WO2015035556A1

WO2015035556A1 - Recommendation method and device

Info

Publication number: WO2015035556A1
Application number: PCT/CN2013/083218
Authority: WO
Inventors: 张洪波; 格卢霍夫瓦列里
Original assignee: 华为技术有限公司
Priority date: 2013-09-10
Filing date: 2013-09-10
Publication date: 2015-03-19
Also published as: CN104854580A; CN104854580B

Abstract

Embodiments of the present invention provide a recommendation method and a device. Recommendation efficiency of a recommendation system in a massive data environment is improved through parallel calculation, and a recommendation effect of the recommendation system is improved through considering implicit feedbacks of users, wherein the recommendation method comprises: placing scoring data in a scoring data set to at least one data layer respectively; based on a preset recommendation system model and the scoring data in the data layers, computing parameters of the recommendation system model in the data layers in parallel and using parameters of each layer of the data layers as initial values of a next layer of the data layers until optimal parameters of the recommendation system model are obtained; based on the optimal parameters and the recommendation system model, obtaining a predicted scoring value of each user on each product; and recommending products to the users based on the predicted scoring value.

Description

A recommended method and device

Technical field

The present invention relates to the field of information processing, and in particular, to a recommendation method and device.

Background technique

With the continuous development of the network, the amount of network information has also exploded. The emergence of information recommendation methods enables users to obtain information of interest in massive network information. The existing information recommendation method is to display feedback from users. The information and implicit feedback information are fused in the data model. The optimization problem of minimizing the loss function is solved by the stochastic gradient descent algorithm. The parameters in the data model are solved by serial calculation, and the products that have not yet been scored by the user are based on this parameter. For products that are of interest to users who make predictions and make recommendations for users, this method has higher recommendation accuracy because both the user's display feedback and implicit feedback are considered. The inventors have found that the existing information recommendation method is not suitable for data processing in a massive data environment because it adopts a serial computing method, and is extremely inefficient in a mass data environment, and also affects the effect recommended to the user.

Summary of the invention

The embodiment of the invention provides a recommendation method and device, and improves the recommendation efficiency of the recommendation system in the mass data environment by parallel computing, and improves the recommendation effect of the recommendation system by considering the implicit feedback of the user.

In order to achieve the above objectives, embodiments of the present invention adopt the following technical solutions:

In a first aspect, an embodiment of the present invention provides a recommendation method, including: placing score data in a score data set to at least two data layers, where the score data corresponds to a user and a product respectively, and each The user and the product corresponding to any two of the score data in one of the data layers are different; Calculating parameters of the recommended system model in the data layer in parallel according to the preset recommendation system model and the scoring data in the data layer, and using the parameters of each layer of the data layer as initial values of the corresponding next layer of data layers And obtaining an optimal parameter of the recommendation system model; wherein the recommendation system model is a correspondence between a score prediction value of each product for each user and a parameter of the average score and the recommendation system model And obtaining, according to the optimal parameter and the recommendation system model, a score prediction value of each product for each product, and recommending a product to the user according to the score prediction value. According to a first possible implementation manner, in combination with the first aspect, the recommendation system model includes a recommendation system model that provides implicit feedback, a recommendation system model that does not provide implicit feedback, and a recommendation system model that considers space-time speciality And a recommended system model for asymmetric potential factors.

According to the second possible implementation manner, in combination with the first aspect or the first possible implementation manner, the recommendation system includes: a first recommendation system model

—丄

j≡N(u) Alternatively, the second recommendation system model indicates, in the first recommendation model and the second recommendation model, a score predicted by the user u for the product i, and μ indicates all the ratings in the score data set The average value of the data, b „ represents the offset of the user u from the average user score, represents the offset of the product i from the average product score, represents the product factor vector, T represents the transpose operator symbol, p „ User factor vector,

Further, in the first recommendation model, |N(w)| represents a collection size of all products in which user u provides an implicit preference, and N(w) represents a collection of all products in which user u provides an implicit preference. ; represents a factor vector associated with product j that is used to characterize implicit feedback information.

According to a third possible implementation manner, in combination with the second possible implementation manner, the method further includes: determining a mean square error of the score prediction value and the score data, and the recommendation system The relationship between the parameters of the unified model results in a cost function of the recommended system model, wherein the cost function comprises: a first cost function

Or, the second cost function

∑ [r _ui -μ- -b - qp _u f + {b _u ² + bf) + ^\\ +\\p _u \\ ) where ||*|| represents the sum of the squares of all the elements of the vector *, And ^ is a regularization factor. According to the fourth possible implementation manner, in combination with any one of the first aspect, the first to the third possible implementation manner, the preset recommendation system model and the rating data in the data layer, The parameters of the recommended system model in the data layer are calculated in parallel, and the parameters of each layer of the data layer are used as initial values of the corresponding next layer of data layers, until the optimal parameters of the recommended system model are obtained, including:

A: calculating an average score of all the score data in the score data set;

B: sequentially calculating the parameters of the data layer of each layer by using parallel computing, and using the parameter calculated by the data layer of each layer as the initial value of the parameter of the next data layer; wherein, the first layer The initial value of the parameter of the data layer is set by the system;

C: judging whether the recommendation system model converges according to the parameter calculated by the data layer of the last layer, if convergence, the calculation ends, and the optimal parameter is obtained; if not, the last layer of the data layer is obtained The calculated parameters are used as the initial values of the parameters of the data layer of the first layer, and the steps B and C are repeated. According to the fifth possible implementation manner, in combination with the fourth possible implementation manner, the step B includes:

B1: obtaining an initial estimated value of the score data of the data layer of the layer according to the initial value of the parameter of the data layer of each layer, and further, according to the score data of the data layer of the layer and the initial The estimated value is obtained by the scoring error of the data layer of the layer; B2: acquiring, according to the scoring error, a parameter calculated by the data layer of the layer;

B3: taking the parameter calculated by the data layer of the layer as the initial value of the parameter of the next layer, and obtaining the parameter calculated by the next layer of the data layer according to steps B1 and B2 until the data layer of the last layer is calculated. The resulting parameters. According to the sixth possible implementation manner, in combination with the fourth or fifth possible implementation manner, determining, according to the parameter calculated by the data layer of the last layer, whether the recommendation system model converges, including: Calculating the calculated parameters of the data layer of the last layer and the parameters calculated by the data layer of the last layer obtained by the previous calculation are all substituted into the cost function for calculation, and if the cost function is substituted for calculation The difference between the results of the calculation is not greater than the preset threshold value, and the parameter calculated by the data layer of the last layer obtained by the current calculation is convergent. Otherwise, the data of the last layer obtained by the current calculation is converged. The parameters calculated by the layer are not convergent. According to the seventh possible implementation manner, in combination with the fifth possible implementation manner, the obtaining the parameter calculated by the layer of the data layer according to the scoring error further includes: expressing the expression in the first recommendation model Let ft+|N( _M ) X be used as an equivalent parameter, and use the auxiliary variable to represent the equivalent parameter, ie =A+|N( _M )|4 ; then according to the gradient of the auxiliary variable Δζ _Μ =2e _M · ^, obtaining the auxiliary variable to obtain the equivalent parameter; and acquiring the parameter q according to the auxiliary variable, ie, +7 ₂ A^,), wherein the symbol represents the updated symbol, that is, the calculated value on the right side of the updated symbol Instead of the variable value to the left of the update symbol, the parameters appearing on the right side of the update symbol are the initial values of the corresponding parameters, and the parameters appearing on the left side of the update symbol are the updated values of the parameters.

The second aspect provides a recommendation device, including: a data placement unit, configured to separately set the score data in the score data set to at least two data layers, where the score data corresponds to the user and the product respectively. And the user and the product corresponding to any two of the score data in each of the data layers are different; the parallel computing unit is configured to perform parallel calculation according to the preset recommendation system model and the score data in the data layer. The parameters of the system model are recommended in the data layer, and the parameters of each layer of the data layer are used as initial values of the corresponding next layer of data layers until the optimal parameters of the recommended system model are obtained; wherein the recommendation system The model is a relationship between a score prediction value of each product for each product and a parameter of the average score and the recommendation system model; a prediction recommendation unit for using the optimal parameter and the recommendation The system model obtains a score prediction value for each product for each user, and recommends a product to the user based on the score prediction value. According to a first possible implementation manner, in combination with the second aspect, the recommendation system model includes a recommendation system model that provides implicit feedback, a recommendation system model that does not provide implicit feedback, and a recommendation system model that considers space-time special characteristics. And a recommended system model for asymmetric potential factors.

According to the second possible implementation manner, in combination with the second aspect or the first possible implementation manner, the recommendation system includes: a first recommendation system model

—丄 j≡N(u)

Alternatively, the second recommendation system model indicates, in the first recommendation model and the second recommendation model, a score predicted by the user u for the product i, and μ represents an average value of all the score data in the score data set, b „ indicates the offset of the user u from the average user score, indicating the offset of the product i from the average product score, representing the product factor vector, T representing the transpose operator symbol, p „ representing the user factor vector,

Further, in the first recommendation model, |N(w)| represents a collection size of all products in which user u provides an implicit preference, and N(w) represents a collection of all products in which user u provides an implicit preference. ; represents a factor vector associated with product j that is used to characterize implicit feedback information. According to a third possible implementation manner, in combination with the second possible implementation manner, the method further includes: a cost function generating unit, configured to: according to the mean square error of the score prediction value and the score data, and the recommendation system model The relationship between the parameters yields a cost function of the recommended system model, wherein the cost function includes: a first cost function

Or, the second cost function

∑ [r _ui - μ - - b - qp _u f + {b _u ² + bf ) + ^\\ + \\p _u \\ ) where |*| represents the sum of the squares of all the elements of the vector *, and ^ Is a regularization factor. According to the fourth possible implementation, in combination with the second aspect, the first to the third possible implementation, the parallel computing unit includes:

An average score calculation sub-unit, configured to calculate an average score of all the score data in the score data set; a hierarchical calculation sub-unit, configured to sequentially calculate a parameter of the data layer of each layer by using a parallel calculation manner, and The parameter calculated by the data layer of each layer is used as a parameter initial value of the data layer of the next layer; wherein, the initial value of the parameter of the data layer of the first layer is set by the system; the convergence determining subunit is used according to The parameter calculated by the data layer of the last layer determines whether the recommendation system model converges, and if it converges, the calculation ends, and the optimal parameter is obtained; if not, the data layer calculated by the last layer is calculated. The parameter is used as a parameter initial value of the data layer of the first layer, and the parameter initial value is transmitted to the hierarchical calculation subunit to perform hierarchical calculation.

According to a fifth possible implementation manner, in combination with the fourth possible implementation manner, the hierarchical calculation sub-unit is further configured to: a score error generating module, configured to initialize an initial value of the data layer according to each layer The recommendation system model obtains an initial estimate of the score data of the data layer of the layer, And obtaining a scoring error of the layer of the data layer according to the scoring data of the layer of the data layer and the initial estimated value;

a parameter calculation module, configured to acquire, according to the scoring error, a parameter calculated by the data layer of the layer; and a calculation control module, configured to use the parameter calculated by the data layer of the layer as an initial value of a parameter of a data layer of a next layer And obtaining, by the scoring error generating module and the parameter calculating module, the parameters calculated by the next layer of the data layer until the parameter calculated by the data layer of the last layer is obtained.

According to the sixth possible implementation manner, in combination with the fourth or the fifth possible implementation manner, the convergence determining subunit is further configured to: calculate the parameter calculated by the last layer of the data layer obtained by the current calculation And the parameters calculated by the data layer of the last layer obtained by the previous calculation are all substituted into the cost function for calculation, and if the difference between the results of the calculation performed by the cost function is not greater than a preset threshold, The parameters calculated by the data layer of the last layer obtained by the second calculation are convergent. Otherwise, the parameters calculated by the data layer of the last layer obtained in this calculation are not convergent.

According to the seventh possible implementation, in combination with the fifth possible implementation, the parameter calculation module is further configured to use the expression ft + |N( _M ) X in the first recommendation model as an equivalent a parameter, and an auxiliary variable to represent the equivalent parameter, ie, = A + |N( _M )|4 ; and then obtaining the auxiliary variable according to the gradient Δζ _Μ = 2e _M · ^ of the auxiliary variable Equivalent parameter; and obtaining the parameter q according to the auxiliary variable, ie, + 7 ₂ A^, ), wherein the symbol represents an update symbol, that is, replacing the variable value on the left side of the update symbol with the calculated value on the right side of the update symbol, updating the right side of the symbol The parameters that appear are the initial values of the corresponding parameters, and the parameters that appear to the left of the update symbol are the updated values of the parameters.

In a third aspect, a recommendation device is provided, including a processor and a memory, where the processor is configured to separately place the score data in the score data set into at least two data layers, where the score data is associated with the user and Product separately - corresponding, And the user and the product corresponding to any two of the score data in each of the data layers are different; and calculating the data layer in parallel according to the preset recommendation system model and the score data in the data layer Recommending parameters of the system model, and taking the parameters of each layer of the data layer as initial values of the corresponding next layer of data layers until obtaining the optimal parameters of the recommended system model; wherein the recommended system model is for each user a correspondence between the score prediction value of each product and the average score and the parameters of the recommendation system model;

And obtaining a score prediction value of each product for each product according to the optimal parameter and the recommendation system model, and recommending a product to the user according to the score prediction value; the memory is used to save the score data set and The program executed by the processor and the result of the execution. According to a first possible implementation manner, in combination with the third aspect, the recommendation system model includes a recommendation system model that provides implicit feedback, a recommendation system model that does not provide implicit feedback, and a recommendation system model that considers space-time speciality And a recommended system model for asymmetric potential factors.

According to the second possible implementation manner, in combination with the third aspect or the first possible implementation manner, the recommendation system includes: a first recommendation system model

—丄

Further, in the first recommendation model, |N(w)| represents a set size of all products in which user u provides an implicit preference, and N(w) represents a place in which user u provides an implicit preference. There is a collection of products; represents a factor vector associated with product j that is used to characterize implicit feedback information.

According to a third possible implementation manner, in combination with the second possible implementation manner, the processor is further configured to: according to the mean square error of the score prediction value and the score data, and the parameter of the recommended system model The relationship between the costs of the recommended system model, wherein the cost function comprises: a first cost function

Or, the second cost function

Where |*| represents the sum of the squares of all the elements of the vector *, and ^ is the regularization factor. According to a fourth possible implementation manner, in combination with any one of the third aspect, the first to the third possible implementation manner, the processor is configured to:

A: calculating an average score of all the score data in the score data set;

B: calculating the parameters of the data layer of each layer in a parallel calculation manner, and using the parameter calculated by the data layer of each layer as the initial value of the parameter of the next data layer; wherein, the first layer The initial value of the parameter of the data layer is set by the system;

C: determining whether the recommended system model converges according to the parameter calculated by the data layer of the last layer. If convergence, the calculation ends, and the optimal parameter is obtained; if not, the last layer of the data layer is obtained. The calculated parameters are used as the initial values of the parameters of the data layer of the first layer, and the steps B and C are repeated. According to a fifth possible implementation, in combination with the fourth possible implementation, the processor is further configured to:

B 1 : obtaining an initial estimated value of the score data of the layer of the data layer according to the parameter initial value of the data layer of each layer and the recommendation system model, and further according to the score data of the layer of the layer and the The initial estimate obtains a score error of the data layer of the layer; B 2: obtaining, according to the scoring error, a parameter calculated by the data layer of the layer;

B 3: taking the parameter calculated by the data layer of the layer as the initial value of the parameter of the next layer of data, and obtaining the parameter calculated by the next layer of the data layer according to steps B 1 and B 2 until the last layer is obtained. The data layer calculates the parameters. According to the sixth possible implementation manner, in combination with the fourth or fifth possible implementation manner, the processor is configured to calculate a parameter calculated by the last layer of the data layer obtained by the current calculation and a previous calculation The obtained parameters of the data layer calculated in the last layer are all substituted into the cost function for calculation. If the difference between the results calculated by the cost function is not greater than a preset threshold, then the calculation is performed. The obtained parameters of the data layer calculated in the last layer are converged. Otherwise, the parameters calculated by the data layer of the last layer obtained in this calculation are not converged. According to the seventh possible implementation, in combination with the fifth possible implementation, the processor is configured to acquire, according to the scoring error, the calculated parameter of the layer of the data layer, further comprising: the processor: The expression + |N( _M ) X in the first recommendation model is eW(") as an equivalent parameter, and the auxiliary variable is used to represent the equivalent parameter, ie

"

Then obtaining the auxiliary parameter according to the gradient Δζ _Μ = 2e _M · ^ of the auxiliary variable, and obtaining the parameter q according to the auxiliary variable; ie, the parameter q is + 7 ₂ A^, ) The symbol indicates the update symbol, that is, the variable value on the left side of the update symbol is replaced by the calculated value on the right side of the update symbol. The parameters appearing on the right side of the update symbol are the initial values of the corresponding parameters, and the parameters appearing on the left side of the update symbol are the updated values of the parameters. The recommendation method and device provided by the embodiments of the present invention improve the recommendation efficiency of the recommendation system in the mass data environment by parallel computing, and improve the recommendation effect of the recommendation system by considering the user implicit feedback. BRIEF DESCRIPTION OF THE DRAWINGS In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings to be used in the embodiments or the description of the prior art will be briefly described below, and obviously, in the following description The drawings are only some of the embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without any creative work. FIG. 1 is a schematic flowchart of a recommendation method according to an embodiment of the present invention;

2 is a detailed flowchart of a recommendation method according to an embodiment of the present invention; FIG. 3A is a flow chart of a method for placing rating data according to an embodiment of the present invention;

FIG. 3B is a flow chart of another method for placing rating data according to an embodiment of the present invention; FIG.

4 is a schematic flowchart of a solution for an optimal parameter according to an embodiment of the present invention; FIG. 5 is a schematic diagram of a method for updating a parameter according to an embodiment of the present invention; FIG. 5B is another schematic diagram of an embodiment of the present invention. FIG. 6 is a structural diagram of a recommended device according to an embodiment of the present invention; FIG. 8 is a structural diagram of another recommended device according to an embodiment of the present invention; A hardware device diagram of a recommended device. The technical solutions in the embodiments of the present invention are clearly and completely described in the following with reference to the accompanying drawings in the embodiments of the present invention. It is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. example. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts are within the scope of the present invention.

As shown in FIG. 1 , it is a schematic flowchart of a recommendation method provided by an embodiment of the present invention, including:

S 1 0 1: The scoring data in the scoring data set is respectively placed in at least two data layers, Wherein, the scoring data is corresponding to the user and the product respectively, and the user and the product corresponding to any two scoring data in each data layer are different;

S 1 02: Calculate parameters of the recommended system model in the data layer in parallel according to the preset recommendation system model and the scoring data in the data layer, and use the parameters of each layer of the data layer as the initial value of the corresponding next layer of the data layer. Until the optimal parameters of the recommended system model are obtained; wherein the recommendation system model is a correspondence between the predicted value of each product for each product and the average score and the parameters of the recommended system model; exemplary, recommended system model It may include a recommendation system model that provides implicit feedback, a recommendation system model that does not provide implicit feedback, a recommendation system model that considers spatiotemporal characteristics, and a recommendation system model that models asymmetric potential factors; for example, the recommendation system model may also include : First recommendation system model

—丄 j≡N(u) or, the second recommended system model

In the first recommendation model and the second recommendation model, the user u predicts the score of the product i, μ represents the average value of all the score data in the score data set, and b indicates the offset of the user u from the average user score. Indicates the offset of product i from the average product score, indicating the product factor vector, and T represents the transpose operator symbol, representing the user factor vector;

Further, in the first recommendation model, |N(w)| represents a collection size of all products in which user u provides an implicit preference, and N(w) represents a collection of all products in which user u provides an implicit preference; A factor vector associated with product j that is used to characterize implicit feedback information.

Correspondingly, the relationship between the predicted value obtained from the recommended system model and the mean square error of the score data and the parameters of the recommended system model are obtained as a cost function of the recommended system model, wherein the cost function comprises: a first cost function ∑ [r _tl -μ-b _u -b - _q (p _u +\N(u ∑ j,)] ²

≡N

2 ² )+ iu ₊ ≡ xN i if) or, second cost function

∑ [r _ui -μ-Κ-b - qp + {b _u ² + bf) + ^f + ||^ ) where |*| ² represents the sum of the squares of all the elements of the vector *, and ^ is the regularization factor . Exemplarily, according to the preset recommendation system model and the scoring data in the data layer, the parameters of the recommended system model in the data layer are calculated in parallel, and the parameters of each layer of the data layer are used as the initial values of the corresponding next layer of the data layer. Until the optimal parameters for the recommended system model are obtained, including:

A: Calculate the average score of all the score data in the score data set;

B: The parameters of each layer of the data layer are calculated by using parallel computing in turn, and the parameters calculated by each layer of the data layer are used as initial parameters of the data layer of the next layer; wherein the initial values of the parameters of the first layer of the data layer Set by the system; further, step B, including:

B1: obtaining an initial estimated value of the score data of the data layer of the layer according to the initial value of the parameter of each layer of the data layer and the recommendation system model, and further obtaining the score error of the data layer of the layer according to the score data of the data layer of the layer and the initial estimated value ;

B2: obtaining a parameter calculated by the data layer of the layer according to the score error;

B3: The parameter calculated by the data layer of the layer is used as the initial value of the parameter of the next layer of the data layer, and the parameters calculated by the next layer of the data layer are obtained according to steps B1 and B2 until the parameter calculated by the last layer of the data layer is obtained.

C: judge whether the recommended system model converges according to the parameters calculated by the last layer of the data layer. If it converges, the calculation ends and the optimal parameters are obtained; the optimal parameters are obtained, and if not, the data layer of the last layer is calculated. The parameters are used as the initial values of the parameters of the first layer of the data layer, and steps B and C are repeated. Exemplarily, determining whether the recommended system model converges according to the parameter calculated by the last layer of the data layer includes: The parameters calculated by the last layer of the data layer obtained in this calculation and the parameters calculated by the last layer of the data layer obtained in the previous calculation are all substituted into the cost function, and the difference between the results calculated by the cost function is calculated. If the threshold is not greater than the preset threshold, the parameters calculated in the last layer of the data layer obtained by this calculation are convergent. Otherwise, the parameters calculated in the last layer of the data layer obtained in this calculation are not convergent. .

S 1 03: obtaining a score prediction value of each product for each product according to the optimal parameter and the recommendation system model, and recommending the product to the user according to the score prediction value; the embodiment provides a recommendation method, which is improved by parallel calculation. The recommended efficiency of the system is recommended in a mass data environment, and the recommendation effect of the recommendation system is improved by considering user implicit feedback.

As shown in FIG. 2, a detailed flowchart of a recommended method provided in this embodiment includes:

S 2 01 : The scoring data in the scoring data set is respectively placed into at least two data layers; exemplarily, the scoring data set can be obtained by obtaining the user's scoring data of the product, or by browsing and purchasing the record information of the user. The user can not only obtain the user's feedback on the product, but also obtain the implicit feedback of the user's preference. The embodiment of the present invention does not impose any limitation on this. Preferably, the score data set can be obtained by the user's rating data of the product. The skilled person can understand that the user's rating data of the product not only shows the user's evaluation of the product, but also implicitly feedbacks the user's preference for the product through the user's behavior of rating the product. Preferably, in this embodiment The score data set may be represented by a matrix form, wherein different rows of the matrix represent different users, and different columns of the matrix represent different products; further, the score data corresponds to the user and the product respectively, and can be understood by those skilled in the art. , the data layer placed by the rating data It can be represented by a matrix form, and the scoring data set can also be represented by a matrix form, wherein different rows of the matrix represent different users, different columns of the matrix represent different products, and thus the scoring data in the scoring data set can be respectively placed To at least two data layer matrices, and all data layer matrices have the same number of rows and the same number of columns as the matrix of the scoring data set, when any two scoring data in each data layer corresponds to the user And the products are different, that is, when any two scoring data in each data layer matrix are not in the same row and are not in the same column, all the scoring data in the same layer of data layer can be satisfied that there is no dependency between each other. Therefore, the scoring data in the same data layer can be calculated in parallel. The specific placement steps are not limited in any embodiment of the present invention. Any two scoring data in each data layer matrix may not be in the same row and are not in the same The placement method of the columns is within the protection scope of the embodiment of the present invention;

Optionally, as shown in FIG. 3A, a preferred placement method in this embodiment includes:

3A 01 : Select one of the scoring data in the scoring data set and place it in the position of the user and product corresponding to the scoring data in the first data layer matrix. At this time, the last layer is set to / _max = l, The manner of selecting the method is not limited in this embodiment, and the manner of selecting data from the score data set is the same as that of the first selection, and is not described herein;

3A 02 : Select the next scoring data, and start from the first layer of the data layer matrix to the corresponding data layer matrix of the last layer, and compare with all the scoring data in turn, whether the corresponding position of the scoring data in the data layer matrix is satisfied. There is no score data in the row and column. If it is satisfied, the score data is placed in the first satisfied data layer matrix. If the data layer matrix corresponding to the last layer/ _max is not satisfied, the last layer is updated to _{Max max} + i , wherein the symbol indicates that the data on the left side of the symbol is updated to the data on the right side of the symbol, the same below, and the score data is placed in the updated data layer of the last layer;

3A 03 : Repeat step 302 for the remaining score data in the score data set until all the score data in the score data ^ are placed.

E.g ,

3 2

2 2 1

4 5 5 2

A = 2 3 *

1 5

1 2 3

5 2 Through the placement method as shown in FIG. 3A, the exemplary operation process is as follows: Select a score data r _{M in the} score data set, put it into the first layer, and then select the second score data to judge "' and "Is it the same user and whether 'and' is the same product. When "'≠" and '·'≠, the rating data is placed in the first layer, otherwise it is placed in the second layer. Then take out the third rating data ^, if the user "' and product" corresponding to the rating data are different from the users and products placed on the first layer, then put it into the first layer, otherwise The user and product corresponding to the score data of the second layer are compared. When ^ _≠% ^ ₂ and ' _{≠ 2} are satisfied, they are placed in the second layer, otherwise they are placed in the third layer, where " ₂ represents the second layer. User, „ ₂ indicates the product of the second layer. And so on, until all the score data of the set is placed. You can get 5 data layers in turn, from the first data layer to the last layer, which is the fifth data layer. as followed:

u ₂ , i ₂ 2, , 3) [ ₅ , , 5) l _x {Layer 1) ζ·4, 2) ( ₂ , ₅ , l) , 4) ( ₄ , 2) ( ₅ , i ₂ , \ ) / ₂ {Layer 2)

( ₃ , ₅ , 5) l) (w _? / ₃ {Layer 3) z ₆ 2) ( ₆ , ₂ , 2) Z ₄ J / ₄ {Layer 4)

_{_{(7, i 5, 2)}} / 5 {Layer 5) may be represented by a matrix form:

The * appearing in the matrix indicates that the user corresponding to the location does not score the product corresponding to the location, and A = +i ₂ +i ₃ +i ₄ +i ₅ can be obtained. Optionally, as shown in FIG. 3B, another preferred placement method in this embodiment includes:

3B01: Select one of the scoring data in the scoring data set and place it in the position of the user and product corresponding to the scoring data in the first data layer. At this time, the last layer is =1.

3B02: Starting from the second scoring data selected in the scoring data set, the scoring data is sequentially compared from the last data layer to the first data layer until it can be found from the last layer to the first data layer. The last one that satisfies the score data in the row and column of the corresponding position in the data layer matrix has no other scoring data, and H 'j the scoring data is placed at the corresponding position of the last satisfied data layer, if the last layer is If the first layer has no satisfied data scoring matrix, the last layer number is updated to ^/ _max +l, and the scoring data is placed in the data layer corresponding to the updated last layer number;

3B03: Repeat step 3B02 for the remaining score data in the score data set until all the score data in the score data set are placed.

For example, still taking the matrix A as an example, through the placement method as shown in FIG. 3B, the exemplary operation process is as follows: Select a rating data r _M from the score data set, put it into the first layer, and then Select the second rating data to determine whether _M 'and _M are the same user and 'and whether it is the same product. When '≠« and' ≠, the rating data r _{MY is} placed in the first layer, otherwise it will be Put in the second layer. Then take out the third rating data ^ and compare it with the scored data already placed on the second layer (if there is a second layer), if the rating data corresponds to the user _M 'and the product' and placed in the second When the user _M 'and the product' of the layer are different, they will continue to compare it with the score data already placed on the first layer. When the score data already placed with the first layer also satisfies _M " _{≠ M} and then it will be Put in the first layer, otherwise place the second layer, and when the user and product corresponding to the scoring data and the user _M ' and product ' placed on the second layer satisfy _M ' _{= M} ' or = any condition, then Put it directly into the third layer. By analogy, until all the score data in the score data set is placed, you can get 7 data layers in turn, from the first data layer to the last layer, the seventh data layer. as followed: l _x {Layer 1)

^{W Ζ} ·4, 2) ₅ , ₂ , 1 ) / ₂ {Layer 2)

^W 2 ^Ζ · 4 2 ) ( ₃ , ₃ , 5) ( / ₃ {Layer 3)

2 ₅ 1 ) ( ₄ , ₃ , 2) ₆ , ₂ , 2) / ₄ {Layer 4)

^W 3 ^Z 5 5 ) (W ₄ , 3) ( / ₅ {Layer 5)

^W 3 ^Z 6 2) ( ₄ , ₅ , 3) ( ζ·4 5 ) / ₆ {Layer 6)

( ₇ , ₅ , 2) / ₇ {Layer 7) Expressed in matrix form as:

5

2 where * appears in the matrix, indicating that the user corresponding to the location does not score the product corresponding to the location, and can obtain

S202: According to a preset recommendation system model and rating data in a data layer, and Calculate the parameters of the recommended system model in the data layer, and take the parameters of each layer of the data layer as the initial values of the corresponding next layer of data until the optimal parameters of the recommended system model are obtained; exemplary, preset system The model may be a latent factor model considering the implicit feedback of the user, or may be a latent factor model based only on the actual feedback, and may also include a recommendation system model considering the spatiotemporal characteristics, an asymmetric latent factor model, etc. Without further limitation, further, in the present embodiment, an improved first recommendation system model considering user implicit feedback, such as equation (1), and a second recommendation system that does not consider user implicit feedback, is constructed. Model, (2):

= μ + + δ _{ί +} q (p _u + J,) ( 1 )

In equations (1) and (2), the user u predicts the score of product i, μ represents the average of all the score data in the score data set, and ^ represents the offset of user u from the average user score, indicating the product The offset of i from the average score of the product, q _t represents the product factor vector, T represents the transpose operator symbol, p „ represents the user factor vector; further in equation ( 1 ), |N(w)| represents that the user u provides The aggregate size of all products implicitly preferred, N(w) represents the set of all products that user u provides implicit preference; represents the factor vector associated with product j, which is used to characterize implicit feedback information. , b. , q and the unknown parameters of the recommended system model for the user; exemplarily, the relationship between the predicted value obtained from the recommended system model and the mean square error of the score data and the parameters of the recommended system model are obtained by the recommended system model a cost function, and the scoring data of the hierarchical data matrix obtained according to step S 2 01, obtained by solving the cost function optimization problem related to the above model Knowing the parameters, that is, calculating the parameters of the recommended system model in the data layer in parallel, and taking the parameters of each layer of the data layer as the initial values of the corresponding next layer of data layers, until the optimal parameters of the recommended system model are obtained, wherein, the optimal The parameter is the optimal value of the unknown parameter of the above model. Specifically, in this embodiment, the cost function related to the first recommendation system model and the second recommendation system model can be expressed as the first cost function and expression of the equation (3), respectively. The second cost function of (4): ∑ [r _tl -μ-b _u -b - _q (p _u +\N(u ∑ j,)] ²

(",'■) j≡N(u) ( 2 )

2 ² )+ iu ₊ xN(u) i if)

∑ [r _ui -μ-Κ-b - qp _u f +

(4) where |*| ² represents the sum of the squares of all the elements of the vector *, and ^ is the regularization factor. It can be understood by those skilled in the art that the optimization problem is solved for the equations (3) and (4), and the specific solution process has similarities, and will not be described again, as shown in FIG. 4, for the equation (3) The solution to the optimization problem can include:

401. Calculate an average score of all the score data in the score data set;

402. Calculate the parameters of each layer of the data layer by using parallel computing in turn, and use the parameter calculated by each layer of the data layer as the initial value of the parameter of the next layer of the data layer;

Exemplarily, in this embodiment, before the first calculation, the parameters b _M , b., q and the initial values of ^ and ^ of the first layer data layer may be randomly set. For the sake of simplicity, the parameter b „, b., q and ^ initial values can be set to scalar parameters as

0, the vector parameter is a vector δ, wherein the arrow symbol above the number 0 is a vector symbol, and ^ and ^ can be arbitrarily set to a relatively small positive value, which is not limited by the embodiment of the present invention, and is used to indicate the regularity For example, starting from the first layer of data, the values of the parameters b _u , bq and the data layer of the layer are calculated in parallel, and the calculated parameter values are used as parameters of the next layer of data. , the initial value of bq and , until the parameters of the last layer of the data layer b„, bqp„ and ; are obtained; the parameters calculated by the last layer of the data layer are obtained; the specific calculation process is as follows: According to the parameters of the first layer of the data layer The value is updated for the unknown parameters b _M , b., q _t , and the corresponding score data contained in the first layer data layer. Since there is no interdependent relationship between the score data of the same layer matrix, the same layer matrix The update calculations of the unknown parameters b„, b., q _t , and sum corresponding to different scoring data can be performed in parallel, and the updated parameters are used as the initial values of the parameters of the next layer of data layer for the next layer of data layer. The parameters corresponding to the included scoring data are updated and calculated, until the parameters corresponding to the scoring data included in the last layer of the data layer are updated and calculated, and the last layer of data is obtained. Preferably, in this embodiment, the update calculation can be performed by the gradient descent method. Since the update methods of the parameters corresponding to all the scoring data are the same, those skilled in the art can understand that by the parameter b _M corresponding to one scoring data, b., q and ^ After the update calculation step is described, the calculation step can be applied to other score data without creatively, as shown in Fig. 4, the specific steps are as follows:

4041. Obtain an initial estimated value of the score data r _ui according to the initial value of the parameter of each layer of the data layer and the recommended system model represented by the formula (1), and obtain the score according to the score data of the data layer of the layer and the initial estimated value f _ui Layer data layer score error e _ui ;

4022. Obtain an updated value of the parameter calculated by the layer of the data layer in the negative gradient direction according to the scoring error ^, as shown in the schematic diagram of FIG. 5A, including:

(5)

(6) (7)

The symbol represents the update symbol, that is, the value of the variable on the left side of the update symbol is replaced by the calculated value on the right side of the update symbol. In this embodiment, the parameters appearing on the right side of the update symbol of the formula (5)-formula (9) are corresponding parameters. The initial value, the parameters appearing on the left of the update symbol are the updated values of the parameters, ₇₁ and ₂ are the iteration step lengths. In this embodiment, it is preferable to set a suitable constant, that is, ₇₁ and ₂ can be set. A constant, or it can be an amount related to the number of iterations, that is, the iteration step is gradually reduced as the number of iterations increases. Such as: ^ * o.96 ^NumOJIter and 7 ₂ * 0.96 ^NumOJIter , where NumO f Iter represents the number of iterations. This embodiment of the present invention does not limit this. It can be understood by those skilled in the art that the formula (5) - (8) can be updated according to the corresponding calculation formula, and the calculation process will not be described again. Correspondingly, the formula (4) The solution to the optimization problem only needs to perform the update calculation of equation (5) - equation (8), and obtain the parameters b„, b., qp„;

However, the calculation of the formula (9) also needs to be combined with the scoring data in the scoring data set to perform serial calculations in sequence, which reduces the computational efficiency. Therefore, the embodiment of the present invention provides a calculation method for the formula (9), as shown in FIG. 5A. The schematic diagram shown is as follows:

When there are multiple calculations of the same layer matrix, the user corresponding to the product j represented by the score data set is used to calculate the corresponding gradient Ay in parallel, and then all the gradients Ay corresponding to the product j are aggregated to obtain the layer. Values, such as in the scoring data set, User 1, User 2, and User 6 score product 4 ( j = 4 ), then the gradient Ay of product j corresponding to User 1, User 2, and User 6 can be calculated in parallel. Ayf> , Ayf> , where Ay; ¹ ) ^e _n -2^- _qi - λ ₂ γ) , A e ₂₂ · 3— ^ · q ₂ ― λ, γ) , Ay^ e ₆₄ · · q ₄ - X ₂ y); Then the updated value of the layer is obtained by +y ₂ (A) + Ayf)), where is the initial value at the time of calculation in the layer. In addition, the embodiment of the present invention provides another equivalent calculation method for the equations (8) and (9), as shown in the schematic diagram of FIG. 5B. The specific implementation manner is: the expression _ft in the first recommendation model +|N( _M )|4 X is used as an equivalent parameter, and the auxiliary variable is used to represent the equivalent parameter, ie

Where the auxiliary variable is equivalent to the expression _ft +|N( _M )|4∑, in which case the first recommended model is equivalent to r _ui =μ + δ _η +b _i + q _i ^T · z _u , its parameters Is 3⁄4, 6,·, q _i , z _u ;

Then according to the gradient Δζ _Μ = 2e _{ui of the} auxiliary variable z _u and thus the update formula of the auxiliary variable ₊ y ₂ . (2 _eM i- ), where is the initial value at the time of calculation of the layer.

With this method, the update process of the parameter sum can be replaced only by the update process of the auxiliary variable. Furthermore, since the auxiliary variable is introduced, the update of the parameter q _t of the formula (8) can also be changed accordingly, that is, The update formula of q _t becomes ^■^ ^ + 7 ₂ · (^ - where. is the initial value at the time of calculation in this layer. It can be seen that the introduction of the auxiliary variable simplifies the calculation on the one hand, and the parameter of the recommended model on the other hand because the calculation of ^ is eliminated. The solution reduces the inner loop of one layer, thus greatly improving the speed of the operation while ensuring the same accuracy.

4 02 3 , the parameter calculated by the data layer of the layer is used as the initial value of the data layer of the next layer, and the parameters calculated by the next layer of the data layer are obtained according to steps 4 02 1 and 4 022 until the last layer of data is obtained. Calculate the resulting parameters.

4 0 3 : Determine whether the recommended system model converges according to the parameters calculated by the data layer of the last layer. If it converges, the calculation ends and the optimal parameters are obtained; the optimal parameters are obtained, and if it does not converge, the last layer of data is calculated. The obtained parameters are used as the initial values of the parameters of the first layer of the data layer, and steps 4 02, 4 03 are repeated; exemplary, the parameters calculated by the last layer of the data layer obtained by the current calculation and the last one obtained by the previous calculation The parameters calculated by the layer data layer are substituted into the cost function of equation (3) or (4). If the difference between the two calculation results is not greater than the preset threshold, the final result of this calculation can be considered. The parameter calculated by one layer of data layer is the optimal parameter. If it is greater than the preset threshold value, the parameter calculated by the last layer of data layer obtained by this calculation is not convergent, and the calculation system will be The obtained parameter of the last layer of the data layer is used as the initial value of the parameter of the first layer of the data layer to be calculated next, and the parameters of the recommended system model are continuously calculated.

S 2 03 : According to the optimal parameter and the recommendation system model, each user's score prediction value for each product is obtained, and the product is recommended to the user according to the score prediction value.

Exemplarily, in the embodiment, if the obtained optimal parameter is brought into the formula (1) or the formula (2), the predicted value of each user for each product can be obtained, and the same user can be used for all products. The ranking predictors are ranked, and the preset number of products with the highest score prediction value is selected for recommendation to the user.

S 2 04 : After the product recommendation, the new scoring data made by the user to the product is input into the recommendation system, so that the recommendation system can ensure the real-time of the high-precision and high-efficiency recommendation according to the parameters of the system model recommended in real-time update. .

This embodiment provides a recommendation method for improving massive data through parallel computing. The recommended efficiency of the system is recommended in the environment, and the recommendation effect of the recommendation system is improved by considering the implicit feedback of the user.

The embodiment of the present invention provides a recommendation device 60. As shown in FIG. 6, the method includes: a data placement unit 601, configured to separately set the score data in the score data set to at least two data layers, where the score data and the user and The products are respectively corresponding, and the users and products corresponding to any two scoring data in each data layer are different;

The parallel computing unit 602 is configured to calculate parameters of the recommended system model in the data layer in parallel according to the preset recommendation system model and the scoring data in the data layer, and use the parameters of each layer of the data layer as the corresponding next layer of data layer. The initial value until the optimal parameter of the recommendation system model is obtained; wherein, the recommendation system model is a correspondence between the predicted value of the score for each product and the average score of each product and the parameters of the recommended system model; The unit 603 is configured to obtain, according to the optimal parameter and the recommendation system model, a predicted value of each user for each product, and recommend the product to the user according to the predicted value of the score.

Exemplarily, the scoring data set can be obtained by obtaining the user's scoring data of the product, or by browsing and purchasing the record information of the user, not only obtaining the user's display feedback on the product, but also obtaining implicit feedback of the user's preference. The embodiment of the present invention does not impose any limitation on this. Preferably, the scoring data set can be obtained by the user's scoring data of the product. As can be understood by those skilled in the art, the user's scoring data of the product not only displays the feedback of the user to the product. The evaluation, and also implicitly feedback the user's preference for the product by the user's behavior of rating the product. Preferably, in this embodiment, the score data set can be represented by a matrix form, wherein different rows of the matrix represent different users The different columns of the matrix represent different products; further, the score data is respectively corresponding to the user and the product - correspondingly, those skilled in the art can understand that the data layer placed by the score data can be represented by a matrix form, and the score data is simultaneously Sets can also pass through the matrix The form is represented, wherein different rows of the matrix represent different users, different columns of the matrix represent different products, and thus the scoring data in the scoring data set can be respectively placed to at least two data layer moments Array, and all the data layer matrices have the same number of rows and the same number of columns as the matrix of the scoring data set, when the user and the product of any two scoring data in each data layer are different, that is, each data When any two scoring data in the layer matrix are not in the same row and are not in the same column, all the scoring data in the same data layer can be satisfied without any dependence on each other, so the score in the same data layer can be scored. The data is subjected to parallel computing, and the specific placement steps are not limited in any embodiment of the present invention. Any placement method capable of making any two of the score data in each data layer matrix not in the same row and not in the same column is in the embodiment of the present invention. Within the scope of protection;

Further, in this embodiment, the data placement unit 6 0 1 may be used to implement the placement method as shown in FIG. 3, including: the data placement unit 6 0 1 selects one rating data in the score data set, and places it in the first In the layer data layer matrix, the location of the user and the product corresponding to the score data, at this time, the last layer number / _max = l, wherein the selected mode is not limited in this embodiment, and the data is selected from the score data set. The method is the same as the first selection method, and is not described here; the data placement unit 6 0 1 selects the next score data, and starts from the first layer of the data layer matrix to the corresponding data layer matrix of the last layer, and sequentially All the score data are compared, and whether the row and the column of the corresponding position of the score data in the data layer matrix are not scored data, if satisfied, the score data is placed in the first satisfied data layer matrix, if until the last layer / _max data corresponding to a matrix layer is not satisfied, the last layer of the number of update / _{_max} / _max + l, where the symbol represents the number of symbols on the left Update to the data on the right side of the symbol, the same below, and place the score data in the updated last layer of data layer; the data placement unit 6 0 1 repeats the above process of placing the score data in sequence for the remaining score data in the score data set until the score All scoring data in the data set is placed.

For example, the scoring data is shown in matrix A: 3 * * 2 * *

2 2 1

4 5 5 2

A 2 3

1 5

1 2 3

3 5 2 The data placement unit 601 can be exemplified by the placement method as shown in FIG. 3A. The specific operation process is as follows: Select a score data _{r∞ in the} score data set, put it into the first layer, and then select the first Two scoring data ^ , , judge whether "' and " are the same user and whether ' and ' is the same product. When "and" ,, the scoring data r _{MY is} placed in the first layer, otherwise it is placed Into the second layer. Then take the third rating data ^, if the user corresponding to the rating data and the product are not the same as the users and products placed on the first layer, then put it into the first layer Otherwise, the user and the product corresponding to the score data of the second layer are compared. When ² ^'^ and ' is satisfied, the second layer is placed, otherwise the third layer is placed, where _{Μ 2} indicates the second The user of the layer, representing the product of the second layer. And so on, until all the score data of the set is placed. You can get 5 data layers in turn, from the first data layer to the last layer, the fifth data layer. as followed:

ζ·ι, 3) (" ,2) ( ^3 3 5 ) ( ₄ , ₅ , 3) ( l _x {Layer 1)

ζ·4, 2) (M ,l) ( 3 ^ζ ι 4 ) ( ₄ , ₃ ,2) ( ₅ , ₂ , 1 ) / ₂ {Layer 2)

("3, ,5) ( ( _{7 3} , 3 ) / ₃ {Layer 3) z ₆ 2) ("6,'·2, ² ) ( ^W 7 , ^Ζ · 4 , 5 ) / ₄ {Layer 4)

( ₇ , ₅ , 2) / ₅ {Layer 5) The representation in matrix form can be:

The * appearing in the matrix indicates that the user corresponding to the location does not score the product corresponding to the location, and ^+^+^+^+^. Further, in this embodiment, the data placement unit 601 can also be used to implement the placement method as shown in FIG. 3B, including: the data placement unit 601 selects one rating data in the score data set, and places it in the first layer data layer. In the position of the user and product corresponding to the score data, at this time, the last layer number / _max = l,

The data placing unit 601 starts from the second scoring data selected in the scoring data set, and sequentially compares the scoring data from the data layer corresponding to the last layer to the first data layer until the first layer can be counted to the first In the layer data layer, if there is no other scoring data in the row and the column of the column that satisfies the corresponding position of the scoring data in the data layer, the scoring data is placed in the corresponding position of the last satisfied data layer, if from the last one If the number of layers is not satisfied by the first layer, the last layer is updated to / _max / _max +l, and the score data is placed in the data layer corresponding to the last layer of the update; data placement unit 601 Repeat the above-mentioned process of placing the second score data on the remaining score data in the score data set until all the score data in the score data set are placed.

For example, still taking the matrix A as an example, through the placement method as shown in FIG. 3B, the exemplary operation process is as follows: Select a rating data r _M from the score data set, put it into the first layer, and then Select the second rating data to determine whether _M 'and _M are the same user and 'and whether it is the same product. When '≠« and' ≠, the rating data r _{MY is} placed in the first layer, otherwise it will be Put in the first Second floor. Then take out the third scoring data ^ and compare it with the scored data already placed on the second layer (if there is a second layer), if the rating data corresponds to the user _Μ 'and the product ζ·' When the user _M 'and the product' of the second layer are different, they continue to compare it with the score data already placed on the first layer. When the score data already placed with the first layer also satisfies ^ _≠ «And' ≠ Then put it into the first layer, otherwise place the second layer, and when the rating data corresponds to the user _M "and the product and the user _M ' placed on the second layer and the product 'satisfies _M ' _{= M} ' or = any When a condition is met, it is placed directly in the third layer. By analogy, until all the score data in the score data set is placed, 7 data layers can be obtained in turn, from the first data layer to the last layer. The seventh layer of data layers are:

, ^w i, ^ζ ·ι ,3) l _x {Layer 1)

, ^w i, ^ζ · 4, 2) ₅ , ₂ , 1 ) / ₂ {Layer 2)

, w ₂ , ₄ , 2) ( ₃ , ₃ , 5) ( / ₃ {Layer 3)

, w ₂ , ₅ , l) ( ₄ , ₃ , 2) ₆ , ₂ , 2) / ₄ {Layer 4)

, w ₃ , ₅ , 5) (W ₄ , 3) ( / ₅ {Layer 5)

, w ₃ , ₆ , 2) ( ₄ , ₅ , 3) ( ζ · 4 5 ) / ₆ {Layer 6)

( ₇ , ₅ , 2) / ₇ {Layer 7) Expressed in matrix form as:

₇

5

2 where * indicates that the user corresponding to the location does not score the product corresponding to the location, and ^=/; +4 +/;+/: +/;+/+/; can be obtained. Exemplarily, the preset system model may be a latent factor model considering user implicit feedback, or a latent factor model based only on real feedback, and may include a recommendation system model considering a spatiotemporal characteristic, and an asymmetric latent factor model. The embodiment of the present invention does not limit this,

Further, in the embodiment, an improved first recommendation system model considering user implicit feedback is constructed, such as equation (1), and a second recommendation system model that does not consider user implicit feedback, such as 2); In equations (1) and (2), the predicted value of the user u for the product i, μ is the average of all the score data in the score data set, and b indicates the deviation of the user u from the average user score. Shift, which represents the offset of product i from the average product score, q _t represents the product factor vector, T represents the transpose operator symbol, p „ represents the user factor vector; further in equation ( 1 ), |N(w)| Represents the collection size of all products for which user u provides an implicit preference, N(w) represents a collection of all products that user u provides for implicit preference; represents a factor vector associated with product j that is used to characterize implicit feedback information. Moreover, b„, b., q and unknown parameters of the recommended system model of the user; exemplarily, the recommendation device 60 may further include a cost function generating unit 604 for estimating the predicted value and the scoring data according to the recommended system model. The relationship between the square error and the parameters of the recommended system model is obtained by the cost function of the recommended system model and the score data of the hierarchical data matrix obtained from the data placement unit 601, by solving the cost function optimization problem related to the above model. The unknown parameters of the model, that is, the parameters of the recommended system model in the parallel computing data layer, and the parameters of each layer of the data layer As the initial value of the corresponding next layer of the data layer, until the optimal parameter of the recommended system model is obtained, wherein the optimal parameter is the optimal value of the unknown parameter of the model, specifically, in this embodiment, the first recommendation system The cost function related to the model and the second recommendation system model can be expressed as Equation (3) and Equation (4), respectively, where |*|| ² represents the sum of squares of all elements of the vector *, and ^ is a regularization factor. As can be understood by those skilled in the art, the parallel computing unit 602 solves the optimization problem of the equations (3) and (4) according to the above model, and the specific solution process has similarities, and will not be described again. (3) As an example, as shown in FIG. 7, the parallel computing unit 602 may include:

The average score calculation sub-unit 6021 is configured to calculate an average score of all the score data in the score data set; the hierarchical calculation sub-unit 6022 calculates the parameters of each layer of the data layer in a parallel calculation manner, and each layer of the data layer The calculated parameter is used as the initial value of the parameter of the next layer of data; exemplarily, in this embodiment, the parameters b _M , b., q , and ^ of the first layer of the data layer may be before the first calculation. The initial value of ^ is randomly set. For the sake of simplicity, the initial values of the parameters b„, b., q and ^ can be set to scalar parameters.

0, the vector parameter is a vector δ, wherein the arrow symbol above the number 0 is a vector symbol, and 4 and ^ can be arbitrarily set to a relatively small positive value, which is not limited by the embodiment of the present invention, and is used to indicate the regularity. For example, starting from the first layer of the data layer, the values of the parameters b _u , bi, q _t , ? „ and ^. of the layer of the data layer are sequentially calculated in parallel, and the calculated parameter values are taken as The parameters of the data layer b„, bq and the initial value of a layer of data, until the parameters b„, bq and ^ of the last layer of the data layer are calculated, the parameters calculated by the last layer of the data layer are obtained; the specific calculation process is as follows: The layer calculation sub-unit 6022 performs an update calculation on the unknown parameters b„, b., q corresponding to the score data included in the first layer data layer according to the parameter initial value of the first layer data layer, because the score data of the same layer matrix is There is no interdependence relationship between them, so the update calculation of the unknown parameters b _M , b., q and ^ corresponding to different score data of the same layer matrix can be performed in parallel, and the updated parameters are used as the parameters of the next layer of data layer. The initial value of the number is updated and calculated for the parameter corresponding to the score data included in the next layer of the data layer, until the parameter corresponding to the score data included in the last layer of the data layer is updated and calculated, and the parameter calculated by the last layer of the data layer is obtained. Preferably, in this embodiment, the update calculation can be performed by the gradient descent method. Since the update methods of the parameters corresponding to all the scoring data are the same, those skilled in the art can understand that the parameter b _M corresponding to one scoring data is understood. After the b, q and ^ are performed, the calculation steps are described, and the calculation step can be applied to other score data without any creativity. The calculation step of the hierarchical calculation sub-unit 6022 for the formula (1) can be as follows: First, The initial estimated value of the scoring data r _ui is obtained according to the initial value of the parameter of each layer of the data layer and the recommended system model shown by the formula (1), and the layer data is obtained according to the scoring data of the data layer of the layer and the initial estimated value f _ui score layer error e _ui; Next, error rates, and ^ of formula (5) - to give the formula (9) the negative gradient direction The updated value of the parameter, as can be understood by those skilled in the art, the formula (5) - (8) can be updated according to the corresponding calculation formula, and the meaning of the symbol and the parameter appearing in the calculation process is no longer meaningful. To sum up, correspondingly, the solution to the optimization problem of equation (4) only needs to perform the update calculation of equation (5) - equation (8), and obtain the parameters b„, bq;

However, the calculation of the formula (9) also needs to be combined with the scoring data in the scoring data set to perform serial calculations in sequence, which reduces the computational efficiency. Therefore, the embodiment of the present invention provides a calculation method for the formula (9), and the specific implementation manner For: When there are multiple ^s in the same layer matrix, the user corresponding to the product j represented in the score data set is used to calculate the corresponding gradient Ay in parallel, for example, in the score data set, user 1, user 2, and user 6 If the product 4 ( j=4 ) is scored, the gradients Δ^ , Ay^ of the product j corresponding to the user 1, the user 2, and the user 6 can be calculated in parallel.

^{_{Wherein, Ay; 1) e i 2-}} , Ayf ¾ · 3-, Ay ¾ · 3- ^ -;. And then polymerizing the layer obtained value, i.e.,

Where is the initial value of ^. In addition, the embodiment of the present invention provides another method for the equations (8) and (9). The specific calculation method is as follows: the expression _ft +|N( _M )|4 X in the first recommendation model is taken as an equivalent parameter, and the auxiliary parameter is used to represent the equivalent parameter, that is,

, where the auxiliary variable is equivalent to the expression _ft +|N( _M )|4∑, in which case the first recommended model is equivalent to eW(")

r _ui =μ + δ _η +b _i + q _i ^T · z _u , whose parameters are 3⁄4, 6,·, q _i , z _u ;

With this method, the update process of the parameter sum can be replaced only by the update process of the auxiliary variable. Furthermore, due to the introduction of the auxiliary variable, the update of the parameter q _t of equation (8) can also be changed accordingly, ie The update formula of q _t becomes ^■^^ + 7 ₂ · (^- where. is the initial value at the time of calculation in this layer.

It can be seen that the introduction of auxiliary variables simplifies the calculation on the one hand. On the other hand, the elimination of the calculation of ^, the parameter solution of the recommended model reduces the inner loop, thus greatly improving the operation speed while ensuring the same accuracy. Then, an updated value of the parameter of the first layer data layer can be obtained, and then the updated value is used as the initial value of the parameter of the second layer data layer, and the updated value of the parameter of the second layer data layer is calculated according to the above method, This type of push gives the updated value of the parameters of the last layer of the data layer. The convergence determining sub-unit 6023 is configured to determine whether the recommended system model converges according to the parameter calculated by the last layer of the data layer. If the convergence, the calculation ends, and the optimal parameter is obtained; if the optimal parameter is obtained, if not, the last parameter is obtained. The parameter calculated by the layer data layer is used as the initial value of the parameter of the first layer data layer, and the next hierarchical calculation is continued through the hierarchical calculation sub-unit 6022.

Preferably, the convergence determination sub-unit 6023 can substitute the parameter calculated by the last layer of the data layer obtained by the current calculation and the parameter calculated by the last layer of the data layer obtained by the previous calculation into the formula (3) or the formula (3) or 4) The cost function is calculated. If the difference between the two calculation results is not greater than the preset threshold value, the parameter calculated in the last layer of the data layer obtained in this calculation can be considered as the optimal parameter. Set Threshold value, H 'j will not converge the parameters calculated by the last layer of data layer obtained in this calculation, and the parameters calculated by the last layer of data layer obtained in this calculation are used as hierarchical calculation The unit 6022 continues the parameter initial value of the first layer data layer of the next hierarchical calculation, and continues to calculate the parameters of the recommended system model. Exemplarily, the prediction recommending unit 603 brings the obtained optimal parameters into the formula (1) or the formula (2), and can obtain the predicted value of each user for each product, which can be obtained by the same user for all products. The score prediction values are arranged, and a predetermined number of products with the highest score prediction value are selected for recommendation to the user.

Exemplarily, as shown in FIG. 7, the recommendation device 60 may further include: a feedback unit 605, after the product recommendation, inputting new rating data made by the user to the product into the recommendation system, so that the recommendation system can be Real-time updates recommend parameters of the system model to ensure high-precision and high-efficiency recommendations for real-time performance.

The present embodiment provides a recommendation device 60, which improves the recommendation efficiency of the recommendation system in a massive data environment by parallel computing, and improves the recommendation effect of the recommendation system by considering user implicit feedback.

The present embodiment provides a recommendation device 60, as shown in FIG. 8, comprising: at least one processor 801, a memory 802, and at least one communication bus 803 for implementing connection and mutual communication between the devices, wherein

The communication bus 803 may be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus, or an Extended Indus try Standard Architecture (EISA). ) Bus, etc. The bus 803 can be divided into an address bus, a data bus, a control bus, and the like. For ease of representation, only one thick line is shown in Figure 8, but it does not mean that there is only one bus or one type of bus.

The memory 802 is used to store executable program code and processing results of the processor 801, the program code including computer operating instructions. The memory 802 may include a high speed RAM memory and may also include a non-volatile memory. For example at least one disk storage. The processor 801 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more configured to implement the embodiments of the present invention. integrated circuit. The processor 801 is configured to execute executable program code stored in the memory 704, such as a computer program, to execute a program corresponding to the executable code. The processor 801 is configured to: separately record the score data in the score data set to at least one data layer, where the score data corresponds to the user and the product respectively, and the user corresponding to any two of the score data in each data layer and The products are all different;

And calculating the parameters of the recommended system model in the data layer in parallel according to the preset recommendation system model and the scoring data in the data layer, and using the parameters of each layer of the data layer as the initial value of the corresponding next layer of the data layer until obtaining Recommending an optimal parameter of the system model; wherein, the recommendation system model is a correspondence between each user's score prediction value and the average score of each product and the parameters of the recommendation system model; and the optimal parameter and recommendation system The model obtains a score prediction value for each product for each user, and recommends the product to the user based on the score prediction value. Exemplarily, the scoring data set can be obtained by obtaining the user's scoring data of the product, or by browsing and purchasing the record information of the user, not only obtaining the user's display feedback on the product, but also obtaining implicit feedback of the user's preference. The embodiment of the present invention does not impose any limitation on this. Preferably, the scoring data set can be obtained by the user's scoring data of the product. As can be understood by those skilled in the art, the user's scoring data of the product not only displays the feedback of the user to the product. The evaluation, and also implicitly feedback the user's preference for the product by the user's behavior of rating the product. Preferably, in this embodiment, the score data set can be represented by a matrix form, wherein different rows of the matrix represent different users The different columns of the matrix represent different products; further, the score data is respectively corresponding to the user and the product - correspondingly, those skilled in the art can understand that the data layer placed by the score data can be represented by a matrix form, and the score data is simultaneously Sets can also pass through the matrix Form of representation, Wherein, different rows of the matrix represent different users, different columns of the matrix represent different products, and the processor 801 can then place the scoring data in the scoring data set into at least two data layer matrices, and all the data layer matrices and the scoring data set The matrix has the same number of rows and the same number of columns. When any two scores in each data layer correspond to different users and products, that is, any two score data in each data layer matrix are not the same. When the rows are not in the same column, all the scoring data in the same layer of data layer can be mutually independent, so that the scoring data in the same layer of data can be calculated in parallel, and the specific placement steps are implemented in the present invention. The method is not limited, and any method for placing any two of the data in the data layer matrix in the same row and not in the same column is within the protection scope of the embodiment of the present invention; In an example, the processor 801 can be used to implement the placement method as shown in FIG. 3A, including Processor 801 to select a set of data rates of data rates, the first layer on which data layer of the matrix corresponding to the position of the user data rates on the product and, at this time, the number is set to the last layer / _max = l, The method for selecting the method is not limited in this embodiment, and the manner of selecting data from the score data set is the same as that of the first selection, and is not described herein; the processor 801 selects the next score data and the data from the first layer. The layer matrix begins to the corresponding data layer matrix of the last layer, and is sequentially compared with all the score data therein, whether there is no score data for the row and column where the corresponding position of the score data in the data layer matrix is satisfied, and if so, the The scoring data is placed in the first satisfied data layer matrix. If the data layer corresponding to the last layer/ _max is not satisfied, the last layer number is updated to the data indicating that the data on the left side of the symbol is updated to the data on the right side of the symbol. Same, and the rating data is placed in the updated last layer of data; processor 8 01 is left in the score data set Score data sequentially repeating the above process until placing all rates dataset ratings data are placed. For example, the scoring data is shown in matrix A: 3 * * 2 * *

2 2 1

4 5 5 2

A 2 3

1 5

1 2 3

3 5 2 Through the placement method as shown in FIG. 3A, the exemplary operation process is as follows: Select a score data r _{M in the} score data set, put it into the first layer, and then select the second score data to judge "I and " are the same user and whether 'and' is the same product. When "'≠" and '·'≠, the rating data is placed in the first layer, otherwise it is placed in the second layer. Then take out the third rating data ^, if the user "' and product" corresponding to the rating data are different from the users and products placed on the first layer, then put it into the first layer, otherwise The user and product corresponding to the score data of the second layer are compared. When ₂ and ^ ₂ are satisfied, they are placed in the second layer, otherwise they are placed in the third layer, where " ₂ indicates the user of the second layer, „ ₂ indicates The second layer of products. And so on, until all the score data of the collection is placed. You can get 5 data layers in turn, from the first data layer to the last layer, that is, the fifth data layer:

(MJ, J, 3) (u ₂ , i ₂ , 2) (t3, 3, 5) ( ₄ , , 3) (" ₅ , ' ₆ , 5) [ ₆ , i ₄ , l _x {Layer 1)

[u _l ,i _A ,2) (w ₂ , ₅ ,l) (w ₃ , j,4) [u , ,2) ( ₅ ,i ₂ ,\) / ₂ {Layer 2)

(u ₂ , i ₄ , 2) u ₃ , i ₅ , 5) (Μ ₆ , [, 1) (U ₇ . / ₃ {Layer 3)

/ ₄ {Layer 4)

(" ₇ , '· ₅ , ² ) / ₅ {Layer 5) The representation in matrix form can be:

The * appearing in the matrix indicates that the user corresponding to the location does not score the product corresponding to the location, and A = +i ₂ +i ₃ +i ₄ +i ₅ can be obtained. Further, in this embodiment, the processor 801 can also be used to implement the placement method as shown in FIG. 3B, including: the processor 801 selects one rating data in the score data set, and places it in the first layer data layer. At the location of the user and product of the rating data, at this time, the last layer number / =1;

The processor 801 starts from the second score data selected in the score data set, and compares the score data from the data layer corresponding to the last layer to the first data layer until the first layer can be compared to the first layer. If there is no other scoring data in the data layer to find the last row and column of the corresponding position of the scoring data in the data layer, the scoring data is placed at the corresponding position of the last satisfied data layer, if from the last layer Counting the data scoring matrix that is not satisfied by the first layer, the last layer number is updated to _{max max} +l, and the scoring data is placed in the data layer corresponding to the updated last layer; processor 801 pairs the scoring data The remaining data in the set repeats the above-described placement process for the second score data until all the score data in the score data set is placed.

, ^w i, ^ζ ·ι ,3) l _x {Layer 1)

, ^w i, ^ζ · 4, 2) ₅ , ₂ , 1 ) / ₂ {Layer 2)

, w ₂ , ₄ , 2) ( ₃ , ₃ , 5) ( / ₃ {Layer 3)

, w ₂ , ₅ , l) ( ₄ , ₃ , 2) ₆ , ₂ , 2) / ₄ {Layer 4)

, w ₃ , ₅ , 5) (W ₄ , 3) ( / ₅ {Layer 5)

, w ₃ , ₆ , 2) ( ₄ , ₅ , 3) ( ζ · 4 5 ) / ₆ {Layer 6)

( ₇ , ₅ , 2) / ₇ {Layer 7) Expressed in matrix form as:

Ghost

5 2 where * appears in the matrix, indicating that the user corresponding to the location does not score the product corresponding to the location, and can obtain ^=/; +/; +/;+/:+/;+/ +/;. Exemplarily, the preset system model may be a latent factor model considering user implicit feedback, or a latent factor model based only on real feedback, and may include a recommendation system model considering a spatiotemporal characteristic, and an asymmetric latent factor model. The embodiment of the present invention does not limit this. Further, in the embodiment, the processor 801 constructs an improved first recommendation system model considering user implicit feedback, such as equation (1), and a A second recommendation system model that does not consider user implicit feedback, as in equation (2),

In equations (1) and (2), the user u predicts the score of product i, μ represents the average of all the score data in the score data set, and ^ represents the offset of user u from the average user score, indicating the product The offset of i from the average score of the product, q _t represents the product factor vector, T represents the transpose operator symbol, p „ represents the user factor vector; further in equation ( 1 ), |N(w)| represents that the user u provides All collection sizes of implicit preferences, N(w) represents the set of all products that user u provides for implicit preference

'Represents the factor vector associated with product j, which is used to characterize the implicit feedback signal and b _M , bq and ^ are unknown examples of the user's recommendation system model, and the processor 801 can also obtain predictions based on the recommended system model. The relationship between the mean square error of the value and the score data and the parameters of the recommended system model is obtained by the cost function of the recommended system model, and the score data of the hierarchical data matrix, which is obtained by solving the cost function optimization problem related to the above model. The unknown parameters of the model, that is, the parameters of the recommended system model in the parallel computing data layer, and the parameters of each layer of the data layer are used as the initial values of the corresponding next layer of data layers, until the optimal parameters of the recommended system model are obtained, wherein The optimal parameter is the optimal value of the unknown parameter of the above model. Specifically, In this embodiment, the cost functions related to the first recommendation system model and the second recommendation system model may be expressed as equations (3) and (4), respectively; wherein |*| ² represents the sum of squares of all elements of the vector * , and ^ is a regularization factor. Exemplarily, according to the above model and the scoring data of the hierarchical data matrix, the processor 801 can obtain the unknown parameters of the above model by solving the cost function optimization problem related to the above model, that is, the recommended system model in the parallel computing data layer. The parameters of each layer of the data layer are taken as the initial values of the corresponding data layer of the next layer until the optimal parameters of the recommended system model are obtained, and those skilled in the art can understand the formula (3) and the formula ( 4) Solving the optimization problem, the specific solution process has similarity, and will not be described again. Specifically, taking equation (3) as an example, as shown in FIG. 7, the processor 801 is further used to:

Calculating the average score of all the score data in the score data set; and calculating the parameters of each layer of the data layer by using the parallel calculation method in turn, and using the parameter calculated by each layer of the data layer as the initial value of the parameter of the next layer of the data layer; Exemplarily, in this embodiment, before the first calculation, the parameters b _M , b. , q and the initial values of ^ and ^ of the first layer data layer may be randomly set. For the sake of simplicity, the parameter b The initial values of „, b. , q and ^ can be set to scalar parameters as

0, the vector parameter is a vector δ, wherein the arrow symbol above the number 0 is a vector symbol, and ^ and ^ can be arbitrarily set to a relatively small positive value, which is not limited by the embodiment of the present invention, and is used to indicate the regularity. For example, the processor 801 starts from the first layer of the data layer, sequentially calculates the values of the parameters b„, bq?, and ^. of the layer of the data layer in parallel, and takes the calculated parameter value as the next step. The parameters of the layer data layer b„, bq and the initial value, until the parameters of the last layer of data layer b _u , bq and ; are calculated; the parameters calculated by the last layer of the data layer are obtained; the specific calculation process is as follows: Processor 801 According to the initial value of the parameter of the first layer data layer, the unknown parameters b„, b., q _t , and the corresponding parameters corresponding to the scoring data included in the first layer of the data layer are updated, because there is no mutual mutuality between the scoring data of the same layer matrix. dependent relationship, the different data rates corresponding to the same layer as the matrix unknown parameters _{b ", b., q t} , can be calculated and updated in parallel, and the updated parameter as a parameter in one data layer Value for the next The parameters corresponding to the scoring data included in one layer of the data layer are updated and calculated, until the parameters corresponding to the scoring data included in the last layer of the data layer are updated and calculated, and the parameters calculated by the last layer of the data layer are obtained; preferably, In this embodiment, the processor 801 can perform the update calculation by the gradient descent method. Since the update methods of the parameters corresponding to all the scoring data are the same, those skilled in the art can understand that the parameters b _M , b corresponding to one scoring data are understood. ., qp _u and updates the calculation step will be described later, can be calculated without inventive step of applying other score data, the processor 801 may specifically calculate the step of formula (1): the following mouth ¾: first, the processor 801 obtain an initial estimated value of the scoring data r _ui according to the initial value of the parameter of each layer of the data layer and the recommended system model represented by the formula (1), and then find the layer data according to the scoring data and the initial estimated value of the data layer of the layer. Layer score error e _ui ;

Next, the processor 801 obtains an updated value of the parameter in the negative gradient direction by the scoring error ^ and the equation (5) - (9), and those skilled in the art can understand that the equations (5) - (8) can be corresponding according to the corresponding The calculation of the parameters of the calculation, the calculation process and the meaning of the symbols and parameters appearing in the various equations are not repeated, correspondingly, the solution to the optimization problem of equation (4) only needs to be carried out (5) - (8 Update calculation, get the parameter b„, bq P„;

However, for the calculation of the formula (9), the processor 801 also needs to perform serial calculation on the basis of the score data in the score data set, which reduces the computational efficiency. Therefore, the embodiment of the present invention provides a calculation method for the formula (9). The specific implementation manner is: when there are multiple calculations of the same layer matrix, the processor 801 uses the user corresponding to the product j represented in the score data set to calculate the corresponding gradient in parallel, for example, in the score data set, the user 1, User 2, and User 6 have produced product 4 ( j = 4 ).

The processor 801 then aggregates to obtain the value of the layer ^, ie +y ₂ (4) + Ay + A)), where is the initial value. In addition, the embodiment of the present invention provides another method for the equations (8) and (9). The specific calculation method is as follows: The expression _ft +|N( _M )|4 X in the first recommendation model is taken as an equivalent parameter, and the equivalent parameter is used to represent the equivalent parameter, that is, =A+| V ( )|4 , where the auxiliary variable is equivalent to the expression _ft +|N( _M )|4∑. At this time, the first recommended model is equivalent to eW(")

With this method, the update process of the parameter sum can be replaced only by the update process of the auxiliary variable. Furthermore, since the auxiliary variable is introduced, the update of the parameter q _t of the formula (8) can also be changed accordingly, that is, The update formula of q _t becomes ^■^^ + 7 ₂ · (^- where. is the initial value at the time of calculation in this layer.

It can be seen that the introduction of auxiliary variables simplifies the calculation on the one hand. On the other hand, the elimination of the calculation of ^, the parameter solution of the recommended model reduces the inner loop, thus greatly improving the operation speed while ensuring the same accuracy. Then, the processor 801 can use the parameter calculated by the layer data layer as the parameter initial value of the next layer of the data layer, and then the processor 801 uses the updated value as the initial value of the parameter of the second layer data layer, and according to the above The method calculates an updated value of the parameters of the second layer of data layers, and so on, and the processor 801 obtains an updated value of the parameters of the last layer of the data layer. Exemplarily, the processor 801 is further configured to determine, according to the parameter calculated by the last layer of the data layer, whether the recommended system model converges. If the convergence is performed, the calculation ends, and the optimal parameter is obtained; if the optimal parameter is obtained, if not, The parameter calculated by the last layer of the data layer is used as the initial value of the parameter of the first layer of the data layer, and the calculation of the recommended system model is continued.

Preferably, the processor 801 can calculate the obtained parameter of the last layer of the data layer obtained by the current calculation and the parameter calculated by the last layer of the data layer obtained by the previous calculation into the equation (3) or (4). The cost function is calculated. If the difference between the two calculation results is not greater than the preset threshold value, the parameter calculated in the last layer of the data layer obtained by the current calculation may be regarded as the optimal parameter, if it is greater than the preset threshold. Limit, then The parameters calculated in the last layer of the data layer obtained in this calculation are not convergent, and the parameters calculated in the last layer of the data layer obtained in this calculation are used as the parameters of the first layer of the data layer to be calculated next time. Value, continue to calculate the parameters of the recommended system model.

Exemplarily, the processor 801 brings the obtained optimal parameter into the formula (1) or the formula (2), and can obtain the predicted value of each user for each product, and can score the same user for all products. The predicted values are arranged, and a predetermined number of products with the highest score prediction value are selected for recommendation to the user. Exemplarily, the processor 801 can also be used to input new scoring data made by the user on the product into the recommendation system after the product recommendation, so that the recommendation system can ensure high according to the parameters of the system model in real-time update recommendation. Accuracy and high efficiency are recommended for real-time. The present embodiment provides a recommendation device 60, which improves the recommendation efficiency of the recommendation system in a massive data environment by parallel computing, and improves the recommendation effect of the recommendation system by considering user implicit feedback.

In the several embodiments provided by the present application, it should be understood that the disclosed system, apparatus, and method may be implemented in other manners. For example, the device embodiments described above are merely illustrative. For example, the division of cells is only a logical function division. In actual implementation, there may be another division manner. For example, multiple units or components may be combined or integrated. Go to another system, or some features can be ignored, or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interface, device or unit, and may be in an electrical, mechanical or other form. The units described as separate components may or may not be physically separate, and the components displayed as the units may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of the embodiment. In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may be physically included separately, or two or more units may be integrated into one unit. The above integrated unit can be implemented in the form of hardware or in the form of hardware plus software functional units. The above-described integrated unit implemented in the form of a software functional unit can be stored in a computer readable storage medium. The software functional unit described above is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, server, or network device, etc.) to perform portions of the steps of various embodiments of the present invention. The foregoing storage medium includes: a USB flash drive, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk, and the like, which can store program codes. Medium. It should be noted that the above embodiments are only for explaining the technical solutions of the present invention, and are not intended to be limiting; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those skilled in the art that: The technical solutions described in the foregoing embodiments are modified, or some of the technical features are equivalently replaced. The modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

claims

1. A recommended method, characterized by including:

Place the rating data in the rating data set into at least two data layers respectively, where the rating data corresponds to users and products respectively, and any two of the rating data in each of the data layers correspond to users And the products are not the same;

Based on the preset recommendation system model and the rating data in the data layer, the parameters of the recommendation system model in the data layer are calculated in parallel, and the parameters of each data layer are used as the initial values of the corresponding next data layer. , until the optimal parameters of the recommendation system model are obtained; wherein, the recommendation system model is the corresponding relationship between the predicted score of each user for each product and the average score and the parameters of the recommendation system model ;

Obtain the predicted score of each user for each product based on the optimal parameters and the recommendation system model, and recommend products to the user based on the predicted score.

2. The method according to claim 1, wherein the recommendation system model includes a recommendation system model that provides implicit feedback, a recommendation system model that does not provide implicit feedback, a recommendation system model that considers spatiotemporal characteristics, and a non-recommendation system model. A symmetric latent factor model for recommender systems.

3. The method according to claim 1 or 2, characterized in that the recommendation system model includes:

The first recommendation system model

Or, in the first recommendation model and the second recommendation model, the second recommendation system model represents the predicted rating value of user u for product i, μ represents the average value of all rating data in the rating data set, b „ represents the offset of the user u relative to the average user rating, b represents the offset of the product i relative to the average product rating, represents the product factor vector, T represents the transposition operator symbol, represents the user factor vector,

Further, in the first recommendation model, |N(w)| represents the set size of all products for which user u has provided implicit preferences, and N(w) represents the set of all products for which user u has provided implicit preferences. ; Represents the factor vector associated with product j, which is used to characterize implicit feedback information.

4. The method according to claim 3, further comprising: obtaining the recommendation system based on the relationship between the rating prediction value, the mean square error of the rating data and parameters of the recommendation system model. The cost function of the model, where the cost function includes:

first cost function

Or, the second cost function

Among them, |*|| represents the sum of squares of all elements of vector *, and ^ is the regularization factor.

5. The method according to any one of claims 1 to 4, characterized in that, based on the preset recommendation system model and the rating data in the data layer, the recommendation system model in the data layer is calculated in parallel. parameters, and use the parameters of each data layer as the initial value of the corresponding next data layer until the optimal parameters of the recommendation system model are obtained, including:

A: Calculate the average score of all the rating data in the rating data set;

B: Use parallel computing to calculate the parameters of each data layer in turn, and use the parameters calculated by each data layer as the initial parameter values of the next data layer; wherein, the first layer The initial parameter values of the data layer are set by the system;

C: Determine whether the recommendation system model converges based on the parameters calculated by the last layer of the data layer. If it converges, the calculation ends and the optimal parameters are obtained; if it does not converge, the last layer of the data layer The calculated parameters are used as the initial parameter values of the first data layer, and steps B and C are repeated.

6. The method according to claim 5, characterized in that step B includes:

B 1: Obtain the initial estimated value of the scoring data of the data layer of each layer based on the initial parameter value of the data layer and the recommendation system model, and then obtain the initial estimated value of the scoring data of the data layer based on the scoring data of the data layer and the recommendation system model. The initial estimation value obtains the scoring error of the data layer of the said layer;

B2: Obtain the parameters calculated by the data layer of this layer according to the scoring error; B3: Use the parameters calculated by the data layer of this layer as the initial value of the parameters of the next data layer, according to steps B1 and B 2 gets the parameters calculated by the next data layer until it is The parameters calculated by the data layer to the last layer.

7. The method according to claim 4 or 6, characterized in that, judging whether the recommendation system model converges based on the parameters calculated by the last layer of the data layer includes:

Substituting the parameters calculated from the last layer of the data layer obtained in this calculation and the parameters calculated from the last layer of the data layer obtained in the previous calculation into the cost function for calculation, if the cost is substituted The difference between the results calculated by the function is not greater than the preset threshold value, then the parameters calculated by the last layer of the data layer obtained by this calculation are convergent, otherwise, the parameters of the last layer obtained by this calculation are converged. The parameters calculated by the data layer are not convergent.

8. The method according to claim 6, characterized in that: obtaining the parameters calculated by the data layer of the layer according to the scoring error further includes: converting the expression ft + in the first recommendation model _| _N ₍ _M ₎ , -

Obtain the auxiliary variable, that is, obtain the equivalent parameter; and obtain the parameter q according to the auxiliary variable, that is, + 7 ₂ A^, ), where the symbol represents the update symbol, that is, replaced by the calculated value on the right side of the update symbol The variable value on the left side of the update symbol, the parameters appearing on the right side of the update symbol are the initial values of the corresponding parameters, and the parameters appearing on the left side of the update symbol are the updated values of the parameters.

9. A recommended device, characterized by including:

A data placement unit is used to place the rating data in the rating data set into at least two data layers respectively, wherein the rating data corresponds to users and products respectively, and any two of the data in each of the data layers The above rating data corresponds to different users and products;

A parallel computing unit, configured to calculate parameters of the recommendation system model in the data layer in parallel based on the preset recommendation system model and the rating data in the data layer, and use the parameters of each data layer as the corresponding next The initial value of the layer data layer until the optimal parameters of the recommendation system model are obtained; wherein, the recommendation system model is a parameter for each user for each product. The corresponding relationship between the predicted score of the product and the average score and the parameters of the recommendation system model;

A prediction and recommendation unit, configured to obtain each user's predicted score for each product based on the optimal parameters and the recommendation system model, and recommend products to the user based on the predicted score value.

10. The device according to claim 9, wherein the recommendation system model includes a recommendation system model that provides implicit feedback, a recommendation system model that does not provide implicit feedback, a recommendation system model that considers spatiotemporal characteristics, and Asymmetric latent factor model for recommender systems.

11. The device according to claim 9 or 10, characterized in that the recommendation system model includes:

The recommendation system model includes:

The first recommendation system model

1 2. The device according to claim 1 1, further comprising: a cost function generation unit, configured to generate the mean square error between the rating prediction value and the rating data and the parameters of the recommendation system model. The relationship between them obtains the cost function of the recommendation system model, where the cost function includes:

first cost function ∑ [r _tl -μ-b _u -b - _q (p _u +\N(uj,)] ² ^{2 2} )+ iu ₊ xi if)

≡N

Or, the second cost function

∑ [r _ui -μ-Κ-b - qp + {b _u ² + bf) + ^f + ||^ ) where, if represents the sum of squares of all elements of the vector *, and ^ is the regularization factor.

13. The device according to any one of claims 9-12, characterized in that the parallel computing unit includes:

The average score calculation subunit is used to calculate the average score of all the score data in the score data set;

The hierarchical calculation subunit is used to calculate the parameters of each layer of the data layer in a parallel computing manner, and use the parameters calculated by the data layer of each layer as the initial value of the parameters of the next layer of data layer; where , the initial value of the parameters of the first layer and the data layer is set by the system;

The convergence judgment subunit is used to judge whether the recommendation system model has converged based on the parameters calculated by the last layer of the data layer. If it converges, the calculation ends and the optimal parameters are obtained; if it does not converge, the last one is The parameters calculated by the data layer of the first layer are used as the initial parameter values of the data layer of the first layer, and the initial parameter values are transmitted to the hierarchical calculation subunit to repeat the hierarchical calculation.

14. The device according to claim 13, characterized in that the hierarchical computing subunit is further used to,

A rating error generation module, configured to obtain an initial estimate of the rating data of the data layer based on the initial parameter value of each data layer and the recommendation system model, and then based on the rating of the data layer. The data and the initial estimate are used to obtain the scoring error of the data layer of the layer;

A parameter calculation module, used to obtain the parameters calculated by the data layer of the layer according to the scoring error;

The calculation control module is used to use the parameters calculated by the data layer of this layer as the initial parameter values of the next data layer, and obtain the parameters calculated by the next data layer through the scoring error generation module and the parameter calculation module. parameters until the parameters calculated by the last layer of the data layer are obtained.

15. The device according to claim 13 or 14, characterized in that the convergence judgment subunit is further used to,

16. The device according to claim 14, characterized in that the parameter calculation module is further configured to use the expression ft + |N( _M ) ∑ in the first recommendation model as an equivalent parameter , and use auxiliary variables to represent the equivalent parameters, that is = _ft + |N( _M )|4; Then according to the gradient of the auxiliary variables Δζ _M = 2e _M · q _t -

Obtain the auxiliary variable, that is, obtain the equivalent parameter; and obtain the parameter q according to the auxiliary variable, that is, + 7 ₂ where, the symbol represents the update symbol, that is, the calculated value on the right side of the update symbol is used to replace the variable on the left side of the update symbol value, the parameters appearing on the right side of the update symbol are the initial values of the corresponding parameters, and the parameters appearing on the left side of the update symbol are the updated values of the parameters.

1 7. A recommended device, including a processor and a memory, where,

The processor is configured to place the rating data in the rating data set into at least two data layers respectively, wherein the rating data corresponds to users and products respectively, and any two of the data layers in each of the data layers. The above rating data corresponds to different users and products;

And based on the preset recommendation system model and the rating data in the data layer, the parameters of the recommendation system model in the data layer are calculated in parallel, and the parameters of each data layer are used as the initial parameters of the corresponding next data layer. value until the optimal parameters of the recommendation system model are obtained; wherein, the recommendation system model is the correspondence between the predicted score of each user for each product and the average score and the parameters of the recommendation system model relation;

and obtain each user's response to each user based on the optimal parameters and the recommendation system model. Predicted rating values of each product, and recommend products to the user based on the predicted rating values; the memory is used to save the rating data set and the program executed by the processor and the results of the execution.

18. The device according to claim 17, wherein the recommendation system model includes a recommendation system model that provides implicit feedback, a recommendation system model that does not provide implicit feedback, and a recommendation system model that considers spatiotemporal characteristics. and asymmetric latent factors in recommender system models.

19. The device according to claim 17 or 18, the recommendation system model includes: The recommendation system model includes:

The first recommendation system model

Further, in the first recommendation model, |N(w)| represents the set size of all products for which user u has provided implicit preferences, and N(w) represents the set size of all products for which user u has provided implicit preferences. ; Represents the factor vector associated with product j, which is used to characterize implicit feedback information.

20. The device according to claim 19, characterized in that the processor is further configured to calculate the mean square error between the rating prediction value and the rating data and the parameters of the recommendation system model. The cost function of the recommendation system model is obtained by the relationship, where the cost function includes:

first cost function

Or, the second cost function ∑ [r _ui -μ- -b - qp _u f + {b _u ² + bf) + ^\\ +\\p _u \\ ) where, if represents the sum of squares of all elements of the vector *, and ^ is regular transformation factor.

21. The device according to any one of claims 18-20, characterized in that the processor is used to,

A: Calculate the average score of all the rating data in the rating data set;

22. The device according to claim 21, characterized in that, the processor is used to,

B1: Obtain the initial estimated value of the rating data of the data layer of each layer based on the initial parameter value of the data layer and the recommendation system model, and then based on the rating data of the data layer of the layer and the initial The estimated value is the scoring error of the data layer in question;

B2: Obtain the parameters calculated by the data layer of this layer according to the scoring error; B3: Use the parameters calculated by the data layer of this layer as the initial value of the parameters of the next data layer, obtained according to steps B1 and B2 parameters calculated by the next data layer until the parameters calculated by the last data layer are obtained.

23. The device according to claim 21 or 22, characterized in that the processor is used to,

24. The device according to claim 22, wherein the processor is configured to obtain parameters calculated by the data layer of the layer according to the scoring error, further comprising: Includes:

The processor takes the expression _ft +|N( _M )|4 X in the first recommendation model as an equivalent parameter, and uses auxiliary variables to represent the equivalent parameter, that is

Z y.; Then according to the gradient of the auxiliary variable Δζ _M = 2e _M · q _t

Obtain the auxiliary variable, that is, obtain the equivalent parameter; and obtain the parameter q according to the auxiliary variable, that is, +7 ₂ A^,), where the symbol represents the update symbol, that is, replaced by the calculated value on the right side of the update symbol The variable value on the left side of the update symbol, the parameters appearing on the right side of the update symbol are the initial values of the corresponding parameters, and the parameters appearing on the left side of the update symbol are the updated values of the parameters.