CN115640454A - Product recommendation method, system, electronic device and storage medium - Google Patents
Product recommendation method, system, electronic device and storage medium Download PDFInfo
- Publication number
- CN115640454A CN115640454A CN202211101897.8A CN202211101897A CN115640454A CN 115640454 A CN115640454 A CN 115640454A CN 202211101897 A CN202211101897 A CN 202211101897A CN 115640454 A CN115640454 A CN 115640454A
- Authority
- CN
- China
- Prior art keywords
- user
- target
- product
- recommendation
- information
- 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
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application provides a product recommendation method, a product recommendation system, electronic equipment and a storage medium, which can be applied to the field of big data or the field of finance, and target similar users of target users are determined from various users by utilizing a clustering algorithm according to target label data of the target users; constructing a product portrait of a target user according to target shopping information in the target tag data; constructing a behavior feature matrix of a target user by using the target shopping information and the target transaction behavior information in the target label data; inputting the product portrait of the target user, the behavior feature matrix of the user and the basic information and the transaction behavior information of each target similar user into a hybrid recommendation model, so that the hybrid recommendation model carries out product prediction according to the product portrait of the target user, the behavior feature matrix of the user and the basic information and the transaction behavior information of each similarity user to obtain a target product recommendation set; and recommending corresponding products to the target user according to the target product recommendation set.
Description
Technical Field
The present invention relates to the field of internet technologies, and in particular, to a method and a system for recommending a product, an electronic device, and a storage medium.
Background
With the rapid development of internet technology, products based on scene and individuation are continuously abundant in bank enterprises, and the bank enterprises can recommend various products to users to improve the business performance of the bank enterprises.
In the prior art, a corresponding bank counter is usually set in a bank outlet, and a corresponding product is recommended to a user through the bank counter in a manual manner. However, this method cannot perform customized and personalized product recommendation for the user, and is difficult to attract the user and poor in user experience.
Disclosure of Invention
The invention provides a product recommendation method, a product recommendation system, electronic equipment and a storage medium, which are used for realizing customized and personalized product recommendation for users, attracting the users and improving the user experience.
The invention discloses a product recommendation method in a first aspect, which comprises the following steps:
determining target similar users of the target users from all users according to target label data of the target users by using a clustering algorithm; generating target label data of the target user according to the target user data of the target user; the target tag data at least includes target shopping information and the target transaction behavior information of the target user.
Constructing a product portrait of the target user according to the target shopping information;
constructing a behavior feature matrix of the target user by using the target shopping information and the target transaction behavior information;
inputting the product portrait of the target user, the behavior feature matrix of the user and the basic information and the transaction behavior information of each target similar user into a hybrid recommendation model, so that the hybrid recommendation model carries out product prediction according to the product portrait of the target user, the behavior feature matrix of the user and the basic information and the transaction behavior information of each similarity user to obtain a target product recommendation set; the mixed recommendation model is obtained by training the mixed recommendation model to be trained by using the label data of each user;
and recommending corresponding products to the target user according to the target product recommendation set.
Optionally, the hybrid recommendation model to be trained is constructed according to a collaborative filtering algorithm recommended based on user recommendation and a collaborative filtering algorithm recommended based on content, and the hybrid recommendation model to be trained is trained by using the tag data of each user to obtain a hybrid recommendation model, including:
acquiring label data of each user; wherein the tag data of the user comprises basic information, historical shopping information and historical transaction behavior information of the user;
for each user, determining similar users of the user from the users according to basic information and historical transaction behaviors of the user by using a clustering algorithm;
constructing a product portrait of the user according to the historical shopping information of the user, and constructing a behavior feature matrix of the target user by using the historical shopping information and the historical transaction behavior information of the user;
inputting the product portrait, the behavior characteristic matrix of the user, and the basic information and the historical transaction behavior information of each similar user into a hybrid recommendation model to be trained;
predicting according to the basic information and the historical transaction behavior information of each similar user through the collaborative filtering algorithm based on user recommendation in the hybrid recommendation model to be trained to obtain a first product recommendation set; predicting according to the product portrait and the behavior characteristic matrix of the user through the collaborative filtering algorithm based on the content recommendation in the hybrid recommendation model to be trained to obtain a second product recommendation set, determining product recommendations according to the first product recommendation set and the second product recommendation set, taking the final product recommendation of which the product recommendations approach the user as a training target, and adjusting parameters of the collaborative filtering algorithm based on the user recommendation and the collaborative filtering algorithm based on the content recommendation until the collaborative filtering algorithm based on the user recommendation and the collaborative filtering algorithm based on the content recommendation converge to obtain the hybrid recommendation model.
Optionally, the inputting the product representation of the target user, the behavior feature matrix of the user, and the basic information and the transaction behavior information of each target similar user into a hybrid recommendation model to make the hybrid recommendation model perform product prediction according to the product representation of the target user, the behavior feature matrix of the user, and the basic information and the transaction behavior information of each similarity user, so as to obtain a target product recommendation set, includes:
inputting the product portrait of the target user, the behavior characteristic matrix of the user and the basic information and the transaction behavior information of each target similar user into a mixed recommendation model;
predicting according to the basic information and the transaction behavior information of each target similar user through the collaborative filtering algorithm based on user recommendation in the mixed recommendation model to obtain a target first product recommendation set; and predicting according to the target product portrait of the target user and the target behavior characteristic matrix by the content recommendation-based collaborative filtering algorithm in the hybrid recommendation model to obtain a target second product recommendation set, and determining target product recommendation according to the target first product recommendation set and the target second product recommendation set.
Optionally, the target tag data further includes target basic information of the target user, and determining users with similar purposes of the target user from the users according to the target tag data of the target user by using a clustering algorithm, including:
and respectively calculating the user similarity between the target user and each user according to the target basic information, the target transaction behavior information, the basic information of each user and the transaction behavior information of each user by using a clustering algorithm, and determining the user corresponding to the user similarity larger than a preset similarity threshold value as the target similar user of the target user.
Optionally, the constructing a product representation of the target user according to the target historical shopping information includes:
and constructing a product portrait of the target user according to the historical shopping information through a clustering algorithm in preset software.
The second aspect of the present invention discloses a product recommendation system, which comprises:
the target similar user determining unit is used for determining target similar users of the target users from all the users according to target label data of the target users by utilizing a clustering algorithm; generating target label data of the target user according to the target user data of the target user; the target tag data at least includes target shopping information and the target transaction behavior information of the target user.
The product portrait construction unit is used for constructing a product portrait of the target user according to the target shopping information;
the behavior feature matrix construction unit is used for constructing a behavior feature matrix of the target user by utilizing the target shopping information and the target transaction behavior information;
the prediction unit is used for inputting the product portrait of the target user, the behavior feature matrix of the user and the basic information and the transaction behavior information of each target similar user into a hybrid recommendation model so that the hybrid recommendation model carries out product prediction according to the product portrait of the target user, the behavior feature matrix of the user and the basic information and the transaction behavior information of each similarity user to obtain a target product recommendation set; the mixed recommendation model is obtained by training a mixed recommendation model to be trained by a training unit by using the label data of each user;
and the recommending unit is used for recommending corresponding products to the target user according to the target product recommending set.
Optionally, the hybrid recommendation model to be trained is constructed according to a collaborative filtering algorithm based on user recommendation and a collaborative filtering algorithm based on content recommendation, and the training unit includes:
a tag data acquisition unit configured to acquire tag data of each of the users; wherein the tag data of the user comprises basic information, historical shopping information and historical transaction behavior information of the user;
a similar user determining unit, configured to determine, for each user, a similar user of the user from the users according to basic information and historical transaction behaviors of the user by using a clustering algorithm;
the construction unit is used for constructing a product portrait of the user according to historical shopping information of the user and constructing a behavior characteristic matrix of the target user by using the historical shopping information and the historical transaction behavior information of the user;
the first input unit is used for inputting the product portrait, the behavior characteristic matrix of the user, and the basic information and the historical transaction behavior information of each similar user into a hybrid recommendation model to be trained;
the training subunit is used for predicting according to the basic information and the historical transaction behavior information of each similar user through the collaborative filtering algorithm based on user recommendation in the hybrid recommendation model to be trained to obtain a first product recommendation set; predicting according to the product portrait and the behavior characteristic matrix of the user through the collaborative filtering algorithm based on the content recommendation in the hybrid recommendation model to be trained to obtain a second product recommendation set, determining product recommendations according to the first product recommendation set and the second product recommendation set, taking the final product recommendation of the product recommendation approaching the user as a training target, and adjusting parameters of the collaborative filtering algorithm based on the user recommendation and the collaborative filtering algorithm based on the content recommendation until the collaborative filtering algorithm based on the user recommendation and the collaborative filtering algorithm based on the content recommendation converge to obtain the hybrid recommendation model.
Optionally, the prediction unit includes:
the second input unit is used for inputting the product portrait of the target user, the behavior characteristic matrix of the user and the basic information and the transaction behavior information of each target similar user into a mixed recommendation model;
the forecasting sub-unit is used for forecasting according to the basic information and the transaction behavior information of each target similar user through the collaborative filtering algorithm based on user recommendation in the mixed recommendation model to obtain a target first product recommendation set; and predicting according to the target product portrait of the target user and the target behavior characteristic matrix by the content recommendation-based collaborative filtering algorithm in the hybrid recommendation model to obtain a target second product recommendation set, and determining target product recommendation according to the target first product recommendation set and the target second product recommendation set.
A third aspect of the present invention discloses an electronic apparatus comprising: the system comprises a processor and a memory, wherein the processor and the memory are connected through a communication bus; the processor is used for calling and executing the program stored in the memory; the memory is used for storing a program for implementing the product recommendation method as disclosed in the first aspect of the present invention.
In a fourth aspect of the present invention, a computer-readable storage medium is disclosed, having stored thereon computer-executable instructions for performing the product recommendation method as disclosed in the first aspect of the present invention.
The invention provides a product recommendation method, a product recommendation system, electronic equipment and a storage medium, which can be used for training a hybrid recommendation model to be trained by using label data of each user in advance to obtain the hybrid recommendation model, generating corresponding target label data according to target user data of a target user, and determining a target similar user of the target user from each user according to the target label data of the target user by using a clustering algorithm; the method comprises the steps of constructing a product portrait of a target user according to target shopping information in target label data, constructing a behavior feature matrix of the target user by using the target shopping information and target transaction behaviors, conducting product prediction according to the product portrait of the target user, the behavior feature matrix of the user, basic information of each similarity user and the transaction behavior information through a mixed recommendation model to obtain a target product recommendation set, and recommending corresponding products to the target user according to the target product recommendation set. According to the technical scheme provided by the invention, the label data of each user can be used for training the mixed recommendation model to be trained in advance to obtain the mixed recommendation model, and then the mixed recommendation model can be used for predicting the product according to the product portrait of the target user, the behavior characteristic matrix of the user, and the basic information and the transaction behavior information of each similarity user, so that the product which the target user may be interested in and like is recommended to the target user, and the purposes of attracting the user and improving the user experience are achieved.
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, it is obvious that the drawings in the following description are only embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.
Fig. 1 is a schematic flowchart of a product recommendation method according to an embodiment of the present invention;
fig. 2 is a schematic flowchart of a method for training a hybrid recommendation model to be trained by using label data of each user to obtain the hybrid recommendation model according to an embodiment of the present invention;
FIG. 3 is a schematic structural diagram of a product recommendation system according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
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 only a part of the embodiments of the present invention, and not all of the embodiments. 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.
The term "include" and variations thereof as used herein are open-ended, i.e., "including but not limited to". The term "based on" is "based, at least in part, on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". Relevant definitions for other terms will be given in the following description.
It should be noted that the terms "first", "second", and the like in the present disclosure are only used for distinguishing different devices, modules, or units, and are not used for limiting the order or interdependence of the functions performed by the devices, modules, or units.
It is noted that references to "a", "an", and "the" modifications in the disclosure are exemplary rather than limiting, and that those skilled in the art will understand that "one or more" unless the context clearly dictates otherwise.
It should be noted that the product recommendation method provided by the invention can be used in the technical field of cloud computing or the financial field. The above is only an example and does not limit the application field of the product recommendation provided by the present invention.
Referring to fig. 1, a flowchart of a product recommendation method provided by an embodiment of the present invention is shown, where the product recommendation method specifically includes the following steps:
s101: and determining target similar users of the target users from the users according to the target label data of the target users by using a clustering algorithm.
The target label data of the target user is generated according to the target user data of the target user; the target tag data includes at least target shopping information and target transaction behavior information of the target user.
In the embodiment of the application, after the target user data of the target user can be acquired, the target user data of the target user is preprocessed, and a piece of complete target label data which accords with the service logic is constructed by utilizing the preprocessed target user data. And then determining target similar users of the target users from the users according to the target label data of the target users by using a clustering algorithm.
Optionally, in this embodiment of the application, a clustering algorithm is used, and according to the target basic information of the target user, the target transaction behavior information, the basic information of each user, and the transaction behavior information of each user, user similarities between the target user and each user are respectively calculated, and a user corresponding to the user similarity greater than a preset similarity threshold is determined as a target similar user of the target user.
It should be noted that the clustering algorithm may be a clustering algorithm in R software. Wherein, the R software is a set of complete data processing system.
It should be noted that the basic information of the target user at least includes basic information such as age and sex of the target user; the shopping information of the target user at least comprises information of online purchased books, video members and the like.
S102: and constructing a product portrait of the target user according to the target shopping information.
In the specific execution process of step S102, after the tag data of the target user is obtained, a product portrait of the target user may be constructed according to the target shopping information of the tag data of the target user through a clustering algorithm in preset software.
S103: and constructing a behavior feature matrix of the target user by using the target shopping information and the target transaction behavior information.
In the specific implementation of step S103, each target product held or purchased by the target user may be determined from the target shopping information of the target user, and the target products may be sorted. And then, a clustering algorithm in preset software can be utilized to construct a behavior characteristic matrix of the target user according to the sorted target products and the target transaction behavior information of the target user.
It should be noted that the preset software may be R software.
S104: and inputting the product representation of the target user, the behavior feature matrix of the user and the basic information and the transaction behavior information of all target similar users into a mixed recommendation model, so that the mixed recommendation model carries out product prediction according to the product representation of the target user, the behavior feature matrix of the user and the basic information and the transaction behavior information of all similar users to obtain a target product recommendation set.
In the embodiment of the application, after the hybrid recommendation model to be trained is constructed in advance according to the collaborative filtering algorithm based on user recommendation and the collaborative filtering algorithm based on content recommendation, the hybrid recommendation model to be trained is trained by using the label data of each user to obtain the hybrid recommendation model. And further, the product portrait of the target user, the behavior characteristic matrix of the user and the basic information and the transaction behavior information of each target similar user can be input into the mixed recommendation model.
Predicting according to the basic information and the transaction behavior information of each target similar user through a collaborative filtering algorithm based on user recommendation in a mixed recommendation model to obtain a target first product recommendation set; and predicting according to the target product portrait of the target user and the target behavior characteristic matrix by a collaborative filtering algorithm based on content recommendation in the hybrid recommendation model to obtain a target second product recommendation set, and determining target product recommendation according to the target first product recommendation set and the target second product recommendation set.
It should be noted that, after the hybrid recommendation model predicts a target first product recommendation set and a target second product recommendation set according to the product representation of the target user, the behavior feature matrix of the user, and the basic information and the transaction behavior information of each target similar user, the target first product recommendation set and the target second product recommendation set may be adjusted in a weight adjustment manner, so as to obtain the target product recommendation. Wherein the target product recommendation set includes a plurality of products to be recommended.
Referring to fig. 2, a schematic flow chart of a method for training a hybrid recommendation model to be trained by using label data of each user to obtain the hybrid recommendation model provided in the embodiment of the present invention is shown, where the hybrid recommendation model to be trained is constructed according to a collaborative filtering algorithm based on user recommendation and a collaborative filtering algorithm based on content recommendation, and the method specifically includes the following steps:
s201: acquiring label data of each user; the tag data of the user comprises basic information, historical shopping information and historical transaction behavior information of the user.
In the embodiment of the application, the user data of each user can be obtained, for each user, the user data of the user can be preprocessed, and a piece of complete label data which accords with the service logic is constructed by using the preprocessed user data. Wherein the tag data of the user at least comprises basic information, historical shopping information and historical transaction behavior information of the user.
It should be noted that the basic information of the user at least includes basic information such as the age and sex of the user; the historical shopping information of the user at least comprises information such as online purchasing video members and the like.
S202: and aiming at each user, determining similar users of the users from the users according to the basic information and historical transaction behaviors of the users by using a clustering algorithm.
In the embodiment of the application, after the tag data of each user is obtained, for each user, the user similarity between the user and each other user may be respectively calculated by using a clustering algorithm according to the basic information of the user, the historical transaction behavior information, the basic information of each other user, and the historical transaction behavior information of each other user, and the other users corresponding to the user similarity larger than a preset similarity threshold are determined as similar users of the user.
It should be noted that the clustering algorithm may be a clustering algorithm in R software. Wherein, the R software is a set of complete data processing system.
The embodiment of the present invention is not limited, and the present invention may select a corresponding clustering algorithm according to practical applications.
S203: and constructing a product picture of the user according to the historical shopping information of the user, and constructing a behavior characteristic matrix of the user by using the historical shopping information and the historical transaction behavior information of the user.
In the specific execution process of step S203, after the tag data of each user is obtained, a product portrait of each user may be constructed according to historical shopping information of the tag data of the user through a clustering algorithm in preset software.
Meanwhile, each product held or purchased by the user can be determined from the historical shopping information of the user, and each product is ranked. And then, a clustering algorithm in preset software can be utilized to construct a behavior characteristic matrix of the user according to the sorted products and the historical transaction behavior information of the user.
It should be noted that the preset software may be R software.
S204: and inputting the product portrait of the user, the behavior characteristic matrix, and the basic information and the historical transaction behavior information of each similar user into a hybrid recommendation model to be trained.
S205: predicting according to basic information and historical transaction behavior information of each similar user through a collaborative filtering algorithm based on user recommendation in a hybrid recommendation model to be trained to obtain a first product recommendation set; predicting according to the product portrait and the behavior characteristic matrix of the user by a collaborative filtering algorithm based on content recommendation in a hybrid recommendation model to be trained to obtain a second product recommendation set, determining product recommendation according to the first product recommendation set and the second product recommendation set, taking final product recommendation of which the product recommendation approaches to the user as a training target, adjusting parameters of the collaborative filtering algorithm based on the user recommendation and the collaborative filtering algorithm based on the content recommendation until the collaborative filtering algorithm based on the user recommendation and the collaborative filtering algorithm based on the content recommendation converge, and obtaining the hybrid recommendation model.
In the embodiment of the application, a hybrid recommendation model to be trained can be constructed in advance according to a collaborative filtering algorithm based on user recommendation and a collaborative filtering algorithm based on content recommendation, and for each user, after the similar user of the user, the product portrait of the user and the behavior feature matrix of the user are determined, the product portrait, the behavior feature matrix of the user and the basic information and the historical transaction behavior information of each similar user related to the user can be input into the hybrid recommendation model to be trained.
Predicting according to basic information and historical transaction behavior information of each similar user through a collaborative filtering algorithm based on user recommendation in a hybrid recommendation model to be trained to obtain a first product recommendation set; and predicting according to the product portrait and the behavior characteristic matrix of the user by a collaborative filtering algorithm based on content recommendation in the hybrid recommendation model to be trained to obtain a second product recommendation set.
And performing weighted calculation on the first product recommendation set and the second product recommendation set to obtain product recommendations, constructing corresponding loss functions according to the product recommendations and the final product recommendations of the users, and further respectively adjusting parameters of a collaborative filtering algorithm based on user recommendations and a collaborative filtering algorithm based on content recommendations by using the constructed loss functions until the collaborative filtering algorithm based on user recommendations and the collaborative filtering algorithm based on content recommendations converge to obtain a mixed recommendation model.
S105: and recommending corresponding products to the target user according to the target product recommendation set.
In the specific execution process of step S105, after the target product recommendation set is obtained, each product to be recommended in the target product recommendation set may be recommended to the target user.
The invention provides a product recommendation method, which can be used for training a hybrid recommendation model to be trained by using label data of each user in advance to obtain the hybrid recommendation model, generating corresponding target label data according to target user data of a target user, and determining target similar users of the target user from each user according to the target label data of the target user by using a clustering algorithm; and after a product portrait of a target user is constructed according to target shopping information in the target label data, and a behavior feature matrix of the target user is constructed by using the target shopping information and the target transaction behavior, performing product prediction according to the product portrait of the target user, the behavior feature matrix of the user, and basic information and transaction behavior information of users with various similarities through a mixed recommendation model to obtain a target product recommendation set, and recommending corresponding products to the target user according to the target product recommendation set. According to the technical scheme provided by the invention, the mixed recommendation model to be trained can be trained by utilizing the label data of each user in advance to obtain the mixed recommendation model, and then the product prediction can be carried out through the mixed recommendation model according to the product portrait of the target user, the behavior feature matrix of the user, and the basic information and the transaction behavior information of the users with different similarities, so that the product which the target user may be interested in and like is recommended to the target user, and the purposes of attracting the user and improving the user experience are achieved.
Based on the product recommendation method disclosed in the embodiment of the present invention, the embodiment of the present invention also discloses a product recommendation device, as shown in fig. 3, the product recommendation device includes:
a target similar user determination unit 31, configured to determine a target similar user of a target user from the users according to target tag data of the target user by using a clustering algorithm; the target label data of the target user is generated according to the target user data of the target user; the target tag data includes at least target shopping information and target transaction behavior information of the target user.
A product representation construction unit 32 for constructing a product representation of the target user based on the target shopping information;
a behavior feature matrix construction unit 33, configured to construct a behavior feature matrix of the target user by using the target shopping information and the target transaction behavior information;
the prediction unit 34 is used for inputting the product representation of the target user, the behavior feature matrix of the user and the basic information and the transaction behavior information of each target similar user into the hybrid recommendation model, so that the hybrid recommendation model carries out product prediction according to the product representation of the target user, the behavior feature matrix of the user and the basic information and the transaction behavior information of each similarity user to obtain a target product recommendation set; the mixed recommendation model is obtained by training a mixed recommendation model to be trained by a training unit by using label data of each user;
and the recommending unit 35 is configured to recommend a corresponding product to the target user according to the target product recommendation set.
The specific principle and the execution process of each unit in the product recommendation device disclosed in the embodiment of the present invention are the same as those of the product recommendation method disclosed in the embodiment of the present invention, and reference may be made to corresponding parts in the product recommendation method disclosed in the embodiment of the present invention, which are not described herein again.
The invention provides a product recommendation method, which can be used for training a hybrid recommendation model to be trained by using label data of each user in advance to obtain the hybrid recommendation model, generating corresponding target label data according to target user data of a target user, and determining target similar users of the target user from each user according to the target label data of the target user by using a clustering algorithm; and after a product portrait of a target user is constructed according to target shopping information in the target label data, and a behavior feature matrix of the target user is constructed by using the target shopping information and the target transaction behavior, performing product prediction according to the product portrait of the target user, the behavior feature matrix of the user, and basic information and transaction behavior information of users with various similarities through a mixed recommendation model to obtain a target product recommendation set, and recommending corresponding products to the target user according to the target product recommendation set. According to the technical scheme provided by the invention, the mixed recommendation model to be trained can be trained by utilizing the label data of each user in advance to obtain the mixed recommendation model, and then the product prediction can be carried out through the mixed recommendation model according to the product portrait of the target user, the behavior feature matrix of the user, and the basic information and the transaction behavior information of the users with different similarities, so that the product which the target user may be interested in and like is recommended to the target user, and the purposes of attracting the user and improving the user experience are achieved.
Optionally, the hybrid recommendation model to be trained is constructed according to a collaborative filtering algorithm based on user recommendation and a collaborative filtering algorithm based on content recommendation, and the training unit includes:
a tag data acquisition unit for acquiring tag data of each user; the tag data of the user comprises basic information, historical shopping information and historical transaction behavior information of the user;
the similar user determining unit is used for determining similar users of the users from the users according to the basic information and the historical transaction behaviors of the users by utilizing a clustering algorithm aiming at each user;
the construction unit is used for constructing a product image of the user according to the historical shopping information of the user and constructing a behavior characteristic matrix of the target user by using the historical shopping information and the historical transaction behavior information of the user;
the first input unit is used for inputting a product portrait, a behavior characteristic matrix of a user, basic information of each similar user and historical transaction behavior information into a hybrid recommendation model to be trained;
the training subunit is used for predicting according to the basic information and the historical transaction behavior information of each similar user through a collaborative filtering algorithm based on user recommendation in a hybrid recommendation model to be trained to obtain a first product recommendation set; predicting according to the product portrait and the behavior characteristic matrix of the user by a collaborative filtering algorithm based on content recommendation in a hybrid recommendation model to be trained to obtain a second product recommendation set, determining product recommendation according to the first product recommendation set and the second product recommendation set, taking final product recommendation of which the product recommendation approaches to the user as a training target, adjusting parameters of the collaborative filtering algorithm based on the user recommendation and the collaborative filtering algorithm based on the content recommendation until the collaborative filtering algorithm based on the user recommendation and the collaborative filtering algorithm based on the content recommendation converge, and obtaining the hybrid recommendation model.
Optionally, the prediction unit includes:
the second input unit is used for inputting the product portrait of the target user, the behavior characteristic matrix of the user and the basic information and the transaction behavior information of each target similar user into the mixed recommendation model;
the forecasting sub-unit is used for forecasting according to the basic information and the transaction behavior information of each target similar user through a collaborative filtering algorithm based on user recommendation in the hybrid recommendation model to obtain a target first product recommendation set; and predicting according to the target product portrait of the target user and the target behavior characteristic matrix by a collaborative filtering algorithm based on content recommendation in the hybrid recommendation model to obtain a target second product recommendation set, and determining target product recommendation according to the target first product recommendation set and the target second product recommendation set.
Optionally, the target tag data further includes target basic information of the target user, and the target similar user determining unit includes:
and the target similar user determination subunit is used for respectively calculating the user similarity between the target user and each user according to the target basic information, the target transaction behavior information, the basic information of each user and the transaction behavior information of each user by using a clustering algorithm, and determining the user corresponding to the user similarity larger than a preset similarity threshold value as the target similar user of the target user.
Optionally, the product portrait building unit includes:
and the product portrait constructing subunit is used for constructing the product portrait of the target user according to the historical shopping information through a clustering algorithm in preset software.
An electronic device according to an embodiment of the present application is provided, as shown in fig. 4, the electronic device includes a processor 401 and a memory 402, where the memory 402 is used to store program codes and data for product recommendation, and the processor 401 is used to call program instructions in the memory to execute steps shown in the method for implementing product recommendation in the foregoing embodiment.
The embodiment of the application provides a storage medium, the storage medium comprises a storage program, and when the program runs, a device where the storage medium is located is controlled to execute the product recommendation method shown in the embodiment.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, the system or system embodiments are substantially similar to the method embodiments and therefore are described in a relatively simple manner, and reference may be made to some of the descriptions of the method embodiments for related points. The above-described system and system embodiments are merely illustrative, wherein units described as separate components may or may not be physically separate, and components shown 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.
Those of skill would further appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that the various illustrative components and steps have been described above generally in terms of their functionality in order to clearly illustrate this interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
The foregoing is only a preferred embodiment of the present invention, and it should be noted that it is obvious to those skilled in the art that various modifications and improvements can be made without departing from the principle of the present invention, and these modifications and improvements should also be considered as the protection scope of the present invention.
Claims (10)
1. A method for recommending products, the method comprising:
determining target similar users of the target users from all users according to target label data of the target users by using a clustering algorithm; generating target label data of the target user according to the target user data of the target user; the target tag data at least comprises target shopping information and target transaction behavior information of the target user;
constructing a product portrait of the target user according to the target shopping information;
constructing a behavior feature matrix of the target user by using the target shopping information and the target transaction behavior information;
inputting the product portrait of the target user, the behavior feature matrix of the user and the basic information and the transaction behavior information of each target similar user into a hybrid recommendation model, so that the hybrid recommendation model carries out product prediction according to the product portrait of the target user, the behavior feature matrix of the user and the basic information and the transaction behavior information of each similarity user to obtain a target product recommendation set; the mixed recommendation model is obtained by training the mixed recommendation model to be trained by using the label data of each user;
and recommending corresponding products to the target user according to the target product recommendation set.
2. The method according to claim 1, wherein the hybrid recommendation model to be trained is constructed according to a collaborative filtering algorithm based on user recommendation and a collaborative filtering algorithm based on content recommendation, and the training of the hybrid recommendation model to be trained by using the label data of each of the users to obtain the hybrid recommendation model comprises:
acquiring label data of each user; wherein the tag data of the user comprises basic information, historical shopping information and historical transaction behavior information of the user;
for each user, determining similar users of the user from the users according to basic information and historical transaction behaviors of the user by using a clustering algorithm;
constructing a product portrait of the user according to the historical shopping information of the user, and constructing a behavior feature matrix of the target user by using the historical shopping information and the historical transaction behavior information of the user;
inputting the product portrait, the behavior characteristic matrix of the user, and the basic information and the historical transaction behavior information of each similar user into a hybrid recommendation model to be trained;
predicting according to the basic information and the historical transaction behavior information of each similar user through the collaborative filtering algorithm based on user recommendation in the hybrid recommendation model to be trained to obtain a first product recommendation set; predicting according to the product portrait and the behavior characteristic matrix of the user through the collaborative filtering algorithm based on the content recommendation in the hybrid recommendation model to be trained to obtain a second product recommendation set, determining product recommendations according to the first product recommendation set and the second product recommendation set, taking the final product recommendation of which the product recommendations approach the user as a training target, and adjusting parameters of the collaborative filtering algorithm based on the user recommendation and the collaborative filtering algorithm based on the content recommendation until the collaborative filtering algorithm based on the user recommendation and the collaborative filtering algorithm based on the content recommendation converge to obtain the hybrid recommendation model.
3. The method of claim 2, wherein the step of inputting the product representation of the target user, the behavior feature matrix of the user, and the basic information and the transaction behavior information of each target similar user into a hybrid recommendation model, so that the hybrid recommendation model performs product prediction according to the product representation of the target user, the behavior feature matrix of the user, and the basic information and the transaction behavior information of each similar user to obtain a target product recommendation set comprises:
inputting the product portrait of the target user, the behavior characteristic matrix of the user and the basic information and the transaction behavior information of each target similar user into a mixed recommendation model;
predicting according to the basic information and the transaction behavior information of each target similar user through the collaborative filtering algorithm based on user recommendation in the mixed recommendation model to obtain a target first product recommendation set; and predicting according to the target product portrait of the target user and the target behavior characteristic matrix by the content recommendation-based collaborative filtering algorithm in the hybrid recommendation model to obtain a target second product recommendation set, and determining target product recommendation according to the target first product recommendation set and the target second product recommendation set.
4. The method of claim 1, wherein the target tag data further includes target basic information of the target user, and determining the target user's similar users from the users according to the target tag data of the target user by using a clustering algorithm comprises:
and respectively calculating the user similarity between the target user and each user according to the target basic information, the target transaction behavior information, the basic information of each user and the transaction behavior information of each user by using a clustering algorithm, and determining the user corresponding to the user similarity larger than a preset similarity threshold value as the target similar user of the target user.
5. The method of claim 1, wherein said constructing a product representation of said target user from said target historical shopping information comprises:
and constructing a product portrait of the target user according to the historical shopping information through a clustering algorithm in preset software.
6. A product recommendation system, the system comprising:
the target similar user determining unit is used for determining target similar users of the target users from all users according to target label data of the target users by using a clustering algorithm; generating target label data of the target user according to the target user data of the target user; the target tag data at least comprises target shopping information and target transaction behavior information of the target user;
the product portrait construction unit is used for constructing a product portrait of the target user according to the target shopping information;
the behavior feature matrix construction unit is used for constructing a behavior feature matrix of the target user by utilizing the target shopping information and the target transaction behavior information;
the prediction unit is used for inputting the product portrait of the target user, the behavior feature matrix of the user and the basic information and the transaction behavior information of each target similar user into a hybrid recommendation model so that the hybrid recommendation model carries out product prediction according to the product portrait of the target user, the behavior feature matrix of the user and the basic information and the transaction behavior information of each similarity user to obtain a target product recommendation set; the mixed recommendation model is obtained by training a mixed recommendation model to be trained by a training unit by using the label data of each user;
and the recommending unit is used for recommending corresponding products to the target user according to the target product recommending set.
7. The system according to claim 6, wherein the hybrid recommendation model to be trained is constructed according to a collaborative filtering algorithm based on user recommendations and a collaborative filtering algorithm based on content recommendations, and the training unit comprises:
a tag data acquisition unit configured to acquire tag data of each of the users; wherein the tag data of the user comprises basic information, historical shopping information and historical transaction behavior information of the user;
a similar user determining unit, configured to determine, for each user, a similar user of the user from the users according to basic information and historical transaction behaviors of the user by using a clustering algorithm;
the construction unit is used for constructing a product portrait of the user according to the historical shopping information of the user and constructing a behavior characteristic matrix of the target user by using the historical shopping information and the historical transaction behavior information of the user;
the first input unit is used for inputting the product portrait, the behavior characteristic matrix of the user, and the basic information and the historical transaction behavior information of each similar user into a hybrid recommendation model to be trained;
the training subunit is used for predicting according to the basic information and the historical transaction behavior information of each similar user through the collaborative filtering algorithm based on user recommendation in the hybrid recommendation model to be trained to obtain a first product recommendation set; predicting according to the product portrait and the behavior characteristic matrix of the user through the collaborative filtering algorithm based on the content recommendation in the hybrid recommendation model to be trained to obtain a second product recommendation set, determining product recommendations according to the first product recommendation set and the second product recommendation set, taking the final product recommendation of the product recommendation approaching the user as a training target, and adjusting parameters of the collaborative filtering algorithm based on the user recommendation and the collaborative filtering algorithm based on the content recommendation until the collaborative filtering algorithm based on the user recommendation and the collaborative filtering algorithm based on the content recommendation converge to obtain the hybrid recommendation model.
8. The system of claim 7, wherein the prediction unit comprises:
the second input unit is used for inputting the product portrait of the target user, the behavior characteristic matrix of the user and the basic information and the transaction behavior information of each target similar user into a mixed recommendation model;
the forecasting sub-unit is used for forecasting according to the basic information and the transaction behavior information of each target similar user through the collaborative filtering algorithm based on user recommendation in the mixed recommendation model to obtain a target first product recommendation set; and predicting according to the target product portrait of the target user and the target behavior characteristic matrix by the content recommendation-based collaborative filtering algorithm in the hybrid recommendation model to obtain a target second product recommendation set, and determining target product recommendation according to the target first product recommendation set and the target second product recommendation set.
9. An electronic device, comprising: the system comprises a processor and a memory, wherein the processor and the memory are connected through a communication bus; the processor is used for calling and executing the program stored in the memory; the memory for storing a program for implementing the product recommendation method according to any one of claims 1 to 5.
10. A computer-readable storage medium having computer-executable instructions stored thereon for performing the product recommendation method of any one of claims 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211101897.8A CN115640454A (en) | 2022-09-09 | 2022-09-09 | Product recommendation method, system, electronic device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211101897.8A CN115640454A (en) | 2022-09-09 | 2022-09-09 | Product recommendation method, system, electronic device and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115640454A true CN115640454A (en) | 2023-01-24 |
Family
ID=84941519
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211101897.8A Pending CN115640454A (en) | 2022-09-09 | 2022-09-09 | Product recommendation method, system, electronic device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115640454A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116523546A (en) * | 2023-06-29 | 2023-08-01 | 深圳市华图测控系统有限公司 | Method and device for intelligent reader behavior analysis and prediction system data acquisition and analysis |
-
2022
- 2022-09-09 CN CN202211101897.8A patent/CN115640454A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116523546A (en) * | 2023-06-29 | 2023-08-01 | 深圳市华图测控系统有限公司 | Method and device for intelligent reader behavior analysis and prediction system data acquisition and analysis |
| CN116523546B (en) * | 2023-06-29 | 2023-12-19 | 深圳市华图测控系统有限公司 | Method and device for intelligent reader behavior analysis and prediction system data acquisition and analysis |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106651542B (en) | Article recommendation method and device | |
| US11004123B2 (en) | Method for trading goods on a computer network based on image processing for interest, emotion and affinity detection | |
| EP4181026A1 (en) | Recommendation model training method and apparatus, recommendation method and apparatus, and computer-readable medium | |
| CN111506801A (en) | Sequencing method and device for sub-applications in application App | |
| CN110851699A (en) | Deep reinforcement learning-based information flow recommendation method, device, equipment and medium | |
| CN114036398B (en) | Content recommendation and ranking model training method, device, equipment and storage medium | |
| US20220261591A1 (en) | Data processing method and apparatus | |
| CN110598120A (en) | Behavior data based financing recommendation method, device and equipment | |
| CN110020022A (en) | Data processing method, device, equipment and readable storage medium storing program for executing | |
| CN112528164A (en) | User collaborative filtering recall method and device | |
| CN115640454A (en) | Product recommendation method, system, electronic device and storage medium | |
| CN113592593B (en) | Training and application method, device, equipment and storage medium of sequence recommendation model | |
| CN114428910A (en) | Resource recommendation method and device, electronic equipment, product and medium | |
| CN113791909A (en) | Server capacity adjusting method and device, computer equipment and storage medium | |
| CN112269943B (en) | Information recommendation system and method | |
| CN111680213A (en) | Information recommendation method, data processing method and device | |
| CN112381236A (en) | Data processing method, device, equipment and storage medium for federal transfer learning | |
| CN114677168A (en) | Resource recommendation method, device, equipment and medium | |
| CN115185606A (en) | Method, device, equipment and storage medium for obtaining service configuration parameters | |
| CN114723568A (en) | Recommendation method, device, server, medium and product for financial product | |
| CN113761352A (en) | Information pushing method, device, equipment and storage medium | |
| CN114169906A (en) | Electronic ticket pushing method and device | |
| CN114722266A (en) | Questionnaire pushing method and device, electronic equipment and readable storage medium | |
| CN112269942A (en) | Method, device and system for recommending object and electronic equipment | |
| CN112015978A (en) | Customized information sending method and device and electronic equipment |
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 |