CN111178951B - Commodity recommendation method and device - Google Patents
Commodity recommendation method and device Download PDFInfo
- Publication number
- CN111178951B CN111178951B CN201911327992.8A CN201911327992A CN111178951B CN 111178951 B CN111178951 B CN 111178951B CN 201911327992 A CN201911327992 A CN 201911327992A CN 111178951 B CN111178951 B CN 111178951B
- Authority
- CN
- China
- Prior art keywords
- user
- purchase
- commodity
- users
- relation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 47
- 239000011159 matrix material Substances 0.000 claims description 31
- 230000015654 memory Effects 0.000 claims description 19
- 238000004422 calculation algorithm Methods 0.000 claims description 11
- 238000004590 computer program Methods 0.000 claims description 2
- 230000000694 effects Effects 0.000 description 7
- 238000012545 processing Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 208000015181 infectious disease Diseases 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 239000007787 solid Substances 0.000 description 2
- 238000003491 array Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0241—Advertisements
- G06Q30/0251—Targeted advertisements
- G06Q30/0255—Targeted advertisements based on user history
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9536—Search customisation based on social or collaborative filtering
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Finance (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- General Engineering & Computer Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides a commodity recommendation method and a commodity recommendation device, wherein the method comprises the following steps: acquiring purchase information of all users for each commodity, wherein the purchase information comprises purchase, non-purchase and purchase time; respectively establishing a time sequence corresponding to each commodity according to the purchase information of the user; reconstructing a social relation network between users according to the time sequence; and recommending corresponding commodities to the user according to the social relation network and the purchase information. The method solves the problem that the current commodity recommendation mode is adopted to judge the current purchasing preference of the user, and the purchasing preference of the user cannot be potentially influenced by social circles around or personal relationships, so that the purchasing tendency is changed to be predicted in advance, and improves the predictability of commodity recommendation.
Description
Technical Field
The invention relates to the field of computer networks, in particular to a commodity recommendation method and device.
Background
With the continuous development of electronic commerce, electronic shopping platforms in the market are more and more, the types of provided commodities are more and more complete, and people are more and more dependent on online shopping. Therefore, commodity recommendation becomes a way of promoting sales for each electronic shopping platform and electronic merchants.
In the related art, commodity recommendation generally uses the purchase history of a customer to make similar commodity recommendation, but only the current purchase preference of a purchaser can be deduced by using the method, in fact, the purchase preference of people can be potentially influenced by social circles around or interpersonal relations, so that the purchase tendency is changed, and the conventional commodity recommendation mode cannot predict the change of the purchase preference of the user and the tendency of the user to make corresponding commodity recommendation in advance.
Disclosure of Invention
Therefore, the technical problem to be solved by the invention is to overcome the defect that commodity recommendation in the prior art cannot be predicted in advance, so as to provide a commodity recommendation method and device.
According to a first aspect, an embodiment of the present invention provides a commodity recommendation method, including the steps of: acquiring purchase information of all users for each commodity, wherein the purchase information comprises purchase, non-purchase and purchase time; respectively establishing a time sequence corresponding to each commodity according to the purchase information of the user; reconstructing a social relation network between users according to the time sequence; and recommending corresponding commodities to the user according to the social relation network and the purchase information.
With reference to the first aspect, in a first implementation manner of the first aspect, the step of establishing a time sequence corresponding to each commodity according to purchase information of the user includes: for each commodity, a row vector is established corresponding to whether each user purchases the commodity, and a time matrix is established by taking the purchased time sequence as a column vector, wherein the time matrix is as follows:
wherein j represents the number of time sequences, N represents the number of users, i represents the ith user, and all values S in the matrix ji Indicating whether the user purchases the commodity S ji =1 means that the user purchases the commodity, S ji =0 indicates that the user did not purchase the commodity.
With reference to the first aspect, in a second implementation manner of the first aspect, the reconstructing a social relationship network between users according to the time sequence includes: selecting the effective time sequence as a node state sequence; solving the node state sequence according to a compressed sensing algorithm, and acquiring a relation variable between nodes, wherein the relation scalar represents the relation between the nodes; and establishing a social relationship network among the users according to the relationship variables.
With reference to the first aspect, in a third implementation manner of the first aspect, the recommending the commodity to the user according to the social relationship network and the purchase information includes: acquiring the purchase condition of a user of any commodity in the social relationship network; acquiring other users associated with the user presence; and recommending the commodity to other users not buying the commodity in the social relation network.
With reference to the second implementation manner of the first aspect, in a fourth implementation manner of the first aspect, the solving the node state sequence according to the compressed sensing algorithm obtains a relationship variable between nodes, including constructing a matrix relationship equation by using compressed sensing:
wherein ,
an average value representing the user purchase status of the ith user in the vicinity of time tm+1; />
Indicating that the ith user is at t m Average value of purchase status near time; lambda (lambda) i Representing the probability of the item purchased by user i; a, a iN Representing the relationship between user i and user N, and taking the value of 0 or 1, when a iN When=1, it means that user i and user N are friends or the purchase preference is the same; when a is iN When=0, it means that user i and user N are not friends and the purchase preference is different.
According to a second aspect, an embodiment of the present invention provides a commodity recommendation apparatus, including: the information acquisition module is used for acquiring purchase information of all users aiming at each commodity, wherein the purchase information comprises purchase, non-purchase and purchase time; the time sequence establishing module is used for respectively establishing a time sequence corresponding to each commodity according to the purchase information of the user; the social relation network reconstruction module is used for reconstructing a social relation network among users according to the time sequence; and the commodity recommending module is used for recommending corresponding commodities to the user according to the social relation network and the purchase information.
With reference to the second aspect, in a first implementation manner of the second aspect, the social relationship network reconstruction module includes: the time sequence selecting module is used for selecting the effective time sequence as a node state sequence; the relation variable acquisition module is used for solving the node state sequence according to a compressed sensing algorithm to acquire relation variables among nodes, wherein the relation variables represent the relation among the nodes; and the social relation network establishing module is used for establishing a social relation network among the users according to the relation variable.
With reference to the second aspect, in a second implementation manner of the second aspect, the commodity recommendation module includes: the user purchase condition acquisition module is used for acquiring the user purchase condition of any commodity in the social relationship network; the associated user acquisition module is used for acquiring other users associated with the user; and the commodity recommending sub-module is used for recommending commodities to other users who do not purchase the commodities in the social relation network.
According to a third aspect, an embodiment of the present invention provides an electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the steps of the merchandise recommendation method of the first aspect or any one of the first aspects when the program is executed.
According to a fourth aspect, an embodiment of the present invention provides a storage medium having stored thereon computer instructions which, when executed by a processor, implement the steps of the merchandise recommendation method of the first aspect or any one of the first aspects.
The technical scheme of the invention has the following advantages:
1. according to the commodity recommendation method provided by the invention, the social relationship network of commodity purchasing users is reconstructed according to the commodity purchasing information, the social relationship network shows which of a plurality of purchasing users are friends or users with the same purchasing preference, when any user purchases a commodity, the same commodity recommendation is carried out on the reconstructed friends or users with the same purchasing preference, the prediction effect on the purchasing preference and purchasing tendency of the user is achieved, the pertinence is stronger when the commodity is recommended, the welcome degree of a commodity in the user is better mastered, and the group effect is utilized to stimulate the user to purchase the commodity.
2. The commodity recommendation method provided by the invention has the advantages that the time sequence established for each commodity is equivalent to information drive, different commodities are utilized to establish the time sequence, so that the number of the time sequences can meet the basic time sequence number of a reconstruction algorithm, and the problem that the reconstructed social relationship network is incomplete because all users in the same social relationship network do not purchase the commodity when only one commodity is used for establishing the time sequence in the whole social relationship network can be prevented, and the time sequences of a plurality of commodities are established so as to facilitate complete reconstruction of the social relationship network.
3. According to the method for establishing the social relationship network between the users by combining the compressed sensing model with the time sequence, the complex social network relationship between the users can be reconstructed by using fewer time sequences, and the data volume for reconstructing the social network relationship is reduced.
4. According to the commodity recommendation method provided by the embodiment, when a certain commodity is purchased by a user in any social relationship network, other users in the social relationship network are recommended, so that commodity recommendation is more targeted, and the effectiveness of commodity recommendation is improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are needed in the description of the embodiments or the prior art will be briefly described, and it is obvious that the drawings in the description below are some embodiments of the present invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flowchart of a specific example of a commodity recommendation method according to an embodiment of the present invention;
FIG. 2 is a flowchart of a specific example of a commodity recommendation method according to an embodiment of the present invention;
FIG. 3 is a schematic block diagram of a specific example of a commodity recommendation device according to an embodiment of the present invention;
FIG. 4 is a schematic block diagram of a specific example of a commodity recommendation device according to an embodiment of the present invention;
fig. 5 is a schematic block diagram of a specific example of an electronic device in an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made apparent and fully in view of the accompanying drawings, in which some, but not all embodiments of the invention are shown. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
In the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. are based on the directions or positional relationships shown in the drawings, are merely for convenience of describing the present invention and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a specific orientation, be configured and operated in a specific orientation, and thus should not be construed as limiting the present invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.
In the description of the present invention, it should be noted that, unless explicitly specified and limited otherwise, the terms "mounted," "connected," and "connected" are to be construed broadly, and may be either fixedly connected, detachably connected, or integrally connected, for example; can be mechanically or electrically connected; the two components can be directly connected or indirectly connected through an intermediate medium, or can be communicated inside the two components, or can be connected wirelessly or in a wired way. The specific meaning of the above terms in the present invention will be understood in specific cases by those of ordinary skill in the art.
In addition, the technical features of the different embodiments of the present invention described below may be combined with each other as long as they do not collide with each other.
The embodiment provides a commodity recommendation method, which is used for recommending commodities for users on the basis of acquiring purchase information of a plurality of users, and as shown in fig. 1, comprises the following steps:
s110, acquiring purchase information of all users for each commodity, wherein the purchase information comprises purchase, non-purchase and purchase time.
The purchase information of all users may be obtained by collecting the purchase data of the users in a networked computer sales system or an e-commerce platform, or collecting the purchase information from the consumption records of the member cards, the loyalty cards or the bank cards of the consumers. The purchase information indicates the purchase time of the user who purchased a particular item for which users have purchased and which users have not purchased the item. The method for acquiring purchase information of all users is not limited in this embodiment, and those skilled in the art can select the purchase information according to needs.
S120, respectively establishing time sequences corresponding to each commodity according to the purchase information of the user.
For example, the manner of establishing the time series for each commodity according to the purchase information of the user may be whether the user purchases in the purchase information and the purchase time. The time sequence is established for each commodity respectively and is equivalent to the information driven process, and the time sequences are established by utilizing different commodities, so that the number of the time sequences can meet the number of the basic time sequences of a reconstruction algorithm.
S130, reconstructing a social relationship network among users according to the time sequence.
The method for reconstructing the social relationship network between the users according to the time sequence can be that a reconstruction formula matched with a compressed sensing formula is established by using a disease dynamics model, and the social relationship network between the users is obtained by combining the reconstruction formula with the time sequence; or a reconstruction formula matched with the compressed sensing model formula is established by using the public opinion propagation dynamics model, and a social relationship network between users is obtained by combining a time sequence according to the reconstruction formula. The method of reconstructing the social relationship network between the users through the time sequence is not limited in this embodiment, and can be determined by a person skilled in the art according to needs.
And S140, recommending corresponding commodities to the user according to the social relation network and the purchase information.
For example, the manner of recommending the corresponding commodity to the user according to the social relationship network and the purchase information may be to know whether the user purchases the commodity according to the purchase information of the user for a certain commodity, when the user does not purchase the commodity, the user who may be friends or purchase preference is determined through the reconstructed social relationship network, whether the user who may be friends or purchase preference is the same as the user, and the ratio of purchasing the commodity are determined according to the purchase information, when the ratio of purchasing the commodity exceeds a predetermined threshold, the commodity is recommended to the user, and the predetermined threshold may be 30% here, and the specific numerical value is not limited. The commodity recommendation mode can be short message recommendation, web advertisement recommendation and E-commerce platform commodity recommendation, and the recommendation mode is not limited in the embodiment, and can be determined by a person skilled in the art according to requirements.
According to the commodity recommendation method provided by the embodiment, the social relationship network of commodity purchasing users is reconstructed according to the commodity purchasing information, the social relationship network shows which of a plurality of purchasing users are friends or users with the same purchasing preference, when any user purchases a commodity, the same commodity recommendation is carried out on the reconstructed friends or users with the same purchasing preference, the predicting effect on the purchasing preference and purchasing tendency of the user is achieved, the pertinence is higher when the commodity is recommended, the welcome degree of a commodity in the user is better mastered, and the group effect is utilized to stimulate the user to purchase the commodity.
As an optional embodiment of the present application, step S120 includes:
for each commodity, a row vector is established corresponding to whether each user purchases the commodity, and a time matrix is established by taking the purchased time sequence as a column vector, wherein the time matrix is as follows:
wherein j represents the number of time sequences, N represents the number of users, i represents the ith user, and all values S in the matrix ji Indicating whether the user purchases the commodity S ji =1 means that the user purchases the commodity, S ji =0 indicates that the user did not purchase the commodity.
Illustratively, the length of the time sequence is determined according to the requirement, and the embodiment is not limited. For a certain commodity, such as a pair of trousers, it is assumed that 8 users' purchase information at 55 days is acquired, and the specific purchase information is shown in table 1 at intervals of five days as a time series.
TABLE 1
According to the method described above, a time sequence is established for the purchase information in table 1, and a reference time sequence is randomly selected, for example, a time sequence shown in table 2 is constructed with respect to the purchase state corresponding to the time sequence of user 6, wherein the non-purchased time sequence is represented by 0 and the purchased time sequence is represented by 1.
TABLE 2
| User 6 | User 1 | User 2 | User 3 | User 4 | User 5 | User 7 | User 8 | |
| t 1 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 0 |
| t 2 | 0 | 1 | 0 | 0 | 1 | 0 | 0 | 0 |
| t 3 | 0 | 1 | 0 | 0 | 1 | 0 | 0 | 1 |
| t 4 | 0 | 1 | 0 | 1 | 1 | 0 | 0 | 1 |
| t 5 | 0 | 1 | 0 | 1 | 1 | 0 | 0 | 1 |
| t 6 | 0 | 1 | 0 | 1 | 1 | 0 | 1 | 1 |
| t 7 | 0 | 1 | 0 | 1 | 1 | 0 | 1 | 1 |
| t 8 | 0 | 1 | 0 | 1 | 1 | 0 | 1 | 1 |
| t 9 | 0 | 1 | 0 | 1 | 1 | 1 | 1 | 1 |
| t 10 | 1 | 1 | 0 | 1 | 1 | 1 | 1 | 1 |
| t 11 | 1 | 1 | 0 | 1 | 1 | 1 | 1 | 1 |
The arrangement into a matrix is expressed as:
the method is equivalent to information driving, and the time sequence established for each commodity is established by utilizing different commodities, so that the number of the time sequences can meet the basic time sequence number of a reconstruction algorithm, and the problem that the reconstructed social relationship network is incomplete because all users in the same social relationship network do not purchase the commodity when only one commodity is established in the whole social relationship network can be prevented, and the time sequences of the plurality of commodities are established so as to facilitate complete reconstruction of the social relationship network.
As an optional embodiment of the present application, after the time series is established in the above step, the social relationship network between the users is reconstructed according to the time series, where the social relationship network refers to finding the users having a common purchasing tendency relationship, that is, finding the users having a common preference or purchasing tendency, as shown in fig. 2, the step S130 specifically includes:
s131, selecting an effective time sequence as a node state sequence.
Illustratively, for reconstructing the social network relationship of the user 6, the effective time series represents a time series from the un-purchased state to the un-purchased state, and from the un-purchased state to the purchased state, i.e., a series of 00 and 01 states, among the time series of the user 6. Taking the data of table 2 as an example, as shown in table 3, the effective time sequence selected is the first 10 rows:
TABLE 3 Table 3
| User 6 | User 1 | User 2 | User 3 | User 4 | User 5 | User 7 | User 8 | |
| t 1 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 0 |
| t 2 | 0 | 1 | 0 | 0 | 1 | 0 | 0 | 0 |
| t 3 | 0 | 1 | 0 | 0 | 1 | 0 | 0 | 1 |
| t 4 | 0 | 1 | 0 | 1 | 1 | 0 | 0 | 1 |
| t 5 | 0 | 1 | 0 | 1 | 1 | 0 | 0 | 1 |
| t 6 | 0 | 1 | 0 | 1 | 1 | 0 | 1 | 1 |
| t 7 | 0 | 1 | 0 | 1 | 1 | 0 | 1 | 1 |
| t 8 | 0 | 1 | 0 | 1 | 1 | 0 | 1 | 1 |
| t 9 | 0 | 1 | 0 | 1 | 1 | 1 | 1 | 1 |
| t 10 | 1 | 1 | 0 | 1 | 1 | 1 | 1 | 1 |
The matrix is expressed as:
s132, solving a node state sequence according to a compressed sensing algorithm, and acquiring a relation variable between nodes, wherein a relation scalar represents a relation between the nodes.
Illustratively, the compressed sensing formula is: y=Φ·x, where x∈r N Representing the original digital signal; y εR M Represented is a compressed representation of the signal. And Φ is a matrix of M x N, representing the measurement matrix acquired. Accurate reconstruction can be achieved by solving the following convex optimization problem:
min||X|| 1 subordinate to Y=φ.X (1)
Wherein, the left formula is X 1 As a norm on the vector X, it can be expressed by the following equation:
i.e. |X|| 1 May be represented by the sum of the number of non-0 elements of the row vector for each row of the N-dimensional matrix vector X.
Taking the SI model as an example, the probability that an arbitrary node i gets an infection from its neighbors at time t can be expressed by the following formula:
wherein ,λi Is the infection rate of node i, and a ij Representing the connection of elements in the adjacency matrix (1 representing connected, 0 representing disconnected), S j (t) represents the state of node j at time t, and then 01 represents the transition of node from the susceptible state to the infected state. For the next operation, the above formula is logarithmized and is written in the following form:
then, if at a different instant t=t 1 ,t 2 ,…,t m State S of node j (t) is available, the above formula can be written in the form of a matrix multiplication of:
i.e. vector Y is defined by ln [1-P ] at different times i 01 (t)]The matrix X includes the connection between nodes and the infection rate of the node i, which is sparse in a network structure. Wherein P is i 01 (t) cannot be directly obtained from node states in a continuous time series, but P is calculated in the time series at time t according to the law of large numbers i 01 (t) can be made up of all S at the appropriate time i The average value of (t+1) is estimated, whereby all the data in the above calculation formula can be obtained from the time series. From the correspondence relationship, a can be obtained ij ,a ij The connection relationship between the node i and the node j is represented.
Calculating according to the selected node state sequence and the law of large numbers to obtain P i 01 The way of (t) may be to set the thresholds delta and phi, both of which are related to the normalized hamming distance, which is defined by S for defining the reference string at time t j (t) and a series of strings adjacent to the reference string. Let the normalized Hamming distance be H, for example, the Hamming distance between a and b be H (a, b), we need Φ to perform preliminary screening on the character string, that is, to satisfy H (a, b)>A, b sequence of phi condition, while delta aids in the next search, H (a, b)<Delta, and the same thing, the retrieved results can then be used to define a series of reference strings for network reconstruction.
Taking the purchase information in the above step S131 as an example, after the first four rows are selected as the node state sequences, further processing is required for the node state sequences. Since the purchase of user i at time t+1 is determined by the purchase of the neighbor of user i at time t, S in the figure i( t+1) is represented by S -i (t), so the first 10 rows selected are combined, as shown in table 4, as:
TABLE 4 Table 4
| User 6 | User 1 | User 2 | User 3 | User 4 | User 5 | User 7 | User 8 | ||
| t 2 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | t 1 |
| t 4 | 0 | 1 | 0 | 0 | 1 | 0 | 0 | 1 | t 3 |
| t 6 | 0 | 1 | 0 | 1 | 1 | 0 | 0 | 1 | t 5 |
| t 8 | 0 | 1 | 0 | 1 | 1 | 0 | 1 | 1 | t 7 |
| t 10 | 1 | 1 | 0 | 1 | 1 | 1 | 1 | 1 | t 9 |
The threshold values delta and phi are set to be 4/7, at t 1 、t 3 、t 5 、t 7 、t 9 The standard Hamming distance between the character strings is calculated in pairs to find that only t 1 and t9 The standard Hamming distance of the sequence of (2) is greater than phi (3/7), i.e. H (S) -i (t 1 ),S -i (t 9 ))＝5/7>Phi, meets the requirement of the first step. Thus further choose t 1 Sequence sum t 9 The sequences being reference sequences, i.e. respectivelyFrom t 1 Sequence sum t 9 The sequence begins to unwind the retrieval of the sample. Then calculate H (S) -i (t 1 ),S -i (t 3 ))＝2/7<Δ，H(S -i (t 1 ),S -i (t 5 ))＝3/7<Δ，H(S -i (t 9 ),S -i (t 7 ))＝1/7<Delta. Since at this step we need to choose a string less than Δ, t 1 And t 3 、t 1 and t5 、t 7 and t9 Sequence entry.
Y and X in y=Φx need to be determined. Due to t 2 Heel t 1 T 4 Heel t 3 T 6 Heel t 5 T 8 Heel t 7 T 10 Heel t 9 Is directly related, so that the related sequences need to be operated together, t is selected as shown in Table 5 1 And t 3 Mean value of sequence, t 1 and t5 Mean value of t 7 And t 9 The means of the sequences, together constitute
Values representing columns 2 to 8 of the above matrix as Φ in y=Φ×x; and select t 2 Heel t 4 Mean value of sequence, t 2 Heel t 6 Mean value of sequence, t 8 Heel t 10 Mean value of (A), co-composition->
The value of column 1 in the above matrix is indicated as Y in y=Φ×x.
TABLE 5
The following matrix was established according to table 5:
s133, establishing a social relationship network among users according to the relationship variables.
Illustratively due to P i 01 (t) has been indirectly obtained from the above process, in the above formula (4), a is required to be solved for ij Is 0 or 1, ln (1-lambda) i ) Does not affect the solution a ij Therefore, a can be obtained ij When a is the value of ij When=1, it indicates that the user i has a social relationship with the user j, otherwise, the user i has no social relationship with the user j. The users with social relationship representation may be friends or the same buying hobbies; no social relationship characterizes that the users are not friends and the buying preferences are not the same.
The social relationship network between users is established by combining the compressed sensing model with the time sequence, and the relationship with the same friends or buying hobbies between complex users can be reconstructed by using a small data volume, so that the commodity recommendation method has more pertinence.
As an optional implementation manner of the application, after the relationship of the users is obtained, product recommendation is performed according to the relationship of the users, and for friends or users with the same purchase preference, the intention of purchasing the same product is larger, so that the product is recommended to the users with the relationship for the product purchased by the user, the higher the success rate is, and recommendation is performed based on the product recommendation. Step S140, including:
firstly, the purchasing condition of a user of any commodity in a social relationship network is obtained.
For example, the method for obtaining the purchase condition of the user of any commodity in the social relationship network may be to obtain the purchase information of the user, and the method for obtaining the purchase information of the user is shown in step S110, which is not described herein, and the purchase condition of the user for a certain commodity is screened from the obtained purchase information. The method for acquiring the purchasing condition of any commodity in the social relationship network is not limited in this embodiment, and a person skilled in the art can determine the purchasing condition according to the requirement.
Finally, other users associated with the user existence are obtained; and recommending the commodity to other users not buying the commodity in the social relation network.
For example, the acquisition mode of other users associated with the user can be obtained in the social relationship network reconstructed by the time sequence; the determining manner of other users who do not purchase the commodity in the social relationship network may be to acquire purchase information of the user, where the purchase information includes purchase information. And for a specific commodity, recommending the commodity to the user when the user does not purchase the commodity and the user who is possibly friends or has the same purchase preference in the reconstructed social network purchases the commodity. The recommended manner is S140, and will not be described again.
According to the commodity recommendation method provided by the embodiment, when a certain commodity is purchased by a user in any social relationship network, other users in the social relationship network are recommended, so that commodity recommendation is more targeted, and the effectiveness of commodity recommendation is improved.
As an optional embodiment of the present application, step S132 includes constructing a matrix relation using compressed sensing:
wherein ,
an average value representing the user purchase status of the ith user in the vicinity of time tm+1; />
Indicating that the ith user is at t m Average value of purchase status near time; lambda (lambda) i Representing the probability of the item purchased by user i; a, a iN Representing the relationship between user i and user N, and taking the value of 0 or 1, when a iN When=1, it means that user i and user N are friends or the purchase preference is the same; when a is iN When=0, it means that user i and user N are not friends and the purchase preference is different. />
Illustratively, the matrix relationship may be expressed as follows, corresponding to the matrix formed in table 5 in step S132:
wherein ln (1-lambda) i ) The calculation is not influenced, and can be-1, the matrix relation can be established through a plurality of sequences to obtain a 61 、a 62 、a 63 、a 64 、a 65 、a 67 、a 68 The social network relationship between the user 6 and the rest of the users can be obtained. By selecting the purchase states corresponding to the time sequences of the other users as the reference, the social network relationship among the other users can be obtained, so that the social network relationship among all the users is reconstructed.
The present embodiment provides a commodity recommendation device, as shown in fig. 3, including:
the information acquisition module 210 is configured to acquire purchase information of all users for each commodity, where the purchase information includes purchase, non-purchase, and purchase time. The specific implementation manner is described in the related description of the embodiment S110, and will not be described herein.
The time sequence establishing module 220 is configured to establish a time sequence for each commodity according to purchase information of the user. The specific implementation manner is described in the related description of the embodiment S120, and will not be described herein.
The social relationship network reconstructing module 230 is configured to reconstruct a social relationship network between users according to the time sequence. The specific implementation manner is described in the related description of the embodiment S130, which is not repeated here.
And the commodity recommending module 240 is used for recommending corresponding commodities to the user according to the social relation network and the purchase information. The specific implementation manner is described in the related description of the embodiment S140, which is not repeated here.
According to the commodity recommending device provided by the invention, the social relation network of commodity purchasing users is reconstructed according to the commodity purchasing information, the social relation network shows which of a plurality of purchasing users are friends or users with the same purchasing preference, when any user purchases a commodity, the same commodity is recommended to the reconstructed friends or users with the same purchasing preference, the predicting effect on the purchasing preference and purchasing tendency of the user is achieved, the pertinence is stronger when the commodity is recommended, the welcome degree of the commodity in the users is better mastered, and the group effect is utilized to ensure that some users who do not purchase the commodity are influenced by adjacent users and the user is stimulated to purchase the commodity.
As an optional embodiment of the present application, the time sequence establishing module 220 specifically includes:
the time sequence establishing sub-module is used for establishing a row vector corresponding to whether each commodity is purchased or not by each user, and establishing a time matrix by taking the purchased time sequence as a column vector, wherein the time matrix is as follows:
wherein j represents the number of time sequences, N represents the number of users, i represents the ith user, and all values S in the matrix ji Indicating whether the user purchases the commodity S ji =1 means that the user purchases the commodity, S ji =0 indicates that the user did not purchase the commodity. The specific implementation manner is described in the related description of the commodity recommendation method in this embodiment, and will not be described in detail herein.
As an optional embodiment of the present application, as shown in fig. 4, the social relationship network reconstruction module 230 specifically includes:
a time sequence selection module 231, configured to select an effective time sequence as a node state sequence; the specific implementation manner is described in the related description of the implementation step S131, and will not be described herein.
The relation variable obtaining module 232 is configured to solve the node state sequence according to a compressed sensing algorithm, obtain a relation variable between nodes, where the relation variable represents a relation between nodes; the specific implementation manner is described in the step S132 of this embodiment, and will not be described herein.
The social relationship network establishing module 233 is configured to establish a social relationship network between users according to the relationship variables. The specific implementation manner is described in the step S133 of this embodiment, and will not be described herein.
As an optional embodiment of the present application, the commodity recommendation module 240 described above includes:
the user purchase condition acquisition module is used for acquiring the user purchase condition of any commodity in the social relationship network; the specific implementation manner is described in the related description of the commodity recommendation method in this embodiment, and will not be described in detail herein.
The associated user acquisition module is used for acquiring other users associated with the user; the specific implementation manner is described in the related description of the commodity recommendation method in this embodiment, and will not be described in detail herein.
And the commodity recommending sub-module is used for recommending commodities to other users who do not purchase the commodities in the social relationship network. The specific implementation manner is described in the related description of the commodity recommendation method in this embodiment, and will not be described in detail herein.
As an optional embodiment of the present application, the above-mentioned relation variable obtaining module 232 specifically includes: constructing a matrix relation by using compressed sensing:
wherein ,
indicating that the ith user is at t m Average value of user purchase status near +1; />
Indicating that the ith user is at t m Average value of purchase status near time; lambda (lambda) i Representing the probability of the item purchased by user i; a, a iN Representing the relationship between user i and user N, and taking the value of 0 or 1, when a iN When=1, it means that user i and user N are friends or the purchase preference is the same; when a is iN When=0, it indicates user i and userN is not a friend and purchase preference is different. The specific implementation manner is described in the related description of the commodity recommendation method in this embodiment, and will not be described in detail herein.
Embodiments of the present application also provide an electronic device, as shown in fig. 5, a processor 510 and a memory 520, where the processor 510 and the memory 520 may be connected by a bus or other means.
The processor 510 may be a central processing unit (Central Processing Unit, CPU). Processor 510 may also be a chip such as other general purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), field programmable gate arrays (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or a combination thereof.
The memory 520 is used as a non-transitory computer readable storage medium for storing non-transitory software programs, non-transitory computer executable programs, and modules, such as program instructions/modules corresponding to the commodity recommendation method in the embodiments of the present invention. The processor executes various functional applications of the processor and data processing by running non-transitory software programs, instructions, and modules stored in memory.
Memory 520 may include a storage program area that may store an operating system, at least one application program required for functionality, and a storage data area; the storage data area may store data created by the processor, etc. In addition, the memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, memory 520 may optionally include memory located remotely from the processor, such remote memory being connectable to the processor through a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The one or more modules are stored in the memory 520 that, when executed by the processor 510, perform the merchandise recommendation method in the embodiment shown in fig. 1.
The details of the above electronic device may be understood correspondingly with respect to the corresponding related descriptions and effects in the embodiment shown in fig. 1, which are not repeated herein.
The present embodiment also provides a computer storage medium storing computer executable instructions that can execute the commodity recommendation method in any of the above method embodiments. Wherein the storage medium may be a magnetic Disk, an optical Disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), a Flash Memory (Flash Memory), a Hard Disk (HDD), or a Solid State Drive (SSD); the storage medium may also comprise a combination of memories of the kind described above.
It is apparent that the above examples are given by way of illustration only and are not limiting of the embodiments. Other variations or modifications of the above teachings will be apparent to those of ordinary skill in the art. It is not necessary here nor is it exhaustive of all embodiments. While still being apparent from variations or modifications that may be made by those skilled in the art are within the scope of the invention.
Claims (7)
1. The commodity recommending method is characterized by comprising the following steps of:
acquiring purchase information of all users for each commodity, wherein the purchase information comprises purchase, non-purchase and purchase time;
respectively establishing a time sequence corresponding to each commodity according to the purchase information of the user;
reconstructing a social relation network between users according to the time sequence;
recommending corresponding commodities to the user according to the social relation network and the purchase information;
the step of reconstructing the social relation network among the users according to the time sequence comprises the following steps: selecting the effective time sequence as a node state sequence; solving the node state sequence according to a compressed sensing algorithm, and acquiring a relation variable between nodes, wherein the relation scalar represents the relation between the nodes; establishing a social relationship network between the users according to the relationship variables;
the method comprises the steps of solving the node state sequence according to a compressed sensing algorithm to obtain a relation variable among nodes, wherein the method comprises the steps of constructing a matrix relation by using compressed sensing:
wherein ,
indicating that the ith user is at t m Average value of user purchase status near +1; />
Indicating that the ith user is at t m Average value of purchase status near time; lambda (lambda) i Representing the probability of the item purchased by user i; a, a iN Representing the relationship between user i and user N, and taking the value of 0 or 1, when a iN When=1, it means that user i and user N are friends or the purchase preference is the same; when a is iN When=0, it means that user i and user N are not friends and the purchase preference is different.
2. The method of claim 1, wherein the step of establishing a time series for each commodity according to the purchase information of the user comprises:
for each commodity, a row vector is established corresponding to whether each user purchases the commodity, and a time matrix is established by taking the purchased time sequence as a column vector, wherein the time matrix is as follows:
wherein j represents the number of time sequences, N represents the number of users, i represents the ith user, and all values S in the matrix ji Indicating whether the user purchases the commodity S ji =1 means that the user purchases the commodity, S ji =0 indicates that the user did not purchase the commodity.
3. The method of claim 1, wherein the recommending items to the user based on the social relationship network and the purchase information comprises:
acquiring the purchase condition of a user of any commodity in the social relationship network;
acquiring other users associated with the user presence;
and recommending the commodity to other users not buying the commodity in the social relation network.
4. A commodity recommendation device, comprising:
the information acquisition module is used for acquiring purchase information of all users aiming at each commodity, wherein the purchase information comprises purchase, non-purchase and purchase time;
the time sequence establishing module is used for respectively establishing a time sequence corresponding to each commodity according to the purchase information of the user;
the social relation network reconstruction module is used for reconstructing a social relation network among users according to the time sequence;
the commodity recommending module is used for recommending corresponding commodities to the user according to the social relation network and the purchase information;
the social relation network reconstruction module comprises:
the time sequence selecting module is used for selecting the effective time sequence as a node state sequence; the relation variable acquisition module is used for solving the node state sequence according to a compressed sensing algorithm to acquire relation variables among nodes, wherein the relation variables represent the relation among the nodes; the social relation network establishing module is used for establishing a social relation network among the users according to the relation variable;
the relation variable acquisition module is used for constructing a matrix relation formula by using compressed sensing:
wherein ,
indicating that the ith user is at t m Average value of user purchase status near +1; />
Indicating that the ith user is at t m Average value of purchase status near time; lambda (lambda) i Representing the probability of the item purchased by user i; a, a iN Representing the relationship between user i and user N, and taking the value of 0 or 1, when a iN When=1, it means that user i and user N are friends or the purchase preference is the same; when a is iN When=0, it means that user i and user N are not friends and the purchase preference is different.
5. The apparatus of claim 4, wherein the merchandise recommendation module comprises:
the user purchase condition acquisition module is used for acquiring the user purchase condition of any commodity in the social relationship network;
the associated user acquisition module is used for acquiring other users associated with the user;
and the commodity recommending sub-module is used for recommending commodities to other users who do not purchase the commodities in the social relation network.
6. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of the merchandise recommendation method of any one of claims 1-3 when the program is executed by the processor.
7. A storage medium having stored thereon computer instructions which, when executed by a processor, implement the steps of the merchandise recommendation method of any one of claims 1 to 3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911327992.8A CN111178951B (en) | 2019-12-20 | 2019-12-20 | Commodity recommendation method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911327992.8A CN111178951B (en) | 2019-12-20 | 2019-12-20 | Commodity recommendation method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111178951A CN111178951A (en) | 2020-05-19 |
| CN111178951B true CN111178951B (en) | 2023-05-16 |
Family
ID=70653981
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911327992.8A Active CN111178951B (en) | 2019-12-20 | 2019-12-20 | Commodity recommendation method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111178951B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112150195A (en) * | 2020-09-15 | 2020-12-29 | 珠海格力电器股份有限公司 | Target item recommendation method, device, equipment and computer readable medium |
| CN112561644B (en) * | 2020-12-16 | 2023-08-01 | 深圳市网联安瑞网络科技有限公司 | Commodity recommendation method and device based on link prediction and related equipment |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102346894A (en) * | 2010-08-03 | 2012-02-08 | 阿里巴巴集团控股有限公司 | Output method, system and server of recommendation information |
| CN104915734A (en) * | 2015-06-25 | 2015-09-16 | 深圳市腾讯计算机系统有限公司 | Commodity popularity prediction method based on time sequence and system thereof |
| CN108898459A (en) * | 2018-06-25 | 2018-11-27 | 中国联合网络通信集团有限公司 | A kind of Method of Commodity Recommendation and device |
| CN109102127A (en) * | 2018-08-31 | 2018-12-28 | 杭州贝购科技有限公司 | Method of Commodity Recommendation and device |
| CN110415063A (en) * | 2018-07-31 | 2019-11-05 | 北京京东尚科信息技术有限公司 | Method of Commodity Recommendation, device, electronic equipment and readable medium |
-
2019
- 2019-12-20 CN CN201911327992.8A patent/CN111178951B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102346894A (en) * | 2010-08-03 | 2012-02-08 | 阿里巴巴集团控股有限公司 | Output method, system and server of recommendation information |
| CN104915734A (en) * | 2015-06-25 | 2015-09-16 | 深圳市腾讯计算机系统有限公司 | Commodity popularity prediction method based on time sequence and system thereof |
| CN108898459A (en) * | 2018-06-25 | 2018-11-27 | 中国联合网络通信集团有限公司 | A kind of Method of Commodity Recommendation and device |
| CN110415063A (en) * | 2018-07-31 | 2019-11-05 | 北京京东尚科信息技术有限公司 | Method of Commodity Recommendation, device, electronic equipment and readable medium |
| CN109102127A (en) * | 2018-08-31 | 2018-12-28 | 杭州贝购科技有限公司 | Method of Commodity Recommendation and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111178951A (en) | 2020-05-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Baskaran et al. | The Heckscher–Ohlin model and the network structure of international trade | |
| JP5277307B2 (en) | Information recommendation method, system thereof, and server | |
| CN106355449A (en) | User selecting method and device | |
| US11676194B2 (en) | Faceted item recommendation system | |
| CN111178951B (en) | Commodity recommendation method and device | |
| WO2017028735A1 (en) | Method and device for selecting and recommending display object | |
| US9489638B2 (en) | Method and apparatus for propagating user preference information in a communications network | |
| Tatang et al. | The impact of website design quality, service quality, and enjoyment on repurchase intention through satisfaction and trust at Zalora | |
| US11270323B2 (en) | Embedding web page content based on an individual level learning mechanism | |
| CN108305181B (en) | Social influence determination method and device, information delivery method and device, equipment and storage medium | |
| US20210166179A1 (en) | Item substitution techniques for assortment optimization and product fulfillment | |
| Guo et al. | Aspect-aware point-of-interest recommendation with geo-social influence | |
| US20210256547A1 (en) | Computation of item targeting on an electronic processing platform | |
| US20220222706A1 (en) | Systems and methods for generating real-time recommendations | |
| WO2014201333A1 (en) | Systems, apparatuses and methods for providing a price point to a consumer for products in an electronic shopping cart of the consumer | |
| Kiran et al. | Understanding the influence of utilitarian and hedonic factors on buying behaviour of gen Y while purchasing smartphones | |
| KR20130118597A (en) | Item recommendation system and method | |
| CN115271884A (en) | Multi-source data-based commodity selection method and device and electronic equipment | |
| CN107870939B (en) | Mode mining method and device | |
| CN113297467A (en) | Recommendation method, recommendation device and computer storage medium | |
| US10652096B2 (en) | Methods and systems for inferring network structure from cascades | |
| CN108230042B (en) | Demand identification method and device, electronic equipment and computer readable storage medium | |
| Chen et al. | A recommendation approach dealing with multiple market segments | |
| CN111144986B (en) | Social electronic commerce website commodity recommendation method and device based on sharing behavior | |
| US11776016B2 (en) | Systems and methods for determining user personas for a user |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |