CN116628348A - Service recommendation system, method, device, electronic equipment and storage medium - Google Patents
Service recommendation system, method, device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN116628348A CN116628348A CN202310893802.9A CN202310893802A CN116628348A CN 116628348 A CN116628348 A CN 116628348A CN 202310893802 A CN202310893802 A CN 202310893802A CN 116628348 A CN116628348 A CN 116628348A
- Authority
- CN
- China
- Prior art keywords
- service
- data
- target
- service recommendation
- embedded layer
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000012545 processing Methods 0.000 claims abstract description 35
- 238000012549 training Methods 0.000 claims description 51
- 238000000605 extraction Methods 0.000 claims description 39
- 230000006399 behavior Effects 0.000 claims description 30
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 10
- 230000006870 function Effects 0.000 description 7
- 238000013528 artificial neural network Methods 0.000 description 6
- 230000003993 interaction Effects 0.000 description 6
- 230000008569 process Effects 0.000 description 4
- 238000004891 communication Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 238000002347 injection Methods 0.000 description 3
- 239000007924 injection Substances 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000006835 compression Effects 0.000 description 2
- 238000007906 compression Methods 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013475 authorization Methods 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000003780 insertion Methods 0.000 description 1
- 230000037431 insertion Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 230000008707 rearrangement Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 239000000243 solution Substances 0.000 description 1
- 238000010897 surface acoustic wave method Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5061—Partitioning or combining of resources
- G06F9/5072—Grid computing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2209/00—Indexing scheme relating to G06F9/00
- G06F2209/50—Indexing scheme relating to G06F9/50
- G06F2209/502—Proximity
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application relates to the technical field of artificial intelligence, and particularly provides a service recommendation system, a service recommendation method, a service recommendation device, electronic equipment and a storage medium. A method for recommending service comprises the steps of sending service operation data to a data server; receiving target embedded layer characteristics returned by a data server based on service operation data; generating service recommendation data based on service operation data and target embedded layer characteristics by adopting a pre-trained service recommendation target model; and executing the service recommending operation based on the service recommending data. In the embodiment of the application, under the condition that the terminal equipment resources are limited, the complex data processing operation is performed through the data server, so that the resources consumed by the data processing of the terminal equipment are reduced, the service recommendation efficiency is improved, the service recommendation can be performed by adopting the service recommendation target model with high complexity, and the accuracy of the service recommendation is improved.
Description
Technical Field
The application relates to the technical field of artificial intelligence, in particular to a service recommendation system, a service recommendation method, a service recommendation device, electronic equipment and a storage medium.
Background
With the development of internet technology and intelligent terminal technology, users typically browse various services, such as video, of clients through terminal devices, such as mobile phones. To improve the user experience, the terminal device typically makes service recommendations to the user.
In the prior art, the terminal device generally determines service recommendation data through a local service recommendation model, and executes user recommendation operation based on the service recommendation data.
However, since the storage space and the computational resources of the terminal device are usually limited, when a complex service recommendation model is operated, the operation speed of other services in the terminal device may be affected due to relatively hard service recommendation efficiency, and since the available space resources of the service recommendation model are limited, the service recommendation is usually performed by using a relatively simplified service recommendation model, for example, a mode of losing part of an algorithm for optimizing the service recommendation is usually adopted, and the service recommendation model is simplified, which may result in poor accuracy of the service recommendation.
Disclosure of Invention
The embodiment of the application aims to provide a service recommendation method, a device, electronic equipment and a storage medium, which are used for improving the efficiency and accuracy of service recommendation when the service recommendation is carried out.
In a first aspect, an embodiment of the present application provides a service recommendation system, including a terminal device and a data server, where,
the terminal equipment is used for sending the business operation data to the data server; the business operation data are acquired based on the operation behavior of the user aiming at the target business; receiving target embedded layer characteristics returned by a data server based on service operation data; generating service recommendation data based on service operation data and target embedded layer characteristics by adopting a pre-trained service recommendation target model; based on the service recommendation data, executing service recommendation operation;
the data server is used for receiving service operation data sent by the terminal equipment; carrying out data processing on the business operation data to obtain target embedded layer characteristics; and returning the target embedded layer characteristics to the terminal equipment.
In one embodiment, a data server includes an edge node;
the terminal equipment is used for sending the business operation data to the edge node;
and the edge node is used for carrying out feature extraction on the business operation data by adopting a feature extraction model to obtain the features of the target embedded layer.
In one embodiment, a data server includes a center node and an edge node;
The edge node is used for obtaining the universal embedded layer characteristics based on the training data and sending the universal embedded layer characteristics to the center node; the training data is acquired based on the operation behavior of the user for the service;
and the central node is used for carrying out model training on the service recommendation initial model based on the universal embedded layer characteristics and training data to obtain a trained service recommendation target model.
In a second aspect, an embodiment of the present application provides a method for service recommendation, where the method includes:
transmitting the business operation data to a data server; the business operation data are acquired based on the operation behavior of the user aiming at the target business;
receiving target embedded layer characteristics returned by a data server based on service operation data;
generating service recommendation data based on service operation data and target embedded layer characteristics by adopting a pre-trained service recommendation target model;
and executing the service recommending operation based on the service recommending data.
In one embodiment, the target embedded layer features are obtained by extracting features from service operation data by using a feature extraction model.
In a third aspect, an embodiment of the present application provides a method for service recommendation, where the method includes:
Receiving service operation data sent by terminal equipment; the business operation data are acquired based on the operation behavior of the user aiming at the target business;
carrying out data processing on the business operation data to obtain target embedded layer characteristics;
and returning the target embedded layer characteristics to the terminal equipment, so that the terminal equipment adopts a pre-trained service recommendation target model, generates service recommendation data based on the service operation data and the target embedded layer characteristics, and executes service recommendation operation based on the service recommendation data.
In one embodiment, the data processing for the business operation data to obtain the target embedded layer feature includes:
and adopting a feature extraction model to extract features of the business operation data to obtain the features of the target embedded layer.
In one embodiment, the business recommendation target model is obtained by the following steps:
obtaining universal embedded layer characteristics based on training data; the training data is acquired based on the operation behavior of the user for the service;
based on the general embedded layer characteristics and training data, model training is carried out on the service recommendation initial model, and a trained service recommendation target model is obtained.
In a fourth aspect, an embodiment of the present application provides a service recommendation apparatus, including:
The sending unit is used for sending the business operation data to the data server; the business operation data are acquired based on the operation behavior of the user aiming at the target business;
the receiving unit is used for receiving the target embedded layer characteristics returned by the data server based on the business operation data;
the generating unit is used for generating service recommendation data based on the service operation data and the target embedded layer characteristics by adopting a pre-trained service recommendation target model;
and the execution unit is used for executing the service recommendation operation based on the service recommendation data.
In one embodiment, the target embedded layer features are obtained by extracting features from service operation data by using a feature extraction model.
In a fifth aspect, an embodiment of the present application provides a service recommendation apparatus, including:
a receiving unit, configured to receive service operation data sent by a terminal device; the business operation data are acquired based on the operation behavior of the user aiming at the target business;
the processing unit is used for carrying out data processing on the business operation data to obtain the characteristics of the target embedded layer;
and the return unit is used for returning the target embedded layer characteristics to the terminal equipment, so that the terminal equipment adopts a pre-trained service recommendation target model, generates service recommendation data based on the service operation data and the target embedded layer characteristics, and executes service recommendation operation based on the service recommendation data.
In one embodiment, the processing unit is configured to:
and adopting a feature extraction model to extract features of the business operation data to obtain the features of the target embedded layer.
In one embodiment, the return unit is for:
the business recommendation target model is obtained by adopting the following steps:
obtaining universal embedded layer characteristics based on training data; the training data is acquired based on the operation behavior of the user for the service;
based on the general embedded layer characteristics and training data, model training is carried out on the service recommendation initial model, and a trained service recommendation target model is obtained.
In a sixth aspect, an embodiment of the present application provides an electronic device, including:
a processor; and
a memory storing computer instructions for causing a processor to perform the method of any embodiment of the second or third aspects.
In a seventh aspect, an embodiment of the present application provides a storage medium storing computer instructions for causing a computer to perform the method of the second aspect or any embodiment of the third aspect.
In the system, the method, the device, the electronic equipment and the storage medium for recommending the service provided by the embodiment of the application, service operation data are sent to a data server; the business operation data are acquired based on the operation behavior of the user aiming at the target business; receiving target embedded layer characteristics returned by a data server based on service operation data; generating service recommendation data based on service operation data and target embedded layer characteristics by adopting a pre-trained service recommendation target model; and executing the service recommending operation based on the service recommending data. In the embodiment of the application, the data server generates the target embedded layer characteristics based on the service operation data, and the service recommendation target model is adopted in the terminal equipment to conduct service recommendation based on the service operation data and the target embedded layer characteristics, so that the characteristic extraction operation which consumes a large amount of computing resources and storage space resources is transferred to the data server for processing, a large amount of computing resources and storage space consumed by the terminal equipment in the aspect of data characteristic extraction are reduced, more computing resources and storage space resources can be applied to the service recommendation subtasks in the service recommendation target model, and the accuracy and efficiency of service recommendation can be improved.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings that are needed in the description of the embodiments or the prior art will be briefly described, and it is obvious that the drawings in the description below are some embodiments of the present application, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flow chart of a method of service recommendation in accordance with some embodiments of the present application.
Fig. 2 is a schematic architecture diagram of a system for service recommendation according to some embodiments of the present application.
Fig. 3 is a diagram illustrating an example of the structure of a central node according to some embodiments of the present application.
Fig. 4 is a block diagram of an apparatus for service recommendation according to some embodiments of the present application.
Fig. 5 is a block diagram of another service recommendation device according to some embodiments of the present application.
Fig. 6 is a schematic diagram of an electronic device according to some embodiments of the application.
Detailed Description
The following description of the embodiments of the present application will be made more apparent and fully hereinafter with reference to the accompanying drawings, in which some, but not all embodiments of the application are shown. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present application without making any inventive effort, are intended to fall within the scope of the present application. In addition, the technical features of the different embodiments of the present application described below may be combined with each other as long as they do not collide with each other.
Some of the terms involved in the embodiments of the present application will be described first to facilitate understanding by those skilled in the art.
Terminal equipment: the mobile terminal, stationary terminal or portable terminal may be, for example, a mobile handset, a site, a unit, a device, a multimedia computer, a multimedia tablet, an internet node, a communicator, a desktop computer, a laptop computer, a notebook computer, a netbook computer, a tablet computer, a personal communications system device, a personal navigation device, a personal digital assistant, an audio/video player, a digital camera/camcorder, a positioning device, a television receiver, a radio broadcast receiver, an electronic book device, a game device, or any combination thereof, including the accessories and peripherals of these devices, or any combination thereof. It is also contemplated that the terminal device can support any type of interface (e.g., wearable device) for the user, etc.
And (3) a server: the cloud server can be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, and can also be a cloud server for providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, basic cloud computing services such as big data and artificial intelligent platforms and the like.
Embedding layer (Embedding): essentially, "compression" describes higher-dimensional n-dimensional features with redundant information with lower-dimensional k-dimensional features, or higher-dimensional n-dimensional memory with lower-dimensional k-dimensional memory. k and n are positive integers. Embedding is a way to transform discrete variables into a continuous vector representation. In neural networks, ededing is very useful because it not only reduces the storage space dimension of a discrete variable, but also meaningfully represents the variable.
In an actual application scenario, a user usually browses various services of a client through a terminal device such as a mobile phone. For example, video of a short video client is played. In order to improve user experience, the client typically determines service recommendation data through a local service recommendation model, and performs a user recommendation operation based on the service recommendation data. For example, an application scenario may determine service recommendation data for a client through a local service recommendation model, and rearrange or refresh display content based on the service recommendation data.
However, since the terminal devices have diversity, the computing power and the available storage space for the service recommendation are usually different between different terminal devices, when the low-end device (i.e. the device with smaller computing power resource and storage space resource) runs the complex service recommendation model, the low-end device may be relatively hard, and may affect the running speed of other services in the terminal device, and because the available space resource of the service recommendation model is limited, the service recommendation is usually performed by adopting a relatively simplified service recommendation model, for example, a mode of losing part of the algorithm for optimizing the service recommendation is usually adopted, so that the service recommendation model is simplified, which may result in poor accuracy and efficiency of the service recommendation.
Based on the defects of the related art, the embodiment of the application provides a service recommendation system, a method, a device, electronic equipment and a storage medium, aiming at improving service recommendation performance when service recommendation is carried out.
The embodiment of the application provides a method for recommending service, which can be applied to electronic equipment, the type of the electronic equipment is not limited, and the method can be any equipment type suitable for implementation, such as a smart phone, a tablet computer and the like, and the application is not repeated.
Referring to fig. 1, a flowchart of a method for recommending services according to some embodiments of the present application is shown, and the method is described below with reference to fig. 1, where a specific implementation procedure of the method is as follows:
step 100: and the terminal equipment sends the service operation data to the data server.
Specifically, the client in the terminal device may collect service operation data in real time or periodically, and send the service operation data to the data server in real time or periodically.
The client may be an application program, such as a short video application program, that is currently operated by the user through the terminal device.
Alternatively, the business operations data may be obtained based on the user's operations with respect to the target business, such as click data, collection data, and order data.
Wherein the business operation data comprises user operation behavior data. Optionally, the business operation data may further include target business data.
The user operation behavior can be browsing, exposing, clicking, collecting and the like. The user operation behavior data can be browsing duration, browsing frequency, clicking times, praise, attention, collection and the like aiming at the target service. The target service may be short video, article, etc. The target business data can be short videos and the business information such as the subjects, labels and types of articles.
In practical application, the specific content and specific collection mode (such as collection time and collection frequency) of the service operation data can be set according to the practical application scenario, and the method is not limited herein.
Therefore, the operation behavior of the user aiming at the target service can be perceived in real time, and the perceived service operation data is uploaded to the data server, so that the data processing can be carried out through the data server in the subsequent steps, and the resource and time cost consumed by the terminal equipment for processing the service operation data locally are reduced.
Step 101: and the data server receives the service operation data sent by the terminal equipment.
Alternatively, the data server may be one server, or may be a plurality of servers, that is, a server cluster.
In one embodiment, to increase data processing efficiency, the data server may be a distributed computing network, i.e., a distributed system, that is configured based on edge nodes and central nodes. As an example, the central node may be an internet data center (Internet Data Center, IDC) and may also be a cloud server.
In one embodiment, the terminal device collects service operation data in real time and sends the service operation data to the edge node. And the edge node receives the service operation data sent by the terminal equipment.
Among other things, a distributed system has the following advantages, 1) high availability (i.e., low fault tolerance), and one important advantage in a distributed system is reliability. The system breakdown of one server does not affect other servers, and can still normally provide services to the outside. 2) With scalability, the ever-increasing external demands can be accommodated by linearly increasing machine resources. 3) The resource sharing can be realized, and the sharing of data can be realized. 4) The distributed system is highly flexible, so that the distributed system is easy to install, implement and debug the new service 5) with higher speed, can be deployed more often, routes user requests geographically to the nearest machine room for processing, and has higher processing speed than other systems due to the computing power of multiple computers. 6) Is an open system, and since it is an open system, the service is accessible locally or remotely. 7) Having higher performance may provide higher performance and better cost performance than centralized computer network clusters.
Step 102: and the data server performs data processing on the service operation data to obtain the characteristics of the target embedded layer.
Specifically, the data server adopts a feature extraction model to extract features of the business operation data, and the features of the target embedded layer are obtained.
In one embodiment, the data server includes an edge node, and the edge node uses a feature extraction model to perform feature extraction on the service operation data to obtain the target embedded layer feature.
Alternatively, the features in the target embedded layer feature may include a User feature (User and item, U2I) feature (feature). Where U2I generally refers to items (items) of recommended services to users based on matrix factorization.
The feature extraction model is a model for extracting features of data, and is generally any general model for extracting features of data. The feature extraction model comprises an Embedding model. Embedding is essentially "compression", using lower-dimensional k-dimensional features to describe higher-dimensional n-dimensional features with redundant information, or using lower-dimensional k-dimensional memory to describe higher-dimensional n-dimensional memory. Embedding is a way to transform discrete variables into a continuous vector representation. In neural networks, ededing is very useful because it not only reduces the storage space dimension of a discrete variable, but also meaningfully represents the variable.
In order to distinguish the embedded layer features at different stages, in the embodiment of the present application, the embedded layer features obtained in the process of service recommendation application are referred to as target embedded layer features, and the embedded layer features obtained in the training process of the service recommendation target model in the subsequent step are referred to as general embedded layer features.
Step 103: the data server returns the target embedded layer feature to the terminal device.
In one embodiment, the data server includes an edge node, and the edge node returns the target embedded layer feature to the client of the terminal device.
Step 104: and the terminal equipment receives the target embedded layer characteristics returned by the data server based on the service operation data.
In this way, the data server adopts the feature extraction model to perform feature extraction on the business operation data to generate the target embedded layer feature, so that the terminal equipment can directly obtain the target embedded layer feature without consuming a large amount of computing power resources to perform data processing.
Step 105: the terminal equipment adopts a pre-trained service recommendation target model, and generates service recommendation data based on service operation data and target embedded layer characteristics.
Specifically, the business recommendation target model is constructed based on deep learning.
Considering that the types of the services are generally various, the service recommendation target model can perform corresponding content recommendation on the current target service based on the user operation behavior data and the target service data. In one embodiment, the service operation data includes user operation behavior data and target service data, and the target embedded layer feature may be injected into the service recommendation target model, and the user operation behavior data and the target service data may be input into the service recommendation target model into which the target embedded layer feature is injected, so as to output the service recommendation data.
Injection is understood to mean insertion, for example, into a knowledge plug (knowledge plugging).
Therefore, a large amount of data feature extraction operations are transferred to the data server for processing, so that the terminal equipment can directly inject the target embedded layer features obtained through the feature extraction processing of the data server into the service recommendation target model, and therefore, a large amount of computing resources in the terminal equipment are not required to be consumed for performing the feature extraction operations when the service recommendation target model is used for service recommendation, consumed computing power and storage space resources are reduced, and data processing efficiency is improved. And the computational effort and space resources originally used for the feature extraction operation in the terminal equipment can be changed to improve the complexity of the service recommendation subtask so as to optimize the service recommendation subtask, thereby improving the efficiency and accuracy of service recommendation.
In one embodiment, the business recommendation target model is obtained by training the following steps:
s1051: the data server obtains generic embedded layer features based on the training data.
The training data is acquired based on the operation behavior of the user for the service. The training data may be obtained using the same principle as the obtaining of the service operation data, which will not be described here in detail. As one example, the training data may be historical business operations data.
The general embedded layer feature may be obtained based on the same principle as the target embedded layer feature is obtained, and will not be described herein.
S1052: and the data server carries out model training on the service recommendation initial model based on the universal embedded layer characteristics and training data to obtain a trained service recommendation target model.
Specifically, the business recommendation initial model is constructed based on deep learning.
In one embodiment, a data server includes an edge node and a center node, the edge node obtains generic embedded layer features based on training data and transmits the generic embedded layer features to the center node. And the central node performs model training on the service recommendation initial model based on the universal embedded layer characteristics and training data to obtain a trained service recommendation target model.
In this way, a distributed computing network for embedded layer feature generation and model training can be built based on the edge nodes and the central nodes (e.g., cloud servers), so that cloud computing and edge computing are combined, and the efficiency of embedded layer generation and model training is improved.
Step 106: and the terminal equipment executes service recommendation operation based on the service recommendation data.
In some embodiments, the service recommendation operation may include a video stream rearrangement display, a column recommendation display, a page information recommendation display, and the like.
For example, the payment precious client in the terminal equipment updates the recommendation column content at the home page column end of the payment precious client according to the service recommendation data.
For another example, the short video client in the terminal device rearranges the displayed video stream in real time according to the service recommendation data.
For another example, the payment precious client in the terminal device updates the displayed recommended content on the payment result page of the payment precious client according to the service recommended data, for example, if the user pays for the purchased product, the shopping commodity page recommended according to the product is displayed.
Considering that the feature extraction processing in the service recommendation model in the traditional manner generally occupies a large amount of computing power (for example, generally occupies more than 70% of the model consumption computing power), and most of the data features are universal in nature, the data feature extraction process in the service recommendation model can be transferred to the data server for processing, so that a large amount of computing power and storage space resources consumed by the terminal equipment in the feature extraction process can be reduced, and further the computing power and storage space resources of the service recommendation model can be more reasonably distributed. In this way, repeated calculation of different service recommendation tasks on the general task is reduced, so that the client does not need to consume a large amount of computing power resources and space resources to perform feature extraction, and the feature extraction process is extracted, so that the complexity of the service recommendation subtasks in the service recommendation target model can be improved under the limited storage space to optimize the service recommendation, the logic of the service recommendation subtasks in the respective fields can be focused more, the processing capacity of the subtasks is greatly improved, the accuracy and the efficiency of the service recommendation are improved, the performance requirements of terminal equipment are met, the service recommendation performance is improved, and the development of the service effect is met.
In an embodiment of the present application, a system for service recommendation applying the above embodiment is provided, and the above embodiment is further described below with reference to fig. 2. Referring to fig. 2, a schematic architecture diagram of a system for recommending services is shown. Fig. 2 includes a terminal device 2100 and a data server 2200. The data server 2200 includes an edge node 2210 and a center node 2220.
In one embodiment, the terminal device 2100 is configured to transmit the business operation data to the data server 2200; receiving target embedded layer characteristics returned by the data server 2200 based on the business operation data; generating service recommendation data based on service operation data and target embedded layer characteristics by adopting a pre-trained service recommendation target model; and executing the service recommending operation based on the service recommending data.
Specifically, the terminal device 2100 includes a service operation data module 2101 and a service recommendation module 2102.
Wherein the business operations data module 2101: for acquiring service operation data and uploading the service operation data to the edge node 2210 and the service recommendation module 2102, respectively.
Wherein, the service recommendation module 2102: the service recommendation method comprises the steps of receiving a target embedded layer characteristic returned by an edge node 2210 based on service operation data and a service recommendation target model returned by a center node 2220, generating service recommendation data based on the service operation data and the target embedded layer characteristic by adopting the service recommendation target model, and executing service recommendation operation based on the service recommendation data.
In one embodiment, the data server 2200 is configured to receive service operation data transmitted from the terminal device 2100; carrying out data processing on the business operation data to obtain target embedded layer characteristics; the target embedded layer feature is returned to the terminal device 2100.
The edge node 2210 includes a feature extraction module 2211 and an embedded layer acquisition module 2212.
The feature extraction module 2211 is configured to perform feature extraction on the service operation data by using a feature extraction model to obtain a target embedded layer feature.
The embedded layer acquisition module 2212 is configured to return the generated target embedded layer characteristics to the service recommendation module 2102 of the terminal device 2100. For example, the embedded features (i.e., target embedded features) in the embedded layer acquisition module 2212 may include K1 … … Ku. Wherein, K1 is User feature, K2 is U2I task (task) 1, and K3 is U2Itask2.K represents a feature, u is a positive integer, and a sequence number of the feature is represented.
Further, in the model training phase, the embedded layer obtaining module 2212 is further configured to obtain a general embedded layer feature based on the training data, and send the general embedded layer feature to the central node 2220.
In this way, generic embedded layer features for model training and target embedded layer features for business recommendation applications can be generated by the edge node 2210, respectively.
The central node 2220 includes an embedded layer injection module 2221 and a model training module 2222. The specific structure of each module in the central node 2220 will be described below with reference to fig. 3. Fig. 3 is a diagram illustrating a structure of the central node 2220. The model training module 2222 in the central node 2220 includes a data input module, a service embedding module, a feature interaction module, and a neural network module.
Embedded layer injection module 2221: and the universal embedded layer feature module is used for receiving the universal embedded layer features sent by the embedded layer acquisition module 2212 and injecting the universal embedded layer features into the service recommendation initial model.
Model training module 2222: the method is used for carrying out model training on the service recommendation initial model based on training data and the general embedded layer characteristics to obtain a trained service recommendation target model.
The data input module is used for inputting training data. Alternatively, the training data may include User operation behavior data (User Behaviors) and target business data (Item features).
The service Embedding module may be an Embedding Layer (Embedding Layer) for extracting features of target service data in the training data. Feature interaction module (feature Interaction): the method is used for interacting the output characteristics of the business embedding module to obtain characteristic interaction data. Neural network module: and the method is used for carrying out service recommendation based on the feature interaction data by combining the injected universal embedded layer features. Alternatively, the neural network module may employ a multi-layer perceptron (Multilayer Perceptron, MLP). As an example, the neural network module adopts MLP, obtains comprehensive feature data h'm based on the feature interaction data hm and the general embedded layer features, and performs service recommendation based on the comprehensive feature data h'm.
Specifically, for the execution steps of each module in the service recommendation system, refer to the specific steps of the steps 100 to 106, which are not described herein.
In the embodiment of the application, the universal embedded layer characteristics are generated based on the training data, the model training is carried out based on the universal embedded layer characteristics and the training data, the service recommendation target model is obtained, and the target embedded layer characteristics are generated based on the service operation data through the data server, so that the terminal equipment can directly inject the target embedded layer characteristics into the service recommendation target model to carry out service recommendation. In this way, repeated calculation of different service recommendation tasks on the general task is reduced, so that the client does not need to consume a large amount of computing power resources and space resources to perform feature extraction, and the feature extraction process is extracted, so that the complexity of the service recommendation subtasks in the service recommendation target model can be improved under the limited storage space to optimize the service recommendation, the logic of the service recommendation subtasks in the respective fields can be focused more, the processing capacity of the subtasks is greatly improved, the accuracy and the efficiency of the service recommendation are improved, the performance requirements of terminal equipment are met, the service recommendation performance is improved, and the development of the service effect is met.
The user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, presented data, etc.) related to the present application are information and data authorized by the user or fully authorized by each party, and the collection, use and processing of related data is required to comply with the relevant laws and regulations and standards of the relevant country and region, and is provided with corresponding operation entries for the user to select authorization or rejection.
Based on the same inventive concept, the embodiment of the application also provides a service recommendation device, and because the principle of solving the problem by the device and the equipment is similar to that of a service recommendation method, the implementation of the device can refer to the implementation of the method, and the repetition is omitted. The device can be applied to electronic equipment, the type of the electronic equipment is not limited by the application, and the device can be any equipment type suitable for implementation, such as a smart phone, a tablet computer and the like, and the application is not repeated.
Referring to fig. 4, a block diagram of an apparatus for recommending services according to some embodiments of the present application is shown. In some embodiments, the service recommending device of the present application includes:
A transmitting unit 401, configured to transmit service operation data to a data server; the business operation data are acquired based on the operation behavior of the user aiming at the target business;
a receiving unit 402, configured to receive a target embedded layer feature returned by the data server based on the service operation data;
a generating unit 403, configured to generate service recommendation data based on the service operation data and the target embedded layer characteristics by using a pre-trained service recommendation target model;
and an execution unit 404, configured to execute a service recommendation operation based on the service recommendation data.
In one embodiment, the target embedded layer features are obtained by extracting features from service operation data by using a feature extraction model.
Referring to fig. 5, a block diagram of another service recommendation device according to some embodiments of the present application is shown. In some embodiments, the service recommending device of the present application includes:
a receiving unit 501, configured to receive service operation data sent by a terminal device; the business operation data are acquired based on the operation behavior of the user aiming at the target business;
the processing unit 502 is configured to perform data processing on the service operation data to obtain a target embedded layer feature;
And a returning unit 503, configured to return the target embedded layer feature to the terminal device, so that the terminal device adopts a pre-trained service recommendation target model, generates service recommendation data based on the service operation data and the target embedded layer feature, and performs a service recommendation operation based on the service recommendation data.
In one embodiment, the processing unit 502 is configured to:
and adopting a feature extraction model to extract features of the business operation data to obtain the features of the target embedded layer.
In one embodiment, the return unit 503 is configured to:
the business recommendation target model is obtained by adopting the following steps:
obtaining universal embedded layer characteristics based on training data; the training data is acquired based on the operation behavior of the user for the service;
based on the general embedded layer characteristics and training data, model training is carried out on the service recommendation initial model, and a trained service recommendation target model is obtained.
In the system, the method, the device, the electronic equipment and the storage medium for recommending the service provided by the embodiment of the application, service operation data are sent to a data server; the business operation data are acquired based on the operation behavior of the user aiming at the target business; receiving target embedded layer characteristics returned by a data server based on service operation data; generating service recommendation data based on service operation data and target embedded layer characteristics by adopting a pre-trained service recommendation target model; and executing the service recommending operation based on the service recommending data. In the embodiment of the application, the data server generates the target embedded layer characteristics based on the service operation data, and the service recommendation target model is adopted in the terminal equipment to conduct service recommendation based on the service operation data and the target embedded layer characteristics, so that the characteristic extraction operation which consumes a large amount of computing resources and storage space resources is transferred to the data server for processing, a large amount of computing resources and storage space consumed by the terminal equipment in the aspect of data characteristic extraction are reduced, more computing resources and storage space resources can be applied to the service recommendation subtasks in the service recommendation target model, and the accuracy and efficiency of service recommendation can be improved.
The embodiment of the application provides electronic equipment, which comprises:
a processor; and
and a memory storing computer instructions for causing the processor to perform the method of any of the embodiments described above.
The present application provides a storage medium storing computer instructions for causing a computer to perform the method of any of the above embodiments.
Fig. 6 shows a schematic structural diagram of an electronic device 6000. Referring to fig. 6, an electronic device 6000 includes: the processor 6010 and the memory 6020 may further include a power supply 6030, a display unit 6040, and an input unit 6050, as an option.
The processor 6010 is a control center of the electronic device 6000, connects respective components using various interfaces and lines, and performs various functions of the electronic device 6000 by running or executing software programs and/or data stored in the memory 6020.
In the embodiment of the present application, the processor 6010 executes the steps of the above embodiment when calling the computer program stored in the memory 6020.
Optionally, processor 6010 may include one or more processing units; preferably, the processor 6010 may integrate an application processor and a modem processor, wherein the application processor mainly processes an operating system, a user interface, an application, etc., and the modem processor mainly processes wireless communication. It will be appreciated that the modem processor described above may not be integrated into the processor 6010. In some embodiments, the processor, memory, may be implemented on a single chip, and in some embodiments, they may be implemented separately on separate chips.
The memory 6020 may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, various applications, and the like; the storage data area may store data created according to the use of the electronic device 6000, and the like. In addition, memory 6020 may comprise high-speed random access memory and may also comprise non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid state storage device, and the like.
The electronic device 6000 also includes a power supply 6030 (e.g., a battery) for powering the various components, which may be logically connected to the processor 6010 by a power management system that performs functions such as managing charge, discharge, and power consumption.
The display unit 6040 may be used to display information input by a user or information provided to the user, various menus of the electronic device 6000, and the like, and is mainly used to display a display interface of each application in the electronic device 6000 and objects such as texts and pictures displayed in the display interface in the embodiment of the present application. The display unit 6040 may include a display panel 6041. The display panel 6041 may be configured in the form of a liquid crystal display (Liquid Crystal Display, LCD), an Organic Light-Emitting Diode (OLED), or the like.
The input unit 6050 may be used to receive information such as numbers or characters input by a user. The input unit 6050 may include a touch panel 6051 and other input devices 6052. Wherein the touch panel 6051, also referred to as a touch screen, may collect touch operations thereon or thereabout by a user (e.g., operations of the user on the touch panel 6051 or thereabout using any suitable object or accessory such as a finger, stylus, etc.).
Specifically, the touch panel 6051 may detect a touch operation by a user, detect a signal caused by the touch operation, convert the signal into a touch point coordinate, send the touch point coordinate to the processor 6010, and receive and execute a command sent from the processor 6010. In addition, the touch panel 6051 may be implemented in various types such as resistive, capacitive, infrared, and surface acoustic wave. Other input devices 6052 may include, but are not limited to, one or more of a physical keyboard, function keys (e.g., volume control keys, on-off keys, etc.), a trackball, mouse, joystick, etc.
Of course, the touch panel 6051 may cover the display panel 6041, and when the touch panel 6051 detects a touch operation thereon or thereabout, the touch operation is transmitted to the processor 6010 to determine the type of the touch event, and then the processor 6010 provides a corresponding visual output on the display panel 6041 according to the type of the touch event. Although in fig. 6, the touch panel 6051 and the display panel 6041 are two independent components to realize the input and output functions of the electronic device 6000, in some embodiments, the touch panel 6051 and the display panel 6041 may be integrated to realize the input and output functions of the electronic device 6000.
The electronic device 6000 may also include one or more sensors, such as pressure sensors, gravitational acceleration sensors, proximity light sensors, and the like. Of course, the electronic device 6000 may also include other components such as a camera, as needed in a specific application, and these components are not shown in fig. 6 and will not be described in detail since they are not the components that are important in the embodiments of the present application.
It will be appreciated by those skilled in the art that fig. 6 is merely an example of an electronic device and is not meant to be limiting and that more or fewer components than shown may be included or certain components may be combined or different components.
For convenience of description, the above parts are described as being functionally divided into modules (or units) respectively. Of course, the functions of each module (or unit) may be implemented in the same piece or pieces of software or hardware when implementing the present application.
It should be apparent that the above embodiments are merely examples for clarity of illustration and are not limiting of the embodiments. Other variations or modifications of the above teachings will be apparent to those of ordinary skill in the art. It is not necessary here nor is it exhaustive of all embodiments. While still being apparent from variations or modifications that may be made by those skilled in the art are within the scope of the application.
Claims (12)
1. A service recommendation system comprises a terminal device and a data server, wherein,
the terminal equipment is used for sending the business operation data to the data server; the service operation data is acquired based on the operation behavior of a user aiming at a target service; receiving target embedded layer characteristics returned by the data server based on the service operation data; generating service recommendation data based on the service operation data and the target embedded layer characteristics by adopting a pre-trained service recommendation target model; based on the service recommendation data, executing service recommendation operation;
the data server is used for receiving service operation data sent by the terminal equipment; performing data processing on the business operation data to obtain the characteristics of the target embedded layer; and returning the target embedded layer characteristics to the terminal equipment.
2. The system of claim 1, the data server comprising an edge node;
the terminal equipment is used for sending the service operation data to the edge node;
and the edge node is used for carrying out feature extraction on the business operation data by adopting a feature extraction model to obtain the target embedded layer features.
3. The system of claim 1 or 2, the data server comprising a central node and an edge node;
the edge node is used for obtaining the universal embedded layer characteristics based on training data and sending the universal embedded layer characteristics to the central node; the training data is acquired based on the operation behavior of the user for the service;
and the center node is used for carrying out model training on the service recommendation initial model based on the universal embedded layer characteristics and the training data to obtain the trained service recommendation target model.
4. A method of business recommendation, the method comprising:
transmitting the business operation data to a data server; the service operation data is acquired based on the operation behavior of a user aiming at a target service;
receiving target embedded layer characteristics returned by the data server based on the service operation data;
generating service recommendation data based on the service operation data and the target embedded layer characteristics by adopting a pre-trained service recommendation target model;
and executing service recommendation operation based on the service recommendation data.
5. The method of claim 4, wherein the target embedded layer features are obtained by feature extraction of the business operation data using a feature extraction model.
6. A method of business recommendation, the method comprising:
receiving service operation data sent by terminal equipment; the service operation data is acquired based on the operation behavior of a user aiming at a target service;
performing data processing on the business operation data to obtain the characteristics of the target embedded layer;
and returning the target embedded layer characteristics to the terminal equipment, so that the terminal equipment adopts a pre-trained service recommendation target model, generates service recommendation data based on the service operation data and the target embedded layer characteristics, and executes service recommendation operation based on the service recommendation data.
7. The method of claim 6, wherein the data processing the business operation data to obtain the target embedded layer feature comprises:
and adopting a feature extraction model to extract features of the business operation data to obtain the features of the target embedded layer.
8. The method of claim 6 or 7, wherein the business recommendation target model is obtained by:
obtaining universal embedded layer characteristics based on training data; the training data is acquired based on the operation behavior of the user for the service;
And carrying out model training on the service recommendation initial model based on the universal embedded layer characteristics and the training data to obtain the trained service recommendation target model.
9. An apparatus for service recommendation, the apparatus comprising:
the sending unit is used for sending the business operation data to the data server; the service operation data is acquired based on the operation behavior of a user aiming at a target service;
the receiving unit is used for receiving the target embedded layer characteristics returned by the data server based on the service operation data;
the generating unit is used for generating service recommendation data based on the service operation data and the target embedded layer characteristics by adopting a pre-trained service recommendation target model;
and the execution unit is used for executing the service recommendation operation based on the service recommendation data.
10. An apparatus for service recommendation, the apparatus comprising:
a receiving unit, configured to receive service operation data sent by a terminal device; the service operation data is acquired based on the operation behavior of a user aiming at a target service;
the processing unit is used for carrying out data processing on the business operation data to obtain the target embedded layer characteristics;
And the sending unit is used for returning the target embedded layer characteristics to the terminal equipment, so that the terminal equipment adopts a pre-trained service recommendation target model, generates service recommendation data based on the service operation data and the target embedded layer characteristics, and executes service recommendation operation based on the service recommendation data.
11. An electronic device, comprising:
a processor; and
memory storing computer instructions for causing the processor to perform the method of any one of claims 4-5 or 6-8.
12. A storage medium storing computer instructions for causing a computer to perform the method of any one of claims 4-5 or 6-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310893802.9A CN116628348A (en) | 2023-07-19 | 2023-07-19 | Service recommendation system, method, device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310893802.9A CN116628348A (en) | 2023-07-19 | 2023-07-19 | Service recommendation system, method, device, electronic equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116628348A true CN116628348A (en) | 2023-08-22 |
Family
ID=87642158
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310893802.9A Pending CN116628348A (en) | 2023-07-19 | 2023-07-19 | Service recommendation system, method, device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116628348A (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111241395A (en) * | 2020-01-07 | 2020-06-05 | 支付宝(杭州)信息技术有限公司 | Authentication service recommendation method and device |
| CN112633962A (en) * | 2020-12-03 | 2021-04-09 | 北京道隆华尔软件股份有限公司 | Service recommendation method and device, computer equipment and storage medium |
| CN114756737A (en) * | 2021-01-08 | 2022-07-15 | 腾讯科技（北京）有限公司 | Feature extraction method, device, equipment and medium applied to service recommendation |
| CN114840759A (en) * | 2022-05-11 | 2022-08-02 | 北京奇艺世纪科技有限公司 | Recommendation method and device and computer readable storage medium |
-
2023
- 2023-07-19 CN CN202310893802.9A patent/CN116628348A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111241395A (en) * | 2020-01-07 | 2020-06-05 | 支付宝(杭州)信息技术有限公司 | Authentication service recommendation method and device |
| CN112633962A (en) * | 2020-12-03 | 2021-04-09 | 北京道隆华尔软件股份有限公司 | Service recommendation method and device, computer equipment and storage medium |
| CN114756737A (en) * | 2021-01-08 | 2022-07-15 | 腾讯科技（北京）有限公司 | Feature extraction method, device, equipment and medium applied to service recommendation |
| CN114840759A (en) * | 2022-05-11 | 2022-08-02 | 北京奇艺世纪科技有限公司 | Recommendation method and device and computer readable storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 塞巴斯蒂安•拉施卡等: "《智能科学与技术丛书Python机器学习原书第3版》", 北京：机械工业出版社, pages: 378 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103024585B (en) | Program recommendation system, program recommendation method and terminal equipment | |
| CN107491560B (en) | Page picture display method and device, storage medium, server and mobile terminal | |
| CN110287412B (en) | Content recommendation method, recommendation model generation method, device, and storage medium | |
| CN102939579A (en) | Method and apparatus for binding user interface elements and granular reflective processing | |
| CN104115142B (en) | For the file specific to equipment of bookmark synchronization | |
| CN107209905A (en) | For personalized and task completion service, correspondence spends theme and sorted out | |
| CN108549567B (en) | Animation display method, device, terminal, server and storage medium | |
| CN103608811A (en) | Context aware application model for connected devices | |
| CN106534281A (en) | Data request responding method, apparatus and system | |
| Jain | Eventweb: Developing a human-centered computing system | |
| CN112307352B (en) | Content recommendation method, system, device and storage medium | |
| CN108810132B (en) | Animation display method, device, terminal, server and storage medium | |
| CN112052420A (en) | Page sharing picture generation method and device and page sharing method and device | |
| CN103677519A (en) | Method for collecting multimedia resource, terminal and server | |
| CN109462777B (en) | Video heat updating method, device, terminal and storage medium | |
| JP2022530725A (en) | Information push methods, devices, equipment and storage media | |
| CN116628348A (en) | Service recommendation system, method, device, electronic equipment and storage medium | |
| CN108874976B (en) | Search content recommendation method and device, terminal device and storage medium | |
| CN113962316A (en) | Model training method and device and electronic equipment | |
| CN113515701A (en) | Information recommendation method and device | |
| CN115687745A (en) | Multimedia data recommendation method and device, storage medium and computer equipment | |
| CN113704596A (en) | Method and apparatus for generating a set of recall information | |
| CN109561146A (en) | Document down loading method, device, terminal device | |
| CN112748835A (en) | Terminal, server, recent task list display method and application recommendation method | |
| CN111858395B (en) | Data management method and device |
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 |