CN112102039A

CN112102039A - Article pushing method, computer system and storage medium

Info

Publication number: CN112102039A
Application number: CN202011011963.3A
Authority: CN
Inventors: 徐勐; 张珊; 王诚
Original assignee: Hangzhou Miya Information Technology Co ltd
Current assignee: Hangzhou Miya Information Technology Co ltd
Priority date: 2020-09-23
Filing date: 2020-09-23
Publication date: 2020-12-18
Anticipated expiration: 2040-09-23
Also published as: CN112102039B

Abstract

The present disclosure provides an article pushing method, a computer system and a storage medium; the article pushing method comprises the following steps: determining historical user information and historical article information; determining a target object corresponding to a target user according to the historical user information and the historical object information; pushing the target item to the target user.

Description

Article pushing method, computer system and storage medium

Technical Field

The present disclosure relates to the field of machine learning technologies, and in particular, to an article pushing method, a computer system, and a storage medium.

Background

Currently, the mainstream commodity recommendation method in the e-commerce field is Collaborative Filtering (Collaborative Filtering). The premise of collaborative filtering is that similar items may be preferred by similar people, assuming similar users may have similar preferences. Collaborative filtering then seeks similar users or similar items.

Collaborative filtering typically has two directions, an item-based collaboration and a user-based collaboration volume. One of the two directions is focused on articles, and the other direction is focused on users, so that the selection of the users and the selection of the users in the situation with large commodity quantity becomes an important problem. In addition, the existing collaborative filtering recommendation method has the following problems: (1) for a newly appeared shopper, if only one purchase record is available, a recommended commodity cannot be provided; (2) some commodities are few in purchase quantity and difficult to obtain recommendations; (3) the subsequent upgrade of the commodity recommendation system adopting collaborative filtering is difficult.

Disclosure of Invention

Technical problem to be solved

In view of the above problems, it is a primary object of the present disclosure to provide an article pushing method, a computer system, and a storage medium, so as to solve at least one of the above problems.

(II) technical scheme

According to an aspect of the present disclosure, there is provided an article pushing method including:

determining historical user information and historical article information;

determining a target object corresponding to a target user according to the historical user information and the historical object information;

pushing the target item to the target user.

Further, the historical user information comprises a relation matrix V of historical users and historical articles^U，IRelation matrix V of historical articles and historical users^I，U(ii) a The historical article information comprises a relation matrix V of the article selected at this time and the article selected at last time^I，LRelationship matrix V of last selected item and present selected item^L，I。

Further, the historical user information also comprises a relation matrix V of the historical users and external data^U，KRelationship matrix V of external data and historical users^K，U(ii) a The historical item information also comprises a relation matrix V of the historical items and external data^I，KRelationship matrix V of external data and historical articles^K，IThe relation matrix V of the selected article and the selected article n times before the last time^I，L+1、V^I，L+2、…、V^I，L+nThe relation matrix V of the article selected n times before the last time and the article selected this time^L+1，I、V^L+2，I、…、V^L+n，I(ii) a Wherein the external data comprises weather, holidays, public opinion data.

Further, determining a target item corresponding to the target user according to the historical user information and the historical item information includes:

according to the relation matrix V of the historical users and the historical articles^U，IRelation matrix V of historical articles and historical users^I，UDetermining a matrix decomposition vector MF;

according to the relation matrix V of the selected article and the article selected last time^I，LRelation matrix V of last selected article and this selected article^L，IDetermining a decomposed non-personalized Markov chain vector (FMC);

determining a probability vector p according to the vector MF and the vector FMC;

and determining a target item corresponding to the target user according to the probability vector p.

Further, according to the relation matrix V of the historical users and the historical articles^U，IRelation matrix V of historical articles and historical users^I，UDetermining a matrix decomposition vector, MF, comprising:

according to the relation matrix V of the historical users and the historical articles^U，IRelation matrix V of historical articles and historical users^I，UDetermining a matrix VUI _ m _ VIU; and

the vector MF is determined from the matrix VUI _ m _ VIU.

Further, according to the relation matrix V of the selected article and the selected article last time^I，LRelation matrix V of last selected article and this selected article^L，IDetermining a vector FMC, comprising:

according to the relation matrix V of the selected article and the article selected last time^I，LRelation matrix V of last selected article and this selected article^L，IDetermining a matrix VIL _ m _ VLI; and

the vector FMC is determined from the matrix VIL _ m _ VLI.

Further, the matrix VIL _ m _ VLI, the matrix VUI _ m _ vu, the vector MF, the vector FMC, and the probability vector p satisfy the following relations:

VIL_m_VLI＝V^I，L·V^L，I；

VUI_m_VIU＝V^U，I·V^I，U；

MF＝VUI_m_VIU_i；

FMC＝mean(VIL_m_VLI，)＝mean(VIL_m_VLI，{x，y，z})；

p＝MF+FMC；

the item set selected this time is represented, where { x, y, z }, and x, y, z represent each element in the item set selected this time.

Further, determining a target item corresponding to the target user according to the probability vector p includes:

presetting a probability threshold;

and determining the item corresponding to the element with the probability larger than the probability threshold value in the vector p as a target item corresponding to a target user.

According to another aspect of the present disclosure, there is provided a computer system including:

one or more processors;

a storage device for storing one or more programs,

wherein the one or more programs, when executed by the one or more processors, cause the one or more processors to perform the method.

According to another aspect of the disclosure, there is provided a computer-readable storage medium having stored thereon executable instructions that, when executed by a processor, cause the processor to perform the method described.

(III) advantageous effects

According to the technical scheme, the article pushing method, the computer system and the storage medium disclosed by the invention have at least one of the following beneficial effects:

(1) the method and the system are suitable for the shopper who only has one-time purchase record, and can recommend relatively reasonable target articles to a new shopper.

(2) The present disclosure also has the possibility of obtaining a high purchase probability for goods with a low purchase quantity, without completely shielding the goods with a low purchase quantity.

(3) The method and the system can supplement external data information and multiple shopping information before the shopping as required, and improve the accuracy of recommendation.

Drawings

The accompanying drawings, which are included to provide a further understanding of the disclosure, illustrate embodiments of the disclosure and together with the description serve to explain the disclosure and are not to limit the disclosure. In the drawings:

fig. 1 schematically shows a flow chart of an item recommendation method according to an embodiment of the present disclosure.

Fig. 2 schematically illustrates a flowchart of determining a target item corresponding to a target user according to the historical user information and the historical item information according to an embodiment of the present disclosure.

FIG. 3 schematically shows a relationship matrix V according to historical users and historical items according to an embodiment of the disclosure^U，IRelation matrix V of historical articles and historical users^I，UA flow chart of a matrix decomposition vector MF is determined.

FIG. 4 schematically illustrates a relationship matrix V according to a present selection of an item and a last selection of an item according to an embodiment of the disclosure^I，LRelation matrix V of last selected article and this selected article^L，IFlow chart for determining vector FMC.

FIG. 5 schematically illustrates a flow chart for determining a target item corresponding to a target user based on a probability vector.

Fig. 6 schematically shows a block diagram of a computer system suitable for implementing the above described method according to an embodiment of the present disclosure.

Detailed Description

Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. It should be understood that the description is illustrative only and is not intended to limit the scope of the present disclosure. In the following detailed description, for purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of the embodiments of the disclosure. It may be evident, however, that one or more embodiments may be practiced without these specific details. Moreover, in the following description, descriptions of well-known structures and techniques are omitted so as to not unnecessarily obscure the concepts of the present disclosure.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. The terms "comprises," "comprising," and the like, as used herein, specify the presence of stated features, steps, operations, and/or components, but do not preclude the presence or addition of one or more other features, steps, operations, or components.

All terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art unless otherwise defined. It is noted that the terms used herein should be interpreted as having a meaning that is consistent with the context of this specification and should not be interpreted in an idealized or overly formal sense.

Where a convention analogous to "at least one of A, B and C, etc." is used, in general such a construction is intended in the sense one having skill in the art would understand the convention (e.g., "a system having at least one of A, B and C" would include but not be limited to systems that have a alone, B alone, C alone, a and B together, a and C together, B and C together, and/or A, B, C together, etc.). Where a convention analogous to "A, B or at least one of C, etc." is used, in general such a construction is intended in the sense one having skill in the art would understand the convention (e.g., "a system having at least one of A, B or C" would include but not be limited to systems that have a alone, B alone, C alone, a and B together, a and C together, B and C together, and/or A, B, C together, etc.).

The embodiment of the disclosure provides an article recommendation method. Fig. 1 schematically shows a flow chart of an item recommendation method according to an embodiment of the present disclosure. As shown in fig. 1, the method includes operations S101 to S103.

In operation S101, historical user information and historical item information are determined.

According to an embodiment of the present disclosure, the historical user information is completed (deal) information related to the user. The historical item information is completed (committed) item-related information. The historical user information and the historical item information may be determined by, for example, a historical order.

Illustratively, the user is a shopper and the item is a commodity.

In operation S102, a target item corresponding to a target user is determined according to the historical user information and the historical item information.

According to the embodiment of the disclosure, the historical user information can comprise a relation matrix V of historical users and historical articles^U，IRelation matrix V of historical articles and historical users^I，U(ii) a The historical article information comprises a relation matrix V of the article selected at this time and the article selected at last time^I，LRelationship matrix V of last selected item and present selected item^L，I。

According to the embodiment of the disclosure, the historical user information further comprises a relation matrix V of the historical users and external data^U，KRelationship matrix V of external data and historical users^K，U(ii) a The historical item information also comprises a relation matrix V of the historical items and external data^I，KRelationship matrix V of external data and historical articles^K，IThe relation matrix V of the selected article and the selected article n times before the last time^I，L+1、V^I，L+2、…、V^I，L+nThe relation matrix V of the article selected n times before the last time and the article selected this time^L+1，I、V^L+2，I、…、V^L+n，I(ii) a The external data, i.e. externally publicly available data, may include, for example, weather, holidays, public opinion, etc.

According to the embodiment of the disclosure, the relation matrix V of the historical users and the historical articles^U，IE.g. a plurality of items corresponding to each user, a relation matrix V of historical items and historical users^I，UE.g. for multiple users for each item, visible V^U，IAnd V^I，UAnd are not equivalent. Similarly, V^I，LAnd V^L，INor is it equal, V^K，IAnd V^I，KNor is it equivalent.

In operation S103, the target item is pushed to the target user.

Fig. 2 schematically illustrates a flowchart of determining a target item corresponding to a target user according to the historical user information and the historical item information according to an embodiment of the present disclosure. The historical user information can comprise a relation matrix V of historical users and historical articles^U，IRelation matrix V of historical articles and historical users^I，U(ii) a The historical article information comprises a relation matrix V of the article selected at this time and the article selected at last time^I，LRelationship matrix V of last selected item and present selected item^L，IIn the case of (3), as shown in fig. 2, operations S201 to S204 are included for determining a target item corresponding to the target user from the historical user information and the historical item information.

In operation S201, a relationship matrix V between historical users and historical items is used^U，IRelation matrix V of historical articles and historical users^I，UDetermining a Matrix Factorization (MF) vector;

in operation S202, a relationship matrix V between the selected article and the selected article is determined^I，LRelation matrix V of last selected article and this selected article^L，IA decomposed non-personalized Markov chain vector (factoring unpersonalized markov chain) FMC is determined.

In operation S203, a probability vector P is determined from the vector MF and the vector FMC.

In operation S204, a target item corresponding to the target user is determined according to the probability vector P.

FIG. 3 schematically shows a relationship matrix V according to historical users and historical items according to an embodiment of the disclosure^U，IRelation matrix V of historical articles and historical users^I，UA flow chart of a matrix decomposition vector MF is determined. As shown in FIG. 3, a relation matrix V is formed according to historical users and historical articles^U，IRelation matrix V of historical articles and historical users^I，UDetermining the matrix decomposition vector MF involves operations S301-S302.

In operation S301, a relationship matrix V between historical users and historical items is used^U，IRelation matrix V of historical articles and historical users^I，UDetermining a matrix VUI _ m _ VIU;

in operation S302, a vector MF is determined according to the matrix VUI _ m _ vu.

According to the embodiment of the disclosure, the matrix VUI _ m _ vu and the vector MF respectively satisfy the following relational expressions: VUI _ m _ vu ═ V^U，I·V^I，U；MF＝VUI_m_VIU_i。

FIG. 4 schematically illustrates a relationship matrix V according to a present selection of an item and a last selection of an item according to an embodiment of the disclosure^I，LRelation matrix V of last selected article and this selected article^L，IFlow chart for determining vector FMC. As shown in FIG. 4, the relationship matrix V is based on the relationship between the selected article and the previously selected article^I，LRelation matrix V of last selected article and this selected article^L，IDetermining the vector FMC includes operations S401 to S402.

In operation S401, a relationship matrix V between the selected article and the selected article is determined^I，LRelation matrix V of last selected article and this selected article^L，IDetermining a matrix VIL _ m _ VLI;

in operation S402, a vector FMC is determined according to the matrix VIL _ m _ VLI.

According to the embodiment of the disclosure, the matrix VIL _ m _ VLI and the vector FMC respectively satisfy the following relations: VIL _ m _ VLI ═ V^I，L·V^L，I(ii) a FMC mean (VIL _ m _ VLI,) (VIL _ m _ VLI, { x, y, z }); the item set selected this time is represented, where { x, y, z }, and x, y, z represent each element in the item set selected this time.

FIG. 5 schematically illustrates a flow chart for determining a target item corresponding to a target user based on a probability vector. As shown in fig. 5, determining a target item corresponding to a target user according to a probability vector includes operations S501 to S502.

In operation S501, a probability threshold is preset;

in operation S502, it is determined that the item corresponding to the element of the probability vector p whose probability is greater than the probability threshold is the target item corresponding to the target user.

According to an embodiment of the present disclosure, the probability vector p satisfies the following relation: p ═ MF + FMC.

According to the embodiment of the present disclosure, the number f of target items may also be preset, and then the items corresponding to the f elements with the highest probability in the vector p are determined as the target items corresponding to the target users.

An application example of the article recommendation method of the present disclosure is described below.

(1) Initializing a matrix

Firstly, 4 initialization matrixes V are established in a database^U，I，V^I，U，V^I，L，V^L，IRespectively represent:

V^U，I: relationship matrix of shoppers and commodities

V^I，U: relationship matrix of goods and shoppers

V^I，L: relationship matrix of current purchased commodity and last purchased commodity

V^L，I: relationship matrix of last purchased commodity and current purchased commodity

Where the initial states are all set to a matrix with a mean of 0 and a standard deviation of 1. Column n _ factor of the current landing intercept matrix is 32, namely matrix V^I，U，V^I，L，V^L，IAre all M_IX 32 dimensional matrix, matrix V^U，IIs M_UX 32 dimensional matrix, where M_IRepresenting the number of all the commodities in the sakazao house, M_URepresenting the number of shoppers owned by the sakazawa house. These 4 matrices are tabulated:

miya_data_analysis.bzw_commodity_recommendation_model_vui；

miya_data_analysis.bzw_commodity_recommendation_model_viu；

miya_data_analysis.bzw_commodity_recommendation_model_vil；

miya_data_analysis.bzw_commodity_recommendation_model_vli；

the data are as follows:

factor0	factor1	factor2	factor3	factor4	……	factor30	factor31	id
									-0.0131	0.001041	-0.00314	-0.0022	0.002793	……	0.000922	0.013006	0
0.011415	0.003253	0.006173	-0.01078	-0.00405	……	-0.00548	-0.00491	1
									0.006789	-0.01996	0.009549	-0.02649	0.014538	……	-0.00499	-0.00149	2
-0.03015	0.014498	-0.01689	0.008037	0.009024	……	0.002212	-0.00787	3
									0.011947	0.015364	-0.02246	-0.01074	-0.00754	……	0.018706	-0.00408	4
-0.01071	0.003366	-0.01338	0.008451	-0.01494	……	0.001225	0.000686	5
									0.012215	-0.00684	0.012505	0.002067	0.015239	……	0.014465	-0.00417	6
0.006332	-0.01039	0.008635	0.003894	-0.00951	……	0.004672	-0.00727	7
									……	……	……	……	……	……	……	……	……

(2) model learning sample iteration matrix

The shopping basket list of the shopper and the list of the last shopping basket are processed before the model:

miya_data_analysis.bzw_buyer_basket_and_last_basket；

the data are as follows:

it can be seen that the two columns of baskets, which store shopping basket lists in list form, have different numbers representing different items. Each line is sample data for model learning.

For example: the first sample is:

the shopper with reference number 1772 purchases two products at this time, each being [145, 21153], and purchases 5 products at the last time, each being [578, 5498, 20998, 21041, 21073], and executes the following steps:

1. get matrix V^I，LLine 145, denoted as vector V^I，L[145]Are each independently of V^L，I578, 5498, 20998, 21041, 21073]The rows are vector multiplied and summed as χ_fmc. The specific formula is as follows:

2. get matrix V^U，I1772, as vector V^U，I[1772]And V and^I，U145 th line of (1), denoted as vector V^I，U[145]Make vector product, and is marked as χ_mfTool for measuringThe volume formula is as follows:

χ_mf＝V^U，I[1772]·V^I，U[145]

3. order to

4. Randomly selecting one commodity j, requiring that 2 commodities [145, 21153] purchased at this time are not included]Calculating by using a calculation method of 1-3

5. Order to

6. The following indices were calculated:

VUI_update＝α·(Δ·(V^I，U[145]-V^I，U[f])-λ·V^U，I[1772])

VIU_update，145＝α·(Δ·V^U，I[1772]-λ·V^I，U[145])

VIU_update，j＝α·(Δ·V^U，I[1772]-λ·V^I，U[j])

7. updating the matrix:

V^U，I[1772]_new＝V^U，I[1772]+VUI_update

V^I，U[145]_new＝V^I，U[145]+VIU_update，145

V^I，U[j]_new＝V^I，U[145]+VIU_update，j

8. computingThe following criteria: will matrix V^L，I578, 5498, 20998, 21041, 21073]Lines were taken and averaged:

η＝mean(V^L，I[578，5498，20998，21041，21073]，axis＝0)

VIL_update，145＝α·(Δ·η-λ·V^I，L[145])

VIL_update，j＝α·(Δ·η-λ·V^I，L[j])

9. updating the matrix:

V^I，L[145]_new＝V^I，L[145]+VIL_update，145

V^I，L[j]_new＝V^I，L[145]+VIL_update，j

V^L，I[578，5498，20998，21041，21073]_new

＝V^L，I[578，5498，20998，21041，21073]+VLI_update

10. and (3) applying the algorithm, and repeatedly calculating the steps 1-9 to update the matrix according to each sample to complete the task of model training.

(3) Model prediction

Reading in a list of shoppers to be predicted:

miya_data_analysis.bzw_buyer_basket_list；

the data are as follows:

assuming that it is predicted what merchandise will be purchased by shopper 39493 the next time, shopper 39493 can be found from the table as the shopping list for this time [593, 19862, 21129]

All M for sakazao house_IThe commodity application algorithm is as follows:

1. get matrix V^I，LLine i (i ∈ M)_I) Is recorded as a vector V^I，L[i]Are each independently of V^L，I593, 19862, 21129]The rows are vector multiplied and summed as χ_fmc. The specific formula is as follows:

2. get matrix V^U，I39493 th line, denoted as vector V^U，I[39493]And V and^I，Uline i of (1), denoted as vector V^I，U[i]Make vector product, and is marked as χ_mfThe concrete formula is as follows:

χ_mf＝V^U，I[39493]·V^I，U[i]

3. the probability of purchasing item i is:

repeatedly applying the steps 1-3 to calculate all M_IAnd sorting the commodities according to the calculated probability, and preferentially recommending the commodities with high probability.

In order to verify the effect of the method disclosed by the invention, 5 crowd bags are selected for testing, and a commodity with a high purchase probability of each user top10 is found out, about 10% -30% of users can directly purchase at least one commodity, and the specific test results are as follows:

test crowd bag	Hit rate
		test1	22.22％
test2	22.39％
		test3	28.74％
test4	14.29％
		test5	11.11％

Fig. 6 schematically shows a block diagram of a computer system suitable for implementing the above described method according to an embodiment of the present disclosure. The computer system illustrated in FIG. 6 is only one example and should not impose any limitations on the scope of use or functionality of embodiments of the disclosure.

As shown in fig. 6, a computer system 600 according to an embodiment of the present disclosure includes a processor 601, which can perform various appropriate actions and processes according to a program stored in a Read Only Memory (ROM)602 or a program loaded from a storage section 608 into a Random Access Memory (RAM) 603. Processor 601 may include, for example, a general purpose microprocessor (e.g., a CPU), an instruction set processor and/or associated chipset, and/or a special purpose microprocessor (e.g., an Application Specific Integrated Circuit (ASIC)), among others. The processor 601 may also include onboard memory for caching purposes. Processor 601 may include a single processing unit or multiple processing units for performing different actions of a method flow according to embodiments of the disclosure.

In the RAM603, various programs and data necessary for the operation of the system 600 are stored. The processor 601, the ROM 602, and the RAM603 are connected to each other via a bus 604. The processor 601 performs various operations of the method flows according to the embodiments of the present disclosure by executing programs in the ROM 602 and/or RAM 603. It is to be noted that the programs may also be stored in one or more memories other than the ROM 602 and RAM 603. The processor 601 may also perform various operations of the method flows according to embodiments of the present disclosure by executing programs stored in the one or more memories.

According to an embodiment of the present disclosure, system 600 may also include an input/output (I/O) interface 605, input/output (I/O) interface 605 also connected to bus 604. The system 600 may also include one or more of the following components connected to the I/O interface 605: an input portion 606 including a keyboard, a mouse, and the like; an output portion 607 including a display such as a Cathode Ray Tube (CRT), a Liquid Crystal Display (LCD), and the like, and a speaker; a storage section 608 including a hard disk and the like; and a communication section 605 including a network interface card such as a LAN card, a modem, or the like. The communication section 609 performs communication processing via a network such as the internet. The driver 610 is also connected to the I/O interface 605 as needed. A removable medium 611 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is mounted on the drive 610 as necessary, so that a computer program read out therefrom is mounted in the storage section 608 as necessary.

According to embodiments of the present disclosure, method flows according to embodiments of the present disclosure may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable storage medium, the computer program containing program code for performing the method illustrated by the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network through the communication section 609, and/or installed from the removable medium 611. The computer program, when executed by the processor 601, performs the above-described functions defined in the system of the embodiments of the present disclosure. The systems, devices, apparatuses, modules, units, etc. described above may be implemented by computer program modules according to embodiments of the present disclosure.

The present disclosure also provides a computer-readable storage medium, which may be contained in the apparatus/device/system described in the above embodiments; or may exist separately and not be assembled into the device/apparatus/system. The computer-readable storage medium carries one or more programs which, when executed, implement the method according to an embodiment of the disclosure.

According to an embodiment of the present disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium. Examples may include, but are not limited to: a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present disclosure, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.

For example, according to embodiments of the present disclosure, a computer-readable storage medium may include the ROM 602 and/or RAM603 described above and/or one or more memories other than the ROM 602 and RAM 603.

The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams or flowchart illustration, and combinations of blocks in the block diagrams or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions. Those skilled in the art will appreciate that various combinations and/or combinations of features recited in the various embodiments and/or claims of the present disclosure can be made, even if such combinations or combinations are not expressly recited in the present disclosure. In particular, various combinations and/or combinations of the features recited in the various embodiments and/or claims of the present disclosure may be made without departing from the spirit or teaching of the present disclosure. All such combinations and/or associations are within the scope of the present disclosure.

The embodiments of the present disclosure have been described above. However, these examples are for illustrative purposes only and are not intended to limit the scope of the present disclosure. Although the embodiments are described separately above, this does not mean that the measures in the embodiments cannot be used in advantageous combination. The scope of the disclosure is defined by the appended claims and equivalents thereof. Various alternatives and modifications can be devised by those skilled in the art without departing from the scope of the present disclosure, and such alternatives and modifications are intended to be within the scope of the present disclosure.

Claims

1. An item pushing method comprising:

determining historical user information and historical article information;

pushing the target item to the target user.

2. The method of claim 1, wherein the historical user information comprises a relationship matrix V of historical users to historical items^U，IRelation matrix V of historical articles and historical users^I，U(ii) a The historical article information comprises a relation matrix V of the article selected at this time and the article selected at last time^I，LRelationship matrix V of last selected item and present selected item^L，I。

3. According to claim 2The method, the historical user information also includes a relation matrix V of the historical users and external data^U，KRelationship matrix V of external data and historical users^K，U(ii) a The historical item information also comprises a relation matrix V of the historical items and external data^I，KRelationship matrix V of external data and historical articles^K，IThe relation matrix V of the selected article and the selected article n times before the last time^I，L+1、V^I，L+2、…、V^I，L+nThe relation matrix V of the article selected n times before the last time and the article selected this time^L+1，I、V^L+2，I、…、V^L+n，I(ii) a Wherein the external data comprises weather, holidays, public opinion data.

4. The method of claim 2, determining a target item corresponding to a target user from the historical user information and historical item information, comprising:

5. The method of claim 4, wherein the relationship matrix V is based on historical user and historical item^U，IRelation matrix V of historical articles and historical users^I，UDetermining a matrix decomposition vector, MF, comprising:

the vector MF is determined from the matrix VUI _ m _ VIU.

6. The method of claim 5, wherein the relationship matrix V is based on the current selected item versus the last selected item^I ^，LRelation matrix V of last selected article and this selected article^L，IDetermining a vector FMC, comprising:

the vector FMC is determined from the matrix VIL _ m _ VLI.

7. The method of claim 6, wherein the matrix VIL _ m _ VLI, the matrix VUI _ m _ vu, the vector MF, the vector FMC, and the probability vector p satisfy the following relationships, respectively:

VIL_m_VLI＝V^I，L·V^L，I；

VUI_m_VIU＝V^U，I·V^I，U；

MF＝VUI_m_VIU_i；

FMC＝mean(VIL_m_VLI_，)＝mean(VIL_m_VLI_{，{x，y，z}})；

p＝MF+FMC；

8. The method of claim 4, determining a target item corresponding to a target user from a probability vector p, comprising:

presetting a probability threshold;

9. A computer system, comprising:

one or more processors;

a storage device for storing one or more programs,

wherein the one or more programs, when executed by the one or more processors, cause the one or more processors to perform the method of any of claims 1-8.

10. A computer readable storage medium having stored thereon executable instructions which, when executed by a processor, cause the processor to perform the method of any one of claims 1 to 8.