CN113868522A - Object recommendation method and device, electronic equipment and storage medium - Google Patents

Object recommendation method and device, electronic equipment and storage medium Download PDF

Info

Publication number
CN113868522A
CN113868522A CN202111127592.XA CN202111127592A CN113868522A CN 113868522 A CN113868522 A CN 113868522A CN 202111127592 A CN202111127592 A CN 202111127592A CN 113868522 A CN113868522 A CN 113868522A
Authority
CN
China
Prior art keywords
information
recommendation
objects
index
resource consumption
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.)
Pending
Application number
CN202111127592.XA
Other languages
Chinese (zh)
Inventor
谢淼
郭远
陈丛铮
张弛
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 Dajia Internet Information Technology Co Ltd
Original Assignee
Beijing Dajia Internet Information 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 Dajia Internet Information Technology Co Ltd filed Critical Beijing Dajia Internet Information Technology Co Ltd
Priority to CN202111127592.XA priority Critical patent/CN113868522A/en
Publication of CN113868522A publication Critical patent/CN113868522A/en
Pending legal-status Critical Current

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
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • 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
    • 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
    • G06Q30/0273—Determination of fees for advertising

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Physics & Mathematics (AREA)
  • Development Economics (AREA)
  • Strategic Management (AREA)
  • Finance (AREA)
  • General Physics & Mathematics (AREA)
  • Accounting & Taxation (AREA)
  • Databases & Information Systems (AREA)
  • General Business, Economics & Management (AREA)
  • Marketing (AREA)
  • Data Mining & Analysis (AREA)
  • Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • General Engineering & Computer Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Molecular Biology (AREA)
  • Biomedical Technology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Mathematical Physics (AREA)
  • Computing Systems (AREA)
  • Artificial Intelligence (AREA)
  • General Health & Medical Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The disclosure relates to an object recommendation method, an object recommendation device, an electronic device and a storage medium. The method comprises the following steps: acquiring a plurality of objects to be recommended and respective initial recommendation parameter information of the objects; determining index information of the plurality of objects under a preset recommendation index; determining recommended parameter adjustment information corresponding to the plurality of objects according to the index information; recommending the plurality of objects to a target user based on the initial recommendation parameter information and the recommendation parameter adjustment information. According to the technical scheme provided by the disclosure, the object recommendation effect and the object cold start passing rate can be improved.

Description

Object recommendation method and device, electronic equipment and storage medium
Technical Field
The present disclosure relates to the field of internet application technologies, and in particular, to an object recommendation method and apparatus, an electronic device, and a storage medium.
Background
Recommendation services in internet applications are receiving increasing attention, such as recommendation services for advertisements. In the related technology, generally, the recommendation priority of an object is set according to the bid of an advertisement or historical interaction information, the recommendation mode is relatively fixed, the dependence degree on the interaction information is large, most objects in internet application are short in life cycle and fast to update, some newly created objects lack interaction information, and the bid difference of the objects under different categories is large, so that the object recommendation effect based on the existing recommendation mode is poor, and the scene applicability is poor.
Disclosure of Invention
The present disclosure provides an object recommendation method, an object recommendation apparatus, an electronic device, and a storage medium, so as to at least solve a problem of how to improve an object recommendation effect in related technologies. The technical scheme of the disclosure is as follows:
according to a first aspect of the embodiments of the present disclosure, there is provided an object recommendation method, including:
acquiring a plurality of objects to be recommended and respective initial recommendation parameter information of the objects;
determining index information of the plurality of objects under a preset recommendation index;
determining recommended parameter adjustment information corresponding to the plurality of objects according to the index information;
recommending the plurality of objects to a target user based on the initial recommendation parameter information and the recommendation parameter adjustment information.
In a possible implementation manner, the preset recommendation index includes a conversion prediction index of a preset conversion rate, a resource consumption index associated with the preset conversion rate, an importance index of a user to an object at the preset conversion rate, and a compensation index of the preset conversion rate; the step of determining the index information of each of the plurality of objects under the preset recommendation index includes:
acquiring user association characteristics of the target user and object characteristics of the plurality of objects;
determining first recommendation index information of each object under the conversion prediction index according to the user association characteristics and the object characteristics;
acquiring resource consumption statistical information under object categories corresponding to the objects, and taking the resource consumption statistical information as second recommendation index information of the objects under the resource consumption indexes;
determining third recommendation index information of the target user under the importance degree indexes for the plurality of objects according to the first recommendation index information and recommendation index distribution information associated with the plurality of objects;
determining accumulated resource consumption information associated with the preset conversion rates corresponding to the plurality of objects respectively;
and determining fourth recommendation index information of the plurality of objects under the compensation index according to the accumulated resource consumption information.
In a possible implementation manner, the step of determining, according to the user-associated feature and the object feature, first recommendation indicator information of each of the plurality of objects under the conversion prediction indicator includes:
inputting the user correlation characteristics and the object characteristics into a click rate prediction model and a conversion rate prediction model respectively to obtain click rate prediction information and conversion rate prediction information;
and multiplying the click rate prediction information and the conversion rate prediction information to obtain the first recommendation index information.
In a possible implementation manner, the step of determining, according to the first recommendation index information and recommendation index distribution information associated with the plurality of objects, third recommendation index information of the target user under the importance degree index for the plurality of objects includes:
determining target distribution information of the first recommendation index information in the recommendation index distribution information;
determining the number of objects corresponding to the target distribution information and the total number of objects included in the recommendation index distribution information;
and taking the proportion of the number of the objects in the total number of the objects as the third recommendation index information.
In a possible implementation manner, the step of determining the cumulative resource consumption information associated with the preset conversion rate corresponding to each of the plurality of objects includes:
acquiring historical accumulated resource consumption information of each object under the preset conversion rate;
predicting the predicted accumulated resource consumption information after each object is recommended for a preset number of times;
and determining the accumulated resource consumption information according to the historical accumulated resource consumption information and the predicted accumulated resource consumption information.
In a possible implementation manner, the step of determining, according to the accumulated resource consumption information, fourth recommendation index information of each of the plurality of objects under the compensation index includes:
and mapping the accumulated resource consumption information to a preset value range, and taking a target value obtained by mapping as the fourth recommendation index information.
In a possible implementation manner, the determining, according to the index information, recommended parameter adjustment information corresponding to each of the plurality of objects includes:
and multiplying the first recommendation index information, the second recommendation index information, the third recommendation index information and the fourth recommendation index information of the plurality of objects to obtain recommendation parameter adjustment information corresponding to each of the plurality of objects.
In a possible implementation manner, the recommending the plurality of objects to the target user corresponding to the user identifier based on the initial recommendation parameter information and the recommendation parameter adjustment information includes:
sequencing the plurality of objects based on the initial recommendation parameter information and the recommendation parameter adjustment information to obtain a sequencing result;
screening out a preset number of objects from the plurality of objects as target objects according to the sorting result;
and recommending the target object to the target user.
According to a second aspect of the embodiments of the present disclosure, there is provided an object recommendation apparatus including:
the system comprises an initial recommendation parameter information acquisition module, a recommendation processing module and a recommendation processing module, wherein the initial recommendation parameter information acquisition module is configured to acquire a plurality of objects to be recommended and respective initial recommendation parameter information of the objects;
an index information determination module configured to perform determining index information of each of the plurality of objects under a preset recommendation index, the preset recommendation index representing expected conversion rate information of the plurality of objects;
a recommended parameter adjustment information determination module configured to perform determining recommended parameter adjustment information corresponding to each of the plurality of objects according to the index information;
a recommending module configured to recommend the plurality of objects to a target user corresponding to the user identifier based on the initial recommendation parameter information and the recommendation parameter adjustment information.
In a possible implementation manner, the preset recommendation index includes a conversion prediction index of a preset conversion rate, a resource consumption index associated with the preset conversion rate, an importance index of a user to an object at the preset conversion rate, and a compensation index of the preset conversion rate; the index information determination module includes:
a user associated feature and object feature acquisition unit configured to perform acquisition of a user associated feature of the target user and object features of the plurality of objects;
a first recommendation index information determination unit configured to perform determining first recommendation index information of each of the plurality of objects under the conversion prediction index according to the user-associated feature and the object feature;
a second recommendation index information acquisition unit configured to perform acquisition of resource consumption statistical information under object categories to which the plurality of objects correspond, as second recommendation index information under the resource consumption index for each of the plurality of objects;
a third recommendation index information determination unit configured to perform determination of third recommendation index information of the target user under the importance degree index for the plurality of objects according to the first recommendation index information and recommendation index distribution information associated with the plurality of objects;
an accumulated resource consumption information determination unit configured to perform determination of accumulated resource consumption information associated with the preset conversion rate corresponding to each of the plurality of objects;
a fourth recommendation index information determination unit configured to perform determining fourth recommendation index information of each of the plurality of objects under the compensation index according to the accumulated resource consumption information.
In one possible implementation manner, the first recommendation index information determining unit includes:
the prediction subunit is configured to input the user-associated feature and the object feature into a click rate prediction model and a conversion rate prediction model respectively to obtain click rate prediction information and conversion rate prediction information;
a first recommendation index information determination subunit configured to perform multiplication of the click rate prediction information and the conversion rate prediction information to obtain the first recommendation index information.
In one possible implementation manner, the third recommendation index information determining unit includes:
a target distribution information determination subunit configured to perform determination of target distribution information of the first recommendation index information in the recommendation index distribution information;
a quantity determining subunit configured to perform determination of a quantity of objects corresponding to the target distribution information and a total quantity of objects included in the recommendation index distribution information;
a third recommendation index information determining subunit configured to perform, as the third recommendation index information, a ratio of the number of objects in the total number of objects.
In one possible implementation manner, the cumulative resource consumption information determining unit includes:
a historical accumulated resource consumption obtaining subunit configured to perform obtaining of historical accumulated resource consumption information of each object at the preset conversion rate;
a predicted cumulative resource consumption acquisition subunit configured to execute prediction of predicted cumulative resource consumption information after each object is recommended a preset number of times;
an accumulated resource consumption information determination subunit configured to perform determining the accumulated resource consumption information from the historical accumulated resource consumption information and the predicted accumulated resource consumption information.
In one possible implementation manner, the fourth recommendation index information determining unit includes:
and the fourth recommendation index information determining subunit is configured to perform mapping of the accumulated resource consumption information to a preset value range, and use a mapped target value as the fourth recommendation index information.
In one possible implementation manner, the recommended parameter adjustment information determining module includes:
a recommendation parameter adjustment information determination unit configured to perform multiplication processing on the first recommendation index information, the second recommendation index information, the third recommendation index information, and the fourth recommendation index information of the plurality of objects to obtain recommendation parameter adjustment information corresponding to each of the plurality of objects.
In one possible implementation, the recommendation module includes:
a sorting unit configured to perform sorting of the plurality of objects based on the initial recommendation parameter information and the recommendation parameter adjustment information, resulting in a sorting result;
a target object screening unit configured to perform screening of a preset number of objects from the plurality of objects as target objects according to the sorting result;
a recommending unit configured to perform recommending the target object to the target user.
According to a third aspect of the embodiments of the present disclosure, there is provided an electronic apparatus including: a processor; a memory for storing the processor-executable instructions; wherein the processor is configured to execute the instructions to implement the method of any of the first aspects above.
According to a fourth aspect of embodiments of the present disclosure, there is provided a computer-readable storage medium, wherein instructions, when executed by a processor of an electronic device, enable the electronic device to perform the method of any one of the first aspects of the embodiments of the present disclosure.
According to a fifth aspect of embodiments of the present disclosure, there is provided a computer program product comprising computer instructions which, when executed by a processor, cause a computer to perform the method of any one of the first aspects of the embodiments of the present disclosure.
The technical scheme provided by the embodiment of the disclosure at least brings the following beneficial effects:
the recommendation parameter adjustment information is determined through the index information of the object under the preset recommendation index, object recommendation is performed by combining the initial recommendation parameter information, namely the initial recommendation parameter information is adjusted through the initial recommendation parameter information, the object with higher index information can be compensated in the recommendation priority sequence, the recommendation priority of the object with high recommendation potential can be effectively improved, the object in a cold start state can obtain a higher recommendation priority sequence under the condition that the object in a non-cold start state is recommended simultaneously, the object recommendation effect can be improved, the cold start passing rate of the object can be improved, the object with high conversion potential can be accelerated through a cold start stage, and therefore the method can be effectively suitable for the recommendation scene with higher object timeliness requirements.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the disclosure and are not to be construed as limiting the disclosure.
FIG. 1 is a schematic diagram illustrating an application environment in accordance with an exemplary embodiment.
FIG. 2 is a flow diagram illustrating a method of object recommendation, according to an example embodiment.
Fig. 3 is a flowchart illustrating a method for recommending a plurality of objects to a target user based on initial recommendation parameter information and recommendation parameter adjustment information according to an example embodiment.
Fig. 4 is a flowchart illustrating a method for obtaining index information of each of a plurality of objects under a preset recommendation index according to an exemplary embodiment.
FIG. 5 is a flowchart illustrating a method for determining third recommendation index information of a target user under an importance index for a plurality of objects according to recommendation index distribution information associated with a plurality of objects and first recommendation index information, according to an exemplary embodiment.
FIG. 6 is a diagram illustrating a recommendation indicator profile, according to an example embodiment.
FIG. 7 is a block diagram illustrating an object recommendation device according to an example embodiment.
FIG. 8 is a block diagram illustrating an electronic device for object recommendation, according to an example embodiment.
Detailed Description
In order to make the technical solutions of the present disclosure better understood by those of ordinary skill in the art, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below with reference to the accompanying drawings.
It should be noted that the terms "first," "second," and the like in the description and claims of the present disclosure and in the above-described drawings are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the disclosure described herein are capable of operation in sequences other than those illustrated or otherwise described herein. The implementations described in the exemplary embodiments below are not intended to represent all implementations consistent with the present disclosure. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present disclosure, as detailed in the appended claims.
Artificial Intelligence (AI) is a theory, method, technique and application system that uses a digital computer or a machine controlled by a digital computer to simulate, extend and expand human Intelligence, perceive the environment, acquire knowledge and use the knowledge to obtain the best results. The artificial intelligence software technology mainly comprises a computer vision technology, a voice processing technology, a natural language processing technology, machine learning/deep learning and the like.
In recent years, with research and development of artificial intelligence technology, the artificial intelligence technology is widely applied in a plurality of fields, and the scheme provided by the embodiment of the application relates to technologies such as machine learning/deep learning, and is specifically described by the following embodiments:
referring to fig. 1, fig. 1 is a schematic diagram illustrating an application environment according to an exemplary embodiment, which may include a server 01 and a terminal 02, as shown in fig. 1.
In an alternative embodiment, server 01 may be used for the object recommendation process. Specifically, the server 01 may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server providing basic cloud computing services such as a cloud service, a cloud database, cloud computing, a cloud function, cloud storage, a Network service, cloud communication, a middleware service, a domain name service, a security service, a CDN (Content Delivery Network), a big data and artificial intelligence platform, and the like.
In an alternative embodiment, terminal 02 may be used to send an object recommendation request and present the recommended object. Specifically, the terminal 02 may include, but is not limited to, a smart phone, a desktop computer, a tablet computer, a notebook computer, a smart speaker, a digital assistant, an Augmented Reality (AR)/Virtual Reality (VR) device, a smart wearable device, and other types of electronic devices. Optionally, the operating system running on the electronic device may include, but is not limited to, an android system, an IOS system, linux, windows, and the like.
In addition, it should be noted that fig. 1 illustrates only one application environment of the image processing method provided by the present disclosure.
In the embodiment of the present specification, the server 01 and the terminal 02 may be directly or indirectly connected by a wired or wireless communication method, and the present application is not limited herein.
It should be noted that the following figures show a possible sequence of steps, and in fact do not limit the order that must be followed. Some steps may be performed in parallel without being dependent on each other. User information (including but not limited to user device information, user personal information, user behavior information, etc.) and data (including but not limited to data for presentation, training, etc.) to which the present disclosure relates are both information and data that are authorized by the user or sufficiently authorized by various parties.
FIG. 2 is a flow diagram illustrating a method of object recommendation, according to an example embodiment. As shown in fig. 2, the following steps may be included.
In step S201, initial recommendation parameter information of each of a plurality of objects to be recommended and a plurality of objects is acquired.
In the embodiment of the present specification, the object may refer to an object that can be recommended to a user in an internet application platform, and may include, for example, an advertisement, multimedia, and the like.
In practical applications, objects published to a platform in a recent period of time may be acquired as a plurality of objects to be recommended to ensure timeliness of the objects, the plurality of objects may be taken as candidate recommendation objects, and the plurality of objects may include objects in a cold start state, or may include objects in a cold start state and objects in a non-cold start state. As an example, the initial recommendation parameter information may be ecpm (effective cost per minute), and in consideration of different object types of different objects, each object type may be converted into ecpm, so that the initial recommendation parameter information of each object is unified, thereby facilitating subsequent information processing. The object category may refer to a bid category selected by the object for delivering on the platform, for example, the object category may include ecpm (corresponding to cpm), cpc (Cost Per Click), ocpm (Optimized Cost Per mill, Optimized thousand show bids), and occc (Optimized Cost Per Click), wherein cpm (Cost Per mill) refers to thousand show bids.
In step S203, index information of each of the plurality of objects under a preset recommendation index is determined, wherein the preset recommendation index may represent expected click rate and/or expected conversion rate of the plurality of objects.
In practical application, it is considered that the initial recommendation parameter information is generally relatively fixed, so that the initial recommendation parameter cannot effectively represent the recommendation effect of the object or cannot effectively represent the recommendation potential of the object. Based on this, in the embodiment of the present specification, the preset recommendation index is selected and set for the processing of the index information corresponding to the object, so that each object may perform the adjustment of the recommendation priority based on the index information, that is, in the recommendation, not only the initial recommendation parameter information but also the index information of each object under the preset recommendation index is considered.
In one example, the recommendation priority order of the objects may be adjusted by the expected conversion rate, based on which at least one recommendation index capable of characterizing expected conversion rate information of the objects may be set as a preset recommendation index for measuring index information of each object. The expected conversion rate information may refer to prediction information of a preset conversion rate, and the preset conversion rate may include a click conversion rate, a purchase conversion rate, an exposure conversion rate, and the like.
In step S205, recommendation parameter adjustment information corresponding to each of the plurality of objects is determined based on the index information.
In practical application, when the preset recommendation index comprises a recommendation index, the index information can be used as recommendation parameter adjustment information; when the preset recommendation index includes a plurality of recommendation indexes, the sum of the index information and the product of the index information may be used as the recommendation parameter adjustment information. The present disclosure is not limited thereto.
In step S209, a plurality of objects are recommended to the target user based on the initial recommendation parameter information and the recommendation parameter adjustment information.
In this embodiment, the initial recommended parameter information may be adjusted by using the recommended parameter adjustment information. For example, preset operation may be performed on the recommended parameter adjustment information and the initial recommended parameter information, so as to adjust the initial recommended parameter information by the recommended parameter adjustment information. The specific adjustment mode is not limited in the disclosure, as long as the method can effectively help the object with higher recommendation potential to quickly pass through the cold start stage, or can improve the cold start passing rate of the object.
In one example, the sum of the initial recommendation parameter information and the recommendation parameter adjustment information may be used as a target recommendation parameter, so that a target object recommended to a target user may be determined from a plurality of objects according to the target recommendation parameter, and the target object may be issued to a terminal.
Optionally, the object recommendation process may be triggered by the terminal, for example, before the step S201, an object recommendation request sent by the terminal may be received, where the object recommendation request may include a user identifier. In one example, a user of the terminal may trigger an object recommendation request, for example, the object recommendation request may be triggered when the user enters an application, and the object recommendation request may carry a user identifier, so as to determine a recommended target user.
The recommendation parameter adjustment information is determined through the index information of the object under the preset recommendation index, object recommendation is performed by combining the initial recommendation parameter information, namely the initial recommendation parameter information is adjusted through the initial recommendation parameter information, the object with higher index information can be compensated in the recommendation priority sequence, the recommendation priority of the object with high recommendation potential can be effectively improved, the object in a cold start state can obtain a higher recommendation priority sequence under the condition that the object in a non-cold start state is recommended simultaneously, the object recommendation effect can be improved, the cold start passing rate of the object can be improved, the object with high conversion potential can be accelerated through a cold start stage, and therefore the method can be effectively suitable for the recommendation scene with higher object timeliness requirements.
Fig. 3 is a flowchart illustrating a method for recommending a plurality of objects to a target user corresponding to a user identifier based on initial recommendation parameter information and recommendation parameter adjustment information according to an exemplary embodiment. As shown in fig. 3, in a possible implementation manner, the step S209 may include:
in step S301, sorting the plurality of objects based on the initial recommended parameter information and the recommended parameter adjustment information to obtain a sorting result;
in step S303, a preset number of objects are screened out from the plurality of objects as target objects according to the sorting result;
in step S305, a target object is recommended to the target user.
In practical application, the initial recommendation parameter information and the recommendation parameter adjustment information can be quantized information, so that the target recommendation parameter can be determined based on the initial recommendation parameter information and the recommendation parameter adjustment information. For example, the sum of the initial recommendation parameter information and the recommendation parameter adjustment information may be used as the target recommendation parameter. Therefore, the plurality of objects can be sorted based on the target recommendation parameters to obtain a sorting result. And according to the sorting result, a preset number of objects can be screened out from the plurality of objects as target objects, for example, 20 objects can be screened out from the plurality of objects as target objects. Further, the 20 objects may be delivered to a terminal for presentation to a target user.
The plurality of objects are sorted based on the initial recommendation parameter information and the recommendation parameter adjustment information, and the target objects recommended to the target user are screened out according to the sorting result, so that the sorting consideration factors are comprehensive, and the sorting result is accurate; therefore, the recommendation effect of the target object screened out based on the sorting result is better.
In practical applications, objects in a platform may have different states or periods in the platform due to different times of creation or release to the platform. Taking the object as an example of the advertisement, the plurality of advertisements may include an advertisement in a mature period (an advertisement in a non-cold start state) and an advertisement in a cold start period (an advertisement in a cold start state), and the advertisement in the mature period may refer to an advertisement passing through the cold start period. The advertisement for the cold start period may refer to an advertisement created within a certain time period from the current time with a conversion number less than or equal to a conversion number threshold, for example, an advertisement created within 3 days with a conversion number less than or equal to 5. Due to the fact that interactive information of the advertisements in the cold start period is less in the platform, the click rate and the conversion rate are not predicted accurately, and further historical accumulated resource consumption (historical accumulated income) is not high; and the recommendation tendency of the advertisement in the cold start period in the platform recommendation strategy is low, so that the cold start period has no advantage in the priority of recommendation. Based on these circumstances, those advertisements that result in potentially high click-through rates and conversion rates among the advertisements in the cold start period are not recommended, so that the advertisement cold start passing rate is low.
Based on the above consideration, the preset recommendation index may be set to adjust the initial recommendation parameters of the plurality of objects. In a possible implementation manner, the preset recommendation index may include a conversion prediction index of a preset conversion rate, a resource consumption index associated with the preset conversion rate, an importance index of the user to the object at the preset conversion rate, and a compensation index of the preset conversion rate. Based on this, fig. 4 is a flowchart illustrating a method for obtaining index information of each of a plurality of objects under a preset recommendation index according to an exemplary embodiment. As shown in fig. 4, in a possible implementation manner, the step S205 may include:
in step S401, user associated features of a target user and object features of a plurality of objects are acquired.
In this embodiment of the present specification, the user association feature may include a user attribute feature and a context feature, and the context feature may include time information, location information, and the like of the target user triggering the object recommendation request, which is not limited in this disclosure. The object characteristics may include attribute characteristics of the object and interactive characteristics of the object, which may include the number of clicks, the number of forwards, and the like.
In step S403, first recommendation index information of each of the plurality of objects under the conversion prediction index is determined according to the user-associated feature and the object feature.
In one example, the first recommendation indicator information may be determined by a machine learning model. For example, the user-associated feature and the object feature may be input into the click rate prediction model and the conversion rate prediction model, respectively, to obtain click rate prediction information and conversion rate prediction information; and the click rate prediction information and the conversion rate prediction information can be multiplied to obtain first recommendation index information. The click rate prediction model and the conversion rate prediction model may be obtained by training a neural network in advance, and the disclosure does not limit this. The first recommendation index information is determined through the machine learning model, and the efficiency and the precision of the first recommendation index information can be improved.
In step S405, the resource consumption statistical information under the object category corresponding to each of the plurality of objects is acquired as the second recommendation index information under the resource consumption index of each of the plurality of objects.
In this embodiment, the object category may refer to a bid category selected by the object in platform delivery, such as ecpm, cpc, ocpm, etc. Taking the object a as an example, the object class corresponding to the object a can be determined to be ecpm, based on which, the resource consumption information of all objects under the ecpm in the platform can be counted, and the average value of the resource consumption information of all objects can be used as the resource consumption statistical information. For example, the advertisement under the ecpm may include advertisements of multiple article categories, such as a diamond advertisement, a book advertisement, a makeup advertisement, and the like, and an average of bids corresponding to the advertisements of the multiple article categories may be used as the second recommendation index information under the object category ecpm, that is, the second recommendation index information of the object a under the resource consumption index. Optionally, when the object type of the object is non-ecpm (cpc, ocpm, etc.), the resource consumption information of all the objects under the non-ecpm may be obtained, and an average value of the resource consumption information of all the objects may be used as the resource consumption statistical information. Further, the statistical information of resource consumption under non-ecpm may be converted into statistical information of resource consumption under ecpm, and the converted statistical information of resource consumption under ecpm may be used as second recommendation index information.
In step S407, third recommendation index information of the target user under the importance degree index for the plurality of objects is determined according to the first recommendation index information and recommendation index distribution information associated with the plurality of objects.
In one possible implementation manner, fig. 5 is a flowchart illustrating a method for determining third recommendation index information of a target user under an importance index for a plurality of objects according to recommendation index distribution information associated with a plurality of objects and first recommendation index information, according to an exemplary embodiment. As shown in fig. 5, this step S407 can be implemented by:
in step S501, target distribution information of the first recommendation index information in the recommendation index distribution information is determined;
in step S503, determining the number of objects corresponding to the target distribution information and the total number of objects included in the recommendation index distribution information;
in step S505, the ratio of the number of objects to the total number of objects is used as the third recommendation index information.
As shown in fig. 6, the recommendation index distribution information may refer to a recommendation index distribution graph, where the x-axis may be the click conversion rate and the y-axis may be the number of objects. In an example, taking an object a as an example, the range of the x axis may be [0,1], and if the first recommendation index information of the object a is 0.2, the y axis corresponding to 0.2 on the x axis may be found to be 100; if the total number of the objects is determined to be 1000 from the recommendation index distribution map, the ratio of 100 to 1000, which is 0.1, can be used as the third recommendation index information corresponding to the object a. The third recommendation index information may refer to the expected value, i.e., the degree of importance, of the target user for the object a in the request. Or it may be understood that object a may be better than 10% of all objects for the target user. It follows that the third recommendation indicator information may be positively correlated with the recommendation priority.
The third recommendation index information is obtained by setting the importance degree index of the user to the object under the preset conversion rate, so that the third recommendation index information can effectively represent the expected conversion rate of the object, and the recommendation effect can be improved.
In step S409, cumulative resource consumption information associated with the preset conversion rates corresponding to the respective plurality of objects is determined.
In practical application, the profit of each object can be predicted based on an online learning algorithm (e.g., an online bandit algorithm), that is, each object can be scored by using the online bandit algorithm, and the score is participated in the recommended parameter adjustment information as index information, so that the recommended parameter adjustment information can balance historical profit and exploration. Therefore, the objects which are in a cold start period, few in recommendation times and high in transformation potential can be effectively excavated. Specifically, this S409 may be implemented based on the following steps:
acquiring historical accumulated resource consumption information of each object under a preset conversion rate;
predicting the predicted accumulated resource consumption information after each object is recommended for a preset number of times;
and determining the accumulated resource consumption information according to the historical accumulated resource consumption information and the predicted accumulated resource consumption information.
The accumulated resource consumption information may refer to accumulated revenue, and by taking the click conversion rate as an example, the historical accumulated revenue of each object under the click conversion rate may be obtained; and the predicted accumulated profit after each object is recommended for a preset number of times can be predicted based on the online bandit algorithm, which is not limited by the disclosure. Further, the sum of the historical accumulated resource consumption information and the predicted accumulated resource consumption information may be used as the accumulated resource consumption information. By combining the predicted accumulated benefits, the object with high transformation potential can be mined as a recommended object, and the cold start passing rate of the object is accelerated.
In step S411, fourth recommendation index information of each of the plurality of objects under the compensation index is determined based on the accumulated resource consumption information.
In practical application, the step may be regarded as normalization processing of the accumulated resource consumption information, for example, the accumulated resource consumption information may be mapped to a preset value range, and a target value obtained by mapping may be used as the fourth recommendation index information. Wherein the predetermined numerical range may be [0,1 ]. By mapping the accumulated resource consumption information to a preset numerical range, the index information can be conveniently and uniformly processed.
In one example, the mapping described above may be performed by the following equation (1):
Figure BDA0003279158580000121
wherein Y is the upper limit of the preset numerical range, X is the lower limit of the preset numerical range, rmaxTo accumulate the maximum value in the resource consumption information, rminTo accumulate the minimum value in the resource consumption information, riCumulative resource consumption information for the ith object, TiAnd fourth recommendation index information of the ith object.
In another example, a correspondence relationship of the accumulated resource consumption information and a preset numerical range may be set in advance, and based on the correspondence relationship, fourth recommendation index information of each of the plurality of objects under the compensation index may be determined.
Optionally, based on the four pieces of recommendation index information, the step S205 may include: and multiplying the first recommendation index information, the second recommendation index information, the third recommendation index information and the fourth recommendation index information of each object, and taking the multiplication result as recommendation parameter adjustment information corresponding to each object. In practical application, the fourth recommendation index information is obtained based on an online learning algorithm, and when online learning of a certain object tends to converge or when the object is in a non-cold start state, the fourth recommendation index information may tend to 0 or equal to 0, and based on this, through the multiplication processing, the recommendation parameter adjustment information of the object in the non-cold start state and the object in the online learning convergence state may tend to 0 or equal to 0, so that recommendation compensation of the object in the non-cold start state and the object in the online learning convergence state may be reduced or cancelled. The recommendation parameter adjustment information can balance historical income and exploration, so that objects which are in a cold start period and have few recommendation times and high transformation potential are effectively mined, namely, under the compensation index, fourth recommendation index information corresponding to the objects which are in the cold start period and have few recommendation times and high transformation potential is higher.
A plurality of recommendation indexes which are rich in setting and can represent the expected conversion rate of the object are used for determining index information, so that the object with high conversion potential can be fully mined by participating in the adjustment of initial recommendation parameter information, the recommendation effect can be improved, the cold start passing rate of the object can be improved, and the cold start period of the object can be accelerated to pass.
FIG. 7 is a block diagram illustrating an object recommendation device according to an example embodiment. Referring to fig. 7, the apparatus may include:
an initial recommendation parameter information obtaining module 701 configured to perform obtaining initial recommendation parameter information of each of a plurality of objects to be recommended;
an index information determination module 703 configured to perform determining index information of each of the plurality of objects under a preset recommendation index;
a recommended parameter adjustment information determining module 705 configured to determine recommended parameter adjustment information corresponding to each of the plurality of objects according to the index information;
and a recommending module 707 configured to recommend a plurality of objects to the user identification corresponding target users based on the initial recommendation parameter information and the recommendation parameter adjustment information.
The recommendation parameter adjustment information is determined through the index information of the object under the preset recommendation index, object recommendation is performed by combining the initial recommendation parameter information, namely the initial recommendation parameter information is adjusted through the initial recommendation parameter information, the object with higher index information can be compensated in the recommendation priority sequence, the recommendation priority of the object with high recommendation potential can be effectively improved, the object in a cold start state can obtain a higher recommendation priority sequence under the condition that the object in a non-cold start state is recommended simultaneously, the object recommendation effect can be improved, the cold start passing rate of the object can be improved, the object with high conversion potential can be accelerated through a cold start stage, and therefore the method can be effectively suitable for the recommendation scene with higher object timeliness requirements.
In one possible implementation manner, the preset recommendation index includes a conversion prediction index of a preset conversion rate, a resource consumption index associated with the preset conversion rate, an importance index of a user to the object at the preset conversion rate, and a compensation index of the preset conversion rate; the index information determination module 705 may include:
a user associated feature and object feature acquisition unit configured to perform acquisition of a user associated feature of a target user and object features of a plurality of objects;
a first recommendation index information determination unit configured to perform determination of first recommendation index information of each of the plurality of objects under the conversion prediction index according to the user-associated feature and the object feature;
a second recommendation index information acquisition unit configured to perform acquisition of resource consumption statistical information under object categories corresponding to the respective plurality of objects as second recommendation index information under the resource consumption indexes of the respective plurality of objects;
a third recommendation index information determination unit configured to perform determination of third recommendation index information of the target user under the importance degree index for the plurality of objects according to the first recommendation index information and recommendation index distribution information associated with the plurality of objects;
an accumulated resource consumption information determination unit configured to perform determining accumulated resource consumption information associated with preset conversion rates corresponding to the respective plurality of objects;
and a fourth recommendation index information determination unit configured to perform determination of fourth recommendation index information of each of the plurality of objects under the compensation index according to the accumulated resource consumption information.
In one possible implementation manner, the first recommendation indicator information determining unit may include:
the prediction subunit is configured to input the user-associated feature and the object feature into a click rate prediction model and a conversion rate prediction model respectively to obtain click rate prediction information and conversion rate prediction information;
and the first recommendation index information determining subunit is configured to multiply the click rate prediction information and the conversion rate prediction information to obtain first recommendation index information.
In one possible implementation manner, the third recommendation index information determining unit may include:
a target distribution information determination subunit configured to perform determination of target distribution information of the first recommendation index information in the recommendation index distribution information;
the quantity determining subunit is configured to determine the quantity of the objects corresponding to the target distribution information and the total quantity of the objects included in the recommendation index distribution information;
a third recommendation index information determination subunit configured to perform, as third recommendation index information, a ratio of the number of objects in the total number of objects.
In one possible implementation manner, the accumulated resource consumption information determining unit may include:
a historical accumulated resource consumption obtaining subunit configured to perform obtaining of historical accumulated resource consumption information of each object at a preset conversion rate;
a predicted cumulative resource consumption acquisition subunit configured to execute prediction of predicted cumulative resource consumption information after each object is recommended a preset number of times;
an accumulated resource consumption information determination subunit configured to perform determining accumulated resource consumption information from the historical accumulated resource consumption information and the predicted accumulated resource consumption information.
In one possible implementation manner, the fourth recommendation index information determining unit may include:
and the fourth recommendation index information determining subunit is configured to perform mapping of the accumulated resource consumption information to a preset value range, and use a target value obtained through mapping as fourth recommendation index information.
In one possible implementation manner, the recommended parameter adjustment information determining module 705 may include:
the recommendation parameter adjustment information determination unit is configured to perform multiplication processing on first recommendation index information, second recommendation index information, third recommendation index information and fourth recommendation index information of a plurality of objects to obtain recommendation parameter adjustment information corresponding to each of the plurality of objects.
In one possible implementation, the recommending module 709 may include:
the sorting unit is configured to perform sorting on the plurality of objects based on the initial recommendation parameter information and the recommendation parameter adjustment information to obtain a sorting result;
a target object screening unit configured to perform screening of a preset number of objects from the plurality of objects as target objects according to the sorting result;
a recommending unit configured to perform recommending a target object to a target user.
With regard to the apparatus in the above-described embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.
Fig. 8 is a block diagram illustrating an electronic device for object recommendation, which may be a server, according to an exemplary embodiment, and an internal structure thereof may be as shown in fig. 8. The electronic device includes a processor, a memory, and a network interface connected by a system bus. Wherein the processor of the electronic device is configured to provide computing and control capabilities. The memory of the electronic equipment comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The network interface of the electronic device is used for connecting and communicating with an external terminal through a network. The computer program is executed by a processor to implement a method of object recommendation.
Those skilled in the art will appreciate that the architecture shown in fig. 8 is merely a block diagram of some of the structures associated with the disclosed aspects and does not constitute a limitation on the electronic devices to which the disclosed aspects apply, as a particular electronic device may include more or less components than those shown, or combine certain components, or have a different arrangement of components.
In an exemplary embodiment, there is also provided an electronic device including: a processor; a memory for storing the processor-executable instructions; wherein the processor is configured to execute the instructions to implement the object recommendation method as in the embodiments of the present disclosure.
In an exemplary embodiment, there is also provided a computer-readable storage medium in which instructions, when executed by a processor of an electronic device, enable the electronic device to perform an object recommendation method in an embodiment of the present disclosure. The computer readable storage medium may be a ROM, a Random Access Memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like.
In an exemplary embodiment, a computer program product containing instructions is also provided, which when run on a computer, causes the computer to perform the method of object recommendation in embodiments of the present disclosure.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This application is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.
It will be understood that the present disclosure is not limited to the precise arrangements described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims (10)

1. An object recommendation method, comprising:
acquiring a plurality of objects to be recommended and respective initial recommendation parameter information of the objects;
determining index information of the plurality of objects under a preset recommendation index;
determining recommended parameter adjustment information corresponding to the plurality of objects according to the index information;
recommending the plurality of objects to a target user based on the initial recommendation parameter information and the recommendation parameter adjustment information.
2. The object recommendation method according to claim 1, wherein the preset recommendation index includes a conversion prediction index of a preset conversion rate, a resource consumption index associated with the preset conversion rate, an importance index of a user for an object at the preset conversion rate, and a compensation index of the preset conversion rate; the step of determining the index information of each of the plurality of objects under the preset recommendation index includes:
acquiring user association characteristics of the target user and object characteristics of the plurality of objects;
determining first recommendation index information of each object under the conversion prediction index according to the user association characteristics and the object characteristics;
acquiring resource consumption statistical information under object categories corresponding to the objects, and taking the resource consumption statistical information as second recommendation index information of the objects under the resource consumption indexes;
determining third recommendation index information of the target user under the importance degree indexes for the plurality of objects according to the first recommendation index information and recommendation index distribution information associated with the plurality of objects;
determining accumulated resource consumption information associated with the preset conversion rates corresponding to the plurality of objects respectively;
and determining fourth recommendation index information of the plurality of objects under the compensation index according to the accumulated resource consumption information.
3. The object recommendation method according to claim 2, wherein the step of determining the first recommendation index information of each of the plurality of objects under the conversion prediction index according to the user-associated feature and the object feature comprises:
inputting the user correlation characteristics and the object characteristics into a click rate prediction model and a conversion rate prediction model respectively to obtain click rate prediction information and conversion rate prediction information;
and multiplying the click rate prediction information and the conversion rate prediction information to obtain the first recommendation index information.
4. The object recommendation method according to claim 2, wherein the step of determining, according to the first recommendation index information and the recommendation index distribution information associated with the plurality of objects, third recommendation index information of the target user under the importance degree index for the plurality of objects comprises:
determining target distribution information of the first recommendation index information in the recommendation index distribution information;
determining the number of objects corresponding to the target distribution information and the total number of objects included in the recommendation index distribution information;
and taking the proportion of the number of the objects in the total number of the objects as the third recommendation index information.
5. The object recommendation method according to claim 2, wherein the step of determining the cumulative resource consumption information associated with the preset conversion rate corresponding to each of the plurality of objects comprises:
acquiring historical accumulated resource consumption information of each object under the preset conversion rate;
predicting the predicted accumulated resource consumption information after each object is recommended for a preset number of times;
and determining the accumulated resource consumption information according to the historical accumulated resource consumption information and the predicted accumulated resource consumption information.
6. The object recommendation method according to claim 5, wherein the determining fourth recommendation index information for each of the plurality of objects under the compensation index according to the accumulated resource consumption information comprises:
and mapping the accumulated resource consumption information to a preset value range, and taking a target value obtained by mapping as the fourth recommendation index information.
7. An object recommendation apparatus, comprising:
the system comprises an initial recommendation parameter information acquisition module, a recommendation processing module and a recommendation processing module, wherein the initial recommendation parameter information acquisition module is configured to acquire a plurality of objects to be recommended and respective initial recommendation parameter information of the objects;
the index information determining module is configured to determine index information of each of the plurality of objects under a preset recommendation index;
a recommended parameter adjustment information determination module configured to perform determining recommended parameter adjustment information corresponding to each of the plurality of objects according to the index information;
a recommending module configured to recommend the plurality of objects to a target user corresponding to the user identifier based on the initial recommendation parameter information and the recommendation parameter adjustment information.
8. An electronic device, comprising:
a processor;
a memory for storing the processor-executable instructions;
wherein the processor is configured to execute the instructions to implement the object recommendation method of any of claims 1 to 6.
9. A computer-readable storage medium, wherein instructions in the computer-readable storage medium, when executed by a processor of an electronic device, enable the electronic device to perform the object recommendation method of any of claims 1-6.
10. A computer program product comprising computer instructions, characterized in that the computer instructions, when executed by a processor, implement the object recommendation method of any one of claims 1 to 6.
CN202111127592.XA 2021-09-26 2021-09-26 Object recommendation method and device, electronic equipment and storage medium Pending CN113868522A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111127592.XA CN113868522A (en) 2021-09-26 2021-09-26 Object recommendation method and device, electronic equipment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111127592.XA CN113868522A (en) 2021-09-26 2021-09-26 Object recommendation method and device, electronic equipment and storage medium

Publications (1)

Publication Number Publication Date
CN113868522A true CN113868522A (en) 2021-12-31

Family

ID=78994363

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111127592.XA Pending CN113868522A (en) 2021-09-26 2021-09-26 Object recommendation method and device, electronic equipment and storage medium

Country Status (1)

Country Link
CN (1) CN113868522A (en)

Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104504098A (en) * 2014-12-29 2015-04-08 北京奇虎科技有限公司 Information recommending method and device
US20160034952A1 (en) * 2014-04-28 2016-02-04 Pollen Capital Ltd. Control apparatus and accelerating method
US20180268073A1 (en) * 2017-03-15 2018-09-20 Yahoo Holdings, Inc. Online user space exploration for recommendation
CN109657132A (en) * 2017-10-11 2019-04-19 腾讯科技（深圳）有限公司 Recommendation information cost control method, device, computer equipment and storage medium
CN109919651A (en) * 2019-01-17 2019-06-21 阿里巴巴集团控股有限公司 The method for pushing and device of object
CN109948040A (en) * 2017-12-04 2019-06-28 北京京东尚科信息技术有限公司 Storage, recommended method and the system of object information, equipment and storage medium
CN111667311A (en) * 2020-06-08 2020-09-15 腾讯科技（深圳）有限公司 Advertisement delivery method, related device, equipment and storage medium
US20200349610A1 (en) * 2013-09-26 2020-11-05 Mark W. Publicover Providing targeted content based on a user's preferences
CN111932314A (en) * 2020-08-27 2020-11-13 腾讯科技（深圳）有限公司 Method, device and equipment for pushing recommended content and readable storage medium
CN112184293A (en) * 2020-09-17 2021-01-05 北京字节跳动网络技术有限公司 Method, device, equipment and storage medium for adjusting information delivery cost
CN113343024A (en) * 2021-08-04 2021-09-03 北京达佳互联信息技术有限公司 Object recommendation method and device, electronic equipment and storage medium

Patent Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20200349610A1 (en) * 2013-09-26 2020-11-05 Mark W. Publicover Providing targeted content based on a user's preferences
US20160034952A1 (en) * 2014-04-28 2016-02-04 Pollen Capital Ltd. Control apparatus and accelerating method
CN104504098A (en) * 2014-12-29 2015-04-08 北京奇虎科技有限公司 Information recommending method and device
US20180268073A1 (en) * 2017-03-15 2018-09-20 Yahoo Holdings, Inc. Online user space exploration for recommendation
CN109657132A (en) * 2017-10-11 2019-04-19 腾讯科技（深圳）有限公司 Recommendation information cost control method, device, computer equipment and storage medium
CN109948040A (en) * 2017-12-04 2019-06-28 北京京东尚科信息技术有限公司 Storage, recommended method and the system of object information, equipment and storage medium
CN109919651A (en) * 2019-01-17 2019-06-21 阿里巴巴集团控股有限公司 The method for pushing and device of object
CN111667311A (en) * 2020-06-08 2020-09-15 腾讯科技（深圳）有限公司 Advertisement delivery method, related device, equipment and storage medium
CN111932314A (en) * 2020-08-27 2020-11-13 腾讯科技（深圳）有限公司 Method, device and equipment for pushing recommended content and readable storage medium
CN112184293A (en) * 2020-09-17 2021-01-05 北京字节跳动网络技术有限公司 Method, device, equipment and storage medium for adjusting information delivery cost
CN113343024A (en) * 2021-08-04 2021-09-03 北京达佳互联信息技术有限公司 Object recommendation method and device, electronic equipment and storage medium

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
WENTAO OUYANG 等: "Learning Graph Meta Embeddings for Cold-Start Ads in Click-Through Rate Prediction", 《SIGIR \'21: PROCEEDINGS OF THE 44TH INTERNATIONAL ACM SIGIR CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION RETRIEVAL》, 11 July 2021 (2021-07-11), pages 1157, XP059342388, DOI: 10.1145/3404835.3462879 *
YITONG MENG 等: "Wasserstein Collaborative Filtering for Item Cold-start Recommendation", 《UMAP \'20: PROCEEDINGS OF THE 28TH ACM CONFERENCE ON USER MODELING, ADAPTATION AND PERSONALIZATION》, 13 July 2020 (2020-07-13), pages 318, XP059220113, DOI: 10.1145/3340631.3394870 *
佘焕波: "基于淘宝商品行为的向量化内容召回方法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》, no. 06, 15 June 2020 (2020-06-15), pages 138 - 1248 *
王晓倩: "在线旅游个性化推荐方案及策略研究", 《中国博士学位论文全文数据库 经济与管理科学辑》, no. 06, 15 June 2020 (2020-06-15), pages 153 - 3 *

Similar Documents

Publication Publication Date Title
EP3819821B1 (en) User feature generating method, device, and apparatus, and computer-readable storage medium
CN108763502B (en) Information recommendation method and system
RU2018132713A (en) METHOD AND SYSTEM FOR DETERMINING USER-DEPENDING PROPORTIONS OF CONTENTS FOR RECOMMENDATION
CN110609935A (en) User identity tag generation method and device, computer equipment and storage medium
CN114461871B (en) Recommendation model training method, object recommendation device and storage medium
CN113641835B (en) Multimedia resource recommendation method and device, electronic equipment and medium
CN109461053B (en) Dynamic distribution method of multiple recommendation channels, electronic device and storage medium
CN113343024B (en) Object recommendation method and device, electronic equipment and storage medium
CN114201626B (en) Multimedia recommendation method and device, electronic equipment and storage medium
CN113382289B (en) Live broadcast room delivery method and device, electronic equipment and storage medium
CN112905879B (en) Recommendation method, recommendation device, server and storage medium
CN112435070A (en) Method, device and equipment for determining user age and storage medium
CN111651666A (en) User theme recommendation method and device, computer equipment and storage medium
CN114491093B (en) Multimedia resource recommendation and object representation network generation method and device
CN114222146B (en) Data delivery method and device, electronic equipment and storage medium
CN116226537A (en) Layout and display method, device, equipment and medium of page modules in page
CN113868522A (en) Object recommendation method and device, electronic equipment and storage medium
CN114139046B (en) Object recommendation method and device, electronic equipment and storage medium
CN115914363A (en) Message pushing method and device, computer equipment and storage medium
CN113468402B (en) Target object determining method, device and storage medium
CN113850416A (en) Advertisement promotion cooperation object determining method and device
CN110929163B (en) Course recommendation method and device, computer equipment and storage medium
CN113610215A (en) Task processing network generation method, task processing device, electronic equipment and storage medium
CN112883256A (en) Multitasking method and device, electronic equipment and storage medium
CN111275473A (en) Content item delivery method, device, server and storage medium

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