CN108537596B - Method, device and system for recommending vehicle type in search box and memory - Google Patents

Method, device and system for recommending vehicle type in search box and memory Download PDF

Info

Publication number
CN108537596B
CN108537596B CN201810338063.6A CN201810338063A CN108537596B CN 108537596 B CN108537596 B CN 108537596B CN 201810338063 A CN201810338063 A CN 201810338063A CN 108537596 B CN108537596 B CN 108537596B
Authority
CN
China
Prior art keywords
user
vehicle type
search box
interest portrait
down list
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201810338063.6A
Other languages
Chinese (zh)
Other versions
CN108537596A (en
Inventor
车皓阳
洪煦
杜涛
朱劲松
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Chehui Technology Co Ltd
Original Assignee
Beijing Chehui Technology Co Ltd
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 Beijing Chehui Technology Co Ltd filed Critical Beijing Chehui Technology Co Ltd
Priority to CN201810338063.6A priority Critical patent/CN108537596B/en
Priority to CN201910730605.9A priority patent/CN110517067A/en
Publication of CN108537596A publication Critical patent/CN108537596A/en
Application granted granted Critical
Publication of CN108537596B publication Critical patent/CN108537596B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9535—Search customisation based on user profiles and personalisation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0241—Advertisements
    • G06Q30/0251—Targeted advertisements
    • G06Q30/0269—Targeted advertisements based on user profile or attribute
    • G06Q30/0271—Personalized advertisement

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Databases & Information Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Accounting & Taxation (AREA)
  • Development Economics (AREA)
  • Finance (AREA)
  • Strategic Management (AREA)
  • General Physics & Mathematics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • General Business, Economics & Management (AREA)
  • Marketing (AREA)
  • Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The invention discloses a method, a device and a system for recommending vehicle types in a search box and a memory. The method comprises the following steps: inquiring an online cache by a user under the condition that the user does not input information but places a focus in a search box, and displaying an inquired first user vehicle type interest portrait in a drop-down list of the search box to realize vehicle type recommendation, wherein the online cache contains the first user vehicle type interest portrait; and/or generating a second user vehicle type interest portrait according to the information input by the user under the condition that the user inputs the information in the search box, and displaying the generated second user vehicle type interest portrait in a drop-down list of the search box to realize vehicle type recommendation. According to the invention, the personalized competitive product recommendation items are provided in the search box drop-down list, the recommendation results are put forward from the search result list to the search box drop-down list, and the vehicle type search intention of the user is more accurately positioned through two times of accurate recommendation screening of the drop-down list and the recommendation results, so that the personalized search experience of the user is promoted and enriched.

Description

Method, device and system for recommending vehicle type in search box and memory
Technical Field
The present application relates to search boxes, and in particular, to a method, an apparatus, a system, and a storage for recommending a vehicle type in a search box.
Background
Search is widely available in web pages and mobile application software as a main means for acquiring network information. As the domain is subdivided, it is slowly evolving by traditional search engines to personalized searches. The technical solutions used by different search service providers are different, and the display modes are also different. The present application takes a vehicle type search as an example for explanation.
Fig. 1 shows a schematic diagram of using a Baidu search for an audi brand, fig. 2 and 4 show pull-down menu options of home search boxes for easy car networks and cars, and fig. 3 and 5 show pull-down menu options of home search boxes for easy car searches and cars.
In the prior art, a search box is based on search keywords for maximum prefix matching, Google Suggest and AutoComplete technologies are mostly used, the real vehicle type search intention of a searcher is not accurately identified, and the efficiency of recommendation and advertisement delivery is not high.
FIG. 6 shows the front-end and back-end interaction steps in the Google Suggest's technical architecture as follows:
a user inputs vehicle type information, such as 'Cleisler' in a search box, a browser asynchronously sends a request to a Web server through AJAX, and meanwhile, the frequency of continuously inputting and submitting the request is controlled, and the request is submitted in a delayed mode;
after receiving the request, the server forwards the request to a memcached cache cluster, a compressed search tree (trie) is used for judging whether a prefix which can be matched exists in the cache or not, if so, a result is returned to the server, and the result is returned to the front end of the browser through AJAX by the server;
and if the cache does not have corresponding data, the data is obtained from the RDBMS and stored in memcached, and meanwhile, the data is returned to the front end of the browser through the server.
If the Autocomplete in the general search engine is used, the back end is relatively complex, but if the vehicle model information is only automatically supplemented, because the data volume is very small, the cache cluster is not needed at all, and the requirement can be met by using a memcached cache server or directly using an xml file.
In the prior art, the technologies used in a general search engine are not oriented to a specific field, the exclusive characteristics of the field are not strong, the particularity of vehicle type recommendation is not highlighted, and optimization is not carried out aiming at the vehicle type recommendation; at present, GoogleSuggest is applied without combining any searcher feature information, so that only one face of thousands of people can be achieved, but not the face of thousands of people can be achieved; in the prior art, vehicle type information input in a search box is only supplemented or only limited recommendation of similar vehicle types is made, and recommendation of various competitive product vehicle types and hot search vehicle types is lacked; in the prior art, the advertisement position attribute of a pull-down list box is not highlighted, and the presentation information is usually non-advertisement attribute information, not to mention the combination with accurate advertisements; in addition, in the prior art, vehicle type text information is contained in pull-down list boxes, and vivid media forms such as visual vehicle type diagrams and dynamic diagrams are lacked.
Disclosure of Invention
The embodiment of the application provides a method, a device, a system and a memory for accurately recommending vehicle types in a search box, which are used for solving the problem of recommending vehicle types according to user attributes and behaviors.
The application provides a method for recommending vehicle types in a search box, which comprises the following steps:
inquiring an online cache by a user under the condition that the user does not input information but places a focus in a search box, and displaying an inquired first user vehicle type interest portrait in a drop-down list of the search box to realize vehicle type recommendation, wherein the online cache contains the first user vehicle type interest portrait; and/or
And under the condition that the user inputs information in the search box, generating a second user vehicle type interest portrait according to the information input by the user, and displaying the generated second user vehicle type interest portrait in a drop-down list of the search box to realize vehicle type recommendation.
Preferably, the method further comprises:
displaying the advertisement under the condition that the option hovered in the search box drop-down list by the user is an advertisement item;
when the option clicked by the user in the search box drop-down list is the recommended item, jumping to the webpage corresponding to the recommended item; and skipping to the advertisement landing page under the condition that the option clicked by the user in the search box pull-down list is the advertisement.
Preferably, the method further comprises:
and generating a user interest portrait according to the user log.
Preferably, the first user vehicle type interest portrait is stored in an online cache according to the following steps:
receiving the exposure and click behaviors of a drop-down list item in a search box of a user;
pulling a user interest portrait to generate a feature vector of a user;
generating a first user vehicle type interest portrait according to the feature vector of the user and the feature-article correlation matrix;
and storing the interest portrait of the first user vehicle type into an online cache.
Preferably, the generating the first user vehicle type interest portrait according to the feature vector of the user and the feature-item correlation matrix comprises:
acquiring an initial recommended article list according to the feature vector of the user and the feature-article correlation matrix;
obtaining a vehicle type and a competitive product vehicle type which accord with a correlation threshold value from an initial recommended article list according to the user interest portrait;
and sorting the vehicle types meeting the correlation threshold and the vehicle types of the competitive products to generate the interest portrait of the first user vehicle type.
Preferably, the vehicle type meeting the correlation threshold and the vehicle type of the competitive products are vehicle types in the candidate vehicle type set.
Preferably, the method further comprises:
under the condition that the first user vehicle type interest portrait is not inquired in the online cache, inquiring the first user vehicle type interest portrait in the online database, and if the first user vehicle type interest portrait exists, displaying the inquired first user vehicle type interest portrait in a pull-down list of a search box;
under the condition that the first user vehicle type interest portrait is not inquired in the online database, inquiring the first user vehicle type interest portrait in an offline user center, and displaying the inquired first user vehicle type interest portrait in a pull-down list of a search box;
and under the condition that the first user vehicle type interest portrait is not inquired by the offline user center, displaying the hot search vehicle type in a pull-down list of the search box.
Preferably, the generating the second user vehicle type interest representation according to the information input by the user comprises the following steps:
identifying the vehicle type input by the user according to the information input by the user;
and acquiring a recommendation list and/or an advertisement list corresponding to the user, and generating a second user vehicle type interest portrait by using the vehicle type input by the user, the user offline vehicle type interest portrait and the content mixed arrangement in the vehicle type competition list.
The first user vehicle type interest portrait or the second user vehicle type interest portrait is a vehicle purchasing stage portrait of a purchased vehicle user and is used for providing the same brand vehicle type for the purchased vehicle user.
The application provides a device that car type was recommended in search box includes:
the first recommendation module is used for inquiring an online cache under the condition that a user places a focus in a search box but does not input information, and displaying an inquired first user vehicle type interest portrait in a drop-down list of the search box to realize vehicle type recommendation, wherein the online cache contains the first user vehicle type interest portrait; and/or
And the second recommending module is used for generating a second user vehicle type interest portrait according to the information input by the user under the condition that the user inputs the information in the search box, and displaying the generated second user vehicle type interest portrait in a pull-down list of the search box to realize vehicle type recommendation.
Preferably, the apparatus further comprises an advertisement module, configured to, in a case that an option hovered in the search box drop-down list by the user is an advertisement item, present an advertisement;
preferably, the device further comprises a skipping module, which is used for skipping to a webpage corresponding to the recommended item when the option clicked in the search box drop-down list by the user is the recommended item; and skipping to the advertisement landing page under the condition that the option clicked by the user in the search box pull-down list is the advertisement.
Preferably, the apparatus further comprises a user interest representation generation module for generating a user interest representation from the user log.
Preferably, the first recommending module is further configured to perform drop-down list item exposure and click behaviors in the search box according to the received user; pulling a user interest portrait to generate a feature vector of a user; generating a first user vehicle type interest portrait according to the feature vector of the user and the feature-article correlation matrix; and storing the interest portrait of the first user vehicle type into an online cache.
Generating a first user vehicle type interest portrait according to the feature vector of the user and the feature-item correlation matrix comprises the following steps: acquiring an initial recommended article list according to the feature vector of the user and the feature-article correlation matrix; obtaining a vehicle type and a competitive product vehicle type which accord with a correlation threshold value from an initial recommended article list according to a long-term interest portrait of a user; and sorting the vehicle types meeting the correlation threshold and the vehicle types of the competitive products to generate the interest portrait of the first user vehicle type.
Preferably, the first recommending module is further configured to query the first user vehicle type interest representation in the online database under the condition that the first user vehicle type interest representation is not queried in the online cache, and if the first user vehicle type interest representation exists, display the queried first user vehicle type interest representation in a drop-down list of the search box; under the condition that the first user vehicle type interest portrait is not inquired in the online database, inquiring the first user vehicle type interest portrait in an offline user center, and displaying the inquired first user vehicle type interest portrait in a pull-down list of a search box; and under the condition that the first user vehicle type interest portrait is not inquired by the offline user center, displaying the hot search vehicle type in a pull-down list of the search box.
Preferably, the second recommending module is further configured to identify a vehicle type input by the user according to the information input by the user; and acquiring a recommendation list and/or an advertisement list corresponding to the user, and generating a second user vehicle type interest portrait by using the vehicle type input by the user, the user offline vehicle type interest portrait and the content mixed arrangement in the vehicle type competition list.
The application provides a system for accurately recommending vehicle types in a search box, a memory and a processor;
a memory for storing a program;
a processor for executing a program to implement the method.
A memory is provided having stored thereon a program for implementing the method when executed by a processor.
According to the method and the device, the personalized competitive product recommendation items are provided in the search box pull-down list, the recommendation results are put forward from the search result list to the search box pull-down list, and through two times of accurate recommendation screening of the pull-down list and the recommendation results, the vehicle type search intention of the user is more accurately positioned, and the personalized search experience of the user is improved and enriched. According to the method and the device, a plurality of vehicle type advertisement positions are added in the original search box pull-down list without vehicle type advertisement positions, the number of the advertisement positions of the platform at the mobile end is effectively expanded, and a new possibility is provided for the increase of the advertising win and win of the mobile platform.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:
FIG. 1 is a schematic illustration of using a hundred degree search for an Audi vehicle type brand;
FIG. 2 is a schematic diagram of a pull-down menu option of a home search box of the easy-to-drive web page;
FIG. 3 is a schematic view of the vehicle easy search homepage pull-down menu options;
FIG. 4 is a diagram of a drop down menu option for a home page search box of an automobile;
FIG. 5 is a diagram of a home search home pull-down menu option for a car;
FIG. 6 is a schematic diagram of a Google Suggest technical solution;
FIG. 7 is a schematic diagram of a process for providing a user interest representation according to the present invention;
FIG. 8 is a schematic view of a personalized recommendation/advertisement process provided by the present invention;
FIG. 9 is a detailed flowchart of a method for accurately recommending vehicle types in a search box according to the present invention;
fig. 10 is a schematic diagram of an apparatus for accurately recommending vehicle types in a search box according to the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the technical solutions of the present application will be described in detail and completely with reference to the following specific embodiments of the present application and the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
The invention utilizes the mature personalized recommendation algorithm to realize the personalization of the vehicle type recommendation of the search box, binds the behavior and the attribute of the searcher with the vehicle type to implement the recommendation, and helps the searcher to quickly find the target vehicle type matched with the search intention of the searcher, and the realization of the target effect is mainly to see the click rate improvement of the searcher on the pull-down menu option.
In order to achieve the purpose of the invention, the invention is mainly realized by three parts, namely user interest portrait preparation, a personalized recommendation/advertisement engine, a recommendation/advertisement triggering mechanism and the like.
FIG. 7 is a flow chart illustrating a process of preparing a user interest representation according to the present invention. The user interest portrait is obtained by analyzing the user log. The user interest portrait comprises a user long-term interest portrait generated according to behaviors such as historical access and clicking of a user and a user short-term interest portrait generated according to behaviors such as instant access and clicking of the user, the short-term interest portrait is often generated efficiently in a streaming mode, and the long-term interest portrait is often generated off-line every other day in a batch mode. When a user accesses a website or application software where a search box is located for the first time, the website can plant cookies at the browser end of the user, and the application software can record APP and open a log for the first time according to imei/idfa. The server generates an offline user log according to the browsing, accessing and purchasing behaviors of the user. The user center (comprising a user behavior library and a user attribute library) analyzes a user log by taking cookie and imei/idfa as units to form a multi-dimensional user interest portrait related to population attributes, historical behaviors, interest contents, preference tendency and vehicle purchasing stages of a user, wherein the most important user interest portrait is a vehicle type interest portrait and a vehicle purchasing stage portrait for the application, the vehicle type interest portrait comprehensively judges the interest weight of the user on a certain vehicle type according to behaviors of accessing, clicking, placing orders and the like of the user, and the vehicle purchasing stage portrait comprehensively judges which stage the user is in interest, intention, purchasing and purchased vehicles according to multi-channel source data comprising offline data. The vehicle purchasing stage drawing can help the system to accurately capture what kind of stage the user is in relative to the vehicle, so that different recommendation strategies are used for providing different vehicle type recommendations when recommending drop-down list box recommendations, for example, as many competitive vehicle types as possible are provided for the user to select in the interest stage, more accurate closing-up vehicle types of the same brand are provided for selection in the purchasing stage, and vehicle type changing recommendations are provided for the purchased vehicle user to select in the drop-down box. In addition, the access and clue data of all users are counted, and the vehicle type competition list and the hot vehicle searching type list are refreshed. The user behavior library comprises behaviors of data items such as user vehicle type channel browsing, clue submitting, praise, comment and collection, and the user attribute library comprises user gender, age range, academic history, marital status, work category, frequent residence, income level, interests and hobbies.
As shown in FIG. 7, the process of generating a representation of interest of a user includes:
step 705, determining whether a browser or an application software APP is used by a user;
step 710, if the APP is used, recording a first opening log;
step 715, determining whether the user accesses for the first time through the browser;
step 720, under the condition that the user accesses for the first time, the browser types the cookie;
step 725, add the offline user log;
step 730, generating a user interest portrait;
step 735, refreshing a vehicle type competitive product table;
and step 740, refreshing the hot vehicle searching type table.
Fig. 8 is a flow chart of personalized recommendations/advertisements provided by the present invention. Through user operations, such as drop-down list recommendation/advertisement space item exposure when a user puts a cursor in a search box and clicking actions of the user, a user interest portrait is pulled from the center of the user and a feature vector of the user is generated. When the drop-down list is expanded to execute recommendation, vehicle type interests do not need to be used as orientation conditions, for example, regional orientation recommendation is made to the whole users in a certain city, at this time, if only non-behavior features are used, namely user attribute features are used, a vehicle type behavior extraction and behavior feature conversion module is not needed, and the output of the module is the user feature vector. And converting the feature vector of the user into an initial recommended item list by using the feature vector of the user and a pre-stored feature-item correlation matrix. Preferably, the recommended item list can be preprocessed to filter out vehicle types with obvious problems, such as vehicle types with abnormal names, vehicle types with extremely low interest weight, and the like. And obtaining the vehicle type and the competitive product vehicle type with the correlation degree meeting the threshold value according to the long-term interest portrait of the user generated by the offline data, predicting the click through rate CTR according to the trained sequencing model, and selecting the first N as the final recommendation result to obtain the vehicle type interest portrait. The vehicle type interest portrait can be in the form of "(< vehicle type name, vehicle type quotation section > | < vehicle type name, vehicle type page URL, vehicle type quotation section > | < vehicle type name, advertisement type, advertisement material, advertisement landing page URL, vehicle type quotation section >)" and can be periodically brushed into an online cache.
As shown in fig. 8, the personalized recommendation/advertisement flow chart provided by the present invention includes:
step 805, extracting behavior information related to the vehicle type and the like from the user behavior library, wherein the behavior information comprises behavior occurrence time, category, times and the like;
step 810, user behavior feature conversion, namely converting behaviors of browsing, collecting, praise, purchasing and the like of the user into features, namely expression modes of labels;
step 815, generating a user feature vector, wherein the feature vector consists of features and feature weights, and generally, the more recent behavior, the more frequent behavior and the more active behavior weight are generated, the higher the user feature vector is;
step 820, generating a user characteristic-vehicle type correlation matrix according to the user characteristic vector and the candidate vehicle type set, wherein the correlation matrix records the corresponding relation between the characteristics and the vehicle type-weight, the characteristics can be age, gender, favorite color, purchasing power and the like, and the candidate vehicle type exists for the purpose of ensuring that the recommendation result only contains vehicle types in the candidate vehicle type set, such as only hope of providing some competitive vehicle types for recommending to the user or new vehicle types listed in the last three months;
step 825, forming an initial vehicle type recommendation list;
step 830, filtering the initial recommendation list, such as filtering out some bidding vehicle models that have been recommended previously or excluding some domestic vehicles with low prices and low quality;
step 835, generating a vehicle type ranking list of interest of the user based on an offline ranking algorithm, for example, the user can be analyzed for previous clicking behavior in a pull-down menu recommendation area, and the user can be predicted to be interested in which recommendation results, which is a ranking processing method integrating user feedback;
step 840, introducing interpretable strategy rules to strengthen sequencing results, and combining various sequencing strategies such as vehicle type novelty, vehicle type sales volume, user public praise evaluation, time variation diversity and the like in principle;
at step 845, a final recommendation is formed.
FIG. 9 is a detailed flowchart of a method for accurately recommending vehicle types in a search box according to the present invention. When a user puts a focus (such as a cursor) in a search box but does not input information, a server firstly queries an online cache cluster, matching is carried out through compressing a Trie tree, and if a matching result exists, the matching result is directly displayed in a drop-down list box. If the online cache cluster is not queried, the online database can be called to continue matching, and if the matching result exists, the matching result is displayed in a drop-down list box. And if the matching result does not exist in the online database, matching is carried out in the user center, and if the matching result exists, the matching result is displayed in a drop-down list box. If the user center also does not have a match, the hot search type is presented in the drop-down list box. When the user inputs information in the search box, the contained vehicle type can be identified according to the input information, and the browser or the APP sends the vehicle type information to the server. The server calls the recommendation/advertisement list corresponding to the user, the real-time recommendation algorithm is utilized to jointly input vehicle types, long-term interest pictures of the user and content in a vehicle type competition list to be mixed, and the recommendation list aiming at the user is directly returned and displayed in a drop-down list box. Specifically, the logs are analyzed in a streaming manner according to the short-term behaviors and interest characteristics of the users through Kafka + storm real-time analysis.
In the returned search results, if the user hovers over the mouse in the drop-down list box and the corresponding option is an advertisement item, different types such as video advertisements, extensible advertisements, animation advertisements and the like can be triggered according to the advertisement type. If the user clicks the corresponding option in the pull-down list box to be the recommended item, the user jumps to the URL corresponding to the vehicle type page, and if the user clicks the corresponding option in the pull-down list box to be the advertisement item, the user jumps to the advertisement landing page.
As shown in fig. 9, the method for accurately recommending the vehicle type in the search box specifically includes:
step 905, obtaining an input focus by a search box;
step 910, using the search box input content as an index to query the online cache;
step 915, determining whether a prefix matching result exists, if yes, executing step 930, otherwise, executing step 920;
step 920, determining whether there is a matching result in the online database, if yes, executing step 930, otherwise, executing step 925;
step 925, determining whether the user center has a matching result, if yes, executing step 930, otherwise, executing step 935;
step 930, displaying the recommendation result in a drop-down box;
step 935, displaying the hot vehicle searching type in a drop-down frame;
step 940, a user inputs information in the search box, and the background carries out query rewriting;
step 945, recognizing the vehicle type in the input information, wherein the vehicle type in the input information can be realized by a front end, such as a browser or an APP, or a rear end, such as a server;
step 950, calling a user vehicle type interest portrait from a user center in an off-line manner;
step 955, obtaining the competitive product vehicle type corresponding to the identified vehicle type, wherein the calculation scheme of the competitive product vehicle type has a plurality of types such as attention coincidence, clue coincidence, public opinion volume coincidence and the like;
step 960, calculate search results in real time. The real-time calculation engine requests the offline user center to call a recommendation/advertisement list corresponding to the user, and content in a vehicle type input by the user, an interest portrait of the user offline vehicle type and a vehicle type competition list is mixed and arranged by utilizing a real-time recommendation algorithm.
Through the process, the recommendation result can be displayed to the user more accurately according to the interest portrait of the user.
The present application provides a device for recommending vehicle types in a search box, as shown in fig. 10, including:
the first recommendation module is used for inquiring an online cache under the condition that a user places a focus in a search box but does not input information, and displaying an inquired first user vehicle type interest portrait in a drop-down list of the search box to realize vehicle type recommendation, wherein the online cache contains the first user vehicle type interest portrait; and/or
The second recommending module is used for generating a second user vehicle type interest portrait according to the information input by the user under the condition that the user inputs the information in the search box, and displaying the generated second user vehicle type interest portrait in a pull-down list of the search box to realize vehicle type recommendation
Preferably, the apparatus further comprises an advertisement module, configured to, in a case that an option hovered in the search box drop-down list by the user is an advertisement item, present an advertisement;
preferably, the device further comprises a skipping module, which is used for skipping to a webpage corresponding to the recommended item when the option clicked in the search box drop-down list by the user is the recommended item; and skipping to the advertisement landing page under the condition that the option clicked by the user in the search box pull-down list is the advertisement.
Preferably, the apparatus further comprises a user interest representation generation module for generating a user interest representation from the user log.
Preferably, the first recommending module is further configured to perform drop-down list item exposure and click behaviors in the search box according to the received user; pulling a user interest portrait to generate a feature vector of a user; generating a first user vehicle type interest portrait according to the feature vector of the user and the feature-article correlation matrix; and storing the interest portrait of the first user vehicle type into an online cache.
Generating a first user vehicle type interest portrait according to the feature vector of the user and the feature-item correlation matrix comprises the following steps: acquiring an initial recommended article list according to the feature vector of the user and the feature-article correlation matrix; obtaining a vehicle type and a competitive product vehicle type which accord with a correlation threshold value from an initial recommended article list according to a long-term interest portrait of a user; and sorting the vehicle types meeting the correlation threshold and the vehicle types of the competitive products to generate the interest portrait of the first user vehicle type.
Preferably, the first recommending module is further configured to query the first user vehicle type interest representation in the online database under the condition that the first user vehicle type interest representation is not queried in the online cache, and if the first user vehicle type interest representation exists, display the queried first user vehicle type interest representation in a drop-down list of the search box; under the condition that the first user vehicle type interest portrait is not inquired in the online database, inquiring the first user vehicle type interest portrait in an offline user center, and displaying the inquired first user vehicle type interest portrait in a pull-down list of a search box; and under the condition that the first user vehicle type interest portrait is not inquired by the offline user center, displaying the hot search vehicle type in a pull-down list of the search box.
Preferably, the second recommending module is further configured to identify a vehicle type input by the user according to the information input by the user; and acquiring a recommendation list and/or an advertisement list corresponding to the user, and generating a second user vehicle type interest portrait by using the vehicle type input by the user, the user offline vehicle type interest portrait and the content mixed arrangement in the vehicle type competition list.
The preferred embodiments of the present invention have been described in detail with reference to the accompanying drawings, however, the present invention is not limited to the specific details of the above embodiments, and various simple modifications can be made to the technical solution of the present invention within the technical idea of the present invention, and these simple modifications are within the protective scope of the present invention.
It should be noted that the various features described in the above embodiments may be combined in any suitable manner without departing from the scope of the invention. The invention is not described in detail in order to avoid unnecessary repetition.
In addition, any combination of the various embodiments of the present invention is also possible, and the same should be considered as the disclosure of the present invention as long as it does not depart from the spirit of the present invention.

Claims (13)

1. A method for vehicle type recommendation in a search box is characterized by comprising the following steps:
inquiring an online cache by a user under the condition that the user does not input information but places a focus in a search box, and displaying an inquired first user vehicle type interest portrait in a drop-down list of the search box to realize vehicle type recommendation, wherein the online cache comprises the first user vehicle type interest portrait;
generating a user interest portrait according to a user log, and storing the first user vehicle type interest portrait in an online cache according to the following steps:
receiving the exposure and click behaviors of a drop-down list item in a search box of a user;
pulling a user interest portrait to generate a feature vector of a user;
generating a first user vehicle type interest portrait according to the feature vector of the user and the feature-article correlation matrix;
and storing the interest portrait of the first user vehicle type into an online cache.
2. The method of claim 1, further comprising:
displaying the advertisement under the condition that the option hovered in the search box drop-down list by the user is an advertisement item;
when the option clicked by the user in the search box drop-down list is the recommended item, jumping to the webpage corresponding to the recommended item; and skipping to the advertisement landing page under the condition that the option clicked by the user in the search box pull-down list is the advertisement.
3. The method of claim 1, wherein generating the first user vehicle type interest representation according to the feature vector of the user and the feature-item correlation matrix comprises:
acquiring an initial recommended article list according to the feature vector of the user and the feature-article correlation matrix;
obtaining a vehicle type and a competitive product vehicle type which accord with a correlation threshold value from an initial recommended article list according to the user interest portrait;
and sorting the vehicle types meeting the correlation threshold and the vehicle types of the competitive products to generate the interest portrait of the first user vehicle type.
4. The method of claim 1, further comprising:
under the condition that the first user vehicle type interest portrait is not inquired in the online cache, inquiring the first user vehicle type interest portrait in the online database, and if the first user vehicle type interest portrait exists, displaying the inquired first user vehicle type interest portrait in a pull-down list of a search box;
under the condition that the first user vehicle type interest portrait is not inquired in the online database, inquiring the first user vehicle type interest portrait in an offline user center, and displaying the inquired first user vehicle type interest portrait in a pull-down list of a search box;
and under the condition that the first user vehicle type interest portrait is not inquired by the offline user center, displaying the hot search vehicle type in a pull-down list of the search box.
5. The method of claim 3, wherein the vehicle type meeting the relevancy threshold and the racing vehicle type are vehicle types in a candidate vehicle type set.
6. The method of any of claims 1-5, wherein the first user vehicle type interest representation is a vehicle purchase stage representation of a purchased vehicle user for providing same brand vehicle types to the purchased vehicle user.
7. An apparatus for in-search-box vehicle type recommendation, comprising:
the first recommendation module is used for inquiring an online cache under the condition that a user places a focus in a search box but does not input information, and displaying an inquired first user vehicle type interest portrait in a drop-down list of the search box to realize vehicle type recommendation, wherein the online cache contains the first user vehicle type interest portrait;
the device also comprises a user interest portrait generation module, a user interest portrait generation module and a user portrait generation module, wherein the user interest portrait generation module is used for generating a user interest portrait according to a user log; the first recommending module is also used for receiving the exposure and click behaviors of the drop-down list items in the search box by the user; pulling a user interest portrait to generate a feature vector of a user; generating a first user vehicle type interest portrait according to the feature vector of the user and the feature-article correlation matrix; and storing the interest portrait of the first user vehicle type into an online cache.
8. The apparatus of claim 7, further comprising an advertisement module for presenting an advertisement if the option hovered in the search box drop-down list by the user is an advertisement item.
9. The device of claim 7, further comprising a jump module, configured to jump to a web page corresponding to a recommended item if an option clicked in the search box drop-down list by the user is the recommended item; and skipping to the advertisement landing page under the condition that the option clicked by the user in the search box pull-down list is the advertisement.
10. The apparatus of claim 7, wherein the first recommending module is further configured to obtain an initial recommended item list according to the feature vector of the user and the feature-item correlation matrix; obtaining a vehicle type and a competitive product vehicle type which accord with a correlation threshold value from an initial recommended article list according to a long-term interest portrait of a user; and sorting the vehicle types meeting the correlation threshold and the vehicle types of the competitive products to generate the interest portrait of the first user vehicle type.
11. The apparatus of claim 7, wherein the first recommending module is further configured to query the first user vehicle type interest representation in the online database if the first user vehicle type interest representation is not queried in the online cache, and if the first user vehicle type interest representation exists, display the queried first user vehicle type interest representation in a drop-down list of the search box; under the condition that the first user vehicle type interest portrait is not inquired in the online database, inquiring the first user vehicle type interest portrait in an offline user center, and displaying the inquired first user vehicle type interest portrait in a pull-down list of a search box; and under the condition that the first user vehicle type interest portrait is not inquired by the offline user center, displaying the hot search vehicle type in a pull-down list of the search box.
12. A system for accurately recommending vehicle types in a search box is characterized by comprising a memory and a processor;
a memory for storing a program;
a processor for executing a program to implement the method according to any of claims 1-6.
13. Memory having a program stored thereon, characterized in that the program is adapted to carry out the method according to any of claims 1-6 when executed by a processor.
CN201810338063.6A 2018-04-16 2018-04-16 Method, device and system for recommending vehicle type in search box and memory Active CN108537596B (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201810338063.6A CN108537596B (en) 2018-04-16 2018-04-16 Method, device and system for recommending vehicle type in search box and memory
CN201910730605.9A CN110517067A (en) 2018-04-16 2018-04-16 Method, apparatus, system and the memory that vehicle is recommended in search box

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810338063.6A CN108537596B (en) 2018-04-16 2018-04-16 Method, device and system for recommending vehicle type in search box and memory

Related Child Applications (1)

Application Number Title Priority Date Filing Date
CN201910730605.9A Division CN110517067A (en) 2018-04-16 2018-04-16 Method, apparatus, system and the memory that vehicle is recommended in search box

Publications (2)

Publication Number Publication Date
CN108537596A CN108537596A (en) 2018-09-14
CN108537596B true CN108537596B (en) 2020-02-21

Family

ID=63480496

Family Applications (2)

Application Number Title Priority Date Filing Date
CN201810338063.6A Active CN108537596B (en) 2018-04-16 2018-04-16 Method, device and system for recommending vehicle type in search box and memory
CN201910730605.9A Pending CN110517067A (en) 2018-04-16 2018-04-16 Method, apparatus, system and the memory that vehicle is recommended in search box

Family Applications After (1)

Application Number Title Priority Date Filing Date
CN201910730605.9A Pending CN110517067A (en) 2018-04-16 2018-04-16 Method, apparatus, system and the memory that vehicle is recommended in search box

Country Status (1)

Country Link
CN (2) CN108537596B (en)

Families Citing this family (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110263066B (en) * 2019-06-13 2022-11-29 优信拍（北京）信息科技有限公司 Vehicle screening method and device
CN110807049A (en) * 2019-10-23 2020-02-18 上海能塔智能科技有限公司 Test driving vehicle type recommendation method and device and computer readable storage medium
CN110990692A (en) * 2019-11-13 2020-04-10 上海易点时空网络有限公司 Data processing method and device based on portrait analysis
CN111861666A (en) * 2020-07-21 2020-10-30 上海仙豆智能机器人有限公司 Vehicle information interaction method and device
CN112199465B (en) * 2020-09-01 2024-02-06 北京达佳互联信息技术有限公司 Information searching method and device and electronic equipment
CN112148984B (en) * 2020-09-30 2023-11-10 微梦创科网络科技（中国）有限公司 Method and system for capturing instant interests of user
CN117349508A (en) * 2022-06-30 2024-01-05 华为技术有限公司 Searching method, searching system and related device

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105760508A (en) * 2016-02-23 2016-07-13 北京搜狗科技发展有限公司 Information push method and device and electronic equipment
CN107506367A (en) * 2017-07-03 2017-12-22 阿里巴巴集团控股有限公司 It is determined that the method, apparatus and server of application displaying content

Family Cites Families (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20130173398A1 (en) * 2011-12-29 2013-07-04 Microsoft Corporation Search Engine Menu-based Advertising
CN102663627A (en) * 2012-04-26 2012-09-12 焦点科技股份有限公司 Personalized recommendation method
CN103631803A (en) * 2012-08-23 2014-03-12 百度国际科技（深圳）有限公司 Method, device and server for advertisement orientation based on input behaviors
JP6246657B2 (en) * 2013-05-20 2017-12-13 パナソニック インテレクチュアル プロパティ コーポレーション オブ アメリカＰａｎａｓｏｎｉｃ Ｉｎｔｅｌｌｅｃｔｕａｌ Ｐｒｏｐｅｒｔｙ Ｃｏｒｐｏｒａｔｉｏｎ ｏｆ Ａｍｅｒｉｃａ Information providing method and information providing apparatus
US20150242474A1 (en) * 2014-02-27 2015-08-27 Microsoft Corporation Inline and context aware query box
CN103942279B (en) * 2014-04-01 2018-07-10 百度(中国)有限公司 Search result shows method and apparatus
CN105869001A (en) * 2015-01-19 2016-08-17 苏宁云商集团股份有限公司 Customized product recommendation guiding method and system
CN104899315A (en) * 2015-06-17 2015-09-09 百度在线网络技术（北京）有限公司 Method and device for pushing user information
CN105046531A (en) * 2015-08-07 2015-11-11 百度在线网络技术（北京）有限公司 Online delivery information display method and apparatus
CN106296368A (en) * 2016-08-19 2017-01-04 北京好车轰轰电子商务有限公司 A kind of vehicle commending system and method
CN107818103B (en) * 2016-09-12 2023-01-24 阿里巴巴集团控股有限公司 Method and device for providing data object information
CN107545457B (en) * 2017-02-13 2020-01-21 北京车慧科技有限公司 Automobile racing product type determination method and device
CN107038213B (en) * 2017-02-28 2021-06-15 华为技术有限公司 Video recommendation method and device
CN107180078A (en) * 2017-04-21 2017-09-19 河海大学 A kind of method for vertical search based on user profile learning
CN107526828B (en) * 2017-09-01 2020-10-16 阿里巴巴（中国）有限公司 Page information recommendation method and device

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105760508A (en) * 2016-02-23 2016-07-13 北京搜狗科技发展有限公司 Information push method and device and electronic equipment
CN107506367A (en) * 2017-07-03 2017-12-22 阿里巴巴集团控股有限公司 It is determined that the method, apparatus and server of application displaying content

Also Published As

Publication number Publication date
CN108537596A (en) 2018-09-14
CN110517067A (en) 2019-11-29

Similar Documents

Publication Publication Date Title
CN108537596B (en) Method, device and system for recommending vehicle type in search box and memory
US10904117B1 (en) Insights for web service providers
JP6262886B2 (en) Automated click type selection for content performance optimization
US8346791B1 (en) Search augmentation
US10180979B2 (en) System and method for generating suggestions by a search engine in response to search queries
US8738426B2 (en) Taxonomy based targeted search advertising
TWI432980B (en) Dynamic bid pricing for sponsored search
US10198520B2 (en) Search with more like this refinements
US8209331B1 (en) Context sensitive ranking
JP5721818B2 (en) Use of model information group in search
US9141713B1 (en) System and method for associating keywords with a web page
US20090271228A1 (en) Construction of predictive user profiles for advertising
KR101840426B1 (en) User list identification
WO2018040069A1 (en) Information recommendation system and method
CN105718184A (en) Data processing method and apparatus
CN108614832B (en) Method and device for realizing user personalized commodity search
CN104679771A (en) Individual data searching method and device
JP2009531782A (en) Behavioral targeting system
EP2586008A1 (en) Infinite browse
US10853424B1 (en) Content delivery using persona segments for multiple users
WO2013060015A1 (en) Advertisement determination system and method for clustered search results
US11100169B2 (en) Alternative query suggestion in electronic searching
CN111429203A (en) Commodity recommendation method and device based on user behavior data
WO2013023358A1 (en) Using image match technology to improve image advertisement quality
US20100318427A1 (en) Enhancing database management by search, personal search, advertising, and databases analysis efficiently using core-set implementations

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
CB02 Change of applicant information

Address after: 100044 Tengda Building, 168 Xizhimenwai Street, Haidian District, Beijing, 2101-2103 and 2105-2111 on the 21st floor

Applicant after: Beijing Chehui Technology Co., Ltd.

Address before: 100044 Tengda Building, 168 Xizhimenwai Street, Haidian District, Beijing, 2101-2103 and 2105-2111 on the 21st floor

Applicant before: Beijing Che Hui Interactive Advertising Co., Ltd.

CB02 Change of applicant information
GR01 Patent grant
GR01 Patent grant