CN113256368B - Product pushing method and device, computer equipment and storage medium - Google Patents
Product pushing method and device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN113256368B CN113256368B CN202110486268.0A CN202110486268A CN113256368B CN 113256368 B CN113256368 B CN 113256368B CN 202110486268 A CN202110486268 A CN 202110486268A CN 113256368 B CN113256368 B CN 113256368B
- Authority
- CN
- China
- Prior art keywords
- product
- user
- characteristic information
- training
- sample
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
-
- 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
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- 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
- G06F16/958—Organisation or management of web site content, e.g. publishing, maintaining pages or automatic linking
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/50—Network services
- H04L67/55—Push-based network services
Abstract
The application relates to a product pushing method, a product pushing device, computer equipment and a storage medium. The method comprises the following steps: acquiring user characteristic information; inputting the user characteristic information into a trained product pushing model to obtain the weight scores of all product sets corresponding to the user characteristic information, wherein the product pushing model is obtained by training based on the user sample characteristic information and the corresponding product sets; then obtaining products to be pushed corresponding to the user characteristic information according to the weight scores of the product sets; and finally outputting the product to be pushed. By adopting the method, accurate product pushing can be performed according to different types of users.
Description
Technical Field
The present application relates to the field of computer application technologies, and in particular, to a product pushing method and apparatus, a computer device, and a storage medium.
Background
With the development of computer technology, various service providers push various products or services to users through computer networks. The traditional product or service pushing method mainly pushes the same product or service for all users, taking the asset configuration service as an example, and the traditional pushing method pushes all product services to all users.
In the traditional technology, the product pushing is not accurate and cannot meet the requirements of users.
Disclosure of Invention
In view of the above, it is necessary to provide a product pushing method, an apparatus, a computer device and a storage medium capable of pushing products according to different types of users.
A product pushing method comprises the following steps:
acquiring user characteristic information;
inputting the user characteristic information into the trained product push model to obtain the weight scores of all product sets corresponding to the user characteristic information; the product pushing model is obtained based on user sample characteristic information and corresponding product set training;
obtaining products to be pushed corresponding to the user characteristic information according to the weight scores of the product sets;
and outputting the product to be pushed.
In one embodiment, the obtaining of the user characteristic information includes:
acquiring personal attribute characteristics of a user;
acquiring interactive characteristics of a user product;
and integrating the personal attribute characteristics of the user and the interactive characteristics of the user product to obtain the user characteristic information.
In one embodiment, the training mode of the product push model includes:
obtaining a plurality of product sets according to the product characteristics of the products;
acquiring training sets according to a plurality of product sets and user sample characteristic information, wherein each training sample in the training sets comprises user sample characteristic information and a corresponding product set, and each product set corresponds to a plurality of user sample characteristic information;
and inputting each training sample in the training set into an initial deep neural network model for learning training to obtain a trained product pushing model.
In one embodiment, obtaining a training set according to a plurality of product sets and user sample feature information includes:
acquiring user sample characteristic information according to a user sample, wherein the user sample characteristic information comprises user personal attribute characteristics and user product interaction characteristics, and the user product interaction characteristics are used for expressing the mapping relation between the user personal attribute characteristics and products;
determining a product set corresponding to the user sample characteristic information according to the product set and the user product interaction characteristics; integrating the characteristic information of the user sample and a corresponding product set to obtain a training sample;
and acquiring a training set according to the plurality of training samples.
In one embodiment, obtaining the product to be pushed corresponding to the user feature information according to the weight score of each product set includes:
selecting a product set with the weight score larger than a weight threshold value as a product set to be pushed corresponding to the user characteristic information, and obtaining a product to be pushed according to the product set to be pushed;
or selecting a preset number of product sets from high to low according to the weight scores to serve as the product sets to be pushed corresponding to the user characteristic information, and obtaining products to be pushed according to the product sets to be pushed.
A product pusher apparatus, the apparatus comprising:
the characteristic acquisition module is used for acquiring user characteristic information;
the weight processing module is used for inputting the user characteristic information into the trained product pushing model to obtain the weight scores of all product sets corresponding to the user characteristic information; the product pushing model is obtained based on user sample characteristic information and corresponding product set training;
the product determining module is used for obtaining products to be pushed corresponding to the user characteristic information according to the weight scores of the product sets;
and the output module is used for outputting the product to be pushed.
In one embodiment, an apparatus comprises:
the product set acquisition module is used for acquiring a plurality of product sets according to the product characteristics of the products;
the training set acquisition module is used for acquiring a training set according to a plurality of product sets and user sample characteristic information, each training sample in the training set comprises user sample characteristic information and a corresponding product set, and each product set corresponds to a plurality of user sample characteristic information;
and the training module is used for inputting each training sample in the training set into the initial deep neural network model for learning training to obtain a trained product pushing model.
In one embodiment, the training set acquisition module comprises:
the user sample characteristic obtaining sub-module is used for obtaining user sample characteristic information according to a user sample, the user sample characteristic information comprises user personal attribute characteristics and user product interaction characteristics, and the user product interaction characteristics are used for expressing the mapping relation between the user personal attribute characteristics and products;
the training sample acquisition submodule is used for determining a product set corresponding to the user sample characteristic information according to the product set and the user product interaction characteristics; integrating the characteristic information of the user sample and a corresponding product set to obtain a training sample;
and the training sample integration submodule is used for acquiring a training set according to the plurality of training samples.
A computer device comprising a memory and a processor, the memory storing a computer program, the processor implementing the following steps when executing the computer program:
acquiring user characteristic information;
inputting the user characteristic information into the trained product push model to obtain the weight scores of all product sets corresponding to the user characteristic information; the product pushing model is obtained based on user sample characteristic information and corresponding product set training;
obtaining products to be pushed corresponding to the user characteristic information according to the weight scores of the product sets;
and outputting the product to be pushed.
A computer-readable storage medium, on which a computer program is stored which, when executed by a processor, carries out the steps of:
acquiring user characteristic information;
inputting the user characteristic information into the trained product push model to obtain the weight scores of all product sets corresponding to the user characteristic information; the product pushing model is obtained based on user sample characteristic information and corresponding product set training;
obtaining products to be pushed corresponding to the user characteristic information according to the weight scores of the product sets;
and outputting the product to be pushed.
The product pushing method, the product pushing device, the computer equipment and the storage medium acquire the user characteristic information; inputting the user characteristic information into a trained product pushing model to obtain the weight scores of all product sets corresponding to the user characteristic information, wherein the product pushing model is obtained by training based on the user sample characteristic information and the corresponding product sets; then obtaining products to be pushed corresponding to the user characteristic information according to the weight scores of the product sets; and finally outputting the product to be pushed. The product pushing model is obtained through training of the user sample characteristic information and the corresponding product set, the user characteristic information is input into the product pushing model, the product pushing model can output the product set with the highest degree of fit with the user characteristic information, and then the product to be pushed is output, so that the purpose of accurately pushing the product according to different types of users is achieved.
Drawings
FIG. 1 is a flow chart illustrating a product pushing method according to an embodiment;
FIG. 2 is a schematic diagram illustrating a process for obtaining user profile information according to an embodiment;
FIG. 3 is a schematic flow chart illustrating a training mode of the product push model according to an embodiment;
FIG. 4 is a schematic diagram of a process for obtaining a training set in one embodiment;
fig. 5 is a diagram showing a product pushing model structure of a product pushing method according to an embodiment;
FIG. 6 is a block diagram showing the structure of a product pusher according to an embodiment;
FIG. 7 is a diagram illustrating an internal structure of a computer device according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
In an embodiment, as shown in fig. 1, a product pushing method is provided, and this embodiment is illustrated by applying the method to a terminal, and it is to be understood that the method may also be applied to a server, and may also be applied to a system including the terminal and the server, and is implemented by interaction between the terminal and the server. In this embodiment, the method includes the steps of:
and 102, acquiring user characteristic information.
Wherein the user characteristic information comprises user personal attribute characteristics and user product interaction characteristics. The user personal attribute features are used for representing the attribute characteristics of the user. The user product interaction feature is used for representing the mapping relation between the user personal attribute feature and the product.
The user characteristic information is used for distinguishing different users, the user characteristic information is the key for knowing the user appeal points, and the processor can obtain the user appeal points more deeply by analyzing the user characteristic information. Specifically, the processor obtains the user characteristic information through the user database.
Step 104, inputting the user characteristic information into the trained product push model to obtain the weight scores of each product set corresponding to the user characteristic information; the product pushing model is obtained based on user sample characteristic information and corresponding product set training.
Specifically, the processor obtains a group of user sample feature information and a corresponding product set in advance, obtains a training set according to the user sample feature information and the corresponding product set, inputs the training set into an initial deep neural network model for model training, obtains a product pushing model after the training is completed, inputs the user feature information obtained in step 102 into the product pushing model, and the product pushing model outputs weight scores of the product sets corresponding to the user feature information. The weight score is a value reflecting the degree of engagement between the product set and the input user characteristic information, and in general, the higher the weight score of a product set is, the higher the degree of engagement between the product set and the input user characteristic information is, that is, the higher the degree of engagement between the product set and the user is.
And step 106, obtaining products to be pushed corresponding to the user characteristic information according to the weight scores of the product sets.
Wherein each product set contains a plurality of products with the same product characteristics.
Specifically, the processor obtains one or more product sets by sorting the weight scores or comparing the weight thresholds according to the weight scores of the product sets, and then obtains the products to be pushed from the obtained product sets.
And step 108, outputting the product to be pushed.
Specifically, the processor outputs information corresponding to the product to be pushed through the output device. The output device is a terminal device of a computer hardware system, and is used for receiving output display, printing, sound of computer data, controlling peripheral device operation and the like, and simultaneously expressing various calculation result data or information in the form of numbers, characters, images, sound and the like. Common output devices include displays, printers, plotters, video output systems, voice output systems, magnetic recording devices, and the like.
In the product pushing method, the user characteristic information is acquired; inputting the user characteristic information into a trained product pushing model to obtain the weight scores of all product sets corresponding to the user characteristic information, wherein the product pushing model is obtained by training based on the user sample characteristic information and the corresponding product sets; then obtaining products to be pushed corresponding to the user characteristic information according to the weight scores of the product sets; and finally outputting the product to be pushed. The product pushing model is obtained through training of user sample characteristic information and corresponding product sets, the user characteristic information is input into the product pushing model, the product pushing model can output the product set with the highest degree of fit with the user characteristic information, and then the product to be pushed is output, so that the purpose of accurately pushing the product according to different types of users is achieved.
In one embodiment, as shown in fig. 2, the obtaining of the user characteristic information includes:
step 202, obtaining personal attribute characteristics of the user.
Specifically, the processor obtains user personal attribute characteristics, and the user personal attribute characteristics are used for representing the attribute characteristics of the user individuals. For example, when a financial product is pushed, the user personal attribute characteristics may include user basic attribute information (gender, age, occupation, etc.), financial attribute information of the user (user level, risk preference, credit rating, investment preference), user transaction information, user RFM (Recency, Frequency, money) information (interval of the user's last transaction time, number of times the user has transacted in the last period of time, amount of money the user has transacted in the last period of time), user online behavior information, user credit card information, user loan information, and the like.
Further, for example, in the case of multimedia push, the user personal attribute characteristics may include the age, gender, occupation, content preference, registration time, platform member level, and the like of the user. The multimedia may include files such as pictures, audio and video.
And step 204, acquiring the interactive characteristics of the user product.
Specifically, the processor obtains user product interaction characteristics, and the user product interaction characteristics are used for representing the mapping relation between the user personal attribute characteristics and the products. For example, when financial products are pushed, the user product interactive features may include deposit products historically purchased by the user, financial products historically purchased by the user, fund products historically purchased by the user, insurance products historically purchased by the user, and commodity products historically purchased by the user; the deposit products purchased by the user in the last 1 year, the financing products purchased by the user in the last 1 year, the fund products purchased by the user in the last 1 year, the insurance products purchased by the user in the last 1 year, the commodity products purchased by the user in the last 1 year, etc.
Further, for example, in the case of multimedia push, the user product interactive features may include multimedia that is historically played by the user, the type of multimedia that is historically played by the user, multimedia that is historically cached by the user, multimedia that is historically purchased by the user, multimedia members or other paid services that are historically purchased by the user, multimedia that is historically reviewed by the user, and so on.
And step 206, integrating the personal attribute characteristics of the user and the interactive characteristics of the user product to obtain user characteristic information.
Specifically, according to the actual situation, relevant meaningless product information in the user product interaction features is removed, and then the user personal attribute features and the user product interaction features are combined to obtain user feature information.
In one embodiment, as shown in fig. 3, the training mode of the product push model includes:
step 302, a plurality of product sets are obtained according to the product characteristics of the products.
Specifically, according to the actual situation of the product, one or more dimensions of product features are selected, all the products are subjected to product feature division according to each dimension, then the product features of the multiple dimensions are arranged and combined to obtain a combination of multiple groups of product features, and the products with the same combination of the product features are integrated into a product set. For example, when a financing product is pushed, product features can be divided from three dimensions of product category, product risk and product fluidity: the product categories comprise 5 characteristic categories of financing, deposit, fund, insurance and commodity; the product risk has several levels of R1-R5, which are respectively cautious type, steady type, balanced type, access type and aggressive type, and can be set to 3 characteristic categories of low risk, medium risk and high risk in a simplified way; the product fluidity has intervals of less than 14 days, 14 to 35 days, 35 to 96 days, 96 to 180 days, 180 to 370 days, more than 370 days and the like, and can be set as 3 characteristic categories of strong fluidity, medium fluidity and poor fluidity in a simplified way; the product features of the three dimensions are arranged and combined under the condition of meeting the practical situation to obtain the combination of at most 45 product features, for example: [ insurance-low risk-poor fluidity ], [ deposit-low risk-poor fluidity ], [ fund-low risk-strong fluidity ], [ fund-medium risk-poor fluidity ], [ financing-medium risk-poor fluidity ], [ commodity-high risk-strong fluidity ], and the like. Since some combinations of product features, for example [ deposit-high risk-poor flowability ] and [ insurance-high risk-good flowability ] do not match reality, these combinations of product features that do not match reality are rejected. And finally, constructing a product set frame according to the combination of each reserved product characteristic, and respectively putting all products into the product set frame according to the combination of the product characteristics to obtain a product set.
Step 304, obtaining training sets according to the multiple product sets and the user sample feature information, wherein each training sample in the training sets comprises the user sample feature information and a corresponding product set, and each product set corresponds to the multiple user sample feature information.
The user sample characteristic information comprises a user personal attribute characteristic and a user product interaction characteristic, and the user product interaction characteristic is used for representing the mapping relation between the user personal attribute characteristic and a product. The user sample characteristic information is used for distinguishing different user samples, the user sample characteristic information is a key for knowing the appeal points of the user samples, and the deep neural network model can obtain the appeal points of the user samples more deeply by analyzing the user sample characteristic information.
Specifically, user sample characteristic information is obtained according to a user sample; determining a product set corresponding to the user sample characteristic information according to the product set and the user product interaction characteristics; integrating the characteristic information of the user sample and a corresponding product set to obtain a training sample; and acquiring a training set according to the plurality of training samples.
And step 306, inputting each training sample in the training set into the initial deep neural network model for learning training to obtain a trained product pushing model.
Specifically, each training sample in the training set is input into an initial deep neural network model, and each training sample trains the deep neural network model at least once. For example, the initial deep neural network model may select a wide & deep model, and when the model is trained, for each training sample, user sample feature information (user personal attribute features and user product interaction features) is input to the deep side, the user product interaction features in the user sample feature information are independently input to the wide side, the softmax layer of the model outputs a set of weight scores of a product set pushed for the user sample, and parameters of the model are adjusted according to the set of weight scores and the product set corresponding to the user sample feature information, so as to perform one-time training. And after all training samples in the set to be trained train the model for a preset number of times, finishing the model training to obtain a trained product pushing model.
In this embodiment, a plurality of product sets are obtained according to product features of a product; acquiring training sets according to a plurality of product sets and user sample characteristic information, wherein each training sample in the training sets comprises user sample characteristic information and a corresponding product set, and each product set corresponds to a plurality of user sample characteristic information; and inputting each training sample in the training set into an initial deep neural network model for learning training to obtain a trained product pushing model. By inputting the user characteristic information into the product pushing model, the purpose of accurately pushing products according to different types of users can be achieved.
In one embodiment, as shown in fig. 4, the obtaining of the training set according to the plurality of product sets and the user sample feature information includes:
step 402, obtaining user sample characteristic information according to a user sample, wherein the user sample characteristic information comprises user personal attribute characteristics and user product interaction characteristics, and the user product interaction characteristics are used for representing the mapping relation between the user personal attribute characteristics and products.
Specifically, different user samples are selected according to different conditions of the product set, and then the user personal attribute characteristics and the user product interaction characteristics of the user samples are obtained. The user product interaction characteristics represent the mapping relation between the user personal attribute characteristics of the current user sample and the product, and are equivalent to representing the mapping relation between the current user sample and the product. The user sample is essentially the same as the user, and the user sample is a part of the users selected for training the model, so that the user product interaction feature can also be said to represent the mapping relationship between the user personal attribute feature of the current user and the product, and simultaneously represent the mapping relationship between the current user and the product.
For example, when a financial product is pushed, users with rich investment and financial experience or unique investment preference are selected as user samples. Specifically, users with continuous investment experience of 1 year, 2 years and 3 years can be respectively extracted and ranked according to the average profitability, wherein the continuous investment experience can be understood that at least 80% of 1 year has investment behaviors, and the specific proportion can be adjusted according to the data analysis condition; and then respectively extracting a part with higher yield from the users with 1 year, 2 years and 3 years of continuous investment experience as user samples. The specific gravity after extraction can be carried out according to 25%, 35% and 40%, and the specific proportion can be adjusted according to the data analysis result. The selection according to 1 year, 2 years and 3 years is that the experience of the old investment user and the experience of the new investment user are considered. And finally, analyzing the investment behaviors of the user sample in the last 1 year, 2 years and 3 years to obtain the user personal attribute characteristics and the user product interaction characteristics of the user sample.
Step 404, determining a product set corresponding to the user sample feature information according to the product set and the user product interaction features; and integrating the characteristic information of the user sample and the corresponding product set to obtain a training sample.
Specifically, a product set corresponding to the user sample feature information of the current user sample is determined according to the product set and the user product interaction features of the current user sample, and a training sample can be obtained according to the current user sample by integrating the user sample feature information of the current user sample and the corresponding product set. In a training sample, the user sample characteristic information is X, the product set corresponding to the user sample characteristic information is Y, and the X and the Y jointly form the training sample. The user product interaction feature can reflect a product set preferred by the user sample. According to actual requirements, selecting partial characteristics in the user product interaction characteristics of the current user sample, and then determining a preference product set of the current user sample according to the partial characteristics, namely a product set corresponding to the user sample characteristic information of the current user sample. The selection of part of the characteristics can be selected according to dimensions such as time, proportion and the like. For example, most of the funds of the user sample a are invested in the product set of [ financing, risk-liquidity difference ] in the last 1 year, and a higher yield is obtained, so that the user sample a corresponds to the product set of [ financing, risk-liquidity difference ], and the user sample characteristic information of the user sample a and the product set of [ financing, risk-liquidity difference ] form a training sample. If the user sample prefers two product sets, two training samples corresponding to the two product sets respectively are obtained.
Step 406, obtaining a training set according to the plurality of training samples.
Specifically, at least one training sample is generated for each user sample, and all the training samples are integrated to obtain a training set.
In the embodiment, user sample characteristic information is obtained according to a user sample, wherein the user sample characteristic information comprises user personal attribute characteristics and user product interaction characteristics, and the user product interaction characteristics are used for representing the mapping relation between the user personal attribute characteristics and products; determining a product set corresponding to the user sample characteristic information according to the product set and the user product interaction characteristics; integrating the characteristic information of the user sample and a corresponding product set to obtain a training sample; and acquiring a training set according to the plurality of training samples. The deep neural network model is trained through the training set, and the obtained product pushing model can achieve the purpose of accurately pushing products according to different types of users.
In one embodiment, obtaining the product to be pushed corresponding to the user feature information according to the weight score of each product set includes: selecting a product set with the weight score larger than a weight threshold value as a product set to be pushed corresponding to the user characteristic information, and obtaining a product to be pushed according to the product set to be pushed; or selecting a preset number of product sets from high to low according to the weight scores to serve as the product sets to be pushed corresponding to the user characteristic information, and obtaining products to be pushed according to the product sets to be pushed.
Specifically, there are two general methods for obtaining a product to be pushed corresponding to user feature information according to the weight score of each product set, which are as follows: the first method is that a weight threshold value is preset, after the weight scores of all product sets are output, the processor selects the product sets with the weight scores larger than the weight threshold value, and then the products in the product sets are extracted as the products to be pushed; the second method is to preset a ranking threshold, for example, 3, after the weight scores of all product sets are output, the processor sorts the product sets from high to low according to the weight scores, then selects the product set ranked before the ranking threshold, which is the product set with the weight score of 3 above, and then extracts the products in the product sets as the products to be pushed.
In one embodiment, a product pushing method, as shown in fig. 5, is applied to the product pushing model in fig. 5. The method comprises the following steps: acquiring user personal attribute characteristics, inputting the user personal attribute characteristics to a coordinated LEDs layer of a product pushing model, acquiring user product interaction characteristics, inputting the user product interaction characteristics to the LEDs layer for characteristic extraction, then inputting the extracted user product interaction characteristics to the coordinated LEDs layer, and integrating the user personal attribute characteristics and the extracted user product interaction characteristics by the coordinated LEDs layer to obtain user characteristic information; inputting user characteristic information into a trained wide & deep product pushing model in the above mode, namely inputting user characteristic information (user personal attribute characteristics and user product interaction characteristics) into a deep side, independently inputting user product interaction characteristics in the user characteristic information into the wide side, outputting weight scores of each product set corresponding to the user characteristic information by a softmax layer, and obtaining the wide & deep product pushing model by training the wide & deep initial model based on the user sample characteristic information and the corresponding product set; selecting a product set with the weight score larger than a weight threshold value as a product set to be pushed corresponding to the user characteristic information, and obtaining a product to be pushed according to the product set to be pushed; or selecting a preset number of product sets from high to low according to the weight scores to serve as the product sets to be pushed corresponding to the user characteristic information, and obtaining products to be pushed according to the product sets to be pushed; and finally outputting the product to be pushed.
It should be understood that although the various steps in the flow charts of fig. 1-4 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 1-4 may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, which are not necessarily performed in sequence, but may be performed in turn or alternately with other steps or at least some of the other steps.
In one embodiment, as shown in fig. 5, there is provided a product pusher 600 comprising: a feature obtaining module 601, a weight processing module 602, a product determining module 603, and an output module 604, wherein:
the feature obtaining module 601 is configured to obtain user feature information. The method specifically comprises the following steps: acquiring personal attribute characteristics of a user; acquiring interactive characteristics of a user product; and integrating the personal attribute characteristics of the user and the interactive characteristics of the user product to obtain the user characteristic information.
The weight processing module 602 is configured to input the user feature information into the trained product push model to obtain a weight score of each product set corresponding to the user feature information; the product pushing model is obtained based on user sample characteristic information and corresponding product set training.
The product determining module 603 is configured to obtain a product to be pushed corresponding to the user feature information according to the weight score of each product set. The method specifically comprises the following steps: selecting a product set with the weight score larger than a weight threshold value as a product set to be pushed corresponding to the user characteristic information, and obtaining a product to be pushed according to the product set to be pushed; or selecting a preset number of product sets from high to low according to the weight scores to serve as the product sets to be pushed corresponding to the user characteristic information, and obtaining products to be pushed according to the product sets to be pushed.
And the output module 604 is used for outputting the product to be pushed.
In one embodiment, the product pushing device 600 includes:
the product set acquisition module is used for acquiring a plurality of product sets according to the product characteristics of the products;
the training set acquisition module is used for acquiring a training set according to a plurality of product sets and user sample characteristic information, each training sample in the training set comprises user sample characteristic information and a corresponding product set, and each product set corresponds to a plurality of user sample characteristic information;
and the training module is used for inputting each training sample in the training set into the initial deep neural network model for learning training to obtain a trained product pushing model.
In one embodiment, the training set acquisition module comprises:
the user sample characteristic obtaining sub-module is used for obtaining user sample characteristic information according to a user sample, the user sample characteristic information comprises user personal attribute characteristics and user product interaction characteristics, and the user product interaction characteristics are used for expressing the mapping relation between the user personal attribute characteristics and products;
the training sample acquisition submodule is used for determining a product set corresponding to the user sample characteristic information according to the product set and the user product interaction characteristics; integrating the characteristic information of the user sample and a corresponding product set to obtain a training sample;
and the training sample integration submodule is used for acquiring a training set according to the plurality of training samples.
For specific limitations of the product pushing device, reference may be made to the above limitations of the product pushing method, which will not be described herein again. The modules in the product pushing device can be wholly or partially realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as shown in fig. 7. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless communication can be realized through WIFI, an operator network, NFC (near field communication) or other technologies. The computer program is executed by a processor to implement a product push method. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.
Those skilled in the art will appreciate that the architecture shown in fig. 7 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, a computer device is provided, comprising a memory and a processor, the memory having a computer program stored therein, the processor implementing the following steps when executing the computer program:
acquiring user characteristic information;
inputting the user characteristic information into the trained product push model to obtain the weight scores of all product sets corresponding to the user characteristic information; the product pushing model is obtained based on user sample characteristic information and corresponding product set training;
obtaining products to be pushed corresponding to the user characteristic information according to the weight scores of the product sets;
and outputting the product to be pushed.
In one embodiment, the processor, when executing the computer program, further performs the steps of:
acquiring personal attribute characteristics of a user;
acquiring interactive characteristics of a user product;
and integrating the personal attribute characteristics of the user and the interactive characteristics of the user product to obtain the user characteristic information.
In one embodiment, the processor, when executing the computer program, further performs the steps of:
obtaining a plurality of product sets according to the product characteristics of the products;
acquiring training sets according to a plurality of product sets and user sample characteristic information, wherein each training sample in the training sets comprises user sample characteristic information and a corresponding product set, and each product set corresponds to a plurality of user sample characteristic information;
and inputting each training sample in the training set into an initial deep neural network model for learning training to obtain a trained product pushing model.
In one embodiment, the processor, when executing the computer program, further performs the steps of:
acquiring user sample characteristic information according to a user sample, wherein the user sample characteristic information comprises user personal attribute characteristics and user product interaction characteristics, and the user product interaction characteristics are used for expressing the mapping relation between the user personal attribute characteristics and products;
determining a product set corresponding to the user sample characteristic information according to the product set and the user product interaction characteristics; integrating the characteristic information of the user sample and a corresponding product set to obtain a training sample;
and acquiring a training set according to the plurality of training samples.
In one embodiment, the processor, when executing the computer program, further performs the steps of:
selecting a product set with the weight score larger than a weight threshold value as a product set to be pushed corresponding to the user characteristic information, and obtaining a product to be pushed according to the product set to be pushed;
or selecting a preset number of product sets from high to low according to the weight scores to serve as the product sets to be pushed corresponding to the user characteristic information, and obtaining products to be pushed according to the product sets to be pushed.
In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, which when executed by a processor, performs the steps of:
acquiring user characteristic information;
inputting the user characteristic information into the trained product push model to obtain the weight scores of all product sets corresponding to the user characteristic information; the product pushing model is obtained based on user sample characteristic information and corresponding product set training;
obtaining products to be pushed corresponding to the user characteristic information according to the weight scores of the product sets;
and outputting the product to be pushed.
In one embodiment, the computer program when executed by the processor further performs the steps of:
acquiring personal attribute characteristics of a user;
acquiring interactive characteristics of a user product;
and integrating the personal attribute characteristics of the user and the interactive characteristics of the user product to obtain the user characteristic information.
In one embodiment, the computer program when executed by the processor further performs the steps of:
obtaining a plurality of product sets according to the product characteristics of the products;
acquiring training sets according to a plurality of product sets and user sample characteristic information, wherein each training sample in the training sets comprises user sample characteristic information and a corresponding product set, and each product set corresponds to a plurality of user sample characteristic information;
and inputting each training sample in the training set into an initial deep neural network model for learning training to obtain a trained product pushing model.
In one embodiment, the computer program when executed by the processor further performs the steps of:
acquiring user sample characteristic information according to a user sample, wherein the user sample characteristic information comprises user personal attribute characteristics and user product interaction characteristics, and the user product interaction characteristics are used for expressing the mapping relation between the user personal attribute characteristics and products;
determining a product set corresponding to the user sample characteristic information according to the product set and the user product interaction characteristics; integrating the characteristic information of the user sample and a corresponding product set to obtain a training sample;
and acquiring a training set according to the plurality of training samples.
In one embodiment, the computer program when executed by the processor further performs the steps of:
selecting a product set with the weight score larger than a weight threshold value as a product set to be pushed corresponding to the user characteristic information, and obtaining a product to be pushed according to the product set to be pushed;
or selecting a preset number of product sets from high to low according to the weight scores to serve as the product sets to be pushed corresponding to the user characteristic information, and obtaining products to be pushed according to the product sets to be pushed.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include at least one of non-volatile and volatile memory. Non-volatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical storage, or the like. Volatile Memory can include Random Access Memory (RAM) or external cache Memory. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A product pushing method, the method comprising:
acquiring user characteristic information;
inputting the user characteristic information into a trained product pushing model to obtain the weight scores of all product sets corresponding to the user characteristic information; the training mode of the product pushing model comprises the following steps: selecting product features of multiple dimensions, dividing the product features of all products according to each dimension, and then arranging and combining the product features of the multiple dimensions to obtain a plurality of groups of product feature combinations, wherein products with the same product feature combinations are integrated into a product set; acquiring training sets according to a plurality of product sets and user sample characteristic information, wherein each training sample in the training sets comprises user sample characteristic information and a corresponding product set, and each product set corresponds to a plurality of user sample characteristic information; inputting each training sample in the training set into an initial deep neural network model for learning training to obtain the trained product push model; obtaining products to be pushed corresponding to the user characteristic information according to the weight scores of the product sets;
and outputting the product to be pushed.
2. The method of claim 1, wherein the obtaining user characteristic information comprises:
acquiring personal attribute characteristics of a user;
acquiring interactive characteristics of a user product;
and integrating the personal attribute characteristics of the user and the interactive characteristics of the user product to obtain the user characteristic information.
3. The method of claim 1, wherein obtaining a training set from the plurality of product sets and user sample feature information comprises:
acquiring user sample characteristic information according to a user sample, wherein the user sample characteristic information comprises user personal attribute characteristics and user product interaction characteristics, and the user product interaction characteristics are used for representing the mapping relation between the user personal attribute characteristics and the product;
determining a product set corresponding to the user sample characteristic information according to the product set and the user product interaction characteristics; integrating the user sample characteristic information and the corresponding product set to obtain a training sample;
and acquiring a training set according to the plurality of training samples.
4. The method according to claim 1, wherein obtaining the product to be pushed corresponding to the user feature information according to the weight scores of the product sets comprises:
selecting a product set with a weight score larger than a weight threshold value as a product set to be pushed corresponding to the user characteristic information, and obtaining a product to be pushed according to the product set to be pushed;
or selecting a preset number of product sets from high to low according to the weight scores to serve as the product sets to be pushed corresponding to the user characteristic information, and obtaining products to be pushed according to the product sets to be pushed.
5. A product pusher device, characterized in that it comprises:
the characteristic acquisition module is used for acquiring user characteristic information;
the weight processing module is used for inputting the user characteristic information into a trained product pushing model to obtain the weight scores of all product sets corresponding to the user characteristic information; the product pushing model is obtained by training based on user sample characteristic information and a corresponding product set;
the product set acquisition module is used for selecting product features of multiple dimensions, dividing the product features of all products according to each dimension, then arranging and combining the product features of the multiple dimensions to obtain a plurality of groups of product feature combinations, and integrating products with the same product feature combinations into a product set;
the training set acquisition module is used for acquiring a training set according to a plurality of product sets and user sample characteristic information, wherein each training sample in the training set comprises the user sample characteristic information and a corresponding product set, and each product set corresponds to the plurality of user sample characteristic information;
the training module is used for inputting each training sample in the training set into an initial deep neural network model for learning training to obtain the trained product pushing model;
the product determining module is used for obtaining products to be pushed corresponding to the user characteristic information according to the weight scores of the product sets;
and the output module is used for outputting the product to be pushed.
6. The apparatus of claim 5, wherein the feature obtaining module is further configured to obtain a user personal attribute feature; acquiring interactive characteristics of a user product; and integrating the personal attribute characteristics of the user and the interactive characteristics of the user product to obtain the user characteristic information.
7. The apparatus of claim 5, wherein the training set acquisition module comprises:
the user sample characteristic obtaining sub-module is used for obtaining user sample characteristic information according to a user sample, wherein the user sample characteristic information comprises user personal attribute characteristics and user product interaction characteristics, and the user product interaction characteristics are used for representing the mapping relation between the user personal attribute characteristics and the product;
the training sample acquisition sub-module is used for determining a product set corresponding to the user sample characteristic information according to the product set and the user product interaction characteristics; integrating the user sample characteristic information and the corresponding product set to obtain a training sample;
and the training sample integration submodule is used for acquiring a training set according to the plurality of training samples.
8. The device according to claim 5, wherein the product determination module is further configured to select a product set with a weight score greater than a weight threshold as a product set to be pushed corresponding to the user feature information, and obtain a product to be pushed according to the product set to be pushed; or selecting a preset number of product sets from high to low according to the weight scores to serve as the product sets to be pushed corresponding to the user characteristic information, and obtaining products to be pushed according to the product sets to be pushed.
9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, implements the steps of the method of any of claims 1 to 5.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110486268.0A CN113256368B (en) | 2021-04-30 | 2021-04-30 | Product pushing method and device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110486268.0A CN113256368B (en) | 2021-04-30 | 2021-04-30 | Product pushing method and device, computer equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113256368A CN113256368A (en) | 2021-08-13 |
| CN113256368B true CN113256368B (en) | 2022-02-18 |
Family
ID=77223540
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110486268.0A Active CN113256368B (en) | 2021-04-30 | 2021-04-30 | Product pushing method and device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113256368B (en) |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108416625A (en) * | 2018-02-28 | 2018-08-17 | 阿里巴巴集团控股有限公司 | The recommendation method and apparatus of marketing product |
| CN109451038A (en) * | 2018-12-06 | 2019-03-08 | 北京达佳互联信息技术有限公司 | A kind of information-pushing method, device, server and computer readable storage medium |
| US11204973B2 (en) * | 2019-06-21 | 2021-12-21 | Microsoft Technology Licensing, Llc | Two-stage training with non-randomized and randomized data |
| CN111161009B (en) * | 2019-11-19 | 2023-11-07 | 中国电子产品可靠性与环境试验研究所（（工业和信息化部电子第五研究所）（中国赛宝实验室）） | Information pushing method, device, computer equipment and storage medium |
| CN111626832B (en) * | 2020-06-05 | 2023-10-03 | 中国银行股份有限公司 | Product recommendation method and device and computer equipment |
| CN112328909B (en) * | 2020-11-17 | 2021-07-02 | 中国平安人寿保险股份有限公司 | Information recommendation method and device, computer equipment and medium |
| CN112231584B (en) * | 2020-12-08 | 2021-03-26 | 平安科技（深圳）有限公司 | Data pushing method and device based on small sample transfer learning and computer equipment |
-
2021
- 2021-04-30 CN CN202110486268.0A patent/CN113256368B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN113256368A (en) | 2021-08-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109543925B (en) | Risk prediction method and device based on machine learning, computer equipment and storage medium | |
| CN111625632A (en) | Question-answer pair recommendation method, device, equipment and storage medium | |
| CN112633962B (en) | Service recommendation method and device, computer equipment and storage medium | |
| CN111352962A (en) | Client portrait construction method and device | |
| CN110162359A (en) | Method for pushing, the apparatus and system of new hand's guidance information | |
| CN108509458A (en) | A kind of business object recognition methods and device | |
| CN113407854A (en) | Application recommendation method, device and equipment and computer readable storage medium | |
| CN109614982A (en) | Product analysis method, apparatus, computer equipment and storage medium | |
| CN112395499B (en) | Information recommendation method and device, electronic equipment and storage medium | |
| Schoellman | The occupations and human capital of US immigrants | |
| CN111598632B (en) | Method and device for determining equity shares and equity share sequence | |
| CN113781161A (en) | Display page generation method and device, computer equipment and storage medium | |
| CN113256368B (en) | Product pushing method and device, computer equipment and storage medium | |
| CN112487284A (en) | Bank customer portrait generation method, equipment, storage medium and device | |
| CN115809889A (en) | Intelligent passenger group screening method, system, medium and equipment based on marketing effect | |
| CN116402596A (en) | Data analysis method, device, computer equipment and readable storage medium | |
| CN115311001A (en) | Method and system for predicting user change tendency based on multiple voting algorithm | |
| CN113988917A (en) | Self-service data labeling method and device based on service priority | |
| CN113011985A (en) | Financial product push data processing method and device | |
| CN114372199A (en) | Business pushing method and device, storage medium and electronic equipment | |
| CN112766824A (en) | Data processing method and device, electronic equipment and storage medium | |
| TWM615832U (en) | Credit evaluation system | |
| CN112750023A (en) | Consumption financial user income estimation method based on factor analysis | |
| CN114548620A (en) | Logistics punctual insurance service recommendation method and device, computer equipment and storage medium | |
| CN113536103A (en) | Information recommendation method and device, electronic 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |