CN111177581A - Multi-platform-based social e-commerce website commodity recommendation method and device - Google Patents
Multi-platform-based social e-commerce website commodity recommendation method and device Download PDFInfo
- Publication number
- CN111177581A CN111177581A CN201911359245.2A CN201911359245A CN111177581A CN 111177581 A CN111177581 A CN 111177581A CN 201911359245 A CN201911359245 A CN 201911359245A CN 111177581 A CN111177581 A CN 111177581A
- Authority
- CN
- China
- Prior art keywords
- social
- user
- commodity
- commerce platform
- commerce
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 46
- 230000004927 fusion Effects 0.000 claims abstract description 85
- 230000003993 interaction Effects 0.000 claims abstract description 73
- 230000006399 behavior Effects 0.000 claims abstract description 72
- 238000012549 training Methods 0.000 claims abstract description 53
- 230000006870 function Effects 0.000 claims description 52
- 230000002452 interceptive effect Effects 0.000 claims description 41
- 238000004590 computer program Methods 0.000 claims description 9
- 238000007781 pre-processing Methods 0.000 claims description 4
- 230000009467 reduction Effects 0.000 claims description 4
- 239000000126 substance Substances 0.000 claims description 4
- 230000000087 stabilizing effect Effects 0.000 claims description 3
- 230000008569 process Effects 0.000 abstract description 6
- 238000010586 diagram Methods 0.000 description 8
- 230000000694 effects Effects 0.000 description 7
- 239000013598 vector Substances 0.000 description 6
- 238000004891 communication Methods 0.000 description 5
- 238000005457 optimization Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000012163 sequencing technique Methods 0.000 description 3
- 238000012512 characterization method Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 241001122767 Theaceae Species 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9536—Search customisation based on social or collaborative filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/01—Social networking
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Finance (AREA)
- Economics (AREA)
- Accounting & Taxation (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- General Health & Medical Sciences (AREA)
- Tourism & Hospitality (AREA)
- Primary Health Care (AREA)
- Human Resources & Organizations (AREA)
- Health & Medical Sciences (AREA)
- Development Economics (AREA)
- Computing Systems (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the invention provides a multi-platform-based social e-commerce website commodity recommendation method and device, wherein the method comprises the following steps: acquiring commodity interaction behavior data of a user of a common e-commerce platform, commodity interaction behavior data of a user of a social e-commerce platform and social relation data of the user of the social e-commerce platform; and inputting commodity interaction behavior data of the common E-commerce platform user, commodity interaction behavior data of the social E-commerce platform user and social relation data of the social E-commerce platform user into a preset social-common E-commerce platform fusion model to obtain commodity personalized recommendation information. The fusion model loss function is constructed based on the common e-commerce recommendation model and the social e-commerce recommendation model, so that the user commodity interaction behavior data from the common e-commerce platform and the user commodity interaction behavior data from the social e-commerce platform can be effectively utilized in the training process, the user behaviors of the common platform and the social e-commerce platform are fully considered, and the commodity personalized recommendation accuracy is improved.
Description
Technical Field
The invention relates to the technical field of information processing, in particular to a multi-platform-based social e-commerce website commodity recommendation method and device.
Background
With the development of internet technology, Electronic Business (E-Business) is seen everywhere in our daily life and work. Electronic commerce generally refers to a novel business operation mode in which, in wide commercial and trade activities worldwide, in an internet environment open to the internet, buyers and sellers conduct various commercial and trade activities without conspiracy based on a browser/server application mode, and consumer online shopping, online transactions and online electronic payments among merchants, and various commercial activities, transaction activities, financial activities, and related comprehensive service activities are realized.
At present, the existing e-commerce recommendation system usually obtains user preference characteristics according to the historical behavior of the user on the e-commerce website, and then finds out and recommends the commodities related to the user preference to the user.
With the rapid development of the internet and the mobile internet, the amount of network information is huge, and users are puzzled to obtain own requirements from massive information.
Therefore, how to more accurately and effectively implement the recommendation of the social e-commerce website commodities based on multiple platforms has become an urgent problem to be solved in the industry.
Disclosure of Invention
The embodiment of the invention provides a multi-platform-based social e-commerce website commodity recommendation method and device, which are used for solving the technical problems in the background technology or at least partially solving the technical problems in the background technology.
In a first aspect, an embodiment of the present invention provides a multi-platform-based method for recommending a commodity in a social e-commerce website, including:
acquiring commodity interaction behavior data of a user of a common e-commerce platform, commodity interaction behavior data of a user of a social e-commerce platform and social relation data of the user of the social e-commerce platform;
inputting the commodity interaction behavior data of the common E-commerce platform user, the commodity interaction behavior data of the social E-commerce platform user and the social relation data of the social E-commerce platform user into a preset social-common E-commerce platform fusion model to obtain commodity personalized recommendation information;
the preset social-common E-commerce platform fusion model is obtained through training of user commodity sample data with interactive labels.
More specifically, before the step of inputting the commodity interaction behavior data of the general e-commerce platform user, the commodity interaction behavior data of the social e-commerce platform user and the social relationship data of the social e-commerce platform user into a preset social-general e-commerce platform fusion model, the method further includes:
acquiring user commodity sample data with an interactive tag;
randomly acquiring user commodity sample data with a label which is not interacted from an unobserved sample set, and constructing positive and negative sample pair information according to the user commodity sample data with the label which is interacted and the user commodity sample data with the label which is not interacted;
inputting the information of the positive and negative sample pairs into a social-common E-commerce platform fusion model, finishing training when preset training conditions are met, and outputting commodity personalized recommendation information to obtain the preset social-common E-commerce platform fusion model.
More specifically, before the step of inputting the positive and negative samples into the social-common e-commerce platform fusion model for training, the method further includes:
obtaining a common E-commerce recommendation model loss function and a social E-commerce recommendation model loss function;
constructing a fusion model loss function according to the common E-commerce recommendation model loss function and the social E-commerce recommendation model loss function;
and obtaining a social-common E-commerce platform fusion model according to the fusion model loss function.
More specifically, the fusion model loss function is specifically:
L＝L(A)+L(T)
wherein the content of the first and second substances,
indicating whether user u has an interaction with item i,
in the loss function, λ, which represents whether the interaction pair of user u and commodity i is selected for trainingSIs a coefficient of a social regularization term.
More specifically, the step of obtaining the fusion model of the social-general e-commerce platform according to the fusion model loss function specifically includes:
acquiring user commodity sample data with interactive labels and user commodity sample data without interactive labels to construct positive and negative sample pair information;
constructing positive and negative sample pair information through user commodity sample data with interactive labels and user commodity sample data without interactive labels, training a fusion model loss function parameter through a gradient random reduction method, and when preset training conditions are met, stabilizing the fusion model loss function so as to obtain a social-common electronic commerce platform fusion model according to the fusion model loss function.
In a second aspect, an embodiment of the present invention provides a multi-platform-based social e-commerce website commodity recommendation apparatus, including:
the acquisition module is used for acquiring commodity interaction behavior data of a user of the common e-commerce platform, commodity interaction behavior data of a user of the social e-commerce platform and social relation data of the user of the social e-commerce platform;
the recommendation module is used for inputting the commodity interaction behavior data of the common E-commerce platform user, the commodity interaction behavior data of the social E-commerce platform user and the social relation data of the social E-commerce platform user into a preset social-common E-commerce platform fusion model to obtain commodity personalized recommendation information;
the preset social-common E-commerce platform fusion model is obtained through training of user commodity sample data with interactive labels.
More specifically, the apparatus further comprises a pre-processing module;
the preprocessing module is used for acquiring user commodity sample data with an interactive tag;
randomly acquiring user commodity sample data with a label which is not interacted from an unobserved sample set, and constructing positive and negative sample pair information according to the user commodity sample data with the label which is interacted and the user commodity sample data with the label which is not interacted;
inputting the information of the positive and negative sample pairs into a social-common E-commerce platform fusion model, finishing training when preset training conditions are met, and outputting commodity personalized recommendation information to obtain the preset social-common E-commerce platform fusion model.
In a third aspect, an embodiment of the present invention provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the steps of the multi-platform-based social e-commerce website commodity recommendation method according to the first aspect when executing the program.
In a fourth aspect, an embodiment of the present invention provides a non-transitory computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the steps of the multi-platform based social e-commerce website commodity recommendation method according to the first aspect.
According to the social e-commerce website commodity recommendation method and device based on the multiple platforms, the fusion model loss function is constructed based on the common e-commerce recommendation model and the social e-commerce recommendation model, so that the user commodity interaction behavior data from the common e-commerce platform and the social e-commerce platform user commodity interaction behavior data can be effectively utilized in the training process, the user behaviors of the common platform and the social e-commerce platform are fully considered, and the commodity personalized recommendation accuracy is improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
Fig. 1 is a schematic flow chart illustrating a multi-platform-based method for recommending commodities to an e-commerce website according to an embodiment of the present invention;
FIG. 2 is a diagram illustrating a social e-commerce platform and a comparison of common e-commerce items, according to an embodiment of the present invention;
FIG. 3 is a schematic diagram illustrating a modeling of a social-general e-commerce platform fusion model according to an embodiment of the invention;
FIG. 4 is a schematic structural diagram of a multi-platform-based merchandise recommendation device for social E-commerce websites according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Fig. 1 is a schematic flow chart of a multi-platform-based social e-commerce website commodity recommendation method according to an embodiment of the present invention, as shown in fig. 1, including:
step S1, commodity interaction behavior data of a user of the common E-commerce platform, commodity interaction behavior data of a user of the social E-commerce platform and social relation data of the user of the social E-commerce platform are obtained;
step S2, inputting the commodity interaction behavior data of the ordinary E-commerce platform user, the commodity interaction behavior data of the social E-commerce platform user and the social relation data of the social E-commerce platform user into a preset social-ordinary E-commerce platform fusion model to obtain commodity personalized recommendation information;
the preset social-common E-commerce platform fusion model is obtained through training of user commodity sample data with interactive labels.
Specifically, the social e-commerce platform described in the embodiment of the present invention is an e-commerce platform in which a social relationship exists and mutual influence exists between users, the existence of the social relationship is mainly different from that of a general e-commerce platform, and fig. 2 is a comparison diagram of the social e-commerce platform and a general e-commerce platform described in the embodiment of the present invention, as shown in fig. 2, therefore, the general e-commerce platform is an e-commerce platform without a social relationship, the user commodity interaction behavior data described in the embodiment of the present invention refers to interaction records of users on the social e-commerce platform, and the social relationship data refers to friend relationships of users on the social e-commerce platform.
The user commodity interaction behavior data described in the embodiment of the invention specifically refers to interaction behavior data of a user and a merchant, and comprises behavior data of clicking commodities by the user and behavior data of collecting or purchasing the commodities by the user.
The preset social contact-common e-commerce platform fusion model described in the embodiment of the invention is used for realizing personalized commodity recommendation for the social e-commerce platform according to the commodity interaction behavior data of the users of the electrified e-commerce platform, the commodity interaction behavior data of the users of the social e-commerce platform and the social relation data of the users of the social e-commerce platform, and specifically, low-dimensional vectors of the interests of the users on the social e-commerce platform are described according to the preset social contact-common e-commerce platform fusion model, similarity calculation is carried out on the vectors and the characterization vectors of all commodities of the platform, sequencing is carried out according to the similarity from high to low, and K commodities in front of the sequencing list are selected as personalized commodity recommendation information.
The preset social contact-common E-commerce platform fusion model described in the embodiment of the invention is obtained by training user commodity sample data with an interactive tag, and a fusion model loss function is constructed according to a common E-commerce recommendation model loss function and a social E-commerce recommendation model loss function; obtaining a social contact-common E-commerce platform fusion model according to a fusion model loss function, randomly obtaining user commodity sample data with an un-interacted label from an unobserved sample set, and constructing positive and negative sample pair information according to the user commodity sample data with the interacted label and the user commodity sample data with the un-interacted label; inputting the information of the positive and negative sample pairs into a social-common E-commerce platform fusion model for training, finishing the training when a preset training condition is met, and outputting commodity personalized recommendation information to obtain the preset social-common E-commerce platform fusion model.
According to the embodiment of the invention, the fusion model loss function is constructed based on the common E-commerce recommendation model and the social E-commerce recommendation model, so that the user commodity interaction behavior data from the common E-commerce platform and the user commodity interaction behavior data from the social E-commerce platform can be effectively utilized in the training process, the user behaviors of the common platform and the social E-commerce platform are fully considered, and the commodity personalized recommendation accuracy is improved.
On the basis of the above embodiment, before the step of inputting the commodity interaction behavior data of the general e-commerce platform user, the commodity interaction behavior data of the social e-commerce platform user, and the social relationship data of the social e-commerce platform user into a preset social-general e-commerce platform fusion model, the method further includes:
acquiring user commodity sample data with an interactive tag;
randomly acquiring user commodity sample data with a label which is not interacted from an unobserved sample set, and constructing positive and negative sample pair information according to the user commodity sample data with the label which is interacted and the user commodity sample data with the label which is not interacted;
inputting the information of the positive and negative sample pairs into a social-common E-commerce platform fusion model for training, finishing the training when a preset training condition is met, and outputting commodity personalized recommendation information to obtain the preset social-common E-commerce platform fusion model.
Specifically, the interactive tag described in the embodiment of the present invention represents that the user product sample data is data interacted with the user, and the non-interactive tag refers to that the user product sample data is data that has not been interacted with the user.
The user commodity sample data with the interactive label is obtained from user logs collected by a traditional e-commerce platform and a social e-commerce platform, a personalized commodity list is generated for a user, and only data sets of the user and the commodity existing in the two platforms are reserved, so that the user commodity sample data with the interactive label is obtained.
The unobserved sample set described in the embodiment of the present invention specifically refers to data that has not interacted with a user, and the constructing of the positive and negative sample pair information specifically refers to that based on the positive: and the proportion of minus 1:1 takes the user commodity sample data with the interactive label as a positive example, and takes the user commodity sample data without the interactive label as a negative example, so as to construct the user commodity sample data with the interactive label and the user commodity sample data without the interactive label.
The preset training condition described in the embodiment of the present invention may refer to a preset number of training rounds, for example, 300 training rounds, or may refer to a preset training time, for example, a training time of 30 minutes.
According to the embodiment of the invention, the commodity interactive behavior data of the user from the common e-commerce platform and the commodity interactive behavior data of the user from the social e-commerce platform are effectively utilized, and the user behaviors of the common platform and the social e-commerce platform are fully considered to establish the preset social-common e-commerce platform fusion model, so that the commodity personalized recommendation accuracy is improved.
On the basis of the above embodiment, before the step of inputting the positive and negative samples into the social-common e-commerce platform fusion model for training, the method further includes:
obtaining a common E-commerce recommendation model loss function and a social E-commerce recommendation model loss function;
constructing a fusion model loss function according to the common E-commerce recommendation model loss function and the social E-commerce recommendation model loss function;
and obtaining a social-common E-commerce platform fusion model according to the fusion model loss function.
The fusion model loss function is specifically as follows:
L＝L(A)+L(T)
wherein the content of the first and second substances,
indicating whether user u has an interaction with item i,
in the loss function, λ, which represents whether the interaction pair of user u and commodity i is selected for trainingSIs a coefficient of a social regularization term.
Specifically, the loss function of the common electricity quotient recommendation model described in the embodiment of the present invention is L(A)The social e-commerce recommendation model loss function is L(T)The fusion model loss function is L ═ L(A)+L(T)。
FIG. 3 is a schematic diagram illustrating a modeling of a social-general e-commerce platform fusion model according to an embodiment of the present invention, as shown in FIG. 3, that is, a loss function of two parts is combinedTogether, i.e. L ═ L(A)+L(T). For the optimization of the loss function, all model parameters are learned based on random gradient descent. The observed interactions are sparse compared to all user-commodity pairs, while optimizing only the observed interactions may leave the model in local optimality. The embodiment of the invention randomly selects a proper number of user-commodity pairs from unobserved interactions based on random negative sampling as a negative example of sample training. In order to better learn from implicit feedback data, the method uses a sequencing learning mode to carry out optimization, namely, the distance between the predicted value of the positive case and the predicted value of the negative case is optimized, and the method is insensitive to the size of the negative sampling proportion, so that the training of the model is more stable and easy to converge.
And meanwhile, positive and negative sample pair information is constructed through user commodity sample data with interactive labels and user commodity sample data without interactive labels, and loss function parameters of the fusion model are trained through a gradient random reduction method, so that a social-common electronic commerce platform fusion model is finally obtained.
On the basis of the above embodiment, the step of obtaining the fusion model of the social-general e-commerce platform according to the fusion model loss function specifically includes:
acquiring user commodity sample data with interactive labels and user commodity sample data without interactive labels to construct positive and negative sample pair information;
constructing positive and negative sample pair information through user commodity sample data with interactive labels and user commodity sample data without interactive labels, training a fusion model loss function parameter through a gradient random reduction method, and when preset training conditions are met, stabilizing the fusion model loss function so as to obtain a social-common electronic commerce platform fusion model according to the fusion model loss function.
Specifically, training data in a user-positive and negative sample pair information triple form is constructed, the size of tea between a user commodity sample data score of an interactive label and a user commodity sample data score of a non-interactive label is as large as possible, and model parameters are updated in a random gradient descending mode, so that a loss function is smaller and smaller, the difference of partial order learning is larger and larger, and finally convergence to stability is achieved. The update formula is as follows:
wherein the content of the first and second substances,
indicating whether user u has an interaction with item i,
in the loss function, λ, which represents whether the interaction pair of user u and commodity i is selected for trainingSIs a coefficient of a social regularization term.
According to the embodiment of the invention, the fusion model loss function is constructed based on the common E-commerce recommendation model and the social E-commerce recommendation model, so that the user commodity interaction behavior data from the common E-commerce platform and the user commodity interaction behavior data from the social E-commerce platform can be effectively utilized in the training process, the user behaviors of the common platform and the social E-commerce platform are fully considered, and the commodity personalized recommendation accuracy is improved.
On the basis of the above embodiment, a preset social-general e-commerce platform fusion model is constructed by using user logs of a traditional e-commerce platform and a social e-commerce platform collected from 6/1/2018 to 6/30/6, a personalized commodity list is generated for a user, and only data sets of the user and the commodity existing in the two platforms are retained, and statistics are as described in table 1 below:
TABLE 1
| Number of users | 2,623,433 |
| Number of commodities | 1,194,766 |
| Number of friendship | 152,982 |
| Number of records purchased by traditional e-commerce platform | 2,006,887 |
| Number of purchase records of social e-commerce platform | 3,923,367 |
Firstly, training data is constructed, in order to effectively guarantee the online effect of an online model, purchasing records are sorted according to time stamps, the latest record is reserved for each user to serve as a test set, and after one record is randomly selected from other records to serve as a verification set, the rest records serve as training sets.
Selecting an Adam optimizer as a stochastic gradient descent optimizer, based on positive: the negative-1: 1 ratio randomly selects samples from all the unobserved samples as negative examples, and then positive and negative sample pairs for sequence learning are constructed. For the characterization matrix of the user and the commodity, in order to prevent overfitting, an L2 regular term is introduced into an optimization target, and the coefficient of the regular term is set to be 0.01 which is commonly used. And selecting the optimal hyper-parameter combination as a final model by observing the recall performance based on two evaluation indexes of HR and NDCG on the verification set based on grid search for the dimensionality of a vector space, the coefficient of a social regular term, the learning rate of model optimization and other core hyper-parameters.
The trained models are then deployed to a recall phase of the online recommendation system, which is intended to select a medium number of commodities from a large commodity pool for further ranking. Considering the number of massive users and commodities which can be handled by the actual recommendation system, the recommendation model is required to have lower storage cost and higher calculation efficiency at this stage, and the model output by the invention meets the two requirements. Specifically, on one hand, the storage requirement of the model is in a linear relation with the number of users/commodities, and on the other hand, the model is in a vector library form in actual deployment and is combined with a distributed vector recall engine, so that the calculation efficiency is high.
Fig. 4 is a schematic structural diagram of a multi-platform-based social e-commerce website commodity recommendation device according to an embodiment of the present invention, as shown in fig. 4, including: an acquisition module 410 and a recommendation module 420; the obtaining module 410 is configured to obtain commodity interaction behavior data of a user of the general e-commerce platform, commodity interaction behavior data of a user of the social e-commerce platform, and social relationship data of the user of the social e-commerce platform; the recommendation module 420 is configured to input the commodity interaction behavior data of the general e-commerce platform user, the commodity interaction behavior data of the social e-commerce platform user and the social relationship data of the social e-commerce platform user into a preset social-general e-commerce platform fusion model to obtain commodity personalized recommendation information; the preset social-common E-commerce platform fusion model is obtained through training of user commodity sample data with interactive labels.
The apparatus provided in the embodiment of the present invention is used for executing the above method embodiments, and for details of the process and the details, reference is made to the above embodiments, which are not described herein again.
According to the embodiment of the invention, the fusion model loss function is constructed based on the common E-commerce recommendation model and the social E-commerce recommendation model, so that the user commodity interaction behavior data from the common E-commerce platform and the user commodity interaction behavior data from the social E-commerce platform can be effectively utilized in the training process, the user behaviors of the common platform and the social E-commerce platform are fully considered, and the commodity personalized recommendation accuracy is improved.
Fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present invention, and as shown in fig. 5, the electronic device may include: a processor (processor)510, a communication Interface (Communications Interface)520, a memory (memory)530 and a communication bus 540, wherein the processor 510, the communication Interface 520 and the memory 530 communicate with each other via the communication bus 540. Processor 510 may call logic instructions in memory 530 to perform the following method: acquiring commodity interaction behavior data of a user of a common e-commerce platform, commodity interaction behavior data of a user of a social e-commerce platform and social relation data of the user of the social e-commerce platform; inputting the commodity interaction behavior data of the common E-commerce platform user, the commodity interaction behavior data of the social E-commerce platform user and the social relation data of the social E-commerce platform user into a preset social-common E-commerce platform fusion model to obtain commodity personalized recommendation information; the preset social-common E-commerce platform fusion model is obtained through training of user commodity sample data with interactive labels.
Furthermore, the logic instructions in the memory 530 may be implemented in the form of software functional units and stored in a computer readable storage medium when the software functional units are sold or used as independent products. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
An embodiment of the present invention discloses a computer program product, which includes a computer program stored on a non-transitory computer readable storage medium, the computer program including program instructions, when the program instructions are executed by a computer, the computer can execute the methods provided by the above method embodiments, for example, the method includes: acquiring commodity interaction behavior data of a user of a common e-commerce platform, commodity interaction behavior data of a user of a social e-commerce platform and social relation data of the user of the social e-commerce platform; inputting the commodity interaction behavior data of the common E-commerce platform user, the commodity interaction behavior data of the social E-commerce platform user and the social relation data of the social E-commerce platform user into a preset social-common E-commerce platform fusion model to obtain commodity personalized recommendation information; the preset social-common E-commerce platform fusion model is obtained through training of user commodity sample data with interactive labels.
Embodiments of the present invention provide a non-transitory computer-readable storage medium storing server instructions, where the server instructions cause a computer to execute the method provided in the foregoing embodiments, for example, the method includes: acquiring commodity interaction behavior data of a user of a common e-commerce platform, commodity interaction behavior data of a user of a social e-commerce platform and social relation data of the user of the social e-commerce platform; inputting the commodity interaction behavior data of the common E-commerce platform user, the commodity interaction behavior data of the social E-commerce platform user and the social relation data of the social E-commerce platform user into a preset social-common E-commerce platform fusion model to obtain commodity personalized recommendation information; the preset social-common E-commerce platform fusion model is obtained through training of user commodity sample data with interactive labels.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (9)
1. A multi-platform-based social e-commerce website commodity recommendation method is characterized by comprising the following steps:
acquiring commodity interaction behavior data of a user of a common e-commerce platform, commodity interaction behavior data of a user of a social e-commerce platform and social relation data of the user of the social e-commerce platform;
inputting the commodity interaction behavior data of the common E-commerce platform user, the commodity interaction behavior data of the social E-commerce platform user and the social relation data of the social E-commerce platform user into a preset social-common E-commerce platform fusion model to obtain commodity personalized recommendation information;
the preset social-common E-commerce platform fusion model is obtained through training of user commodity sample data with interactive labels.
2. The multi-platform-based social e-commerce website commodity recommendation method according to claim 1, wherein before the step of inputting the general e-commerce platform user commodity interaction behavior data, the social e-commerce platform user commodity interaction behavior data and the social e-commerce platform user social relationship data into a preset social-general e-commerce platform fusion model, the method further comprises:
acquiring user commodity sample data with an interactive tag;
randomly acquiring user commodity sample data with a label which is not interacted from an unobserved sample set, and constructing positive and negative sample pair information according to the user commodity sample data with the label which is interacted and the user commodity sample data with the label which is not interacted;
and inputting the information of the positive and negative sample pairs into a social-common E-commerce platform fusion model for training, and finishing the training when a preset training condition is met to obtain a preset social-common E-commerce platform fusion model.
3. The multi-platform-based social e-commerce website commodity recommendation method according to claim 2, wherein before the step of training the information input into the social-ordinary e-commerce platform fusion model by the positive and negative sample pairs, the method further comprises:
obtaining a common E-commerce recommendation model loss function and a social E-commerce recommendation model loss function;
constructing a fusion model loss function according to the common E-commerce recommendation model loss function and the social E-commerce recommendation model loss function;
and obtaining a social-common E-commerce platform fusion model according to the fusion model loss function.
4. The multi-platform-based social e-commerce website commodity recommendation method according to claim 3, wherein the fusion model loss function is specifically as follows:
L＝L(A)+L(T)
wherein the content of the first and second substances,
indicating whether user u has an interaction with item i,
in the loss function, λ, which represents whether the interaction pair of user u and commodity i is selected for trainingSIs a coefficient of a social regularization term.
5. The multi-platform-based social e-commerce website commodity recommendation method according to claim 4, wherein the step of obtaining the social-general e-commerce platform fusion model according to the fusion model loss function specifically comprises:
acquiring user commodity sample data with interactive labels and user commodity sample data without interactive labels to construct positive and negative sample pair information;
constructing positive and negative sample pair information through user commodity sample data with interactive labels and user commodity sample data without interactive labels, training a fusion model loss function parameter through a gradient random reduction method, and when preset training conditions are met, stabilizing the fusion model loss function so as to obtain a social-common electronic commerce platform fusion model according to the fusion model loss function.
6. A social e-commerce website commodity recommendation device based on multiple platforms is characterized by comprising:
the acquisition module is used for acquiring commodity interaction behavior data of a user of the common e-commerce platform, commodity interaction behavior data of a user of the social e-commerce platform and social relation data of the user of the social e-commerce platform;
the recommendation module is used for inputting the commodity interaction behavior data of the common E-commerce platform user, the commodity interaction behavior data of the social E-commerce platform user and the social relation data of the social E-commerce platform user into a preset social-common E-commerce platform fusion model to obtain commodity personalized recommendation information;
the preset social-common E-commerce platform fusion model is obtained through training of user commodity sample data with interactive labels.
7. The multi-platform-based social merchant website commodity recommendation device of claim 6, further comprising a preprocessing module;
the preprocessing module is used for acquiring user commodity sample data with an interactive tag;
randomly acquiring user commodity sample data with a label which is not interacted from an unobserved sample set, and constructing positive and negative sample pair information according to the user commodity sample data with the label which is interacted and the user commodity sample data with the label which is not interacted;
inputting the information of the positive and negative sample pairs into a social-common E-commerce platform fusion model, finishing training when preset training conditions are met, and outputting commodity personalized recommendation information to obtain the preset social-common E-commerce platform fusion model.
8. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the steps of the multi-platform based social e-commerce website commodity recommendation method according to any one of claims 1 to 5 when executing the program.
9. A non-transitory computer readable storage medium having a computer program stored thereon, wherein the computer program when executed by a processor implements the steps of the multi-platform based social merchant website commodity recommendation method according to any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911359245.2A CN111177581A (en) | 2019-12-25 | 2019-12-25 | Multi-platform-based social e-commerce website commodity recommendation method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911359245.2A CN111177581A (en) | 2019-12-25 | 2019-12-25 | Multi-platform-based social e-commerce website commodity recommendation method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111177581A true CN111177581A (en) | 2020-05-19 |
Family
ID=70654053
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911359245.2A Pending CN111177581A (en) | 2019-12-25 | 2019-12-25 | Multi-platform-based social e-commerce website commodity recommendation method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111177581A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111738802A (en) * | 2020-06-30 | 2020-10-02 | 广东奥园奥买家电子商务有限公司 | Introduction method, device and equipment of E-commerce commodities |
| CN112667907A (en) * | 2021-01-04 | 2021-04-16 | 深圳市华通易点信息技术有限公司 | Multi-platform information delivery and data statistics system and method |
| CN113850649A (en) * | 2021-09-29 | 2021-12-28 | 中国科学院空天信息创新研究院 | Customized recommendation method and recommendation system based on multi-platform user data |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103886074A (en) * | 2014-03-24 | 2014-06-25 | 江苏名通信息科技有限公司 | Commodity recommendation system based on social media |
| CN105740430A (en) * | 2016-01-29 | 2016-07-06 | 大连理工大学 | Personalized recommendation method with socialization information fused |
| US20170193390A1 (en) * | 2015-12-30 | 2017-07-06 | Facebook, Inc. | Identifying Entities Using a Deep-Learning Model |
| US20180137389A1 (en) * | 2016-11-16 | 2018-05-17 | Facebook, Inc. | Deep Multi-Scale Video Prediction |
| CN110222272A (en) * | 2019-04-18 | 2019-09-10 | 广东工业大学 | A kind of potential customers excavate and recommended method |
-
2019
- 2019-12-25 CN CN201911359245.2A patent/CN111177581A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103886074A (en) * | 2014-03-24 | 2014-06-25 | 江苏名通信息科技有限公司 | Commodity recommendation system based on social media |
| US20170193390A1 (en) * | 2015-12-30 | 2017-07-06 | Facebook, Inc. | Identifying Entities Using a Deep-Learning Model |
| CN105740430A (en) * | 2016-01-29 | 2016-07-06 | 大连理工大学 | Personalized recommendation method with socialization information fused |
| US20180137389A1 (en) * | 2016-11-16 | 2018-05-17 | Facebook, Inc. | Deep Multi-Scale Video Prediction |
| CN110222272A (en) * | 2019-04-18 | 2019-09-10 | 广东工业大学 | A kind of potential customers excavate and recommended method |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111738802A (en) * | 2020-06-30 | 2020-10-02 | 广东奥园奥买家电子商务有限公司 | Introduction method, device and equipment of E-commerce commodities |
| CN112667907A (en) * | 2021-01-04 | 2021-04-16 | 深圳市华通易点信息技术有限公司 | Multi-platform information delivery and data statistics system and method |
| CN112667907B (en) * | 2021-01-04 | 2023-10-17 | 深圳市华通易点信息技术有限公司 | Multi-platform information delivery and data statistics system and method |
| CN113850649A (en) * | 2021-09-29 | 2021-12-28 | 中国科学院空天信息创新研究院 | Customized recommendation method and recommendation system based on multi-platform user data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110222272B (en) | Potential customer mining and recommending method | |
| CN111144986B (en) | Social electronic commerce website commodity recommendation method and device based on sharing behavior | |
| CN111008858A (en) | Commodity sales prediction method and system | |
| CN109711955B (en) | Poor evaluation early warning method and system based on current order and blacklist base establishment method | |
| CN108921602B (en) | User purchasing behavior prediction method based on integrated neural network | |
| CN111177581A (en) | Multi-platform-based social e-commerce website commodity recommendation method and device | |
| CN112435067A (en) | Intelligent advertisement putting method and system for cross-e-commerce platform and social platform | |
| EP2545479A2 (en) | Methods, computer-accessible medium and systems for construction of and inference with networked data, for example, in a financial setting | |
| CN107423356A (en) | The processing method and processing device of evaluation information, computer-readable medium, electronic equipment | |
| CN107301592A (en) | The method and device excavated for commodity substitute | |
| JP6976207B2 (en) | Information processing equipment, information processing methods, and programs | |
| CN111949887A (en) | Item recommendation method and device and computer-readable storage medium | |
| US11762819B2 (en) | Clustering model analysis for big data environments | |
| CN111695965A (en) | Product screening method, system and equipment based on graph neural network | |
| CN106600039A (en) | E-commerce commodity information pushing optimization method | |
| CN115409577A (en) | Intelligent container repurchase prediction method and system based on user behavior and environmental information | |
| CN111382843A (en) | Method and device for establishing upstream and downstream relation recognition model of enterprise and relation mining | |
| CN114266443A (en) | Data evaluation method and device, electronic equipment and storage medium | |
| CN113298546A (en) | Sales prediction method and device, and commodity processing method and device | |
| Yin et al. | Prediction of Merchandise Sales on E‐Commerce Platforms Based on Data Mining and Deep Learning | |
| CN111429214A (en) | Transaction data-based buyer and seller matching method and device | |
| CN110968670B (en) | Method, device, equipment and storage medium for acquiring attributes of popular commodities | |
| CN115511582B (en) | Commodity recommendation system and method based on artificial intelligence | |
| CN110889716A (en) | Method and device for identifying potential registered user | |
| CN111639274B (en) | Online commodity intelligent sorting method, device, computer equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |