CN113744021A

CN113744021A - Recommendation method, recommendation device, computer storage medium and recommendation system

Info

Publication number: CN113744021A
Application number: CN202110184528.9A
Authority: CN
Inventors: 王立立; 彭长平; 包勇军; 颜伟鹏
Original assignee: Beijing Wodong Tianjun Information Technology Co Ltd
Current assignee: Beijing Wodong Tianjun Information Technology Co Ltd
Priority date: 2021-02-08
Filing date: 2021-02-08
Publication date: 2021-12-03

Abstract

The embodiment of the application provides a recommendation method, a recommendation device, a computer storage medium and a recommendation system, wherein the method comprises the following steps: acquiring user behavior information of a user to be recommended; determining at least one candidate object to be recommended and attribute characteristics of the at least one candidate object to be recommended from a candidate object set based on the user behavior information; determining at least one newly added object to be recommended from a newly added object set according to the attribute characteristics of the at least one candidate object to be recommended; and recommending the user to be recommended based on the at least one candidate object to be recommended and the at least one newly added object to be recommended. Therefore, the candidate object to be recommended is determined from the candidate object set according to the user behavior information, the new object to be recommended is further determined from the new object set by using the candidate object to be recommended, the exposure of the new commodity can be improved on the basis of considering the user preference, and the purchase rate of the new commodity is increased.

Description

Recommendation method, recommendation device, computer storage medium and recommendation system

Technical Field

The present application relates to the technical field of commodity recommendation, and in particular, to a recommendation method, apparatus, computer storage medium, and system.

Background

For e-commerce websites, in order to improve the user activity, commodities need to be recommended according to the personalized design of users, so as to improve the conversion rate of the commodities. At present, a collaborative filtering algorithm is a commodity recommendation algorithm which is most widely applied, and the collaborative filtering algorithm is to find out similar users of a target user according to commodities which the target user is interested in, and then recommend the commodities which the similar users are interested in to the target user.

However, a considerable amount of newly added commodities exist in the e-commerce website, and the newly added commodities lack corresponding user feedback information and cannot be detected by a commodity recommendation algorithm, so that the situation of no display is very easy to cause, and the purpose of popularization of merchants cannot be achieved.

Disclosure of Invention

The application provides a recommendation method, a recommendation device, a computer storage medium and a recommendation system, which can improve the exposure of newly added commodities on the basis of considering user preference, thereby increasing the purchase rate of the newly added commodities.

The technical scheme of the application is realized as follows:

in a first aspect, an embodiment of the present application provides a recommendation method, where the method includes:

acquiring user behavior information of a user to be recommended;

determining at least one candidate object to be recommended and attribute characteristics of the at least one candidate object to be recommended from a candidate object set based on the user behavior information;

determining at least one newly added object to be recommended from a newly added object set according to the attribute characteristics of the at least one candidate object to be recommended;

and recommending the user to be recommended based on the at least one candidate object to be recommended and the at least one newly added object to be recommended.

In a second aspect, an embodiment of the present application provides a recommendation apparatus, which includes an obtaining unit, a first filtering unit, a second filtering unit, and a recommendation unit, wherein,

the acquisition unit is configured to acquire user behavior information of a user to be recommended;

the first screening unit is configured to determine at least one candidate object to be recommended and attribute features of the at least one candidate object to be recommended from a candidate object set based on the user behavior information;

the second screening unit is configured to determine at least one newly added object to be recommended from a newly added object set according to the attribute characteristics of the at least one candidate object to be recommended;

the recommending unit is configured to recommend to the user to be recommended based on the at least one candidate object to be recommended and the at least one newly added object to be recommended.

In a third aspect, an embodiment of the present application provides a recommendation apparatus, including a memory and a processor; wherein,

the memory for storing a computer program operable on the processor;

the processor is adapted to perform the steps of the method according to the first aspect when running the computer program.

In a fourth aspect, embodiments of the present application provide a computer storage medium storing a recommendation program, which when executed by at least one processor implements the steps of the method according to the first aspect.

In a fifth aspect, an embodiment of the present application provides a recommendation system, which includes at least the recommendation apparatus according to the second aspect or the third aspect.

The embodiment of the application provides a recommendation method, a recommendation device, a computer storage medium and a recommendation system, which are used for acquiring user behavior information of a user to be recommended; determining at least one candidate object to be recommended and attribute characteristics of the at least one candidate object to be recommended from a candidate object set based on the user behavior information; determining at least one newly added object to be recommended from a newly added object set according to the attribute characteristics of the at least one candidate object to be recommended; and recommending the user to be recommended based on the at least one candidate object to be recommended and the at least one newly added object to be recommended. Therefore, the candidate object to be recommended is determined from the candidate objects according to the user behavior information, the candidate object to be recommended is further utilized to determine the new object to be recommended from the new object, the exposure of the new commodity can be improved on the basis of considering the preference of the user, and the purchase rate of the new commodity is increased.

Drawings

Fig. 1 is a schematic flowchart of a recommendation method according to an embodiment of the present application;

fig. 2 is a schematic structural diagram of a recommendation system according to an embodiment of the present application;

FIG. 3 is a schematic flowchart of another recommendation method provided in an embodiment of the present application;

FIG. 4 is a block diagram of a new business support module provided in an embodiment of the present application;

fig. 5 is a schematic diagram of a working process of a recommendation method according to an embodiment of the present application;

fig. 6 is a schematic structural diagram of a recommendation device according to an embodiment of the present application;

FIG. 7 is a schematic diagram of a component structure of another recommendation device according to an embodiment of the present application

Fig. 8 is a schematic hardware structure diagram of a recommendation device according to an embodiment of the present application;

fig. 9 is a schematic structural diagram of another recommendation system provided in an embodiment of the present application.

Detailed Description

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. It is to be understood that the specific embodiments described herein are merely illustrative of the relevant application and are not limiting of the application. It should be noted that, for the convenience of description, only the parts related to the related applications are shown in the drawings.

Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of the present application only and is not intended to be limiting of the application.

In the following description, reference is made to "some embodiments" which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict.

It should be noted that the terms "first \ second \ third" referred to in the embodiments of the present application are only used for distinguishing similar objects and do not represent a specific ordering for the objects, and it should be understood that "first \ second \ third" may be interchanged under specific ordering or sequence if allowed, so that the embodiments of the present application described herein can be implemented in other orders than illustrated or described herein.

Advertisement has gradually formed a huge industry as the first mode of change of internet companies, so advertisement putting is very important in an advertisement system, and a problem that needs to be solved by a current recommendation system is to help merchants to quickly and effectively put goods to be sold into the advertisement system after putting the advertisement.

In the e-commerce scene, the set of objects is very huge, and many new objects which are never shown to the user are included in the set of objects, although the business hopes to rapidly popularize the new objects, because the objects lack sufficient user behavior data, the e-commerce website frequently causes the situation that the new objects are not shown/consumed when being recommended by using a collaborative filtering algorithm, and the purpose of business popularization cannot be achieved.

The embodiment of the application provides a recommendation method, and the basic idea of the method is as follows: acquiring user behavior information of a user to be recommended; determining at least one candidate object to be recommended and attribute characteristics of the at least one candidate object to be recommended from a candidate object set based on the user behavior information; determining at least one newly added object to be recommended from a newly added object set according to the attribute characteristics of the at least one candidate object to be recommended; and recommending the user to be recommended based on the at least one candidate object to be recommended and the at least one newly added object to be recommended. Therefore, the candidate object to be recommended is determined from the candidate objects according to the user behavior information, the new object to be recommended is further determined from the new object by using the candidate object to be recommended, the exposure of the new commodity can be improved on the basis of considering the user preference, and the purchase rate of the new commodity is increased.

Embodiments of the present application will be described in detail below with reference to the accompanying drawings.

In an embodiment of the present application, refer to fig. 1, which shows a schematic flowchart of a recommendation method provided in an embodiment of the present application. As shown in fig. 1, the method may include:

s101: and acquiring user behavior information of the user to be recommended.

It should be noted that the recommendation method in the embodiment of the present application may be applied to any scene related to object sorting, for example, various shopping websites of large electronic commerce, movie and television broadcasting websites, reading websites, and travel websites. In other words, the object in the subsequent "recommendation method" in the embodiment of the present application refers to an article in various shopping scenes or push scenes, for example, an entity article such as jewelry, an automobile, a digital product, and the like, and may also be a virtual article such as a video, a novel, an application, news, a movie, a travel product, and the like. In addition, the objects in the recommendation method may also be referred to as articles, products, commodities, goods, industrial goods, and the like.

Here, for convenience of explanation, the following explanation will be made with the product as an object in the recommendation method.

It should be noted that, in order to perform personalized recommendation, user behavior information of a user to be recommended needs to be acquired. Here, the user behavior information refers to operation behavior information of the user to be recommended on the commodities in the last history period, such as which commodities the user purchased, which commodities the user collected, which commodities the user added a shopping cart, and the like. The duration of the history period may be determined according to an actual usage environment, for example, the history period may be 1 month, 3 months, or 6 months, etc.

Therefore, after the user behavior information is acquired, interest mining can be performed according to the user behavior information so as to recommend the user to be recommended.

S102: and determining at least one candidate object to be recommended and attribute characteristics of the at least one candidate object to be recommended from the candidate object set based on the user behavior information.

It should be noted that, interest mining can be performed on the user to be recommended according to the user behavior information, so as to determine at least one candidate object to be recommended in the candidate object set and determine the attribute characteristics of the at least one candidate object to be recommended,

taking an e-commerce website as an example, the candidate object set may refer to all commodities in the e-commerce website, or may refer to a part of commodities that are pre-screened in the e-commerce website, or may refer to all commodities that have been shown to the user in the e-commerce website, which is not limited herein.

The attribute feature refers to tag information of the candidate object to be recommended, and for example, the attribute feature may be a brand to which the product belongs, a product word to which the product belongs, a category to which the product belongs, a store to which the product belongs, or the like. In addition, in the embodiment of the present application, the attribute feature may only select one type of tag information, or may select multiple types of tag information, for example, considering a brand to which the product belongs, a product word to which the product belongs, a category to which the product belongs, and a store to which the product belongs. For convenience of explanation, the following embodiments will be described by taking the brand name of the commodity as an attribute feature.

Further, in some embodiments, the determining, from the candidate object set, at least one candidate object to be recommended and an attribute feature of the at least one candidate object to be recommended based on the user behavior information may include:

calculating a plurality of candidate objects in the candidate object set by using a preset collaborative filtering algorithm based on the user behavior information, and determining object scores of the plurality of candidate objects;

sorting the candidate objects according to the object scores of the candidate objects to obtain a candidate object sequence;

determining the first N candidate objects in the candidate object sequence as the at least one candidate object to be recommended, and determining the attribute characteristics of the at least one candidate object to be recommended; wherein N is a positive integer.

It should be noted that, taking the collaborative filtering algorithm as an example, according to the user behavior information, the preset collaborative filtering algorithm is used to calculate the multiple candidate objects respectively, so as to obtain respective object scores of the multiple candidate objects, which are equivalent to respective commodity scores of the multiple candidate commodities. Here, the product score is used to indicate a probability value that the candidate product is interested by the user to be recommended, that is, the higher the product score is, the more likely the candidate product is interested by the user to be recommended. In addition, the preset collaborative filtering algorithm is already a mature algorithm in the field of commodity recommendation, and the specific calculation process is not described herein.

After the commodity scores of a plurality of candidate commodities are obtained, sequencing the candidate commodities according to the commodity scores to obtain a candidate commodity sequence; then, the first N candidate commodities in the candidate commodity sequence are determined as the aforementioned "candidate objects to be recommended". Here, N is a positive integer, and a specific value of N is determined by a specific application scenario, for example, N may be 200, 400, and the like.

After the candidate object to be recommended is determined, the attribute characteristics of the candidate object to be recommended also need to be obtained for subsequent calculation.

In this way, at least one candidate commodity to be recommended can be determined according to the user behavior information, but due to the short board of the recommendation algorithm, the determined candidate commodity to be recommended may include a very small amount of or even no additional commodity, so that the exposure of the additional commodity is too low. Therefore, after the candidate product to be recommended is determined, subsequent processing is also required.

S103: and determining at least one newly added object to be recommended from the newly added object set according to the attribute characteristics of the at least one candidate object to be recommended.

It should be noted that, according to the attribute features of the multiple candidate objects to be recommended, at least one new object to be recommended is determined in the new object set. Here, the newly added object set is a set of newly online commodities that have not been presented to the user.

That is to say, in the embodiment of the present application, the recommendation method may be subdivided into two stages, where the first stage is to screen out at least one candidate product to be recommended from a candidate product set according to user behavior information; and the second stage is to determine at least one newly added commodity to be recommended in the newly added commodity set according to the attribute characteristics of the at least one candidate commodity to be recommended screened out in the first stage. Therefore, the second stage is specially used for screening the newly added commodities, and the fairness of displaying the newly added commodities can be ensured.

It should be noted that, when selecting a new product to be recommended from the new product set according to the attribute characteristics of the candidate product to be recommended, the selection may be performed according to various principles. Taking the attribute characteristics as the brand as an example, assuming that the attribute characteristics of a plurality of candidate commodities to be recommended are the brand a and the brand b, then newly added commodities of all the brands a and newly added commodities of all the brands b can be selected from the newly added commodity set as newly added commodities to be recommended; or selecting part of newly added commodities of the brand a and part of newly added commodities of the brand b as newly added commodities to be recommended; a triggering scheme can be designed, a brand a and a brand c are triggered according to the brand a, a brand b and a brand d are triggered according to the brand b, and then the newly added commodities to be recommended are determined. That is, the step of "selecting at least one new object to be recommended from the new object set according to the attribute characteristics of the at least one candidate object to be recommended" may be implemented by various principles.

Further, in some embodiments, before determining at least one additional object to be recommended from the additional object set according to the attribute feature of the at least one candidate object to be recommended, the method may further include:

acquiring a plurality of newly added objects, and determining the service types of the newly added objects and the attribute characteristics of the newly added objects;

dividing the newly added objects according to the service classes of the newly added objects and the attribute characteristics of the newly added objects to obtain at least one service-attribute-newly added object subset;

and determining the at least one service-attribute-newly added object subset as a newly added object set.

It should be noted that, taking the e-commerce website as an example, the number of the newly added commodities is quite large due to the numerous merchants. Therefore, in order to improve the efficiency of the subsequent processing, the newly added product may be preprocessed.

Specifically, the newly added commodities can be preprocessed according to the service category and the attribute characteristics. The service category refers to service division performed by the e-commerce platform to better meet the shopping demand of the user, and for example, the service category may include a new product service category, a flash purchase service category, an auction service category, and the like.

Specifically, all newly added commodities are obtained, the service type and the attribute characteristic of the newly added commodities are determined, and then all the newly added commodities are classified according to the service type and the attribute characteristic of the newly added commodities to obtain at least one service-attribute-newly added object subset.

Thus, the newly added object set exists in the form of a plurality of service-attribute-newly added object subsets, namely, newly added commodities are pre-classified, and therefore subsequent processing is facilitated.

Further, in some embodiments, the dividing the plurality of newly added commodities according to the service categories of the plurality of newly added objects and the attribute characteristics of the plurality of newly added objects to obtain at least one service-attribute-newly added object subset may include:

dividing the newly added commodities according to the service classes of the newly added objects to obtain at least one service-newly added object subset; wherein, different service types correspond to different service-newly added object subsets;

according to the attribute characteristics of the newly added commodities, sequentially dividing each service-newly added object subset in the at least one service-newly added object subset to obtain a service-attribute-newly added object subset corresponding to each service-newly added object subset;

and obtaining the at least one service-attribute-newly added object subset according to the service-attribute-newly added object subset corresponding to each service-newly added object subset.

It should be noted that the preprocessing process of the newly added commodity can be divided into two parts, the first part classifies all newly added objects according to different service classes, so as to obtain newly added commodity subsets under different service classes, which are called service-newly added object subsets; aiming at the service-newly added object subset under each service category, the newly added commodities in the service-newly added object subset are divided according to the attribute characteristics to obtain a plurality of newly added commodity subsets under different attribute characteristics, which are called as service-attribute-newly added object subsets. In this way, the at least one service-attribute-new object subset is finally obtained.

Taking the service category including new product, flash purchase and auction, the attribute feature using the brand keyword as an example, assuming that the newly added product includes: item 1 (brand a, new item business category), item 2 (brand a, new item business category), item 3 (brand b, new item business category), item 4 (brand c, auction business category), item 5 (brand c, auction business category), item 6 (brand d, auction business category), item 7 (brand e, flash purchase business category), item 8 (brand e, flash purchase business category), item 9 (brand f, flash purchase business category).

The first step is to classify the newly added commodities according to the service classes, and obtain a service-newly added object set a corresponding to the service class of the new commodity { commodity 1, commodity 2, commodity 3}, a service-newly added object set B corresponding to the auction service class { commodity 4, commodity 5, commodity 6}, and a service-newly added object set C corresponding to the flash purchase service class { commodity 7, commodity 8, commodity 9 }.

Secondly, aiming at a service-newly added object set A, which is { commodity 1, commodity 2 and commodity 3}, according to attribute characteristics (brands of the products), a service-attribute-newly added object set A1 corresponding to a brand a is { commodity 1 and commodity 2}, and a service-attribute-newly added object set A2 corresponding to a brand b is { commodity 3 };

aiming at the service-new object set B, according to attribute characteristics, the service-attribute-new object set B1 corresponding to the brand c is divided into { product 4, product 5} and service-attribute-new object set B2 corresponding to the brand d is divided into { product 6 };

for the service-added-object set C ═ commodity 7, commodity 8, and commodity 9, according to the attribute characteristics, the service-attribute-added-object set C1 corresponding to the brand e is further divided into { commodity 7, commodity 8}, and the service-attribute-added-object set C2 corresponding to the brand f is further divided into { commodity 9 }. Thus, the service-attribute-added object sets a1, a2, B1, B2, C1 and C2 are "at least one service-attribute-added object subset".

In this way, the newly added commodities are thinned into a plurality of service-attribute-newly added object subsets, so that the newly added commodities to be recommended can be screened subsequently.

Further, in some embodiments, the determining, according to the attribute feature of the at least one candidate object to be recommended, at least one additional object to be recommended from an additional object set may include:

determining a preset filtering rule corresponding to the first service class; wherein the first service class is any one of the service classes of the newly added objects;

filtering the attribute features of the at least one candidate object to be recommended based on the preset filtering rule to obtain at least one target attribute feature;

determining at least one target object subset corresponding to the at least one target attribute characteristic in the service-attribute-newly added object subset corresponding to the first service class;

determining a newly added object to be recommended corresponding to the first service class according to the at least one target object subset;

and after determining the newly added objects to be recommended corresponding to the plurality of preset service types, obtaining the at least one newly added object to be recommended.

It should be noted that after a plurality of candidate objects to be recommended are determined, new objects to be recommended need to be further determined according to the candidate objects to be recommended, and this process needs to be performed according to different service categories respectively.

For convenience of description, if any one of the service classes of the multiple newly added objects is regarded as the first service class, then the process of determining the newly added object to be recommended in the first service class is as follows:

firstly, acquiring a preset filtering rule corresponding to a first service class; secondly, obtaining respective attribute characteristics of at least one candidate object to be recommended, and filtering the determined attribute characteristics by using a preset filtering rule to obtain at least one target attribute characteristic. That is to say, in the embodiment of the application, new added commodities to be recommended are further screened out according to the attribute features of the candidate commodities to be recommended, however, some attribute features of the candidate commodities to be recommended are meaningless for a certain service category, for example, the attribute feature of the candidate commodities to be recommended is a brand a, but for an auction service line category, the category does not include the new added commodities of the brand a, and therefore, inappropriate attribute features are filtered out through a preset filtering rule.

Then, in the service-attribute-newly added object subset corresponding to the first service class, at least one target object subset corresponding to the at least one target attribute feature is determined. For example, if the target attribute features are brand a and brand b, the service-attribute-newly added object subsets corresponding to brand a and brand b in the first service category are respectively determined as target object subsets.

And finally, according to the determined at least one target object subset, determining a plurality of newly added objects to be recommended corresponding to the first service class. And respectively repeating the steps for different service types to obtain the final newly added object to be recommended, namely the final newly added commodity to be recommended.

Further, in some embodiments, the determining, according to the at least one target object subset, a new object to be recommended corresponding to the first service class may include:

performing object screening on the at least one target object subset to obtain at least one target newly-added object and the number of the corresponding target newly-added objects;

judging whether the number of the target newly added objects is larger than or equal to a preset threshold value or not;

if so, determining the at least one target newly-added object as a newly-added object to be recommended corresponding to the first service class;

under the condition that the judgment result is negative, determining at least one complementary newly added object; wherein the number of the at least one complementary newly added object is a difference value between the preset threshold value and the target newly added object number;

and determining the at least one complementary newly added object and the at least one target newly added object as newly added objects to be recommended corresponding to the first service class.

It should be noted that, when determining a plurality of new objects to be recommended, the method may be divided into two mount stages:

in the first mounting stage, at least one target newly-added object is randomly selected from the determined at least one target object subset.

Specifically, M newly added commodities can be randomly selected from each target object subset as target newly added objects, where M is a positive integer. The value of M may be determined according to an actual application scenario, for example, M is 3, 6, 9, and the like. It should be understood that, in this specific embodiment, if the number of the commodities in the target object subset is less than the preset number, all the commodities in the target object subset are determined as the target additional object.

After determining at least one target newly-added object, counting the number of the target newly-added objects, namely the number of the target newly-added objects, judging whether the number of the target newly-added objects is larger than or equal to a preset threshold value, if so, directly determining the at least one target newly-added object as a newly-added object to be recommended corresponding to the first service class without entering a second mounting stage; if not, the second mounting phase needs to be entered.

In a second mount phase, at least one complementary add object is determined. Here, the supplementary newly added object may be randomly determined from a portion of the first service-newly added object excluding the target newly added object, and the number of the supplementary newly added objects is a difference between a preset threshold value and the number of the target newly added objects. The preset threshold may be determined according to an application scenario, for example, the preset threshold may be 200, 500, 1000, or the like.

That is, in order to prevent the problem that the number of the screened target additional commodities is too small, a complementary mechanism is provided. Specifically, after the target newly-added commodity is determined, whether the number of the target newly-added commodities is greater than or equal to a preset threshold value or not is judged, and if the judgment result is yes, all the target newly-added commodities can be directly determined as newly-added commodities to be recommended and corresponding to the first service class; and if the judgment result is negative, determining at least one newly added supplementary commodity, wherein the number of the newly added supplementary commodity is the difference value between the preset threshold value and the target newly added commodity.

In addition to this, the complementary mechanism can also be set as: and judging the total number of the newly added commodities of the target under all the service categories, comparing the total number with a preset threshold value, and complementing according to the comparison result.

It should be further noted that, because the candidate set may include the new object, some candidate objects to be recommended may be the new object itself (with a small probability in practical application), which does not affect the subsequent processing procedure.

Therefore, after the two mounting stages, the newly added objects to be recommended which are not less than the preset threshold value are determined in total, and the problem that the exposure amount of the newly added goods is less due to the fact that the user information corresponding to the newly added goods is too little can be solved.

S104: and recommending the user to be recommended based on the at least one candidate object to be recommended and the at least one newly added object to be recommended.

After determining the new objects to be recommended, recommending the candidate objects to be recommended determined in the first stage and the new objects to be recommended determined in the second stage to the users to be recommended; therefore, the candidate commodities to be recommended mined according to the user behaviors in the first stage and the newly added commodities to be recommended mined by using the candidate commodities to be recommended in the second stage are recommended to the user at the same time, and the problem of low exposure of the newly added commodities can be avoided.

Further, in some embodiments, the obtaining the user behavior information of the user to be recommended may include:

receiving request information sent by the user to be recommended aiming at a target page;

and acquiring the user behavior information of the user to be recommended according to the request information.

It should be noted that, in a specific embodiment, the recommendation method is triggered after the user to be recommended sends the request information to the target page. That is to say, after receiving the request information of the user to be recommended for the target page, the user behavior information of the user to be recommended is acquired. Here, the target page is a commodity display page including an advertisement space, so that the determined new object to be recommended is displayed to the user to be recommended at the advertisement space.

Further, in some embodiments, the recommending the user to be recommended based on the at least one candidate object to be recommended and the at least one newly added object to be recommended may include:

determining the at least one candidate object to be recommended and the at least one newly added object to be recommended as a set of recommended objects to be ranked;

sequencing the recommended object set to be sequenced to determine a target recommended object sequence;

and recommending the target recommendation object sequence to the user to be recommended.

It should be noted that, in order to balance the interest characteristics of the user and the exposure requirements of the merchant on the newly added objects, at least one candidate object to be recommended determined in the first stage and at least one newly added object to be recommended determined in the second stage may be mixed to obtain a set of recommended objects to be sorted;

then, performing sorting calculation on the set of recommended objects to be sorted to obtain a target recommended object sequence; here, the specific method of the ranking calculation may refer to an existing ranking algorithm, which is not described herein.

And finally, recommending the target recommendation object sequence to the user to be recommended.

Further, in some embodiments, the recommending the target recommendation object sequence to the user to be recommended may include:

and displaying the target recommendation object sequence at a preset advertisement position in the target page.

It should be noted that the target recommendation object sequence is displayed at a preset advertisement position on the target page, so that the whole process of recommending to the user to be recommended is completed. Here, if there is more than one preset advertisement position on the target page, a plurality of target recommendation object sub-sequences for different service categories can be determined and then displayed respectively.

It should be noted that, assuming that the object to be recommended is a product and the attribute features refer to a brand to which the product belongs and a store to which the product belongs, the general processing flow is as follows:

firstly, preprocessing a plurality of newly added commodities, and dividing the newly added commodities into a plurality of services-newly added commodity subsets according to different service types; for each business-newly-added commodity subset, performing secondary division by respectively utilizing attribute characteristics 1 (brands to which commodities belong) and attribute characteristics 2 (shops to which commodities belong) to obtain a plurality of business-attribute 1-newly-added commodity subsets and a plurality of business-attribute 2-newly-added commodity subsets;

secondly, as with the above process, screening out a plurality of candidate commodities to be recommended according to the user behavior information, and acquiring target attribute characteristics of the plurality of candidate commodities to be recommended respectively aiming at the attribute characteristics 1 and 2, for example aiming at the attribute characteristic 1, determining brands a to i; for the attribute feature 2, stores 1 to 10 are specified.

Thirdly, aiming at a specific service class, performing attribute feature screening according to a preset screening rule corresponding to the service class, for example, aiming at attribute feature 1, target attribute features are brand a and brand c; for the attribute feature 2, the target attribute features are store 2 and store 9.

And finally, determining two target commodity subsets corresponding to the brands a and c by a plurality of service-attribute 1-newly added commodity subsets under the service category, determining two target commodity subsets corresponding to the stores 2 and 9 by a plurality of service-attribute 2-newly added commodity subsets under the service category, and selecting newly added commodities to be recommended according to the determined target commodity subsets. The technical scheme is also within the protection scope of the embodiment.

In the related technical scheme, a collaborative filtering mining strategy is generally applied to mining a preference commodity set of a user according to behavior data of the user, most of the collaborative filtering mining strategy is applied to behavior information of the user on commodities, but for a newly added commodity set, the user does not produce behaviors on the commodity set of the category, so that the current recommendation system is not specifically optimized when the interest preference of the user and the exposure of a new service line are simultaneously met. Therefore, the embodiment of the application provides a commodity recommendation scheme based on interests, and the commodity recommendation scheme has the following advantages:

(1) carrying out two-stage interest mounting by utilizing the user interest preference mined in the first stage and the attribute characteristics of the newly added commodity; here, mounting corresponds to "recommendation".

(2) The embodiment of the application meets the requirements of merchants on the exposure/consumption of commodities in a new business line on the basis of meeting the requirements of recommending the commodities preferred by users.

(3) The embodiment of the application provides a two-stage mining scheme based on user interest, and meanwhile, diversity of recommended advertising spots is enriched by combining with commodity attribute information.

That is to say, because a plurality of candidate commodities to be recommended are obtained according to the user behavior information of the user to be recommended and belong to interest recommendation for the user side, a plurality of newly added commodities to be recommended are determined in the newly added commodity set according to a plurality of candidate commodities to be recommended, that is, the interest of the user is considered, and the exposure requirement of the newly added commodities is also met, so that the problem that the display quantity of the newly added commodities is small in the existing sorting method is solved.

The embodiment of the application provides a recommendation method, which comprises the steps of obtaining user behavior information of a user to be recommended; determining at least one candidate object to be recommended and attribute characteristics of the at least one candidate object to be recommended from a candidate object set based on the user behavior information; determining at least one newly added object to be recommended from a newly added object set according to the attribute characteristics of the at least one candidate object to be recommended; and recommending the user to be recommended based on the at least one candidate object to be recommended and the at least one newly added object to be recommended. Therefore, the candidate object to be recommended is determined from the candidate objects according to the user behavior information, the candidate object to be recommended is further utilized to determine the new object to be recommended from the new object, the exposure of the new commodity can be improved on the basis of considering the preference of the user, and the purchase rate of the new commodity is increased.

In another embodiment of the present application, an example in which an object is a product will be explained. Referring to fig. 2, a schematic diagram of a component structure of a recommendation system 20 provided in an embodiment of the present application is shown. As shown in fig. 2, the recommendation system 20 may include a recommendation algorithm module 201, a new business support module 202, a Merge (Merge) module 203; the recommendation algorithm module 201 comprises a user behavior unit 2011, a model algorithm recall unit 2012 and a commodity recommendation word list unit 2013; the new business support module 202 includes a commodity preprocessing unit 2021 and a commodity mounting module 2022.

In the recommendation system 20, after user request information sent by a user to be recommended is acquired, the user behavior unit 2011 acquires the user behavior information of the user to be recommended, and the model algorithm recall unit 2012 calculates the user behavior information according to the commodity recommendation vocabulary provided by the commodity recommendation vocabulary unit 2013 to determine a candidate commodity set to be recommended; in addition, the commodity preprocessing unit 2021 in the new service support module 202 has already finished preprocessing all newly added commodities and sends the preprocessing result to the commodity mounting unit 2022; then, the commodity mounting unit 2022 determines a new commodity set to be recommended according to the preprocessing result of the new commodity and the candidate commodity set to be recommended. Thus, the newly added product set to be recommended and the candidate product set to be recommended form an overall recall combination (equivalent to the recommendation object set to be sorted).

Further, based on the composition of the recommendation system 20, refer to fig. 3, which shows a flowchart of another recommendation method provided in the embodiment of the present application. As shown in fig. 3, the method may include:

s301: and acquiring user request information.

It should be noted that the user request information unit receives user request information, and the user request information refers to request information of a user to be recommended for a target page.

S302: the recommendation algorithm module 201 determines a candidate commodity set to be recommended according to the user request information.

It should be noted that, determining the candidate commodity set to be recommended (the candidate commodity set to be recommended is equivalent to the aforementioned multiple candidate objects to be recommended) may include the following steps:

(1) according to a preset collaborative filtering algorithm, a commodity relevance vocabulary is constructed by the commodity recommendation vocabulary unit 2013 and is used for indicating the similarity degree between two commodities, such as S_ijRepresenting the similarity of the commodity i and the commodity j;

(2) the user behavior unit 2011 controls the user traffic server to send user behavior information of a user to be recommended (a user sending user request information), where the user behavior information may include behaviors such as browsing, clicking, purchasing and the like and a commodity targeted by the behavior;

(3) the model algorithm recall unit 2012 retrieves and calculates the product scores of all the candidate products according to the user behavior information and the product relevance vocabulary, thereby ranking all the candidate products.

In a specific embodiment, the product score of the candidate product indicates the similarity between the candidate product and the product of interest of the user to be recommended, which can be S as described above_ijThe representation i can represent candidate goods, and j can represent the sense of the user to be recommendedAn interesting commodity.

(4) And performing stage truncation processing on the sequencing result to obtain a candidate commodity set to be recommended in the first stage.

S303: the new service support module 202 determines a new product set to be recommended according to the candidate product set to be recommended.

It should be noted that, taking three service lines (corresponding to the aforementioned service categories) of new product, auction and flash purchase as an example, it should be noted that, referring to fig. 4, a schematic diagram of a framework of a new service support module provided in an embodiment of the present application is shown. As shown in fig. 4, for a new service product (equivalent to a new added object), firstly, a product preprocessing is performed according to a service line (new product, auction, and flash purchase), then, a product preprocessing result and a user interest mining result (i.e., a candidate object set to be recommended) in the first stage are sent to a product mounting unit, and a new product set to be recommended is determined by the product mounting unit (the new product set to be recommended is equivalent to the aforementioned multiple new added objects to be recommended).

Specifically, the workflow of the new business support module 202 includes:

(1) the commodity preprocessing unit 2021 obtains all new service commodity sets;

(2) according to three service lines of new products, auction and flash purchase, the product preprocessing unit 2021 divides the new service products to obtain three services, namely a newly added product subset; then, under each service line, the newly added commodities are further divided according to attribute information (namely attribute characteristics), wherein the attribute information can be brands, product words and the like, and each service-newly added commodity subset is further divided into service-brand-newly added commodity subsets by taking the brands as an example.

(3) According to the result of the first-stage user interest mining and the result of the commodity preprocessing, the commodity mounting unit 2022 determines a newly added commodity set to be recommended. That is to say, the commodity mounting unit excavates a user one-stage commodity set according to the user interest, and then carries out two-stage commodity mounting according to the one-stage recalled commodity set, and the mounting mode ensures the mounting capacity of the newly added commodity. The method specifically comprises the following steps:

(3.1) acquiring a trigger brand set: sequentially acquiring brand information of the first N commodities from the candidate commodity set to be recommended; here, the value N is 400.

(3.2) brand filtering is carried out on a certain product line according to a preset filtering rule: and (3) aiming at preset filtering rules (or called as trigger types) constructed by different product lines, filtering the brand information acquired in the step (3.1) according to the trigger types to acquire a plurality of target service line brands, and determining a plurality of service lines corresponding to the target service line brands, namely brand new commodity subsets.

And (3.3) sequentially selecting a preset number of cold start commodities in each service line-brand newly-added commodity subset. This phase mainly comprises two random mount schemes:

a first mounting stage: at most, M commodities are randomly mounted in each service line-brand new commodity subset, and at most, M × N cold-starting commodities are obtained, wherein the quantity of the commodities in part of the service line-brand new commodity subsets is less than M, so that enough cold-starting commodities cannot be obtained;

and a second mounting stage: supposing that the number of the cold-starting commodities actually determined in the first mounting stage is X, judging whether X reaches a preset threshold value Y, and if not, randomly obtaining T2 cold-starting commodities again; here, T2 ═ Y — X.

Therefore, the above two random mounting fully ensures the equitable recall of each cold-start commodity and ensures a sufficient number of cold-start commodities.

Through the processes, a newly added commodity recommendation candidate set comprising Y cold-starting commodities is finally obtained.

S304: the merging module 203 merges the candidate commodity set to be recommended and the newly added commodity set to be recommended to obtain an overall recall set.

It should be noted that the merging module 203 merges the candidate product set to be recommended and the newly added product set to be recommended to obtain an overall recalled product set (which is equivalent to the aforementioned recommended object set to be sorted). And sending the subsequent whole recalled commodity set to a sorting module for specific sorting, and then displaying the commodity set to the user to be recommended at the advertisement position of the target page.

The embodiment of the application provides a recommendation system and a recommendation method, and through the detailed explanation of the embodiment, the interest mining is performed according to the user behavior information to obtain candidate commodities to be recommended, then the newly added commodities to be recommended are determined by using the candidate commodities to be recommended, on the basis of considering the preference of the user, the exposure of the newly added commodities is improved, and therefore the purchase rate of the newly added commodities is increased.

In another embodiment of the present application, an object is a product. Referring to fig. 5, a schematic diagram of an operation process of a recommendation method provided in an embodiment of the present application is shown. In fig. 5, the dot-filled boxes represent candidate items to be recommended, and the horizontal line-filled boxes: representing the target brand goods filtered for the service line, black filled squares: and newly added commodities are to be recommended.

As shown in fig. 5, in the workflow of the recommendation system, the operations may be divided into filtering, triggering, attribute mapping (keytype), filtering (keyFilter ()), collecting (Mining ()), and merging (MergeMining Result ()).

Specifically, the operation shown in fig. 5 is as follows:

firstly, a candidate commodity set to be recommended is obtained after the user behavior information of the user to be recommended is screened in the first stage, wherein the candidate commodity set to be recommended comprises a commodity (sku)1, a commodity 2, a commodity 3 and a commodity 4 … ….

Secondly, obtaining respective attribute characteristics of candidate commodities to be recommended, namely a commodity 1 is a brand 1, a commodity 2 is a brand 2, a commodity 3 is a brand 3, and a commodity 4 is a brand 4.

Taking two service lines as an example, filtering the brands obtained by triggering according to the preset filtering rule corresponding to each service line to obtain target recommended brands corresponding to each service line, for example, a service line obtains brand 1 and brand 3, and another service line obtains brand 2, brand 3 and brand 4; and then, collecting the target recommended brand to obtain a newly added commodity set to be recommended.

And finally, merging the newly added commodity set to be recommended and the candidate commodity set to be recommended to obtain an integral recalled commodity set (appendix response ()).

The embodiment of the application provides a recommendation system, which is characterized in that through the detailed explanation of the embodiment, interest mining is performed according to user behavior information to obtain candidate commodities to be recommended, then the newly added commodities to be recommended are determined by utilizing the candidate commodities to be recommended, on the basis of considering user preferences, the exposure of the newly added commodities is improved, and therefore the purchase rate of the newly added commodities is increased.

In a further embodiment of the present application, refer to fig. 6, which shows a schematic structural diagram of a component of a recommendation device 40 provided in an embodiment of the present application. As shown in fig. 6, the recommendation apparatus 40 includes an acquisition unit 401, a first filtering unit 402, a second filtering unit 403, and a recommendation unit 404, wherein,

an obtaining unit 401 configured to obtain user behavior information of a user to be recommended;

a first screening unit 402, configured to determine at least one candidate object to be recommended and attribute features of the at least one candidate object to be recommended from a candidate object set based on the user behavior information;

a second screening unit 403, configured to determine at least one new object to be recommended from the new object set according to the attribute feature of the at least one candidate object to be recommended;

the recommending unit 404 is configured to recommend to the user to be recommended based on the at least one candidate object to be recommended and the at least one newly added object to be recommended.

In some embodiments, the first filtering unit 402 is specifically configured to calculate, based on the user behavior information, a plurality of candidate objects in the candidate object set by using a preset collaborative filtering algorithm, and determine object scores of the plurality of candidate objects; sorting the candidate objects according to the object scores of the candidate objects to obtain a candidate object sequence; determining the first N candidate objects in the candidate object sequence as the at least one candidate object to be recommended, and determining the attribute characteristics of the at least one candidate object to be recommended; wherein N is a positive integer.

In some embodiments, as shown in fig. 7, the recommendation apparatus 40 further includes a preprocessing unit 405 configured to obtain a plurality of new objects, and determine service classes of the new objects and attribute features of the new objects; dividing the newly added objects according to the service classes of the newly added objects and the attribute characteristics of the newly added objects to obtain at least one service-attribute-newly added object subset; and determining the at least one service-attribute-newly added object subset as a newly added object set.

In some embodiments, the preprocessing unit 405 is specifically configured to divide the plurality of newly added objects according to the service classes of the plurality of newly added objects, so as to obtain at least one service-newly added object subset; wherein, different service types correspond to different service-newly added object subsets; according to the attribute characteristics of the newly added objects, sequentially dividing each service-newly added object subset in the at least one service-newly added object subset to obtain a service-attribute-newly added object subset corresponding to each service-newly added object subset; and obtaining the at least one service-attribute-newly added object subset according to the service-attribute-newly added object subset corresponding to each service-newly added object subset.

In some embodiments, the second filtering unit 403 is specifically configured to determine a preset filtering rule corresponding to the first service category; wherein the first service class is any one of the service classes of the newly added objects; filtering the attribute features of the at least one candidate object to be recommended based on the preset filtering rule to obtain at least one target attribute feature; determining at least one target object subset corresponding to the at least one target attribute characteristic in the service-attribute-newly added object subset corresponding to the first service class; determining a newly added object to be recommended corresponding to the first service class according to the at least one target object subset; and after determining the newly added objects to be recommended corresponding to the plurality of preset service types, obtaining the at least one newly added object to be recommended.

In some embodiments, the second screening unit 403 is further configured to perform object screening on the at least one target object subset to obtain at least one target newly added object and the corresponding target newly added object number;

judging whether the number of the target newly added objects is larger than or equal to a preset threshold value or not; if so, determining the at least one target newly-added object as a newly-added object to be recommended corresponding to the first service class; under the condition that the judgment result is negative, determining at least one complementary newly added object; wherein the number of the at least one complementary newly added object is a difference value between the preset threshold value and the target newly added object number; and determining the at least one complementary newly added object and the at least one target newly added object as newly added objects to be recommended corresponding to the first service class.

In some embodiments, the traffic class includes at least one of: a new product service category, an auction service category, and a flash purchase service category; the attribute features include at least one of: brand keywords, category keywords, store keywords, and product keywords.

In some embodiments, the obtaining unit 401 is specifically configured to receive request information sent by the user to be recommended for a target page; and acquiring the user behavior information of the user to be recommended according to the request information.

In some embodiments, the recommending unit 404 is specifically configured to determine the at least one candidate object to be recommended and the at least one newly added object to be recommended as a set of recommended objects to be ranked; sequencing the recommended object set to be sequenced to determine a target recommended object sequence; and recommending the target recommendation object sequence to the user to be recommended.

In some embodiments, the recommending unit 404 is further configured to display the target recommendation object sequence at a preset advertisement position in the target page.

It is understood that in this embodiment, a "unit" may be a part of a circuit, a part of a processor, a part of a program or software, etc., and may also be a module, or may also be non-modular. Moreover, each component in the embodiment may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware or a form of a software functional module.

Based on the understanding that the technical solution of the present embodiment essentially or a part contributing to the prior art, or all or part of the technical solution may be embodied in the form of a software product stored in a storage medium, and include several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) or a processor (processor) to execute all or part of the steps of the method of the present embodiment. And the aforementioned storage medium includes: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.

Accordingly, the present embodiments provide a computer storage medium storing a recommendation program that, when executed by at least one processor, performs the steps of the method of any of the preceding embodiments.

Based on the above-mentioned composition of a recommendation device 40 and the computer storage medium, refer to fig. 8, which shows a specific hardware structure diagram of a recommendation device 40 provided in an embodiment of the present application. As shown in fig. 8, the recommending apparatus 40 may include: a communication interface 501, a memory 502, and a processor 503; the various components are coupled together by a bus device 504. It is understood that bus device 504 is used to enable connected communication between these components. Bus device 504 includes a power bus, a control bus, and a status signal bus in addition to a data bus. But for clarity of illustration the various buses are labeled as bus device 504 in figure 8. The communication interface 501 is used for receiving and sending signals in the process of receiving and sending information with other external network elements;

a memory 502 for storing a computer program capable of running on the processor 503;

a processor 503 for executing, when running the computer program, the following:

acquiring user behavior information of a user to be recommended;

It will be appreciated that the memory 502 in the embodiments of the subject application can be either volatile memory or nonvolatile memory, or can include both volatile and nonvolatile memory. The non-volatile Memory may be a Read-Only Memory (ROM), a Programmable ROM (PROM), an Erasable PROM (EPROM), an Electrically Erasable PROM (EEPROM), or a flash Memory. Volatile Memory can be Random Access Memory (RAM), which acts as external cache Memory. By way of example, but not limitation, many forms of RAM are available, such as Static random access memory (Static RAM, SRAM), Dynamic Random Access Memory (DRAM), Synchronous Dynamic random access memory (Synchronous DRAM, SDRAM), Double Data Rate Synchronous Dynamic random access memory (ddr Data Rate SDRAM, ddr SDRAM), Enhanced Synchronous SDRAM (ESDRAM), Synchronous chained SDRAM (Synchronous link DRAM, SLDRAM), and Direct memory bus RAM (DRRAM). The memory 502 of the apparatus and methods described herein is intended to comprise, without being limited to, these and any other suitable types of memory.

And the processor 503 may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the above method may be performed by integrated logic circuits of hardware or instructions in the form of software in the processor 503. The Processor 503 may be a general-purpose Processor, a Digital Signal Processor (DSP), an APPlication Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic device, or discrete hardware components. The various methods, steps, and logic blocks disclosed in the embodiments of the present application may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of the method disclosed in connection with the embodiments of the present application may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software module may be located in ram, flash memory, rom, prom, or eprom, registers, etc. storage media as is well known in the art. The storage medium is located in the memory 502, and the processor 503 reads the information in the memory 502 and completes the steps of the above method in combination with the hardware thereof.

It is to be understood that the embodiments described herein may be implemented in hardware, software, firmware, middleware, microcode, or any combination thereof. For a hardware implementation, the Processing units may be implemented within one or more APPlication Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), Programmable Logic Devices (PLDs), Field Programmable Gate Arrays (FPGAs), general purpose processors, controllers, micro-controllers, microprocessors, other electronic units configured to perform the functions described herein, or a combination thereof.

For a software implementation, the techniques described herein may be implemented with modules (e.g., procedures, functions, and so on) that perform the functions described herein. The software codes may be stored in a memory and executed by a processor. The memory may be implemented within the processor or external to the processor.

Optionally, as another embodiment, the processor 503 is further configured to perform the steps of the method of any one of the preceding embodiments when running the computer program.

Based on the above-mentioned composition and hardware structure diagram of the recommendation device 40. Referring to fig. 9, a schematic structural diagram of another recommendation system 20 provided in the embodiment of the present application is shown. As shown in fig. 9, the recommendation system 20 at least comprises the recommendation device 40 of any of the previous embodiments.

For the recommendation system 20, the candidate object to be recommended is determined from the candidate object set according to the user behavior information, and the new object to be recommended is further determined from the new object set by using the candidate object to be recommended, so that the exposure of the new commodity can be increased on the basis of considering the user preference, and the purchase rate of the new commodity is increased.

The above description is only a preferred embodiment of the present application, and is not intended to limit the scope of the present application.

It should be noted that, in the present application, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.

The above-mentioned serial numbers of the embodiments of the present application are merely for description and do not represent the merits of the embodiments.

The methods disclosed in the several method embodiments provided in the present application may be combined arbitrarily without conflict to obtain new method embodiments.

Features disclosed in several of the product embodiments provided in the present application may be combined in any combination to yield new product embodiments without conflict.

The features disclosed in the several method or apparatus embodiments provided in the present application may be combined arbitrarily, without conflict, to arrive at new method embodiments or apparatus embodiments.

The above description is only for the specific embodiments of the present application, but the scope of the present application is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present application, and shall be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A recommendation method, characterized in that the method comprises:

acquiring user behavior information of a user to be recommended;

2. The recommendation method according to claim 1, wherein the determining at least one candidate object to be recommended and attribute features of the at least one candidate object to be recommended from a candidate object set based on the user behavior information comprises:

3. The recommendation method according to claim 1, wherein before said determining at least one new object to be recommended from a new object set according to the attribute feature of the at least one candidate object to be recommended, the method further comprises:

4. The recommendation method according to claim 3, wherein the dividing the plurality of newly added objects according to the service classes of the plurality of newly added objects and the attribute features of the plurality of newly added objects to obtain at least one service-attribute-newly added object subset comprises:

dividing the newly added objects according to the service classes of the newly added objects to obtain at least one service-newly added object subset; wherein, different service types correspond to different service-newly added object subsets;

according to the attribute characteristics of the newly added objects, sequentially dividing each service-newly added object subset in the at least one service-newly added object subset to obtain a service-attribute-newly added object subset corresponding to each service-newly added object subset;

5. The recommendation method according to claim 3, wherein the determining at least one new object to be recommended from the new object set according to the attribute feature of the at least one candidate object to be recommended comprises:

6. The recommendation method according to claim 5, wherein the determining, according to the at least one target object subset, the new object to be recommended corresponding to the first service class includes:

7. Recommendation method according to any one of claims 3-6,

the traffic class includes at least one of: a new product service category, an auction service category, and a flash purchase service category;

the attribute features include at least one of: brand keywords, category keywords, store keywords, and product keywords.

8. The recommendation method according to claim 1, wherein the obtaining of the user behavior information of the user to be recommended comprises:

9. The recommendation method according to claim 8, wherein the recommending the user to be recommended based on the at least one candidate object to be recommended and the at least one newly added object to be recommended comprises:

10. The recommendation method according to claim 9, wherein the recommending the target recommendation object sequence to the user to be recommended comprises:

11. A recommendation device is characterized by comprising an acquisition unit, a first screening unit, a second screening unit and a recommendation unit,

12. A recommendation device, characterized in that the recommendation device comprises a memory and a processor; wherein,

the memory for storing a computer program operable on the processor;

the processor, when executing the computer program, is adapted to perform the steps of the method of any of claims 1 to 10.

13. A computer storage medium, characterized in that it stores a recommendation program which, when executed by at least one processor, implements the steps of the method according to any one of claims 1 to 10.

14. A recommendation system, characterized in that it comprises at least a recommendation device according to claim 11 or 12.