CN110689410A

CN110689410A - Data processing method, device, equipment and storage medium

Info

Publication number: CN110689410A
Application number: CN201910937162.0A
Authority: CN
Inventors: 刘华; 王小宇
Original assignee: JD Digital Technology Holdings Co Ltd
Current assignee: JD Digital Technology Holdings Co Ltd; Jingdong Technology Holding Co Ltd
Priority date: 2019-09-29
Filing date: 2019-09-29
Publication date: 2020-01-14
Anticipated expiration: 2039-09-29
Also published as: CN110689410B

Abstract

The application provides a data processing method, a data processing device, data processing equipment and a storage medium, and relates to the technical field of data processing. Wherein, the method comprises the following steps: the method comprises the steps of obtaining a recommended data set aiming at a target user, sorting commodity data in the recommended data set according to historical behavior data of the target user, user association information and a preset sorting algorithm to obtain a data sorting result, and determining at least one piece of commodity data displayed to the target user according to the data sorting result. According to the technical scheme, when the commodity data in the determined recommendation data set are sorted, the historical behavior data of the target user and the user association information of the target user are combined, so that the sorting accuracy of the commodity data is improved, the problem that the existing commodity sorting result is inaccurate is solved, meanwhile, the sorting accuracy in data sorting is improved, and the conversion rate of the commodity data is improved.

Description

Data processing method, device, equipment and storage medium

Technical Field

The present application relates to the field of data processing technologies, and in particular, to a data processing method, apparatus, device, and storage medium.

Background

With the development of internet technology, networks are becoming important sources for people to obtain information. The commodity data display through the webpage/app and the like can attract the attention of the user to a certain extent, explore the subsequent behaviors of the user and determine the personalized requirements of the user so as to create great commercial value under the environment that the internet is popularized.

In the prior art, in order to enable a commodity displayed on a webpage/app to attract the attention of a user to the greatest extent, commodity data which may be interested by the user is generally screened out from massive commodity data, then the commodity data to be displayed is sorted based on the preference degree of the user to the commodity, and finally the commodity data is displayed at a corresponding position on the webpage/app.

However, the commodity data sorting result determined by the method may be influenced by the interference of the characteristics of the commodity, so that the determined commodity data sorting result is inaccurate.

Disclosure of Invention

The application provides a data processing method, a data processing device, data processing equipment and a storage medium, which are used for overcoming the problem that the existing commodity data sorting result is inaccurate.

In a first aspect, a data processing method provided by the present application includes:

acquiring a recommended data set aiming at a target user;

sorting the commodity data in the recommendation data set according to the historical behavior data of the target user, the user association information and a preset sorting algorithm to obtain a data sorting result;

and determining at least one piece of commodity data displayed to the target user according to the data sorting result.

In this embodiment, when sorting the commodity data in the determined recommendation data set, the historical behavior data of the target user and the user association information of the target user are combined, so that the sorting accuracy of the commodity data is improved, the problem of inaccurate sorting result of the existing commodity is solved, the sorting accuracy in data sorting is improved, and the conversion rate of the commodity data is improved.

In a possible design of the first aspect, the sorting the commodity data in the recommended data set according to the historical behavior data of the target user, the user association information, and a preset sorting algorithm to obtain a data sorting result includes:

determining preference information of the target user according to the historical behavior data of the target user, wherein the preference information is used for indicating a selection result of the target user for any two pieces of commodity data in the recommendation data set;

acquiring an affinity value between the target user and a first user according to the user association information of the target user, wherein the first user is any one user in the address book of the target user;

and obtaining a data sorting result according to the preference information of the target user, the affinity values of the target user and the first user and a preset sorting algorithm.

In the embodiment, by introducing the preference information of the user in the address book of the target user, the problem that the data sorting result is inaccurate due to sparse historical behavior data of the target user is avoided.

Optionally, the method further includes:

determining penalty coefficients aiming at any two commodity data in the recommendation data set according to the preference information of the target user, wherein the penalty coefficients are used for correcting the initial preference probability existing between any two commodity data in the recommendation data set;

the obtaining a data sorting result according to the preference information of the target user, the affinity values of the target user and the first user and a preset sorting algorithm comprises:

and obtaining a data sorting result according to the preference information of the target user, the penalty coefficients aiming at any two commodity data in the recommended data set, the affinity values of the target user and the first user and a preset sorting algorithm.

In the embodiment, the influence of the characteristics of the commodity data on the target user can be reduced based on the penalty coefficient, the stickiness of the commodity data and the user can be improved based on the intimacy value of the target user and the first user, the sorting accuracy of the commodity data is improved, and the realization possibility is provided for subsequently improving the recommendation success probability of the commodity data.

Further, the obtaining a data sorting result according to the preference information of the target user, the penalty coefficients for any two pieces of commodity data in the recommended data set, the affinity values of the target user and the first user, and a preset sorting algorithm includes:

determining a first relevance value of each piece of commodity data in the recommendation data set and the target user through the sorting algorithm according to the preference information of the target user, the penalty coefficients aiming at any two pieces of commodity data and the affinity values of the target user and the first user;

and sequencing each commodity data in the recommendation data set according to the sequence of the first relevance value from high to low to obtain the data sequencing result.

In the embodiment, the potential relationship between the user and the commodity data can be well represented based on the first relevance value of each commodity data and the target user, the commodity data in the recommendation data set are sorted based on the obtained first relevance value, more appropriate commodity data can be recommended to the user, and the data sorting accuracy is improved.

In another possible design of the first aspect, the method further includes:

dividing the recommendation data set into at least two groups of recommendation data subsets according to the commodity category attribute of each commodity data in the recommendation data set, wherein the categories of the commodity data in each group of recommendation data subsets are consistent;

for each group of recommended data subset, determining an intra-group penalty coefficient for any two pieces of commodity data in the recommended data subset according to the preference information of the target user, wherein the intra-group penalty coefficient is used for correcting the initial preference probability existing between any two pieces of commodity data in the recommended data subset recommended data set;

obtaining an intra-group data sorting result in each group of recommended data subset according to the preference information of the target user, the intra-group penalty coefficients aiming at any two commodity data in the recommended data subset and a preset sorting algorithm;

determining a second correlation value between each group of recommended data subset and the target user according to the category of the commodity data in each group of recommended data subset and the intimacy value between the target user and the first user;

and obtaining a data sorting result aiming at the recommended data set according to the in-group data sorting result in each group of recommended data subset and the second correlation value between each group of recommended data subset and the target user.

In the embodiment, the recommended data sets are classified, and the commodity data of the same category are divided into one group, so that a more accurate data sorting result aiming at the recommended data sets can be obtained, and the recommendation accuracy is further improved.

In yet another possible design of the first aspect, the obtaining the recommended data set for the target user includes:

acquiring historical behavior data of the target user;

and analyzing the user behavior data to determine the recommended data set suitable for being recommended to the target user.

The user requirements of the target user can be determined by processing the historical behavior data of the target user, and then the recommended data set suitable for being recommended to the target user is determined based on the behavior habits of the target user or interested commodities and the like.

In yet another possible design of the first aspect, the method further includes:

determining the number of recommendation positions on a user interaction interface, wherein each recommendation position is used for presenting one piece of commodity data;

and filling at least one piece of commodity data sequenced at the front in the data sequencing result to each preset recommendation position for presentation, wherein the number of the at least one piece of commodity data is consistent with the number of the preset recommendation positions.

In the embodiment, each piece of data can be guaranteed to correspond to one recommendation bit, so that the corresponding commodity data can be timely and accurately displayed when a user accesses the webpage or the client, accurate recommendation of the commodity data is achieved, and possible conversion rate of the user to the commodity data is improved.

In a second aspect, the present application provides a data processing apparatus comprising: the device comprises an acquisition module, a sorting module and a processing module;

the acquisition module is used for acquiring a recommended data set aiming at a target user;

the sorting module is used for sorting the commodity data in the recommendation data set according to the historical behavior data of the target user, the user association information and a preset sorting algorithm to obtain a data sorting result;

and the processing module is used for determining at least one piece of commodity data displayed to the target user according to the data sorting result.

In a possible design of the second aspect, the ranking module is specifically configured to determine preference information of the target user according to historical behavior data of the target user, where the preference information is used to indicate a selection result of the target user for any two pieces of commodity data in the recommended data set, obtain an affinity value between the target user and a first user according to user association information of the target user, where the first user is any one user in an address book of the target user, and obtain a data ranking result according to the preference information of the target user, the affinity value between the target user and the first user, and a preset ranking algorithm.

Optionally, the sorting module is further configured to determine a penalty coefficient for any two pieces of commodity data in the recommended data set according to the preference information of the target user, where the penalty coefficient is used to correct an initial preference probability existing between any two pieces of commodity data in the recommended data set;

correspondingly, the sorting module is further specifically configured to obtain a data sorting result according to the preference information of the target user, the penalty coefficients for any two pieces of commodity data in the recommended data set, the affinity values of the target user and the first user, and a preset sorting algorithm.

Further, the sorting module is specifically configured to determine, according to the preference information of the target user, penalty coefficients for any two pieces of commodity data in the recommended data set, and affinity values of the target user and the first user, a first relevance value between each piece of commodity data in the recommended data set and the target user through the sorting algorithm, and sort each piece of commodity data in the recommended data set according to a sequence of the first relevance values from high to low, so as to obtain the data sorting result.

In another possible design of the second aspect, the sorting module is further configured to:

the sorting module is further specifically configured to obtain an intra-group data sorting result in each group of recommended data subsets according to the preference information of the target user, the intra-group penalty coefficients for any two pieces of commodity data in the recommended data subsets, and a preset sorting algorithm, determine a second correlation value between each group of recommended data subsets and the target user according to the category to which the commodity data in each group of recommended data subsets belongs and the intimacy value between the target user and the first user, and obtain a data sorting result for the recommended data set according to the intra-group data sorting result in each group of recommended data subsets and the second correlation value between each group of recommended data subsets and the target user.

In a further possible design of the second aspect, the obtaining module is specifically configured to obtain historical behavior data of the target user, analyze the user behavior data, and determine the recommendation data set suitable for being recommended to the target user.

In yet another possible design of the second aspect, the processing module is further configured to determine the number of recommendation bits on the user interaction interface, where each recommendation bit is used to present one piece of commodity data, and fill at least one piece of commodity data sorted before in the data sorting result to each preset recommendation bit for presentation, where the number of the at least one piece of commodity data is consistent with the number of preset recommendation bits.

The apparatus provided in the second aspect of the present application may be configured to perform the method provided in the first aspect, and the implementation principle and the technical effect are similar, which are not described herein again.

In a third aspect, the present application provides an electronic device comprising a processor, a memory, and a computer program stored on the memory and executable on the processor, wherein the processor executes the program to implement the method according to the first aspect as well as possible designs of the first aspect.

In a fourth aspect, the present application provides a computer-readable storage medium having stored therein instructions which, when run on a computer, cause the computer to perform the method as set forth in the first aspect and each possible design of the first aspect.

According to the data processing method, the data processing device, the data processing equipment and the storage medium, a recommended data set for a target user is obtained, the recommended data set is input into a sorting model, a data sorting result in the recommended data set is obtained, the sorting model is obtained according to historical behavior data of all users in a user sample set, user association information of all users and a preset sorting algorithm, and at least one piece of commodity data displayed to the target user is determined according to the data sorting result. According to the technical scheme, when the commodity data are sorted, the historical behavior data of the target user and the user association information of the target user are combined, so that the sorting accuracy of the commodity data is improved, and the problem that the existing commodity data sorting result is inaccurate is solved.

Drawings

Fig. 1 is a schematic flowchart of a first embodiment of a data processing method provided in the present application;

fig. 2 is a schematic flowchart of a second embodiment of a data processing method provided in the present application;

fig. 3 is a schematic flowchart of a third embodiment of a data processing method provided in the present application;

fig. 4 is a schematic flowchart of a fourth embodiment of a data processing method provided in the present application;

FIG. 5 is a schematic diagram illustrating an embodiment of the present application for determining at least one item of merchandise data to be displayed to a user;

fig. 6 is a schematic flowchart of a fifth embodiment of a data processing method provided in the present application;

FIG. 7 is a schematic structural diagram of a data device according to an embodiment of the present application;

fig. 8 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, 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, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.

With the rapid development of information technology, various web pages/client apps in the forms of commodity sales and content browsing are gradually increased, and due to the fact that commodity forms corresponding to commodity data are various and commodity contents are various, information overload is caused, and the time for a user to stay at the web pages/client apps is precious and short, the user can see interesting things at the first time to attract the user, the viscosity of the user is increased, the personalized requirements of the user are found by exploring the behaviors of the user, and therefore the commodity and the content are accurately recommended to the appropriate user.

Generally, the processing of the commodity data is generally divided into two steps, namely firstly screening out commodity data which may be interested by the user from a large quantity of commodity data sets, namely narrowing the range, and secondly sequencing the screened commodity data, so as to determine the sequence presented in the user field of view. It can be understood that the above two steps do not have to exist at the same time, for example, if the number of the commodities in the commodity data set is small, all the commodity data can be directly presented and sorted as the commodity data interested by the user; if the commodity data screened from the commodity data set correspond to a high accuracy rate, that is, the screened commodity data are highly interesting to the user, the determined commodity data do not need to be sorted, and any commodity data arranged at the first position of the presentation position may cause the conversion of the user.

The implementation scheme for sorting the screened commodity data can be realized by a sorting recommendation algorithm. The ranking recommendation algorithm can be roughly divided into three categories, the first category of ranking algorithm is a point pair method (Pointwise Approach), and the first category of ranking algorithm converts the ranking problem into problems such as classification and regression and is realized by processing the problems by using the existing methods such as classification and regression. The second type of ranking algorithm is the Pairwise Approach (Pairwise Approach), in which ranking is converted to a sort of sequences or regression of sequences. The pair ordering is called pair-wise, for example (a, b) indicates that a is earlier than b. The third type of sorting algorithm is the list Approach (Listwise Approach), which deals with the sorting problem in a more straightforward way, taking the sorted list as a sample during both learning and prediction, the sorted group structure being maintained.

The data processing method is mainly used for solving the problem that the commodity data sorting result determined in the prior art is possibly influenced by the interference of the characteristics of commodities, so that the commodity data sorting result is inaccurate. According to the technical scheme, when the commodity data are sorted, the historical behavior data of the target user and the user association information of the target user are combined, and the sorting accuracy of the commodity data is improved.

It can be understood that the execution subject of the embodiment of the present application may be an electronic device, for example, a terminal device such as a computer and a tablet computer, or may also be a server, for example, a background processing platform, and the like. The present embodiment is explained by using an electronic device, and the electronic device is a server in a normal case.

The technical solution of the present application will be described in detail below with reference to specific examples. It should be noted that the following specific embodiments may be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

Fig. 1 is a schematic flowchart of a first embodiment of a data processing method provided in the present application. As shown in fig. 1, the data processing method may include the steps of:

s101, acquiring a recommended data set aiming at a target user.

For example, in the embodiment of the present application, since the data amount on the website is huge and various, it is necessary to screen the commodity data related to the target user from the massive data and recommend the commodity data to the user, so as to improve the conversion rate of the user behavior.

For example, for mass commodity data on an e-commerce website, partial data can be screened from the mass commodity data according to historical behavior data of a user to serve as a recommended data set to be presented to a target user, so that the attention of the user is attracted, and the probability of purchasing the commodity data is improved.

As one implementation manner, the step S101 can be implemented by the following steps:

a1: acquiring historical behavior data of the target user;

a2: and analyzing the user behavior data to determine a recommended data set suitable for being recommended to the target user.

In this embodiment, a user usually searches, browses, purchases, and reviews some goods on a web page or an e-commerce website, and data generated by these behaviors of the user may be referred to as historical behavior data. By acquiring and analyzing the behavior data of the target user before purchasing some commodities, commodity data which may be interested by some users can be screened out from a large amount of commodity data in a targeted manner to serve as a recommendation data set suitable for being recommended to the target user.

For example, the server may analyze background data of the e-commerce website to obtain historical behavior data of the target user, such as information of goods browsed by the target user, retention time during browsing, purchase records of goods, and the like, determine user requirements of the target user by processing the obtained historical behavior data of the target user, for example, counting, classifying, summarizing, mining, and the like, then perform subsequent processing according to the processed data to determine behavior habits or interested goods of the target user, and finally determine a recommended data set suitable for being recommended to the target user.

S102, sorting the commodity data in the recommendation data set according to the historical behavior data of the target user, the user association information and a preset sorting algorithm to obtain a data sorting result.

In practical application, when determining preference relationships of users to different commodity data, because some commodity data may have partial order relationships among themselves, for example, a first commodity data "is naturally" more attractive to eyes of users than a second commodity data, for example, headline party articles in news content often can obtain preferences of most users relative to non-headline party articles, however, the obtained information is not real preference information of the users, but is something stimulating eyes And (5) sorting the product data.

In addition, for the situation that historical behavior data of some users are few, the ordering process of the commodity data is usually close to the cold start state, at this time, the prediction of the users is often not accurate enough, and for the problem, address book friends of the target user are introduced, and it is assumed that the target user and the address book friends have the same preference, so that the ordering accuracy of the commodity data is improved.

Specifically, in this embodiment, the preference degree of the target user for the commodity data may be analyzed according to the historical behavior data of the target user, the preference degree of other users in the address book of the target user for the commodity data is determined according to the user association relationship of the target user, and the commodity data in the recommended data set is sorted by using a preset sorting algorithm in combination with the preference degree of the target user for the commodity data and the preference degree of other users in the address book of the target user for the commodity data, so as to obtain a data sorting result.

Illustratively, in the present embodiment, the representation of the sorting algorithm is various, such as a point-to-point algorithm, a pair algorithm, a list algorithm, and the like. The embodiment of the present application does not limit the specific representation form of the ranking algorithm, and may be determined according to actual situations.

For example, in the present embodiment, the ranking algorithm may be a pairwise algorithm, and specifically, may be a Bayesian Personalized Ranking (BPR) algorithm.

For specific implementation of this step, reference may be made to the following description in the embodiment shown in fig. 2, and details are not described here.

S103, determining at least one piece of commodity data displayed to the target user according to the data sorting result.

Optionally, in this embodiment, for the determined data sorting result, the number of the data items to be displayed may be determined according to the number of the recommendation bits on the display interface, so that at least one data item that needs to be displayed to the target user is screened out from the recommendation data set according to the data sorting result.

According to the data processing method provided by the embodiment of the application, the recommended data set for the target user is obtained, the commodity data in the recommended data set are sequenced according to the historical behavior data of the target user, the user association information and the preset sequencing algorithm to obtain the data sequencing result, and finally at least one piece of commodity data displayed to the target user is determined according to the data sequencing result.

Exemplarily, on the basis of the above embodiments, fig. 2 is a schematic flow diagram of a second embodiment of the data processing method provided by the present application. As shown in fig. 2, in this embodiment, the step S102 may be implemented by:

s201, determining preference information of the target user according to the historical behavior data of the target user.

The preference information is used for indicating the selection result of the target user aiming at any two commodity data in the recommendation data set.

In the embodiment of the application, the historical behavior data of the target user is analyzed, and if the user obtains two commodity data simultaneously, the preference information of the target user is determined according to the selection result of the user in the two commodity data.

For example, if the target user selects the item data i when the item data i and the item data j are presented simultaneously, it may indicate that the target user has a preference degree for the item data i better than that for the item data j.

S202, according to the user association information of the target user, acquiring the intimacy value between the target user and a first user, wherein the first user is any one user in the address book of the target user.

In this embodiment, according to the user association information of the target user, the address book of the target user is determined first, and then the affinity value between the target user and each user in the address book is determined based on the association relationship between the target user and each user in the address book.

For example, the higher the affinity value of the target user and the user in the address book is, the more similar the preference of the target user and the user in the address book is, and therefore, when the historical behavior data amount of the target user is smaller, the historical behavior data of the user in the address book with the highest affinity with the target user can be used as the sorting basis, so that the sorting accuracy of the commodity data can be improved.

S203, obtaining a data sorting result according to the preference information of the target user, the affinity values of the target user and the first user and a preset sorting algorithm.

In the embodiment of the application, after the preference information of the target user is determined according to the historical behavior data of the target user, the preference information of the first user is determined according to the intimacy degree value of the target user and the first user, the preference information of the target user and the preference information of the first user are jointly used as the constraint conditions of a preset sorting algorithm, the recommended data set of the target user is subjected to constraint sorting by the aid of the preset sorting algorithm, and a data sorting result can be obtained.

According to the data processing result provided by the embodiment of the application, the preference information of the target user is determined according to the historical behavior data of the target user, the intimacy value between the target user and the first user is obtained according to the user association information of the target user, the first user is any user in the address book of the target user, and finally the data sorting result is obtained according to the preference information of the target user, the intimacy value between the target user and the first user and a preset sorting algorithm.

Further, on the basis of the embodiment shown in fig. 2, fig. 3 is a schematic flow chart of a third embodiment of the data processing method provided by the present application. As shown in fig. 3, in this embodiment, the method may further include the following steps:

s301, determining a penalty coefficient aiming at any two commodity data according to the preference information of the target user, wherein the penalty coefficient is used for correcting the initial preference probability existing between any two commodity data.

In practical application, as some commodity data on the e-commerce website are easier to attract the attention of users than other commodity data, and the preference degree of the users to the commodities is the same as that of the other commodity data, the method can weaken the difference value of the partial order probability between the commodity data and enable the commodity data to have less difference, so as to more accurately mine the real preference information of the users, can define the penalty coefficient aiming at any two commodity data according to the preference information of the target users, and when the commodity data in the recommendation data set are sorted by using a preset sorting algorithm, the penalty coefficient is used for correcting the initial preference probability existing between any two commodity data.

For example, referring to fig. 3, this step may be located after S201, that is, after determining the preference information of the target user, defining a penalty coefficient based on the preference information of the target user.

Accordingly, the above S203 may be replaced by the following steps:

s302, obtaining a data sorting result according to preference information of the target user, penalty coefficients aiming at any two commodity data, affinity values of the target user and the first user and a preset sorting algorithm.

In this embodiment, after the penalty coefficients for any two pieces of commodity data of the target user are introduced, the preference information of the target user and the intimacy value between the target user and the first user are used as the constraint conditions of the sorting algorithm, and the penalty coefficients for any two pieces of commodity data can also be used as another constraint condition, so as to further improve the output accuracy of the sorting algorithm, thereby obtaining a data sorting result with higher sorting accuracy.

As an example, the step S302 may be specifically implemented as follows:

b1: and determining a first relevance value of each piece of commodity data and the target user in the recommendation data set through the sorting algorithm according to the preference information of the target user, the penalty coefficients aiming at any two pieces of commodity data and the affinity values of the target user and the first user.

B2: and sequencing each commodity data in the recommendation data set according to the sequence of the first relevance value from high to low to obtain a data sequencing result.

In this embodiment, the sorting algorithm may correspond to a data sorting model, and the data sorting model may process the commodity data in the input recommendation data set based on the relevance between the commodity data and the target user, and first determine a first relevance value between each commodity data in the recommendation data set and the target user. The first relevance value can be used for representing the interest degree of each commodity data of the target user, and the higher the first relevance value of the commodity data and the target user is, the more interest of the commodity data of the target user is indicated.

Therefore, in this embodiment, the commodity data in the recommendation data set may be sorted based on the first relevance value of the commodity data and the target user, and specifically, each commodity data in the recommendation data set is sorted from high to low to obtain a data sorting result.

The potential relation between the user and the commodity data can be well represented based on the first relevance value of each commodity data and the target user, the commodity data in the recommendation data set are sorted based on the obtained first relevance value, more appropriate commodity data can be recommended to the user, and the data sorting accuracy is improved.

According to the data processing method provided by the embodiment of the application, when the penalty coefficients aiming at any two commodity data are determined according to the preference information of the target user, the penalty coefficients aiming at any two commodity data, the affinity values of the target user and the first user and the preset sorting algorithm can be used for obtaining the data sorting result, namely, the influence of the characteristics of the commodity data on the target user can be reduced based on the penalty coefficients, the viscosity between the commodity data and the user can be improved based on the affinity values of the target user and the first user, the sorting accuracy of the commodity data is improved, and the possibility of realizing the follow-up improvement of the recommendation success probability of the commodity data is provided.

Further, on the basis of the embodiment shown in fig. 2, fig. 4 is a schematic flow chart of a fourth embodiment of the data processing method provided by the present application. As shown in fig. 4, in this embodiment, the method may further include the following steps:

s401, according to the commodity category attribute of each commodity data in the recommendation data set, dividing the recommendation data set into at least two groups of recommendation data subsets, wherein the categories of the commodity data in each group of recommendation data subsets are consistent.

In this embodiment, because the data volume on the website is huge and the categories are various, the commodity data in the obtained recommendation data set may belong to various categories, and in order to determine the commodity data really interested by the user in the subsequent pertinence, the determined recommendation data set may be further grouped.

For example, the recommendation data set is divided according to the commodity category attribute of each piece of commodity data in the recommendation data set, so that various recommendation data subsets can be obtained, for example, recommendation data subsets of categories such as "clothing", "grain and oil", "food", and the like.

It should be noted that, in this embodiment, the number of groups of the recommended data subset is not limited, and the specific category of the recommended data is not limited, which may be divided according to the actual situation, and is not described herein again.

S402, for each group of recommended data subset, determining punishment coefficients in the group aiming at any two commodity data in the recommended data subset according to the preference information of the target user.

And the penalty coefficient in the group is used for correcting the initial preference probability existing between any two commodity data in the recommended data subset recommended data set.

It is understood that this step is similar to the way of determining the penalty coefficients of any two pieces of commodity data in the recommendation data set in S301, except that S301 is a penalty coefficient for determining any two pieces of commodity data in the recommendation data set, and S402 is a recommendation data subset obtained by dividing the recommendation data set, and is an intra-group penalty coefficient for determining any two pieces of commodity data in the recommendation data subset. For an implementation manner of determining the penalty coefficients in the group in this step, reference may be made to the above description in S301, and details are not described here.

Further, in this embodiment, referring to fig. 4, the step S203 may be replaced by the following steps:

s403, obtaining an intra-group data sorting result in each recommended data subset according to the preference information of the target user, the intra-group penalty coefficients aiming at any two commodity data in the recommended data subset and a preset sorting algorithm.

In this embodiment, after the intra-group penalty coefficients for any two pieces of commodity data in the recommended data subset are introduced, for each group of recommended data subset, the preference information of the target user and the intra-group penalty coefficients of the group of recommended data subset may be used as constraint conditions of a ranking algorithm to jointly determine a ranking result of the intra-group data.

In this embodiment, the sorting algorithm may first determine, based on the preference information of the target user and the intra-group penalty coefficient for any two pieces of commodity data in the recommended data subset, a first relevance value between each piece of commodity data in the recommended data subset and the target user, and then sort the commodity data in the recommended data subset based on the first relevance value between the commodity data and the target user, so as to obtain an intra-group data sorting result.

The specific implementation of this step is similar to the implementation of S302 described above, and is not described here again.

S404, determining a second correlation value between each group of recommended data subset and the target user according to the category of the commodity data in each group of recommended data subset and the intimacy value between the target user and the first user.

In this embodiment, in order to improve the accuracy of the determined data sorting result, a second correlation value between each group of recommended data subsets and the target user may be calculated based on the intimacy values of the target user and the first user. For example, when the target user has substantially the same preference for the "food" recommendation data subset and the "game" recommendation data subset, but if the affinity value of the target user to the first user is higher, and the preference of the first user for the "food" recommendation data subset is much higher than the "game" recommendation data subset, then for the target user, in the "food" recommendation data subset and the "game" recommendation data subset, it may be determined that the second relevance value of the "food" recommendation data subset to the target user is higher than the second relevance value of the "game" recommendation data subset to the target user.

It is understood that the similarity contrast manner is similar for the recommendation data subsets of other categories, and the detailed description is omitted here. The higher the intimacy value between the target user and the first user is, the more accurate the inter-group data sorting result is determined according to the preference of the first user.

S405, obtaining a data sorting result aiming at the recommended data set according to the in-group data sorting result in each group of recommended data subset and the second correlation value between each group of recommended data subset and the target user.

In this embodiment, after the intra-group data sorting result in each group of recommended data subsets and the second correlation value between each group of recommended data subsets and the target user are determined, the commodity data of the recommended data sets may be sorted based on the number of recommendation bits and based on a preset sorting manner.

For example, the sorting manner may be that, according to the sorting result (for example, from high to low) of the components in each group of recommendation subsets, a preset amount of data is sequentially taken out from each group of recommendation subsets to be sorted until the commodity data in all categories of recommendation subsets participate in the sorting.

It should be noted that the quantity of the commodity data taken out from each group of the recommended data subset may be the same or different, and the embodiment of the present application does not limit the quantity of the commodity data, and the quantity of the commodity data may be set according to actual requirements.

The data processing method provided by this embodiment divides the recommended data set into at least two groups of recommended data subsets according to the commodity category attribute of each piece of commodity data in the recommended data set, determines an intra-group penalty coefficient for any two pieces of commodity data in the recommended data subsets according to the preference information of the target user for each group of recommended data subsets, further may obtain an intra-group data sorting result in each group of recommended data subsets according to the preference information of the target user, the intra-group penalty coefficient for any two pieces of commodity data in the recommended data subsets, and a preset sorting algorithm, determines a second correlation value between each group of recommended data subsets and the target user according to the category to which the commodity data in each group of recommended data subsets belong, and the intimacy values between the target user and the first user, and finally may determine a second correlation value between each group of recommended data subsets and the target user according to the intra-group data sorting result in each group of recommended data subsets and the intra-group recommended data subsets and the target data subsets and the And obtaining a data sorting result aiming at the recommended data set by the second relevance value of the user. According to the technical scheme, the recommended data sets are classified, and the commodity data of the same category are divided into one group, so that a more accurate data sorting result aiming at the recommended data sets can be obtained, and the recommendation accuracy is further improved.

Illustratively, on the basis of the above embodiments, fig. 5 is a schematic diagram illustrating that at least one piece of merchandise data displayed to a user is determined in the embodiment of the present application. As shown in fig. 5, in this embodiment, first, historical behavior data of a target user is obtained, preference information and penalty coefficients for any two pieces of commodity data are determined, second, user association information of the target user is obtained, an intimacy degree value between the target user and a first user is determined, the first user is any one user in an address book of the target user, and finally, a preset ranking algorithm is used to obtain a data ranking result by combining the preference information of the target user, the penalty coefficients for any two pieces of commodity data, and the intimacy degree value between the target user and the first user.

On the basis of the above embodiments, when the present embodiment ranks the commodity data in the recommendation data set by using the bayesian personalized ranking BPR algorithm, the specific implementation scheme is as follows:

first, the implementation principle of the BRP algorithm will be explained. Specifically, the BPR algorithm is a pair-wise sorting algorithm, which can solve the preference degree of the user u for the commodity data i from the user u, the commodity data i and the commodity data j. In the BPR algorithm, the relationship between any user u and the commodity data may be labeled, and if the commodity data i and the commodity data j are presented simultaneously, the user u selects the commodity data i, at this time, a triple < u, i, j > may be obtained, which indicates that the preference degree of the user u for the commodity data i is better than that for the commodity data j. Thus, in the present embodiment, the triple < u, i, j > is used to indicate that the target user prefers the item data i among the item data i and the item data j.

By analyzing the historical behavior data of the target user, a plurality of triples related to commodity data sorting can be obtained, correspondingly, preference information of the target user for the commodity data in the recommended data set can be obtained, and at this time, a data sorting model for the target user u can be trained.

It is worth noting that in the embodiment of the present application, two assumption conditions need to be set when training the data ranking model for the target user u based on the BPR algorithm:

assume that 1: the preference behaviors of each user are independent of each other and do not influence each other, namely the preference of the user u between the commodity data i and the commodity data j is irrelevant to other users.

Assume 2: the partial orders of different commodity data of the same user are independent of each other, namely the preference of the user u between the commodity data i and the commodity data j is independent of other articles.

In this embodiment, if the target user u is more interested in the commodity data i than the commodity data j, the target user u may use the commodity data i<u,i,j>Representation while introducing preference compliance>_uThen, then<u,i,j>Can be represented as i>_uj. In this embodiment, all known user preference data of the target user may be represented by the historical behavior data set D, all known user preference data (u, i, j) being present in D.

In the embodiment of the present application, since the user set U and the article set I are both sets with large sample numbers, in order to simplify the computational complexity, the user matrix W corresponding to the user set U may be represented as | U | × k, and the article matrix H corresponding to the article set I may be represented as | I | × k, where | U | is the number of users, | I | is the number of articles, and k is a custom intermediate matrix, the number of columns of the intermediate matrix k is the same as the number of rows of | U |, the number of rows of the intermediate matrix k is the same as the number of columns of | I |, and the dimension of the intermediate matrix k is generally much smaller than | U | and | I |.

Illustratively, the BPR algorithm is the same as matrix decomposition, and the preference ranking result of the user set U and the item set I can be represented by a user preference matrix X obtained by UxI, and correspondingly, a prediction ranking matrix

Can be represented by an article matrix H and a user matrix W, and satisfies

Wherein H^TIs the transpose of matrix H. The goal of this embodiment is to try to make the prediction ordering matrix

Tending towards the actual user preference matrix X.

In this embodiment, the preference degree of the target user u for any item data i may be expressed as:

in the embodiments of the present application, the target user uses u_uIndicates that the first user is u_lIndicates the first user u_lBelonging to a target user u_uContact list (also called social circle) F_uThen, then

The target user u_uWith the first user u_lThe intimacy value of (a) can be expressed as:

wherein s is_ulRepresenting a target user u_uWith the first user u_lCoefficient of intimacy of(s)_ulIs (0,1), w_uRepresenting user u_uFeature vectors in the user matrix W, W_lRepresenting user u_lFeature vectors in the user matrix W, W_ufAnd w_lfRespectively represents w_uAnd w_lK represents the number of the feature vectors in the user matrix W, f represents the serial number of each element in the feature vectors, and the numeric range of f is [1, k ]]。

In an embodiment of the application, the BPR ranking algorithm is based on a maximum a posteriori estimate P (W, H>_u) And solving parameters W and H of the data sequencing model. In this embodiment, the parameters W, H may be represented by a parameter θ, and thus, P (W, H>_u) Can be replaced by P (theta>_u) Correspondingly, according to the Bayesian formula, the target user estimates the P (W, H) according to the maximum a posteriori that the sorting algorithm needs to solve>_u) May be as follows:

wherein P (theta>_u) Represents the probability that the parameter θ holds under the condition that all the user preference data are known, P: (>_u| θ) represents the probability of the occurrence of currently known user preference data under the condition of a parameter θ, P (θ) represents the prior probability of the parameter θ, P: (θ) ((P)>_u) Representing the prior probability of currently known user preference data.

Due to P: (>_u) Is the same for all articles, so P (theta tint)>_u) And P: (>_u| θ) P (θ) has the following relationship:

P(θ|>_u)∝P(>_u|θ)P(θ)

thus, in the present embodiment, the maximum posterior estimate P (W, H>_u) The optimization objective of (2) can be divided into two parts, wherein the first part P (b)>_u| θ) is related to the historical behavior data set D of the target user, and the second portion P (θ) is not related to the historical behavior data set D of the target user.

In the present embodiment, for the first part P: (>_u| θ), since the partial orders of different items by the same user are independent from each other, the maximum bayesian estimation for all users in the user set U has a relationship:

wherein, P (i)>_uj | θ) represents the probability that the target user u is better than the commodity data j than the commodity data i under the condition that the parameter is θ,

wherein (u, i, j) ∈ D indicates that the user preference data (u, i, j) is in the historical behavior data set D of the user u.

Due to the symbol>_uSatisfaction of integrity

And antisymmetry

Thus, n is_u∈UP(>_u| θ) can be simplified as follows:

when target user u_uWhen the behavior data of the user u is sparse, the target user u is selected_uWith the first user u_lS (u) of_u,u_l) Assigning the target user u with the preference data of the largest user l_uThen the above formula is | /)_u∈UP(>_u|θ)＝∏_(u,i,j)∈DP(i>_uj | θ) can be replaced by the following equation:

wherein s (u, l) is s (u) as described above_u,u_l)。

The P (i) in the formula>_uj | θ) and P (i)>_lj | θ) and a prediction ordering matrix

The following formula can be obtained if the element values of (1) are corresponding to each other:

wherein the content of the first and second substances,

representing target users u in a predicted user preference matrix_uThe number of differences between preference data of commodity data i and preference data of commodity data j, sigma is sigmoid function, and can be represented by formula

And (4) showing.

In particular, for those in the above formula

The following properties should be provided: when i is>_uAt the time of j, the number of the first,

on the contrary, when j>_uWhen the time is in the range of i,

the simplest way to express this property is:

wherein the content of the first and second substances,representing target users u in a predicted user preference matrix_uThe total number of the favorite commodity data i,

representing target users u in a predicted user preference matrix_uTotal number of the favorite product data j.

In this embodiment, after determining the penalty coefficient for any two pieces of commodity data according to the preference information of the target user, the initial preference probability existing between any two pieces of commodity data is corrected.

Illustratively, the penalty factor is defined as follows: suppose user u_oThe commodity data i and the commodity data j have a partial order relation i>_uoj, and user u_o∈U_ijAt this time, user u can be obtained_oPenalty coefficients for commodity data i and commodity data j:

wherein, U_ijNum representing a set of users who have contacted the item data i and the item data j at the same time_ijIs U_ijThe number of partial orders of the commodity data i is preferred to the commodity data j.

Thus, the above

Can be expressed by the following formula:

similarly, it can be determined

Representing a predicted user preference matrix user u_lThe number of differences between the preference data of the commodity data i and the preference data of the commodity data j,

representing a predicted user preference matrix user u_lThe total number of the favorite commodity data i,

representing a predicted user preference matrix user u_lTotal number of the favorite product data j.

Correspondingly, in the above-mentioned pair II_u∈UP(>_u| θ) can be expressed by the following formula:

for the second part P (θ), according to bayesian estimation, the probability distribution P (θ) is assumed to conform to a multivariate normal distribution with a mean value of 0, and based on the basic assumptions of BPR, i.e. 1) the preference behavior of each user is independent and 2) the partial order of the same user for different commodities is independent, so the covariance matrix of the distribution is the identity matrix I, i.e. P (θ) -N (0, I), according to the standard multivariate normal distribution function

It can be known that lnP (θ) and | | | θ | | ceiling²Linear relation, without lnP (| | | | θ |) being arranged²Where λ is a constant coefficient and | θ | is the L2 norm of θ.

In summary, Bayesian estimation P (>_u|θ)∝lnP(>_uThe maximum logarithm posterior estimating function corresponding to | θ) P (θ) can be expressed by the following formula:

in this equation, θ is derived by the gradient ascent method to yield:

in this embodiment, since Then

Thereby obtaining the gradient factor of each parameter,

solution method of and

the same is that:

thus, according to

Continuously iterating until convergence to obtain a user matrix W (| U |. x k) and an article matrix H (| I |. x k), and further obtain a target user U_uAnd finally selecting a plurality of commodity data with high sorting scores for the sorting scores of any commodity data i for recommendation.

To sum up, on the basis of the original BPR algorithm, by introducing a punishment coefficient between any two commodity data, some interference factors of the commodity data are eliminated, the accuracy of commodity data preferred by a user is improved, the historical behavior data of a target user is enriched by introducing user intimacy, and the problem that the historical user behavior data is sparse and the interest of the user is not accurately mined is solved.

Further, in any of the above embodiments of the present application, fig. 6 is a schematic flowchart of a fifth embodiment of the data processing method provided by the present application. As shown in fig. 6, the method may further include the steps of:

s601, determining the number of recommendation positions on the user interaction interface, wherein each recommendation position is used for presenting one piece of commodity data.

It should be noted that the electronic device in this embodiment has a user interaction interface, and at least one recommendation bit is provided on the user interaction interface, so that when the commodity data needs to be recommended, the number of the recommendation bits on the user interaction interface is determined first.

Illustratively, the user interaction interface may be a presented web interface, a client interface, or other interface in different forms. Thus, the recommendation bit in this embodiment may be set by the developer when developing the client or designing the web page, which is usually at a fixed location of the client or web page.

S602, filling at least one piece of commodity data sequenced at the front in the data sequencing result to each preset recommendation position for presentation, wherein the number of the at least one piece of commodity data is consistent with the number of the preset recommendation positions.

In this embodiment, in general, each recommendation bit on a user interaction interface such as a web page or a client is used to fill one piece of commodity data, and therefore, after the number of recommendation bits on the user interaction interface is determined, a plurality of pieces of commodity data sorted in the data sorting result in the top may be sequentially filled in the corresponding recommendation bits according to the order in which the recommendation bits are presented to the user or the order in which the recommendation bits are more likely to be concerned by the user, so that when the user accesses the web page or the client, the commodity data filled in the recommendation bits is presented.

According to the data processing method provided by the embodiment of the application, the number of the recommendation positions on the user interaction interface is determined, each recommendation position is used for presenting one piece of commodity data, at least one piece of commodity data sequenced in the data sequencing result before is respectively filled to each preset recommendation position for presentation, the number of the at least one piece of commodity data is consistent with the number of the preset recommendation positions, and therefore each piece of data can be guaranteed to correspond to one recommendation position, a user can be guaranteed to timely and accurately present the corresponding commodity data when accessing the webpage or the client, accurate recommendation of the commodity data is achieved, and the possible conversion rate of the user to the commodity data is improved.

In the above, a specific implementation of the data processing method mentioned in the embodiment of the present application is introduced, and the following is an embodiment of the apparatus of the present application, which can be used to execute the embodiment of the method of the present application. For details which are not disclosed in the embodiments of the apparatus of the present application, reference is made to the embodiments of the method of the present application.

Fig. 7 is a schematic structural diagram of a data device according to an embodiment of the present application. The apparatus may be integrated in or implemented by an electronic device, which may be a server, or a terminal device. As shown in fig. 7, in the present embodiment, the data processing apparatus 70 may include: an obtaining module 701, a sorting module 702 and a processing module 703.

The obtaining module 701 is configured to obtain a recommendation data set for a target user;

the sorting module 702 is configured to sort the commodity data in the recommendation data set according to the historical behavior data of the target user, the user association information, and a preset sorting algorithm, so as to obtain a data sorting result;

the processing module 703 is configured to determine at least one piece of commodity data to be displayed to the target user according to the data sorting result.

In a possible design of the present application, the sorting module 702 is specifically configured to determine, according to the historical behavior data of the target user, preference information of the target user, where the preference information is used to indicate a selection result of the target user for any two pieces of commodity data in the recommended data set, and obtain, according to the user association information of the target user, an affinity value of the target user and a first user, where the first user is any one user in an address book of the target user, and obtain, according to the preference information of the target user, the affinity value of the target user and the first user, and a preset sorting algorithm, a data sorting result.

In this embodiment, the sorting module 72 is further configured to determine a penalty coefficient for the any two pieces of commodity data according to the preference information of the target user, where the penalty coefficient is used to correct an initial preference probability existing between the any two pieces of commodity data;

correspondingly, the sorting module 702 is further specifically configured to obtain a data sorting result according to the preference information of the target user, the penalty coefficients for any two pieces of commodity data, the affinity values of the target user and the first user, and a preset sorting algorithm.

Further, the sorting module 702 is further specifically configured to determine, according to the preference information of the target user, the penalty coefficients for any two pieces of commodity data, and the affinity values of the target user and the first user, a first relevance value between each piece of commodity data in the recommended data set and the target user through the sorting algorithm, and sort each piece of commodity data in the recommended data set according to a sequence of the first relevance values from high to low, so as to obtain the data sorting result.

In another possible design of this embodiment, the sorting module 702 is further configured to:

In yet another possible design of this embodiment, the obtaining module 701 is specifically configured to obtain historical behavior data of the target user, analyze the user behavior data, and determine the recommended data set suitable for being recommended to the target user.

In another possible design of this embodiment, the processing module 703 is further configured to determine the number of recommendation bits on the user interaction interface, where each recommendation bit is used to present one piece of commodity data, and fill at least one piece of commodity data sorted before in the data sorting result to each preset recommendation bit for presentation, where the number of the at least one piece of commodity data is consistent with the number of the preset recommendation bits.

The apparatus provided in the embodiment of the present application may be used to execute the method in the embodiments shown in fig. 1 to fig. 6, and the implementation principle and the technical effect are similar, which are not described herein again.

It should be noted that the division of the modules of the above apparatus is only a logical division, and the actual implementation may be wholly or partially integrated into one physical entity, or may be physically separated. And these modules can be realized in the form of software called by processing element; or may be implemented entirely in hardware; and part of the modules can be realized in the form of calling software by the processing element, and part of the modules can be realized in the form of hardware. For example, the sorting module may be a processing element that is separately set up, or may be implemented by being integrated into a chip of the apparatus, or may be stored in a memory of the apparatus in the form of program code, and a processing element of the apparatus calls and executes the functions of the sorting module. Other modules are implemented similarly. In addition, all or part of the modules can be integrated together or can be independently realized. The processing element described herein may be an integrated circuit having signal processing capabilities. In implementation, each step of the above method or each module above may be implemented by an integrated logic circuit of hardware in a processor element or an instruction in the form of software.

For example, the above modules may be one or more integrated circuits configured to implement the above methods, such as: one or more Application Specific Integrated Circuits (ASICs), or one or more microprocessors (DSPs), or one or more Field Programmable Gate Arrays (FPGAs), among others. For another example, when some of the above modules are implemented in the form of a processing element scheduler code, the processing element may be a general-purpose processor, such as a Central Processing Unit (CPU) or other processor that can call program code. As another example, these modules may be integrated together, implemented in the form of a system-on-a-chip (SOC).

In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loaded and executed on a computer, cause the processes or functions described in accordance with the embodiments of the application to occur, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, from one website site, computer, server, or data center to another website site, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that incorporates one or more of the available media. The usable medium may be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., Solid State Disk (SSD)), among others.

Fig. 8 is a schematic structural diagram of an electronic device according to an embodiment of the present application. The electronic device may be implemented by a server or a terminal device. As shown in fig. 8, the electronic device may include: a processor 801, a memory 802, a communication interface 803 and a system bus 804, wherein the memory 802 and the communication interface 803 are connected with the processor 801 through the system bus 804 and perform communication with each other, the memory 802 is used for storing computer execution instructions, the communication interface 803 is used for communicating with other devices, and the processor 801 executes the computer programs to realize the scheme of the embodiment shown in fig. 1 to 6.

The system bus mentioned in fig. 8 may be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The system bus may be divided into an address bus, a data bus, a control bus, and the like. For ease of illustration, only one thick line is shown, but this does not mean that there is only one bus or one type of bus. The communication interface is used for realizing communication between the database access device and other equipment (such as a client, a read-write library and a read-only library). The memory may comprise Random Access Memory (RAM) and may also include non-volatile memory (non-volatile memory), such as at least one disk memory.

The processor may be a general-purpose processor, including a central processing unit CPU, a Network Processor (NP), and the like; but also 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, discrete hardware components.

Optionally, an embodiment of the present application further provides a computer-readable storage medium, where instructions are stored in the computer-readable storage medium, and when the instructions are executed on a computer, the computer is caused to execute the method according to the embodiment shown in fig. 1 to 6.

Optionally, an embodiment of the present application further provides a chip for executing the instruction, where the chip is used to execute the method in the embodiment shown in fig. 1 to 6.

The embodiment of the present application further provides a program product, where the program product includes a computer program, where the computer program is stored in a storage medium, and the computer program can be read from the storage medium by at least one processor, and when the computer program is executed by the at least one processor, the method of the embodiment shown in fig. 1 to 6 can be implemented.

In the present application, "at least one" means one or more, "a plurality" means two or more. "and/or" describes the association relationship of the associated objects, meaning that there may be three relationships, e.g., a and/or B, which may mean: a exists alone, A and B exist simultaneously, and B exists alone, wherein A and B can be singular or plural. The character "/" generally indicates that the former and latter associated objects are in an "or" relationship; in the formula, the character "/" indicates that the preceding and following related objects are in a relationship of "division". "at least one of the following" or similar expressions refer to any combination of these items, including any combination of the singular or plural items. For example, at least one (one) of a, b, or c, may represent: a, b, c, a-b, a-c, b-c, or a-b-c, wherein a, b, c may be single or multiple.

It is to be understood that the various numerical references referred to in the embodiments of the present application are merely for descriptive convenience and are not intended to limit the scope of the embodiments of the present application. In the embodiment of the present application, the sequence numbers of the above-mentioned processes do not mean the execution sequence, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation to the implementation process of the embodiment of the present application.

Finally, it should be noted that: the above embodiments are only used for illustrating the technical solutions of the present application, and not for limiting the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A data processing method, comprising:

acquiring a recommended data set aiming at a target user;

2. The method according to claim 1, wherein the sorting the commodity data in the recommended data set according to the historical behavior data of the target user, the user association information and a preset sorting algorithm to obtain a data sorting result comprises:

3. The method of claim 2, further comprising:

4. The method according to claim 3, wherein the obtaining a data ranking result according to the preference information of the target user, the penalty coefficients for any two pieces of commodity data in the recommended data set, the affinity values of the target user and the first user, and a preset ranking algorithm comprises:

5. The method of claim 2, further comprising:

6. The method according to any one of claims 1-5, wherein the obtaining of the set of recommendation data for the target user comprises:

acquiring historical behavior data of the target user;

7. The method according to any one of claims 1-5, further comprising:

8. A data processing apparatus, comprising: the device comprises an acquisition module, a sorting module and a processing module;

9. The apparatus of claim 8, wherein the ranking module is specifically configured to determine preference information of the target user according to historical behavior data of the target user, where the preference information is used to indicate a selection result of the target user for any two pieces of commodity data in the recommended data set, obtain an affinity value of the target user and a first user according to user association information of the target user, where the first user is any one of users in an address book of the target user, and obtain a data ranking result according to the preference information of the target user and an affinity preset ranking algorithm of the target user and the first user.

10. The device according to claim 9, wherein the ranking module is further configured to determine a penalty coefficient for any two pieces of commodity data in the recommended data set according to the preference information of the target user, where the penalty coefficient is used to correct an initial preference probability existing between any two pieces of commodity data in the recommended data set;

11. The apparatus according to claim 10, wherein the sorting module is further specifically configured to determine, according to the preference information of the target user, a penalty coefficient for any two pieces of commodity data in the recommended data set, and a affinity value of the target user and the first user, a first relevance value between each piece of commodity data in the recommended data set and the target user through the sorting algorithm, and sort each piece of commodity data in the recommended data set in an order from high to low according to the first relevance value, so as to obtain the data sorting result.

12. The apparatus of claim 9, wherein the ordering module is further configured to:

13. The apparatus according to any one of claims 8 to 12, wherein the obtaining module is specifically configured to obtain historical behavior data of the target user, analyze the user behavior data, and determine the recommendation data set suitable for recommendation to the target user.

14. The apparatus according to any one of claims 8 to 12, wherein the processing module is further configured to determine the number of recommendation bits on the user interaction interface, each recommendation bit is used to present one piece of commodity data, at least one piece of commodity data ranked in the data ranking result before is respectively filled to each preset recommendation bit for presentation, and the number of the at least one piece of commodity data is consistent with the number of preset recommendation bits.

15. An electronic device comprising a processor, a memory and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the method according to any of the claims 1-7 when executing the program.

16. A computer-readable storage medium having stored therein instructions which, when executed on a computer, cause the computer to perform the method of any one of claims 1-7.