A data resource intelligent recommendation system based on heat analysis
Through real-time heat calculation and predictive heat evaluation model combined with user access feature analysis and interest resource clustering, the problem of inaccurate heat analysis of data resource recommendation systems in the existing technology is solved, and the accuracy and diversity of data resource recommendations are achieved.
Patent Information
- Application Number
- CN202210582623.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-05-26
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2042-05-26
AI Technical Summary
The existing data resource recommendation system cannot accurately analyze resource popularity, resulting in a single and untargeted recommendation content, and failing to fully consider the impact of user behavior and future continued popularity.
Using an intelligent data resource recommendation system based on heat analysis, the real-time heat of data resources is calculated through the real-time heat calculation module, and the predictive heat evaluation model is used to predict heat estimate. Combined with user access feature analysis and interest resource clustering, an extended vector is generated for resources. Recommended to achieve accurate analysis of the popularity of data resources and targeted recommendation content.
It achieves accurate analysis of the popularity of data resources and a high degree of compatibility between recommended content, ensures that the recommended content is of high interest to the target users, and enriches the diversity of recommended content.
Smart Images

Figure CN114971289B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to data resource recommendation, and in particular to a data resource intelligent recommendation system based on heat analysis. Background Art
[0002] Power marketing is the process by which power companies provide power products and corresponding services that meet consumer demand in a changing market environment with the purpose of meeting people's power consumption needs. The essence of power marketing is to adjust the demand level and demand time of the power market, meet users' reasonable power demand with good service quality, achieve mutual coordination of power supply and demand, establish a good partnership between power companies and users, and encourage users to actively change their consumption behavior and power usage methods to improve power efficiency.
[0003] With the in-depth development of mobile Internet, the foundation has been laid for the widespread application of data resource recommendation. In power marketing, data resource recommendation systems are also needed to provide users with more interesting service information or facilitate users to quickly search for relevant information of their interest.
[0004] Since resource popularity can directly reflect the impact and importance of data, most data resource recommendation systems use resource popularity as a data resource recommendation standard. However, existing resource popularity analysis statistics usually determine resource popularity by counting the number of visits, without considering the impact of user behavior and future continued popularity, and the resource popularity obtained is not accurate enough. Summary of the invention
[0005] 1. Technical issues to be solved
[0006] In view of the above-mentioned shortcomings of the prior art, the present invention provides a data resource intelligent recommendation system based on heat analysis, which can effectively overcome the defects of the prior art that the data resource heat cannot be accurately analyzed and counted, and the data resource recommendation content is single and lacks pertinence.
[0007] (II) Technical solution
[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions:
[0009] A data resource intelligent recommendation system based on heat analysis, comprising a server, wherein the server obtains data resources through a data resource acquisition module, the server calculates the real-time heat of the data resources through a real-time heat calculation module, and estimates the predicted heat of the data resources based on the real-time heat of the data resources using a resource predicted heat evaluation model through a predicted heat estimation module, the server calculates the heat value of the data resources based on the real-time heat of the data resources and the predicted heat of the data resources through the resource heat calculation module, the server stores the data resources and the corresponding heat values of the data resources through a data resource storage module, and uses a resource heat sorting module to sort the stored data resources based on the heat values of the data resources, and constructs a resource heat table;
[0010] The server obtains an access feature vector based on user access information analysis through a user access feature analysis module, and determines the target resource of the target user in a resource heat table based on the access feature vector of each user through a target resource determination module. The server obtains a target resource feature vector through a target resource feature analysis module, and clusters each target resource based on the target resource feature vector through an interest resource clustering module to obtain an interest resource set. The server generates an extension vector about the interest resource set through an extension vector generation module, and uses an extension resource association module to associate external data resources based on the extension vector to obtain an extension resource set. The server recommends the extension resource set to the target user through an extension resource recommendation module.
[0011] Preferably, the real-time heat calculation module calculates the real-time heat of the data resource by the following formula:
[0012]
[0013] Among them, H1 is the real-time popularity of data resources, N1 is the number of data resources, N2 is the number of user behavior data associated with data resources, ω1 and ω2 represent the weights of the number of data resources and the number of user behavior data respectively, T' is the current time, T0 is the earliest release time of the data resource, and k>0.
[0014] Preferably, it also includes a popularity information acquisition module, which is used to acquire popularity information including the number of data resources, the number of user behavior data associated with the data resources, and the earliest release time of the data resources.
[0015] Preferably, the predicted heat estimation module inputs the real-time heat H1 of the data resource into the resource predicted heat evaluation model, and uses the output result of the resource predicted heat evaluation model as the data resource predicted heat H2;
[0016] Among them, the resource prediction heat evaluation model adopts a deep neural network DNN, and uses the data resource real-time heat-data resource prediction heat training set for model training.
[0017] Preferably, the resource heat calculation module sets weight parameters for the data resource real-time heat H1 and the data resource predicted heat H2 respectively, and performs weighted calculation on the data resource real-time heat H1 and the data resource predicted heat H2 based on the set weight parameters to obtain the data resource heat value H.
[0018] Preferably, the data resource storage module inputs the data resource whose data resource heat value is calculated into a preset resource classification model, and stores the corresponding data resource and its data resource heat value into a corresponding storage block based on the resource type, so as to update the corresponding storage block;
[0019] The data resource storage module sets corresponding storage blocks for different resource types in the system storage area, and adds storage blocks for resource types that do not have corresponding storage blocks.
[0020] Preferably, the resource heat sorting module screens the stored data resources based on the data resource heat value, selects data resources whose data resource heat value is greater than a set threshold from each storage block, performs comprehensive sorting according to the size of the data resource heat value, and enters the corresponding data resource index information into the resource heat table in sequence.
[0021] Preferably, it also includes a user access information acquisition module, the user access information acquisition module is used to obtain user access trace information of each data resource in the resource heat table, and the user access feature analysis module obtains the access feature vector of each user based on the user access trace information analysis;
[0022] The target resource determination module determines the similar user who is closest to the target user's access habits by comparing the distances between the target user and the access feature vectors of each user, and uses the frequently accessed data resources of the similar users in the resource heat table as the target resources of the target user.
[0023] Preferably, the target resource feature analysis module uses the identification information and the resource in the target resource as nodes, and the relationship between the identification information and the resource as an edge, to construct an undirected graph, and obtains multiple wandering paths on the undirected graph, and obtains the target resource feature vector of each target resource according to the multiple wandering paths;
[0024] The interest resource clustering module clusters each target resource according to the target resource feature vector to obtain an interest resource set.
[0025] Preferably, the extension vector generation module determines the target resources in the interest resource set, and processes the target resource feature vector of each target resource by a preset method to generate an extension vector for the interest resource set;
[0026] The extended resource association module associates the external data resources by comparing the distance between the extended vector of the interest resource set and the feature vector of the external data resource to obtain the extended resource set.
[0027] (III) Beneficial effects
[0028] Compared with the prior art, the data resource intelligent recommendation system based on heat analysis provided by the present invention has the following beneficial effects:
[0029] 1) The real-time heat calculation module calculates the real-time heat of data resources. The predicted heat estimation module estimates the predicted heat of data resources based on the real-time heat of data resources using the resource predicted heat evaluation model. The resource heat calculation module calculates the heat value of data resources based on the real-time heat of data resources and the predicted heat of data resources. The data resource storage module stores data resources and corresponding data resource heat values. The resource heat sorting module sorts the stored data resources based on the data resource heat values and constructs a resource heat table. By combining the real-time heat of data resources and the predicted heat of data resources, accurate analysis and statistics of the heat values of data resources are achieved, and a corresponding resource heat table is constructed based on the heat values of data resources.
[0030] 2) The user access feature analysis module obtains the access feature vector based on the user access information analysis. The target resource determination module determines the target resource of the target user in the resource heat table based on the access feature vector of each user, so as to perform targeted screening of the data resource recommendation content based on the target user's access habits, ensuring that the final recommended data resources are highly consistent with the target user's interest level;
[0031] 3) The target resource feature analysis module analyzes and obtains the target resource feature vector, the interest resource clustering module clusters each target resource based on the target resource feature vector to obtain an interest resource set, the extension vector generation module generates an extension vector about the interest resource set, and the extension resource association module associates the external data resource based on the extension vector to obtain an extension resource set, thereby being able to carry out targeted association and expansion of the data resource recommendation content based on the extension vector of the interest resource set, thereby achieving effective enrichment of the data resource recommendation content. BRIEF DESCRIPTION OF THE DRAWINGS
[0032] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0033] Figure 1 It is a schematic diagram of the system of the present invention;
[0034] Figure 2 A schematic diagram of a process for constructing a resource heat table in combination with data resource heat values in the present invention;
[0035] Figure 3 The figure is a flow chart of obtaining an extended resource set based on a resource heat table in the present invention. DETAILED DESCRIPTION
[0036] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0037] An intelligent recommendation system for data resources based on heat analysis, such as Figure 1 and Figure 2 As shown, it includes a server, which obtains data resources through a data resource acquisition module, calculates the real-time heat of the data resources through a real-time heat calculation module, and estimates the predicted heat of the data resources based on the real-time heat of the data resources using a resource prediction heat evaluation model through a prediction heat estimation module, calculates the data resource heat value based on the real-time heat of the data resources and the predicted heat of the data resources through a resource heat calculation module, stores the data resources and the corresponding data resource heat values through a data resource storage module, and uses a resource heat sorting module to sort the stored data resources based on the data resource heat values, and constructs a resource heat table.
[0038] 1) The technical solution of the present application also includes a heat information acquisition module, which is used to obtain heat information including the number of data resources, the number of user behavior data associated with the data resources, and the earliest release time of the data resources.
[0039] The real-time heat calculation module calculates the real-time heat of data resources using the following formula:
[0040]
[0041] Among them, H1 is the real-time popularity of data resources, N1 is the number of data resources, N2 is the number of user behavior data associated with data resources, ω1 and ω2 represent the weights of the number of data resources and the number of user behavior data respectively, T' is the current time, T0 is the earliest release time of the data resource, and k>0.
[0042] 2) The predicted heat estimation module inputs the real-time heat of the data resource H1 into the resource predicted heat evaluation model, and uses the output result of the resource predicted heat evaluation model as the data resource predicted heat H2.
[0043] Among them, the resource prediction heat evaluation model adopts the deep neural network DNN, and uses the data resource real-time heat-data resource prediction heat training set for model training.
[0044] 3) The resource heat calculation module sets weight parameters for the real-time heat of data resources H1 and the predicted heat of data resources H2 respectively, and performs weighted calculation on the real-time heat of data resources H1 and the predicted heat of data resources H2 based on the set weight parameters to obtain the data resource heat value H.
[0045] 4) The data resource storage module inputs the data resource whose data resource heat value is calculated into a preset resource classification model, and stores the corresponding data resource and its data resource heat value into the corresponding storage block based on the resource type to update the corresponding storage block.
[0046] The data resource storage module sets corresponding storage blocks for different resource types in the system storage area, and adds storage blocks for resource types that do not have corresponding storage blocks.
[0047] 5) The resource heat sorting module screens the stored data resources based on the data resource heat value, selects data resources whose data resource heat value is greater than the set threshold from each storage block, performs comprehensive sorting according to the size of the data resource heat value, and enters the corresponding data resource index information into the resource heat table in sequence.
[0048] In the above technical scheme, the real-time heat calculation module calculates the real-time heat of data resources, the predicted heat estimation module estimates the predicted heat of data resources based on the real-time heat of data resources using the resource predicted heat evaluation model, the resource heat calculation module calculates the data resource heat value based on the real-time heat of data resources and the predicted heat of data resources, the data resource storage module stores the data resources and the corresponding data resource heat values, the resource heat sorting module sorts the stored data resources based on the data resource heat values, and constructs a resource heat table. By combining the real-time heat of data resources and the predicted heat of data resources, accurate analysis and statistics of the data resource heat values are achieved, and a corresponding resource heat table is constructed based on the data resource heat values.
[0049] like Figure 1 and Figure 3 As shown, the server obtains an access feature vector based on user access information analysis through a user access feature analysis module, and determines the target resource of the target user in the resource heat table based on the access feature vector of each user through a target resource determination module.
[0050] The technical solution of the present application also includes a user access information acquisition module, which is used to obtain user access trace information of each data resource in the resource heat table.
[0051] The user access feature analysis module obtains the access feature vector of each user based on the user access trace information analysis; the target resource determination module determines the similar users who are closest to the target user's access habits by comparing the distance between the target user and the access feature vectors of each user, and uses the frequently visited data resources of similar users in the resource heat table as the target resources of the target user.
[0052] In the above technical solution, the user access feature analysis module obtains the access feature vector based on the user access information analysis, and the target resource determination module determines the target resource of the target user in the resource heat table based on the access feature vector of each user, so that the data resource recommendation content can be targetedly screened based on the access habits of the target user, ensuring that the final recommended data resources are highly consistent with the interest level of the target user.
[0053] like Figure 1 and Figure 3 As shown, the server obtains the target resource feature vector by analyzing through the target resource feature analysis module, and clusters each target resource based on the target resource feature vector through the interest resource clustering module to obtain the interest resource set. The server generates an extension vector about the interest resource set through the extension vector generation module, and uses the extension resource association module to associate the external data resource based on the extension vector to obtain the extension resource set. The server recommends the extension resource set to the target user through the extension resource recommendation module.
[0054] 1) The target resource feature analysis module takes the identification information and resources in the target resources as nodes, and the relationship between the table identification information and resources as edges, constructs an undirected graph, and obtains multiple wandering paths on the undirected graph. According to the multiple wandering paths, the target resource feature vectors of each target resource are obtained.
[0055] 2) The interest resource clustering module clusters each target resource according to the target resource feature vector to obtain an interest resource set.
[0056] 3) The extension vector generation module determines the target resources in the interest resource set, and processes the target resource feature vector of each target resource by a preset method to generate an extension vector for the interest resource set.
[0057] 4) The extended resource association module associates the external data resources by comparing the distance between the extended vector of the interest resource set and the feature vector of the external data resource to obtain the extended resource set.
[0058] In the above technical scheme, the target resource feature analysis module analyzes to obtain the target resource feature vector, the interest resource clustering module clusters each target resource based on the target resource feature vector to obtain the interest resource set, the extension vector generation module generates an extension vector about the interest resource set, and the extension resource association module associates the external data resource based on the extension vector to obtain the extension resource set, thereby being able to carry out targeted association and expansion of the data resource recommendation content based on the extension vector of the interest resource set, thereby achieving effective enrichment of the data resource recommendation content.
[0059] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A data resource intelligent recommendation system based on heat analysis, characterized in that: The server includes a server, wherein the server obtains data resources through a data resource acquisition module, the server calculates the real-time heat of the data resources through a real-time heat calculation module, and estimates the predicted heat of the data resources based on the real-time heat of the data resources using a resource predicted heat evaluation model through a predicted heat estimation module, the server calculates the heat value of the data resources based on the real-time heat of the data resources and the predicted heat of the data resources through the resource heat calculation module, the server stores the data resources and the corresponding heat values of the data resources through a data resource storage module, and uses a resource heat sorting module to sort the stored data resources based on the heat values of the data resources, and constructs a resource heat table; The server obtains an access feature vector based on user access information analysis through a user access feature analysis module, and determines the target resource of the target user in the resource heat table based on the access feature vector of each user through a target resource determination module. The server obtains a target resource feature vector through a target resource feature analysis module, and clusters each target resource based on the target resource feature vector through an interest resource clustering module to obtain an interest resource set. The server generates an extension vector about the interest resource set through an extension vector generation module, and uses an extension resource association module to associate external data resources based on the extension vector to obtain an extension resource set. The server recommends the extension resource set to the target user through an extension resource recommendation module. The real-time heat calculation module calculates the real-time heat of data resources by the following formula: Among them, H1 is the real-time popularity of data resources, N1 is the number of data resources, N2 is the number of user behavior data associated with data resources, ω1 and ω2 represent the weights of the number of data resources and the number of user behavior data respectively, T' is the current time, T0 is the earliest release time of the data resource, and k>0; Also included is a popularity information acquisition module, which is used to acquire popularity information including the number of data resources, the number of user behavior data associated with the data resources, and the earliest release time of the data resources; The prediction heat estimation module inputs the real-time heat H1 of the data resource into the resource prediction heat evaluation model, and uses the output result of the resource prediction heat evaluation model as the data resource prediction heat H2; The resource prediction heat evaluation model adopts a deep neural network DNN, and uses a data resource real-time heat-data resource prediction heat training set for model training; The resource heat calculation module sets weight parameters for the data resource real-time heat H1 and the data resource predicted heat H2 respectively, and performs weighted calculation on the data resource real-time heat H1 and the data resource predicted heat H2 based on the set weight parameters to obtain the data resource heat value H.
2. The data resource intelligent recommendation system based on heat analysis according to claim 1 is characterized by: The data resource storage module inputs the data resource whose data resource heat value is calculated into a preset resource classification model, and stores the corresponding data resource and its data resource heat value into the corresponding storage block based on the resource type, so as to update the corresponding storage block; The data resource storage module sets corresponding storage blocks for different resource types in the system storage area, and adds storage blocks for resource types that do not have corresponding storage blocks.
3. The data resource intelligent recommendation system based on heat analysis according to claim 2 is characterized by: The resource heat sorting module screens the stored data resources based on the data resource heat value, selects data resources whose data resource heat value is greater than a set threshold from each storage block, performs comprehensive sorting according to the size of the data resource heat value, and enters the corresponding data resource index information into the resource heat table in sequence.
4. The data resource intelligent recommendation system based on heat analysis according to claim 1 is characterized by: It also includes a user access information acquisition module, which is used to obtain user access trace information of each data resource in the resource heat table, and the user access feature analysis module obtains the access feature vector of each user based on the user access trace information analysis; The target resource determination module determines the similar user who is closest to the target user's access habits by comparing the distances between the target user and the access feature vectors of each user, and uses the frequently accessed data resources of the similar users in the resource heat table as the target resources of the target user.
5. The data resource intelligent recommendation system based on heat analysis according to claim 4 is characterized in that: The target resource feature analysis module uses the identification information and resources in the target resource as nodes, and the relationship between the identification information and the resources as edges, to construct an undirected graph, and obtains multiple wandering paths on the undirected graph, and obtains the target resource feature vector of each target resource according to the multiple wandering paths; The interest resource clustering module clusters each target resource according to the target resource feature vector to obtain an interest resource set.
6. The data resource intelligent recommendation system based on heat analysis according to claim 5 is characterized by: The extension vector generation module determines the target resources in the interest resource set, and processes the target resource feature vector of each target resource by a preset method to generate an extension vector for the interest resource set; The extended resource association module associates the external data resources by comparing the distance between the extended vector of the interest resource set and the feature vector of the external data resource to obtain the extended resource set.
Citation Information
Patent Citations
Cache allocation method and device
CN102043732A
Multimedia content hotspot analysis method and multimedia content hotspot analysis device
CN106227883A
Data popularity statistical method and device based on deep learning
CN112883267A
Data popularity analysis method and device, equipment and storage medium
CN113792084A
