CN112667913A - Data recommendation method, device, equipment and storage medium - Google Patents
Data recommendation method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN112667913A CN112667913A CN202110084645.8A CN202110084645A CN112667913A CN 112667913 A CN112667913 A CN 112667913A CN 202110084645 A CN202110084645 A CN 202110084645A CN 112667913 A CN112667913 A CN 112667913A
- Authority
- CN
- China
- Prior art keywords
- matrix
- data
- user data
- clients
- scoring matrix
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 65
- 239000011159 matrix material Substances 0.000 claims abstract description 154
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 30
- 238000000354 decomposition reaction Methods 0.000 claims description 7
- 230000005284 excitation Effects 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 4
- 238000011176 pooling Methods 0.000 claims description 4
- 230000007423 decrease Effects 0.000 claims 1
- 230000000694 effects Effects 0.000 abstract description 6
- 230000006870 function Effects 0.000 description 26
- 238000010586 diagram Methods 0.000 description 9
- 230000008569 process Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 239000013598 vector Substances 0.000 description 4
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 235000019800 disodium phosphate Nutrition 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The application provides a data recommendation method, device, equipment and storage medium, and relates to the technical field of data recommendation. The method comprises the following steps: acquiring user data of at least one client; determining target recommendation data corresponding to each client according to a preset scoring matrix, the user data and the number of the clients; the preset scoring matrix is obtained by expanding an initial scoring matrix by a convolutional neural network according to the user data, the number of the clients and an implicit variable; and returning the target recommendation data to the corresponding client. Compared with the prior art, the problem of inaccurate recommendation effect caused by data sparsity is solved.
Description
Technical Field
The present application relates to the field of data recommendation technologies, and in particular, to a data recommendation method, apparatus, device, and storage medium.
Background
With the rapid development of the internet and the increase of intelligent mobile devices in recent years, people enter an information overload era, and in the face of information on the internet as much as the sea, it is very difficult for users to find information really meeting the interests of the users. The recommendation system can liberate users from massive information, is one of the technical means for solving the problem of information overload, and is widely developed in the research and application fields in recent years.
In the data recommendation method in the prior art, for example, a data system similar to MovieLens stores a scoring matrix in a user-item scoring mode, and subsequently, when data recommendation is performed on each user, an item ranked at the top of the user scoring ranking can be directly determined as a target item according to the scoring matrix for recommendation.
However, such a data recommendation method depends on a scoring matrix due to collaborative filtering, and when the number of articles or users increases, the dimension and sparseness also increase, thereby affecting the accuracy of recommendation.
Disclosure of Invention
An object of the present application is to provide a data recommendation method, apparatus, device and storage medium, aiming at the above deficiencies in the prior art, so as to solve the problem in the prior art that the recommendation effect is inaccurate due to the sparsity of data.
In order to achieve the above purpose, the technical solutions adopted in the embodiments of the present application are as follows:
in a first aspect, an embodiment of the present application provides a data recommendation method, where the method includes:
acquiring user data of at least one client; the user data includes: historical browsing data and/or historical consumption data;
determining target recommendation data corresponding to each client according to a preset scoring matrix, the user data and the number of the clients; the preset scoring matrix is obtained by expanding an initial scoring matrix by a convolutional neural network according to the user data, the number of the clients and an implicit variable;
and returning the target recommendation data to the corresponding client.
Optionally, before determining the target recommendation data corresponding to each client according to the preset scoring matrix, the user data and the number of the clients, the method further includes:
the convolutional neural network decomposes the initial scoring matrix according to the user data, the number of the clients and the hidden variable to obtain a first matrix and a second matrix;
and determining the preset scoring matrix according to the first matrix and the second matrix.
Optionally, decomposing, by the convolutional neural network, the initial scoring matrix according to the user data, the number of the clients, and the hidden variable to obtain a first matrix and a second matrix, including:
determining the first matrix according to the quantity of the user data and the implicit variable;
and determining the second matrix according to the number of the clients and the implicit variable.
Optionally, the determining the preset scoring matrix according to the first matrix and the second matrix includes:
and multiplying the first matrix by the transposed matrix of the second matrix to obtain the preset scoring matrix.
Optionally, the preset scoring matrix C is satisfied
Wherein A is the number of user data, B is the number of clients, N is the number of hidden variables, q is the number of hidden variablesj＝cnn(W，Xj)+εj，
II，JThe method is an indication function with the value of 0 or 1, the function represents that the recommendation is wrong when the value of the function is 0, and represents that the recommendation is correct when the value of the function is 1; CNN is a convolution network, W and X are parameters of the convolution neural network respectively, W is convolution kernel width, and X is a variable;
is represented by rijObey mean value of
Variance is delta2A normal distribution probability density function of (1);
representing a obedient mean of 0 and a variance of
Is a normal distribution probability density function.
Alternatively, the epsilonjSatisfy the Gaussian prior probability
Optionally, the user data satisfies
Where w is the weight and X is the information of the user data.
Optionally, decomposing, by the convolutional neural network, the initial scoring matrix according to the user data, the number of the clients, and the hidden variable to obtain a first matrix and a second matrix, including:
according to
Obtaining a first matrix P and a second matrix Q, wherein
Obtained by solving according to batch gradient descent.
Optionally, the weight satisfies
Optionally, the convolutional neural network comprises: data layer, convolutional layer, excitation layer, pooling layer, and output layer.
In a second aspect, another embodiment of the present application provides a data recommendation apparatus, including: the device comprises an acquisition module, a determination module and a return module, wherein:
the acquisition module is used for acquiring user data of at least one client; the user data includes: historical browsing data and/or historical consumption data;
the determining module is used for determining target recommendation data corresponding to each client according to a preset scoring matrix, the user data and the number of the clients; the preset scoring matrix is obtained by expanding an initial scoring matrix by a convolutional neural network according to the user data, the number of the clients and an implicit variable;
and the return module is used for returning the target recommendation data to the corresponding client.
Optionally, the apparatus further comprises: the decomposition module is used for decomposing the initial scoring matrix by the convolutional neural network according to the user data, the number of the clients and the hidden variable to obtain a first matrix and a second matrix;
the determining module is specifically configured to determine the preset scoring matrix according to the first matrix and the second matrix.
Optionally, the determining module is specifically configured to determine the first matrix according to the number of the user data and the hidden variable; and determining the second matrix according to the number of the clients and the implicit variable.
Optionally, the determining module is specifically configured to multiply the first matrix by the transposed matrix of the second matrix to obtain the preset scoring matrix.
In a third aspect, another embodiment of the present application provides a data recommendation apparatus, including: a processor, a storage medium and a bus, wherein the storage medium stores machine-readable instructions executable by the processor, when the data recommendation device is operated, the processor and the storage medium communicate through the bus, and the processor executes the machine-readable instructions to perform the method of any one of the first aspect.
In a fourth aspect, another embodiment of the present application provides a storage medium having a computer program stored thereon, where the computer program is executed by a processor to perform the steps of the method according to any one of the above first aspects.
The beneficial effect of this application is: by adopting the data recommendation method provided by the application, after the user data of each client is obtained, the target recommendation data corresponding to each client is determined according to the preset scoring matrix, the number of each user data and the number of the clients, and then the determined target recommendation data is returned to the corresponding client.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are required to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present application and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained from the drawings without inventive effort.
Fig. 1 is a schematic flowchart of a data recommendation method according to an embodiment of the present application;
fig. 2 is a schematic flowchart of a data recommendation method according to another embodiment of the present application;
fig. 3 is a schematic structural diagram of a convolutional neural network according to an embodiment of the present application;
fig. 4 is a schematic structural diagram of a data recommendation device according to an embodiment of the present application;
fig. 5 is a schematic structural diagram of a data recommendation device according to another embodiment of the present application;
fig. 6 is a schematic structural diagram of a data recommendation device according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments.
The components of the embodiments of the present application, generally described and illustrated in the figures herein, can be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the present application, presented in the accompanying drawings, is not intended to limit the scope of the claimed application, but is merely representative of selected embodiments of the application. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present application without making any creative effort, shall fall within the protection scope of the present application.
Additionally, the flowcharts used in this application illustrate operations implemented according to some embodiments of the present application. It should be understood that the operations of the flow diagrams may be performed out of order, and steps without logical context may be performed in reverse order or simultaneously. One skilled in the art, under the guidance of this application, may add one or more other operations to, or remove one or more operations from, the flowchart.
The data recommendation method provided by the application is applied to a scene of data recommendation for a user, and in the following embodiments of the application, data recommendation for an application program is taken as an example for explanation; it should be understood that although the embodiments of the present application are mainly explained with reference to data recommendation of an application program, the application scenarios of the method provided by the present application are not limited thereto, and the method provided by the present application may be applied to any scenario that requires data recommendation to a user, for example, data recommendation of a web page, data recommendation of an applet or data recommendation of advertisement placement, and the like.
Before the method provided by the application, in the prior art, the items are usually determined in a user-item scoring form and stored in an initial scoring matrix, and then in the process of data recommendation for each user, the items corresponding to each user and scored ahead can be directly obtained according to the initial scoring matrix as target items, and the target items are recommended to each user.
The data recommendation method provided by the embodiment of the present application is explained below with reference to a plurality of specific application examples. Fig. 1 is a schematic flowchart of a data recommendation method according to an embodiment of the present application, and as shown in fig. 1, the method includes:
s101: user data of at least one client is obtained.
Wherein the user data includes: historical browsing data and/or historical consumption data.
For example, taking the example that a user logs in a first shopping application program through a client installed with the first shopping application program as an example, first, after the user successfully logs in the first shopping application program, user data corresponding to the user may be obtained through historical browsing data or historical consumption data of the user within a preset time on the first shopping application program on the client; the preset time may be, for example, the last week or the last ten days, and may be flexibly adjusted according to the user requirement, which is not limited to the above embodiments.
S102: and determining target recommendation data corresponding to each client according to the preset scoring matrix, the user data and the number of the clients.
Illustratively, in an embodiment of the present application, the preset scoring matrix is obtained by training a convolutional neural network after expanding an initial scoring matrix according to user data, the number of clients, and hidden variables, where the hidden variables are non-observable random variables, and may be browsing records and purchase records of a non-mobile phone owner, for example, a mobile phone owner is a male, but has a consumption record of a female article, and may infer that the owner has a role such as a girlfriend, so that an article such as "send girlfriend gift" may be added in a proper amount, and in statistics, inference may be made on the hidden variables through observable variables such as a historical consumption record or a historical browsing record of a user, so as to determine the hidden variables.
In the embodiment of the application, the preset scoring matrix is obtained after expansion, so that the corresponding data coverage in the preset scoring matrix is more complete, the effect of the target recommendation data determined according to the preset scoring matrix is better, and the possibility that the target recommendation data is data meeting the user requirements is higher; for example, the application program is taken as a shopping application program for illustration, the target recommendation data is data with a high probability of being consumed by the user, and the target recommendation data determined in such a way can increase the use stickiness between the application programs of the user and improve the use experience of the user.
S103: and returning the target recommendation data to the corresponding client.
For example, in an embodiment of the present application, taking an application as a novel reading application as an example, after a user logs in the novel reading application through a client installed with the novel reading application, target recommendation data at this time is a target recommendation novel, after the target recommendation novel of the user is determined according to the method provided by the present application, for example, when the user views the novel recommendation on the novel reading application, or continues to slide down an interface after reading the novel, each target recommendation novel can be recommended on the novel recommendation interface according to the target recommendation novel, for example, recommendation can be performed according to the ranking of each target recommendation novel, the position of the target recommendation novel with the highest ranking displayed on the interface is also front, the position of the target recommendation novel with the lowest ranking displayed on the interface is also rear, so that each recommended novel is determined according to the historical reading record and the historical purchase record of the user, therefore, the probability that the user clicks each target recommendation novel to read is higher, so that the use experience of the user is improved, the use stickiness between the user and the novel reading application program can be increased, the specific method for recommending the target recommendation data can be flexibly adjusted according to the user needs, and the method is not limited to the method provided by the embodiment.
By adopting the data recommendation method provided by the application, after the user data of each client is obtained, the target recommendation data corresponding to each client is determined according to the preset scoring matrix, the number of each user data and the number of the clients, and then the determined target recommendation data are returned to the corresponding clients.
Optionally, on the basis of the above embodiments, the embodiments of the present application may further provide a data recommendation method, and an implementation process of the above method is described as follows with reference to the accompanying drawings. Fig. 2 is a schematic flow chart of a data recommendation method according to another embodiment of the present application, as shown in fig. 2, before S102, the method may further include:
s104: and the convolutional neural network decomposes the initial scoring matrix according to the user data, the number of the clients and the hidden variable to obtain a first matrix and a second matrix.
Illustratively, in one embodiment of the present application, the first matrix P may be determined according to the amount of user data a and the hidden variable NN*A(ii) a Determining a second matrix Q according to the number B of the clients and the hidden variable NN*B。
S105: and determining a preset scoring matrix according to the first matrix and the second matrix.
Illustratively, in one embodiment of the present application, the first matrix P may beN*AAnd a second matrix QN*BThe transposed matrix is multiplied, the preset scoring matrix C is obtained according to the multiplication result, and the preset scoring matrix determined in the mode not only meets the existing value in the initial scoring matrix, but also is filled with the value which is not scored, so that the expansion of the data in the scoring matrix is realized.
In the embodiment of the present application, the scoring matrix C is preset to satisfy
Wherein A is the number of user data, B is the number of clients, and N is the number of hidden variables, for user data, unlike conventional matrix decomposition, the matrix decomposition of the present application incorporates a convolutional neural network, where q is the number of user data, and q is the number of hidden variables, and whereinj＝cnn(W，Xj)+εj，
II，JThe method is an indication function with the value of 0 or 1, the function represents that the recommendation is wrong when the value of the function is 0, and represents that the recommendation is correct when the value of the function is 1;
is represented by rijObey mean value of
Variance is delta2A normal distribution probability density function of (1);
representing a obedient mean of 0 and a variance of
Is a normal distribution probability density function.
Mean of 0 and variance of user
The gaussian prior probability of (a) is as follows:
for each weight W in WkHas a mean value of 0 and a variance of
The gaussian prior probability of (a) is as follows:
therefore, for the user data in the present application, the gaussian prior probability that needs to be satisfied is as follows:
where w is the weight and X is the information of the user data.
Therefore, by the method provided by the application, the information potential vector of the user data which can be obtained from the convolutional neural network is used as the mean value of the Gaussian distribution, and the Gaussian noise in the user data is used as the variance in the Gaussian distribution, which is the core of the convolutional neural network and the initial scoring matrix decomposition cooperation method, and the description information and the scoring information of the user data can be fully mined.
Fig. 3 is a schematic structural diagram of a convolutional neural network according to an embodiment of the present disclosure, as shown in fig. 3, in an embodiment of the present disclosure, the convolutional neural network includes five layers, which may be: data layer, convolutional layer, excitation layer, pooling layer and output layer, wherein:
the data layer is used for converting original information of a description file of user data into a matrix representing file information, then the description file can be regarded as a sequence comprising L word lengths, and then word vectors form an information matrix which can be initialized at will or embedded into a model by using weights obtained by trained words; this way of training the weights obtained before as initialization weights for the model allows faster convergence of the weights and thus optimal results.
The convolutional layer is used to fix the connection weight of each neuron to be regarded as a template, wherein each neuron only concerns one characteristic, and a group of fixed weights and data in different windows are subjected to inner product.
The excitation layer is used for carrying out nonlinear mapping on the output result of the convolutional layer through an excitation function. In one embodiment of the present application, the excitation function may be, for example: an S-type growth curve (Sigmoid) function, a hyperbolic tangent (Tanh) function, a Linear rectification function (RecuU), an activation function Leaky ReLU, an ELU function, and a Maxout function; in one embodiment of the present application, a modified linear element excitation function may be used, for example.
The pooling layer is used to compress the amount of data and parameters, reducing overfitting.
The output layer is used to complete the final output task, and in one embodiment of the present application, for example, the BP algorithm (Error Back Propagation) may be used to multiply the output tasks step by using a chain derivative rule until dW and db are solved. And finally, obtaining a context feature vector s through nonlinear mapping.
From the whole process, the convolutional neural network is equivalent to a function, and by giving a specific document information (i.e. user data), the output value is the feature vector s of the document.
Therefore, in the method provided by the present application, the first matrix, the second matrix and the corresponding weights may be obtained according to a decomposition model of the convolutional neural network on the initial scoring matrix in the following form:
wherein
The method is obtained by solving according to batch gradient descent, so that the regularization is realized according to a loss function, and overfitting is avoided.
In one embodiment of the present application, the loss function may be defined according to the optimal result, and in the training process of the convolutional neural network, a plurality of functions are used for comparison, and finally the loss function with the optimal result is selected.
After the first matrix and the second matrix are determined, multiplying the transpose of the first matrix and the transpose of the second matrix to obtain a preset scoring matrix, wherein the preset scoring matrix comprises a plurality of accurate user-recommendation scores; and accurate recommendation of the target recommendation data is achieved according to the finally obtained preset scoring matrix.
By adopting the data recommendation method provided by the application, the initial scoring matrix can be split according to the convolutional neural network and the user data, the number of the clients and the hidden variable, the expanded preset scoring matrix is obtained according to the split first matrix and the split second matrix, then after the user data of each client is obtained, the target recommendation data corresponding to each client is determined according to the preset scoring matrix, the user data and the number of the clients, and the determined target recommendation data is returned to the corresponding client, because the preset scoring matrix in the application is obtained by expanding the initial scoring matrix according to the user data, the number of the clients and the hidden variable by the neural network, the user data and the initial scoring matrix can be fully mined, so that the data in the preset scoring matrix is denser, and the probability that the recommended target recommendation data is the data required by the user is higher, the problem of poor recommendation effect caused by sparse initial scoring matrix data is solved.
The following explains the data recommendation device provided in the present application with reference to the accompanying drawings, where the data recommendation device can execute the method of any one of the data recommendation devices shown in fig. 1 to 3, and specific implementation and beneficial effects of the method are referred to above, and are not described again below.
Fig. 4 is a schematic structural diagram of a data recommendation device according to an embodiment of the present application, and as shown in fig. 4, the data recommendation device includes: an acquisition module 201, a determination module 202, and a return module 203, wherein:
an obtaining module 201, configured to obtain user data of at least one client; the user data includes: historical browsing data and/or historical consumption data.
The determining module 202 is configured to determine target recommendation data corresponding to each client according to a preset scoring matrix, the user data and the number of the clients; the preset scoring matrix is obtained by expanding the initial scoring matrix by the convolutional neural network according to user data, the number of clients and hidden variables.
And the returning module 203 is used for returning the target recommendation data to the corresponding client.
Fig. 5 is a schematic structural diagram of a data recommendation device according to an embodiment of the present application, and as shown in fig. 5, the device further includes: and the decomposition module 204 is used for decomposing the initial scoring matrix by the convolutional neural network according to the user data, the number of the clients and the hidden variable to obtain a first matrix and a second matrix.
The determining module 202 is specifically configured to determine a preset scoring matrix according to the first matrix and the second matrix.
Optionally, the determining module 202 is specifically configured to determine the first matrix according to the number of the user data and the hidden variable; and determining a second matrix according to the number of the clients and the implicit variable.
Optionally, the determining module 202 is specifically configured to multiply the first matrix by the transposed matrix of the second matrix to obtain a preset scoring matrix.
The above-mentioned apparatus is used for executing the method provided by the foregoing embodiment, and the implementation principle and technical effect are similar, which are not described herein again.
These above modules may be one or more integrated circuits configured to implement the above methods, such as: one or more Application Specific Integrated Circuits (ASICs), or one or more microprocessors (DSPs), or one or more Field Programmable Gate Arrays (FPGAs), among others. For another example, when one of the above modules is implemented in the form of a Processing element scheduler code, the Processing element may be a general-purpose processor, such as a Central Processing Unit (CPU) or other processor capable of calling program code. For another example, these modules may be integrated together and implemented in the form of a system-on-a-chip (SOC).
Fig. 6 is a schematic structural diagram of a data recommendation device according to an embodiment of the present application, where the data recommendation device may be integrated in a server, a terminal device, or a chip of the terminal device.
As shown in fig. 6, the data recommendation apparatus includes: a processor 501, a storage medium 502, and a bus 503, wherein:
the processor 501 is used for storing a program, and the processor 501 calls the program stored in the storage medium 502 to execute the method embodiment corresponding to fig. 1-3. The specific implementation and technical effects are similar, and are not described herein again.
Optionally, the present application also provides a program product, such as a storage medium, on which a computer program is stored, including a program, which, when executed by a processor, performs embodiments corresponding to the above-described method.
In the several embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the units is only one logical division, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional unit.
The integrated unit implemented in the form of a software functional unit may be stored in a computer readable storage medium. The software functional unit is stored in a storage medium and includes several instructions for enabling a computer device (which may be a personal computer, a server, or a network device) or a processor (processor) to perform some steps of the methods according to the embodiments of the present application. And the aforementioned storage medium includes: a U disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
Claims (15)
1. A method for recommending data, the method comprising:
acquiring user data of at least one client; the user data includes: historical browsing data and/or historical consumption data;
determining target recommendation data corresponding to each client according to a preset scoring matrix, the user data and the number of the clients; the preset scoring matrix is obtained by expanding an initial scoring matrix by a convolutional neural network according to the user data, the number of the clients and an implicit variable;
and returning the target recommendation data to the corresponding client.
2. The method of claim 1, wherein before determining the target recommendation data corresponding to each client according to a preset scoring matrix, the user data and the number of the clients, the method further comprises:
the convolutional neural network decomposes the initial scoring matrix according to the user data, the number of the clients and the hidden variable to obtain a first matrix and a second matrix;
and determining the preset scoring matrix according to the first matrix and the second matrix.
3. The method of claim 2, wherein the convolutional neural network decomposing the initial scoring matrix according to the user data, the number of clients, and the hidden variables to obtain a first matrix and a second matrix, comprising:
determining the first matrix according to the quantity of the user data and the implicit variable;
and determining the second matrix according to the number of the clients and the implicit variable.
4. The method of claim 2, wherein said determining the preset scoring matrix from the first matrix and the second matrix comprises:
and multiplying the first matrix by the transposed matrix of the second matrix to obtain the preset scoring matrix.
5. The method of claim 2, wherein the predetermined scoring matrix C is satisfied
Wherein A is the amount of user data, B is the amount of clients, N is the amount of hidden variables,
II，Jis an indicator function with a value of 0 or 1;
is represented by rijObey mean value of
Variance is delta2Normal distribution probability density function of；
Representing a obedient mean of 0 and a variance of
Is a normal distribution probability density function.
6. The method of claim 5, wherein epsilonjSatisfy the Gaussian prior probability
7. The method of claim 5, wherein the user data satisfies
Where w is the weight and X is the information of the user data.
8. The method of claim 2, wherein the convolutional neural network decomposing the initial scoring matrix according to the user data, the number of clients, and the hidden variables to obtain a first matrix and a second matrix, comprising:
according to
Obtaining a first matrix P and a second matrix Q, wherein
Is according to a batchAnd solving the quantity gradient decline.
9. The method of claim 5, wherein the weights are satisfied
10. The method of claim 1, wherein the convolutional neural network comprises: data layer, convolutional layer, excitation layer, pooling layer, and output layer.
11. A data recommendation apparatus, characterized in that the apparatus comprises: the device comprises an acquisition module, a determination module and a return module, wherein:
the acquisition module is used for acquiring user data of at least one client; the user data includes: historical browsing data and/or historical consumption data;
the determining module is used for determining target recommendation data corresponding to each client according to a preset scoring matrix, the user data and the number of the clients; the preset scoring matrix is obtained by expanding an initial scoring matrix by a convolutional neural network according to the user data, the number of the clients and an implicit variable;
and the return module is used for returning the target recommendation data to the corresponding client.
12. The apparatus of claim 11, wherein the apparatus further comprises: the decomposition module is used for decomposing the initial scoring matrix by the convolutional neural network according to the user data, the number of the clients and the hidden variable to obtain a first matrix and a second matrix;
the determining module is specifically configured to determine the preset scoring matrix according to the first matrix and the second matrix.
13. The apparatus according to claim 11, wherein the determining module is specifically configured to determine the first matrix according to the amount of the user data and the hidden variable; and determining the second matrix according to the number of the clients and the implicit variable.
14. A data recommendation device, characterized in that the device comprises: the apparatus comprises: a processor, a storage medium and a bus, the storage medium storing machine-readable instructions executable by the processor, the processor and the storage medium communicating via the bus when the data recommendation device is operating, the processor executing the machine-readable instructions to perform the method of any of claims 1-8.
15. A storage medium, characterized in that the storage medium has stored thereon a computer program which, when being executed by a processor, performs the method of any of the preceding claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110084645.8A CN112667913B (en) | 2021-01-21 | 2021-01-21 | Data recommendation method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110084645.8A CN112667913B (en) | 2021-01-21 | 2021-01-21 | Data recommendation method, device, equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112667913A true CN112667913A (en) | 2021-04-16 |
| CN112667913B CN112667913B (en) | 2024-07-05 |
Family
ID=75415868
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110084645.8A Active CN112667913B (en) | 2021-01-21 | 2021-01-21 | Data recommendation method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112667913B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108090229A (en) * | 2018-01-10 | 2018-05-29 | 广东工业大学 | A kind of method and apparatus that rating matrix is determined based on convolutional neural networks |
| CN110807154A (en) * | 2019-11-08 | 2020-02-18 | 内蒙古工业大学 | Recommendation method and system based on hybrid deep learning model |
-
2021
- 2021-01-21 CN CN202110084645.8A patent/CN112667913B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108090229A (en) * | 2018-01-10 | 2018-05-29 | 广东工业大学 | A kind of method and apparatus that rating matrix is determined based on convolutional neural networks |
| CN110807154A (en) * | 2019-11-08 | 2020-02-18 | 内蒙古工业大学 | Recommendation method and system based on hybrid deep learning model |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112667913B (en) | 2024-07-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111259133B (en) | Personalized recommendation method integrating multiple information | |
| CN105701191B (en) | Pushed information click rate estimation method and device | |
| CN109241424B (en) | A kind of recommended method | |
| CN111061961A (en) | Multi-feature-fused matrix decomposition interest point recommendation method and implementation system thereof | |
| CN111784455A (en) | Article recommendation method and recommendation equipment | |
| CN112307332B (en) | Collaborative filtering recommendation method and system based on user portrait clustering and storage medium | |
| CN106157156A (en) | A kind of cooperation recommending system based on communities of users | |
| CN113033090B (en) | Push model training method, data push device and storage medium | |
| CN111930931B (en) | Abstract evaluation method and device | |
| CN114065048A (en) | Article recommendation method based on multi-different-pattern neural network | |
| CN112214499A (en) | Graph data processing method and device, computer equipment and storage medium | |
| CN112148994A (en) | Information push effect evaluation method and device, electronic equipment and storage medium | |
| CN106651427A (en) | Data association method based on user behavior | |
| CN117670464A (en) | Commodity recommendation method for removing noise interaction signals of social platform | |
| CN112667913A (en) | Data recommendation method, device, equipment and storage medium | |
| Wang et al. | Incorporating auxiliary information in collaborative filtering data update with privacy preservation | |
| CN114238758B (en) | User portrait prediction method based on multi-source cross-border data fusion | |
| CN115221316A (en) | Knowledge base processing method, model training method, computer device and storage medium | |
| CN112650869B (en) | Image retrieval reordering method and device, electronic equipment and storage medium | |
| CN114443948A (en) | Ranking model training method, ranking method and device based on multi-scene data | |
| CN113947456A (en) | Online shop matching method and device, equipment, medium and product thereof | |
| CN113362034A (en) | Position recommendation method | |
| CN113111257A (en) | Collaborative filtering-based recommendation method for fusing multi-source heterogeneous information | |
| CN112734519A (en) | Commodity recommendation method based on convolution self-encoder network | |
| CN111428741A (en) | Network community discovery method and device, electronic equipment and readable storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |