
    响应于对目标搜索结果聚块的导航项的选取请求，
    使所述目标搜索结果聚块的导航项成为所述导航区域中的展示焦点，以及
    使所述目标搜索结果聚块成为所述内容区域中的展示焦点。
  4. 如权利要求2所述的方法，其特征在于，响应于交互请求，对至少部分所述搜索结果聚块的展示焦点与各所述搜索结果聚块的导航项的展示焦点进行联动调整，包括：
    响应于对所述内容区域的浏览请求，
    使目标搜索结果聚块的导航项成为所述导航区域中的展示焦点，以及
    使所述目标搜索结果聚块成为所述内容区域中的展示焦点。
  5. 如权利要求1所述的方法，其特征在于，所述方法还包括：
    响应于对目标搜索结果聚块的选取请求，在搜索结果详情页中以瀑布流形式展示所述目标搜索结果聚块的搜索结果集。
  6. 如权利要求5所述的方法，其特征在于，在所述搜索结果详情页中以瀑布流形式展示所述目标搜索结果聚块的搜索结果集，包括：
    在所述搜索结果详情页中展示筛选项；
    响应于对目标筛选项的选取请求，根据目标筛选项对所述目标搜索结果聚块的搜索结果集进行筛选；
    在所述搜索结果详情页中以瀑布流形式展示所述筛选后的搜索结果集。
  7. 如权利要求1所述的方法，其特征在于，
    所述一个或多个搜索结果聚块各自对应一个搜索意图，
    各所述搜索意图对应一个或多个业务类别。
  8. 如权利要求2所述的方法，其特征在于，
    对于每个所述搜索结果聚块，该搜索结果聚块的预览结果数量是根据该搜索结果聚块的展示优先级以及该搜索结果聚块对应的最大展示数量确定的；
    所述最大展示数量是根据搜索日志中，至少部分下单用户在下单前浏览过的搜索结果数量确定的。
  9. 如权利要求1所述的方法，其特征在于，确定一个或多个所述搜索结果聚块和各所述搜索结果聚块的导航项，包括：
    确定与所述搜索请求对应的一个或多个搜索意图，
    生成所述一个或多个搜索意图各自对应的搜索结果聚块，以及
    生成各所述搜索结果聚块的导航项；
    其中，各所述搜索结果聚块包括预览搜索结果和展示优先级。
  10. 如权利要求9所述的方法，其特征在于，生成所述搜索意图对应的搜索结果聚块，包括：
    确定该搜索意图的搜索结果集；
    根据搜索相关性确定该搜索结果集的结果权重；
    根据该搜索结果集的结果权重确定该搜索意图的意图权重；
    根据该搜索意图的意图权重确定该搜索意图对应的搜索结果聚块的展示优先级。
  11. 如权利要求10所述的方法，其特征在于，确定该搜索意图的搜索结果集，包括：
    通过分布式搜索，从一个或多个索引服务中获取与该搜索意图对应的搜索结果集。
  12. 如权利要求10所述的方法，其特征在于，所述搜索相关性包括如下至少一种：
    搜索请求与搜索结果描述信息的文本相关性，
    用户偏好与搜索结果的相关性，
    搜索场景信息与搜索结果的相关性。
  13. 如权利要求10所述的方法，其特征在于，根据该搜索结果集的结果权重确定该搜索意图的意图权重，包括：
    根据该搜索结果集的结果权重和该搜索意图的初始权重，确定该搜索意图的意图权重。
  14. 一种用于搜索的用户交互数据生成方法，包括：
    获取移动终端发送的搜索请求；
    确定与所述搜索请求对应的一个或多个搜索意图；
    生成所述一个或多个搜索意图各自对应的搜索结果聚块，其中，所述搜索结果聚块各自包括预览搜索结果和展示优先级；
    生成各所述搜索结果聚块的导航项；
    将各所述搜索结果聚块的导航项和至少部分所述搜索结果聚块返回给所述移动终端。
  15. 一种用于搜索的用户交互装置，包括：
    聚块单元，用于响应于搜索请求，确定一个或多个搜索结果聚块和各所述搜索结果聚块的导航项；
    展示单元，用于
    根据各所述搜索结果聚块的展示优先级，展示各所述搜索结果聚块的导航项以及至少部分所述搜索结果聚块；以及
    响应于交互请求，对至少部分所述搜索结果聚块的展示焦点与各所述搜索结果聚块的导航项的展示焦点进行联动调整。
  16. 一种用于搜索的用户交互数据生成装置，包括：
    通信单元，用于获取移动终端发送的搜索请求；
    生成单元，用于
    确定与所述搜索请求对应的一个或多个搜索意图；
    生成所述一个或多个搜索意图各自对应的搜索结果聚块，其中，所述搜索结果聚块各自包括预览搜索结果和展示优先级；
    生成各所述搜索结果聚块的导航项；
    所述通信单元还用于将各所述搜索结果聚块的导航项和至少部分所述搜索结果聚块返回给所述移动终端。
  17. 一种移动终端，包括：
    处理器；以及
    被安排成存储计算机可执行指令的存储器，
    所述可执行指令在被所述处理执行时，使所述处理器执行如权利要求1-13中任一项所述的用于搜索的用户交互方法。
  18. 一种服务器，包括：
    处理器；以及
    被安排成存储计算机可执行指令的存储器，
    所述可执行指令在被所述处理执行时，使所述处理器执行如权利要求14所述的用于搜索的用户交互数据生成方法。
  19. 一种计算机可读存储介质，存储一个或多个程序，
    所述一个或多个程序当被处理器执行时，使所述处理器实现如权利要求1-14中任一项所述的用于搜索的用户交互方法。
PCT/CN2021/081612 2020-03-20 2021-03-18 用于搜索的用户交互 WO2021185326A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010203785.8 2020-03-20
CN202010203785.8A CN111444405A (zh) 2020-03-20 2020-03-20 用于搜索中的用户交互方法、装置、移动终端和存储介质

Publications (1)

Publication Number Publication Date
WO2021185326A1 true WO2021185326A1 (zh) 2021-09-23

Family

ID=71654346

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/081612 WO2021185326A1 (zh) 2020-03-20 2021-03-18 用于搜索的用户交互

Country Status (2)

Country Link
CN (1) CN111444405A (zh)
WO (1) WO2021185326A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20220310084A1 (en) * 2021-03-24 2022-09-29 Adobe Inc. Extensible search, content, and dialog management system with human-in-the-loop curation

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111444405A (zh) * 2020-03-20 2020-07-24 北京三快在线科技有限公司 用于搜索中的用户交互方法、装置、移动终端和存储介质
CN112417299A (zh) * 2020-12-08 2021-02-26 西安联乘智能科技有限公司 一种网页推荐方法、计算机存储介质及计算设备
CN113486252A (zh) * 2021-07-30 2021-10-08 北京字节跳动网络技术有限公司 搜索结果展示方法、装置、设备和介质
CN114168866A (zh) * 2021-12-14 2022-03-11 北京字跳网络技术有限公司 一种搜索结果展现方法、装置、计算机设备和存储介质

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080222145A1 (en) * 2006-01-12 2008-09-11 International Business Machines Corporation Visual method and apparatus for enhancing search result navigation
CN102184230A (zh) * 2011-05-11 2011-09-14 北京百度网讯科技有限公司 一种搜索结果的展示方法及装置
CN103699619A (zh) * 2013-12-18 2014-04-02 北京百度网讯科技有限公司 一种用于提供搜索结果的方法及装置
CN103744866A (zh) * 2013-12-18 2014-04-23 北京百度网讯科技有限公司 一种搜索方法与装置
CN107168991A (zh) * 2017-03-28 2017-09-15 北京三快在线科技有限公司 一种搜索结果展示方法和装置
CN111444405A (zh) * 2020-03-20 2020-07-24 北京三快在线科技有限公司 用于搜索中的用户交互方法、装置、移动终端和存储介质

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106156244B (zh) * 2015-04-28 2020-08-28 阿里巴巴集团控股有限公司 一种信息搜索导航方法及装置
CN105630908A (zh) * 2015-12-21 2016-06-01 北京奇虎科技有限公司 搜索结果的展示方法和装置
CN106202285A (zh) * 2016-06-30 2016-12-07 北京百度网讯科技有限公司 搜索结果展示方法和装置
CN108932253B (zh) * 2017-05-25 2021-01-26 阿里巴巴（中国）有限公司 多媒体搜索结果展示方法及装置
CN110020094B (zh) * 2017-07-14 2023-06-13 阿里巴巴集团控股有限公司 一种搜索结果的展示方法和相关装置
CN107885889A (zh) * 2017-12-13 2018-04-06 聚好看科技股份有限公司 搜索结果的反馈方法、展示方法及装置

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080222145A1 (en) * 2006-01-12 2008-09-11 International Business Machines Corporation Visual method and apparatus for enhancing search result navigation
CN102184230A (zh) * 2011-05-11 2011-09-14 北京百度网讯科技有限公司 一种搜索结果的展示方法及装置
CN103699619A (zh) * 2013-12-18 2014-04-02 北京百度网讯科技有限公司 一种用于提供搜索结果的方法及装置
CN103744866A (zh) * 2013-12-18 2014-04-23 北京百度网讯科技有限公司 一种搜索方法与装置
CN107168991A (zh) * 2017-03-28 2017-09-15 北京三快在线科技有限公司 一种搜索结果展示方法和装置
CN111444405A (zh) * 2020-03-20 2020-07-24 北京三快在线科技有限公司 用于搜索中的用户交互方法、装置、移动终端和存储介质

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20220310084A1 (en) * 2021-03-24 2022-09-29 Adobe Inc. Extensible search, content, and dialog management system with human-in-the-loop curation
US11948566B2 (en) * 2021-03-24 2024-04-02 Adobe Inc. Extensible search, content, and dialog management system with human-in-the-loop curation

Also Published As

Publication number Publication date
CN111444405A (zh) 2020-07-24

Similar Documents

Publication Publication Date Title
WO2021185326A1 (zh) 用于搜索的用户交互
US20230325358A1 (en) Systems and methods for generating and using shared natural language libraries
CN110110203B (zh) 资源信息推送方法及服务器、资源信息展示方法及终端
US20200183966A1 (en) Creating Real-Time Association Interaction Throughout Digital Media
US11995112B2 (en) System and method for information recommendation
US11978093B2 (en) Automated factor generation for decision engines
US11003678B2 (en) Method and system for presenting a search result in a search result card
US9355407B2 (en) Systems and methods for searching cloud-based databases
US10180979B2 (en) System and method for generating suggestions by a search engine in response to search queries
US9208202B1 (en) Systems and methods for determining interest in an item or category of items
US9615136B1 (en) Video classification
US20050171940A1 (en) Dynamic visualization of search results on a user interface
TW201243632A (en) Search assistant system and method
EP2856358A2 (en) Systems and methods for enabling natural language processing
US20130151509A1 (en) System and method for graphical display of search results
US11682060B2 (en) Methods and apparatuses for providing search results using embedding-based retrieval
CN111782951A (zh) 确定展示页面的方法和装置、以及计算机系统和介质
US20160335697A1 (en) Computer implemented method for ranking data & removing dependencies on menus or multiple fields and controls
CN112699311A (zh) 信息推送方法、存储介质及电子设备
US20200380593A1 (en) Systems and Methods for Presenting and Responding to Visual Information and Ranking Results
US10990601B1 (en) Dynamic optimization of variant recommendations
WO2020023140A1 (en) Personalized whole search page organization and relevance
CA2907055C (en) System, method, and computer program product for comparing decision options
US20240127298A1 (en) Systems and methods for individualized restaurant rating
US20240095276A1 (en) Media file recommendations for a search engine

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: 21770696

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: 21770696

Country of ref document: EP

Kind code of ref document: A1