An information matching method, device and equipment, storage medium and program product

Abstract

Embodiments of the present application disclose a kind of information matching method, device and equipment, storage medium, program product, the method comprises: according to original data graph and label library, the graph network of each group in group set and the graph network of each theme in theme set are constructed;The first feature information of the graph network of first group and the second feature information of the graph network of first theme are acquired, first feature information includes one or more of global feature information, object matching information, label matching information and structure matching information of first group, second feature information includes one or more of global feature information, project matching information, label matching information and structure matching information of first theme;The matching degree between first group and first theme is determined based on first feature information and second feature information, and based on matching degree, first theme includes the project of sending to the object included in first group.It can improve the accuracy of information matching by using the present application.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to an information matching method, apparatus, device, storage medium, and program product. Background Technology

[0002] With the rapid development of internet technology, projects on different themes can be matched with individuals in different groups online. For example, music-related projects can be matched with individuals in music-related groups, and game-related projects can be matched with individuals in game-related groups. While these technologies utilize the overall similarity between the theme and the group during information matching, this method lacks sufficient accuracy and fails to meet the needs of the target audience. Therefore, improving the accuracy of information matching is a pressing issue that needs to be addressed. Summary of the Invention

[0003] This application provides an information matching method, apparatus, device, storage medium, and program product, which can improve the accuracy of information matching.

[0004] On the one hand, embodiments of this application provide an information matching method, including:

[0005] Based on the original data graph and the tag library, construct the graph network of each group in the group set and the graph network of each topic in the topic set. The original data graph includes the relationship between each object in each group and each item in each topic.

[0006] Obtain the first feature information of the graph network of the first group and the second feature information of the graph network of the first topic. The first feature information includes one or more of the global feature information, object matching information, label matching information and structure matching information of the first group. The second feature information includes one or more of the global feature information, item matching information, label matching information and structure matching information of the first topic. The first group is any one of the groups included in the group set, and the first topic is any one of the topics included in the topic set.

[0007] Based on the first feature information and the second feature information, the matching degree between the first group and the first topic is determined, and the items included in the first topic are sent to the objects included in the first group based on the matching degree.

[0008] On the other hand, embodiments of this application provide an information matching device, including:

[0009] The acquisition unit is used to acquire the raw data graph and the tag library. The raw data graph includes the relationship between each object in each group and each item in each topic.

[0010] The processing unit is configured to construct a graph network for each group in the group set and a graph network for each topic in the topic set based on the original data graph and the tag library; obtain first feature information of the graph network of the first group and second feature information of the graph network of the first topic, wherein the first feature information includes one or more of global feature information, object matching information, tag matching information and structure matching information of the first group, and the second feature information includes one or more of global feature information, item matching information, tag matching information and structure matching information of the first topic; the first group is any one of the groups included in the group set, and the first topic is any one of the topics included in the topic set; and determine the matching degree between the first group and the first topic based on the first feature information and the second feature information.

[0011] The sending unit is used to send items included in the first topic to the objects included in the first group based on the matching degree.

[0012] Accordingly, this application provides a computer device, including: a memory, a processor, and a network interface. The processor is connected to the memory and the network interface. The network interface is used to provide network communication functions, the memory is used to store program code, and the processor is used to call the program code to execute the method in this application embodiment.

[0013] Accordingly, embodiments of this application provide a computer-readable storage medium, including: a computer program stored in the computer-readable storage medium, which, when executed by a processor, implements the method in the embodiments of this application.

[0014] Accordingly, embodiments of this application provide a computer program product or computer program that includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads and executes the computer instructions from the computer-readable storage medium, causing the computer device to perform the method described in this application.

[0015] By implementing the embodiments of this application, feature information of the group and the topic can be obtained from multiple aspects based on the graph network of the group and the graph network of the topic. Specifically, global feature information, object matching information, tag matching information, and structure matching information of the group, and global feature information, item matching information, tag matching information, and result matching information of the topic can be obtained. Then, based on the feature information of the group and the topic, the matching degree between the group and the topic can be determined, and items in the topic can be sent to objects in the group according to the matching degree. In this way, the complex attribute interactions between the group and the topic can be considered when performing information matching, thereby improving the accuracy of information matching. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the architecture of an information matching system provided in an embodiment of this application;

[0018] Figure 2 This is a flowchart illustrating an information matching method provided in an embodiment of this application;

[0019] Figure 3a This is a flowchart illustrating another information matching method provided in an embodiment of this application;

[0020] Figure 3b This is a flowchart illustrating another information matching method provided in an embodiment of this application;

[0021] Figure 3c This is a flowchart illustrating another information matching method provided in an embodiment of this application;

[0022] Figure 3d This is a flowchart illustrating another information matching method provided in an embodiment of this application;

[0023] Figure 3e This is a flowchart illustrating another information matching method provided in an embodiment of this application;

[0024] Figure 4 This is a flowchart illustrating another information matching method provided in an embodiment of this application;

[0025] Figure 5a This is a schematic diagram of a dataset used to verify the information matching method provided in an embodiment of this application;

[0026] Figure 5b This is a schematic diagram comparing the verification results of the information matching method provided in the embodiments of this application;

[0027] Figure 6 This is a schematic diagram of the structure of an information matching device provided in an embodiment of this application;

[0028] Figure 7 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0029] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0030] The information matching method provided in this application first constructs a graph network for each group in a group set and a graph network for each topic in a topic set based on the original data graph and tag library. Then, it obtains first feature information of the graph network of the first group and second feature information of the graph network of the first topic. The first feature information may include one or more of global feature information of the first group, object matching information, tag matching information, and structure matching information. The second feature information may include one or more of global feature information of the first topic, item matching information, tag matching information, and structure matching information. Finally, it determines the matching degree between the first group and the first topic based on the first and second feature information, and sends items from the first topic to objects in the first group based on the matching degree. Thus, by enriching the feature information of each group and each topic and fully considering the interaction relationship between groups and topics, it can more accurately send items from a specific topic to items in a specific group, improving the accuracy of information matching.

[0031] Please see Figure 1 , Figure 1 This is a schematic diagram of the architecture of an information matching system provided in an embodiment of this application. For example... Figure 1 As shown, the architecture of this information matching system includes a terminal device 10 and a server 12. The terminal device 10 and the server 12 can communicate through a communication network 11. The communication network 11 can be a wired network or a wireless network. The terminal device 10 and the server 12 can be directly or indirectly connected through wired or wireless communication methods, and this application does not impose any restrictions on this.

[0032] In this embodiment, the terminal device 10 can install an application related to information matching. The application can be software, a webpage, a mini-program, etc. The server 12 is the server corresponding to the software, webpage, or mini-program. The information matching method provided in this embodiment can be implemented independently by the terminal device 10, independently by the server 12, or jointly by the terminal device 10 and the server 12; no specific limitation is made here.

[0033] In one embodiment, the information matching method provided in this application embodiment is illustrated by example, where terminal device 10 and server 12 collaboratively implement the method. Figure 1As shown, terminal device 10 can run applications related to information matching, such as instant messaging applications, live streaming applications, and social media applications. When an object opens an application related to information matching running on terminal device 10, terminal device 10 can receive the object's information matching request. After receiving the object's information matching request, terminal device 10 can send the information matching request to server 12 via the network. Server 12 responds to the information matching request, obtains the raw data, and constructs a raw data graph and tag library based on the raw data. The raw data includes the access relationships between each group in the group set to which the object belongs and each topic in the topic set. During information matching, server 12 constructs a graph network for each group in the group set and a graph network for each topic in the topic set based on the original data graph and tag library. It obtains first feature information of the graph network of the first group and second feature information of the graph network of the first topic. The first group may be the group to which an object belongs; in other embodiments, the first group may be a group directly or indirectly associated with the object. The first feature information may include one or more of the following: global feature information of the first group, object matching information, tag matching information, and structure matching information. The second feature information may include one or more of the following: global feature information of the first topic, item matching information, tag matching information, and structure matching information. Based on the first and second feature information, the matching degree between the first group and the first topic can be determined, and items from the first topic are sent to objects in the first group based on the matching degree. Accordingly, objects in the first group can receive items from the first topic sent by server 12 using terminal device 10.

[0034] It should be noted that the terminal devices involved in the embodiments of this application can be personal computers, laptops, smartphones, tablets, smartwatches, smart voice interaction devices, smart home appliances, vehicle terminals, and smart wearable devices, but are not limited to these. The server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0035] The embodiments of this application can be applied to various scenarios, including but not limited to cloud technology, artificial intelligence, smart transportation, and assisted driving.

[0036] It is understood that in the specific implementation of this application, data such as raw data and tag libraries are involved. When the above embodiments of this application are applied to specific products or technologies, the relevant data need to obtain relevant licenses or consents, and the collection, use and processing of the relevant data need to comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0037] For further details, please see Figure 2 , Figure 2 This is a flowchart illustrating an information matching method provided in an embodiment of this application. This information matching method can be executed by a computer device, which can be a terminal device or a server. For ease of understanding, this embodiment of the application will illustrate the method by example, where the method is executed by a server. Figure 2 As shown, the method may include, but is not limited to, the following steps.

[0038] S201: Construct a graph network for each group in the group set and a graph network for each topic in the topic set based on the original data graph and the tag library.

[0039] In this application, the server can construct a graph network for each group in the group set and a graph network for each topic in the topic set based on the original data graph and the tag library. The original data graph includes the relationships between each object in each group and each item in each topic; the graph network for each group includes the relationships between each object in each group and the tags in the tag library, as well as the connections between each group and the objects it contains; the graph network for each topic includes the relationships between each item in each topic and the tags in the tag library, as well as the connections between each topic and the items it contains.

[0040] In one embodiment, a group can consist of multiple objects with shared interests; for example, objects interested in music can form one group, and objects interested in games can form another. A topic can consist of multiple content-related items, such as news, music, videos, etc. Specifically, assuming the topic is weather, this weather topic can consist of multiple weather-related news items, such as a one-week weather forecast or the daily UV index. In one embodiment, if there is a connection between objects and items in the original data graph, it means that there is an association between the objects and items. An association between objects and items can mean that an object has viewed an item or searched for an item. The tag library can include tags for items that are associated with each object in the group; tags can be, for example, keywords, types, or names of items. The graph network of each group and the graph network of each topic, constructed based on the original data graph and the tag library, can interact through tags.

[0041] S202: Obtain the first feature information of the graph network of the first group and the second feature information of the graph network of the first topic.

[0042] In this embodiment of the application, the server can obtain first feature information of the graph network of the first group and second feature information of the graph network of the first topic. The first feature information may include one or more of the global feature information of the first group, object matching information, label matching information and structure matching information. The second feature information may include one or more of the global feature information of the first topic, item matching information, label matching information and structure matching information. The first group may be any one of the groups included in the group set, and the first topic may be any one of the topics included in the topic set.

[0043] In one embodiment, the first feature information of the graph network of the first group can be the feature representation of the graph network of the first group; the global feature information of the first group can be the feature representation of the first group, which can be obtained by aggregating the feature representations of each object in the first group; the object matching information of the first group can be obtained by matching the feature representation of each object in the first group with the feature representation of the first topic; the tag matching information of the first group can be obtained by matching the feature representation of the tag corresponding to the first topic with the feature representation of the first group; the global structure information of the first group can be obtained by matching the feature representation of the first average popularity of each topic associated with the first group with the feature representation of the first target popularity of the first topic, where the first average popularity can refer to the average number of visits to each topic associated with the first group, and the feature representation of the first average popularity can be obtained by transforming the first average popularity; the first target popularity can refer to the number of visits to the first topic, and the feature representation of the first target popularity can be obtained by transforming the first target popularity. The feature representation of the graph network in the first group can be obtained by freely combining any one or more of the global feature information, object matching information, label matching information and structure matching information of the first group. That is, the scale dimension of the feature representation of the graph network in the first group can be set.

[0044] Accordingly, the second feature information of the graph network of the first topic can be the feature representation of the graph network of the first topic; the global feature information of the first topic can be the feature representation of the first topic, which can be obtained by aggregating the feature representations of each item in the first topic; the item matching information of the first topic can be obtained by matching the feature representation of each item in the first topic with the feature representation of the first group; the tag matching information of the first topic can be obtained by matching the feature representation of the tag corresponding to the first group with the feature representation of the first topic; the global structure information of the first topic can be obtained by matching the feature representation of the second average popularity of each group that is related to the first topic with the feature representation of the second target popularity of the first group. The second average popularity can refer to the average number of visits to each item in each group that is related to the first topic. The feature representation of the second average popularity can be obtained by transforming the second average popularity; the second target popularity can refer to the number of visits to the items in the first group. The feature representation of the second target popularity can be obtained by transforming the second target popularity. The feature representation of the graph network for the first topic can be obtained by freely combining any one or more of the global feature information, item matching information, label matching information, and structure matching information of the first topic. That is, the scale dimension of the feature representation of the graph network for the first topic can be set.

[0045] S203: Determine the matching degree between the first group and the first topic based on the first feature information and the second feature information, and send the items included in the first topic to the objects included in the first group based on the matching degree.

[0046] In this embodiment, the server can determine the matching degree between the first group and the first topic based on the first feature information and the second feature information, and send the items included in the first topic to the objects included in the first group based on the matching degree. The matching degree between the first group and the first topic can be a matching score between the first group and the first topic. If the matching score is greater than a threshold, that is, the matching degree is high, then the items included in the first topic can be sent to the objects included in the first group.

[0047] In one embodiment, after determining the matching score between the first group and each topic in the topic set, the matching scores between the first group and each topic in the topic set are sorted from largest to smallest. Then, according to the sorting of matching scores, a specific number of items included in the topics that are ranked before the sorting are sent to the objects included in the first group.

[0048] By implementing the embodiments of this application, for the graph network of the first group and the graph network of the first topic, first feature information of the first group and second feature information of the first topic can be constructed based on the combination of feature information of different dimensions. The matching degree between the first feature information of the first group and the second feature information of the first topic is determined, and items included in the first topic are sent to the objects included in the first group according to the matching degree. Thus, since the first feature information of the first group and the second feature information of the first topic can integrate feature information of different dimensions, both the first feature information of the first group and the second feature information of the first topic are richer. At the same time, the complex attribute interactions between the first group and the first topic are also taken into account, thereby improving the accuracy of information matching.

[0049] For further details, please see Figure 3a , Figure 3a This is a flowchart illustrating another information matching method provided in an embodiment of this application. This information matching method can be executed by a computer device, which can be a terminal device or a server. For ease of understanding, this embodiment of the application uses the method executed by a server as an example for illustration. Figure 3a As shown, the method may include, but is not limited to, the following steps.

[0050] S301: Obtain raw data.

[0051] In this embodiment of the application, the server can obtain raw data, which includes the access relationship between each group in the group set and each topic in the topic set, that is, whether the objects included in each group have browsed the items included in each topic, or whether the objects included in each group have clicked the access link corresponding to the items included in each topic, or whether the objects included in each group have searched the items included in each topic.

[0052] S302: Construct the original data graph based on the access relationships between each group in the group set and each topic in the topic set.

[0053] In this embodiment, the server, based on the access relationships between each group in the group set and each topic in the topic set, can determine the association relationships between each object in each group and each item in each topic, thereby constructing an original data graph. In the original data graph, there are connections between each group and the objects included in each group, connections between each topic and the items included in each topic, and connections between the objects included in each group and the items associated with those objects.

[0054] S303: Get the tags of each item accessed by each object in each group.

[0055] In this embodiment of the application, the server can obtain the tags of each item accessed by each object in each group.

[0056] S304: Build a tag library based on the tags of each item accessed by each object in each group.

[0057] In this embodiment of the application, the server can build a tag library based on the tags of each item accessed by each object in each group.

[0058] S305: Construct a graph network for each group in the group set and a graph network for each topic in the topic set based on the original data graph and the tag library.

[0059] In this embodiment, the server can construct a graph network for each group in the group set and a graph network for each topic in the topic set based on the original data graph and the tag library. In the graph network of each group, there are connections between groups and objects included in groups, connections between the tags of items associated with objects, connections between topics and items included in topics, and connections between the tags of items. Therefore, the graph network of each group and the graph network of each topic can interact through tags.

[0060] S306: Obtain feature information for each object included in the graph network of the first group.

[0061] In this embodiment of the application, the server can obtain the feature information of each object included in the graph network of the first group. The first group can be any one of the groups included in the group set. The feature information of each object can be the feature representation of each object. The feature representation of each object can be obtained based on the feature representation of the label of the item associated with the object.

[0062] In one embodiment, the feature representation of each object can be obtained by the following formula (1-1):

[0063] Z u =f AGG (x i :i∈I u ), u∈V c (1-1);

[0064] Among them, Z u It is a characteristic representation of an object, f AGG It is an aggregate function, x i It is the feature representation of the items associated with the object. The feature representation of the items associated with the object can be obtained from the feature representation of the item's label. The feature representation of the label can be, for example, the word embedding vector corresponding to the label. u It is the set of labels for items selected for aggregation to obtain a feature representation of an object, Vc It is the set of objects and labels included in the graph network of the first group.

[0065] In one embodiment, the aggregation function can be either a summation function or an average function. Items associated with each object can be... Figure 3b Steps S31-S34 shown are determined.

[0066] S31: Obtain the first set of items browsed for each object in the graph network of the first group.

[0067] In this embodiment of the application, the server can obtain a first set of items browsed by each object in the graph network of the first group. The first set of items includes items belonging to the first topic and items that do not belong to the first topic.

[0068] S32: Determine the relevance of each item in the second set of items that does not belong to the first topic to the first topic.

[0069] In this embodiment of the application, items in the first item set that do not belong to the first topic are identified as the second item set, and then the server can determine the relevance between each item in the second item set and the first topic.

[0070] In one embodiment, the server may determine the relevance between each item in the second item set and the first topic in the following manner:

[0071] Method 1: Calculate the cosine similarity score between each item in the second item set and each item in the first item set belonging to the first topic, and determine the cosine similarity score as the relevance between the items in the second item set and the first topic.

[0072] Method 2: Count how many common browsing objects each item in the second item set shares with each item in the first item set belonging to the first theme, thereby determining the closeness between the two items.

[0073] Specifically, such as Figure 3cAs shown, the first set of items browsed by object B includes three items: item a, item b, and item c. Items b and c belong to the first topic (topic H), while item a is an item in the second set of items that does not belong to the first topic. If we use the first method described above to determine the relevance between item a in the second set and the first topic, we can calculate the cosine similarity scores between item a and items b and c respectively. If we use the second method described above to determine the relevance between item a and the first topic, we can count how many common browsing objects item a shares with items b and c. For example, item a and item b share two common browsing objects, item A and item B, so item a and item b have a certain degree of similarity, that is, they have a high degree of relevance to the first topic.

[0074] S33: Determine the target project from the second project set based on relevance.

[0075] In this embodiment of the application, the server can identify projects in the second project set that have a high degree of relevance to the first topic as target projects. The high degree of relevance can refer to the cosine similarity score between the projects in the second project set and the projects in the first topic exceeding a threshold, or it can refer to the number of common browsing objects between the projects in the second project set and the projects in the first topic exceeding a threshold.

[0076] S34: Treat the target project and the projects included in the first topic as individual projects that are associated with each object.

[0077] In this embodiment of the application, the determined target project and the projects included in the first subject can be regarded as the various projects associated with each object.

[0078] In this way, items associated with an object can be identified, and the labels of these associated items can be aggregated to obtain a feature representation of the object. This enriches the feature representation of the object and makes it more accurate.

[0079] S307: Obtain feature information for each item included in the graph network of the first topic.

[0080] In this embodiment of the application, the server can obtain the feature information of each item included in the graph network of the first topic. The first topic can be any one of the topics included in the topic set. The feature information of each item can be the feature representation of each item. The feature representation of each item can be obtained based on the feature representation of the item's label.

[0081] In one embodiment, the feature representation of each item can be obtained by the following formula (1-2):

[0082] Zi =f AGG (y i :i∈I i ), i∈V h (1-2);

[0083] Among them, Z i It is a characteristic representation of the project, f AGG It is an aggregate function, y i It is a characteristic representation of each tag in the project, I i It is an aggregation of tags to obtain a feature representation of the project, V h It is the collection of items and tags included in the graph network of the first theme.

[0084] S308: Determine the global feature information of the first group based on the feature information of each object.

[0085] In this embodiment of the application, the server can determine the global feature information of the first group based on the feature information of each object.

[0086] In one embodiment, the global feature representation of the first group can be determined based on the feature representation of each object, and the global feature representation of the first group can be obtained by the following formula (1-3):

[0087] Z c =PROPAGATE(Z u G c ;K=1) (1-3);

[0088] Among them, Z u It is the feature representation of each object, which can be obtained through formula (1-1), where PROPAGATE is the propagation function, and G... c Used to guide the direction of message passing, K=1 means to aggregate one layer.

[0089] S309: Determine the global feature information of the first topic based on the feature information of each project.

[0090] In this embodiment of the application, the server can determine the global feature information of the first topic based on the feature information of each project.

[0091] In one embodiment, the global feature representation of the first topic can be determined based on the feature representation of each item, and the global feature representation of the first topic can be obtained by the following formula (1-4):

[0092] Z h =PROPAGATE(Z i G h ;K=1) (1-4);

[0093] Among them, Z i It is a feature representation of each project, which can be obtained through formula (1-2), where PROPAGATE is the propagation function, and G... h Used to guide the direction of message passing, K=1 means to aggregate one layer.

[0094] S310: Based on the feature information of each object included in the graph network of the first group and the global feature information of the first topic, determine the matching information between each object and the first topic.

[0095] In this embodiment of the application, based on the feature information of each object included in the graph network of the first group and the global feature information of the first topic, the server can determine the matching information between each object and the first topic.

[0096] In one embodiment, the feature representation of each object included in the graph network of the first group can be matched with the global feature representation of the first topic using the following formulas (1-5) to obtain the matching information between each object and the first topic:

[0097] Z uh =Z u *Z h ,u∈V c (1-5);

[0098] Among them, Z u The feature representation of each object in the graph network of the first group can be obtained by formula (1-1), Z h This is the global feature representation of the first theme, which can be obtained through formula (1-4), where * represents the multiplication of corresponding bits, V c It is the set of objects and labels included in the graph network of the first group.

[0099] This matching method, compared to directly aggregating the feature representations of each object to obtain the feature representation of the first group, further considers the matching situation between each object and the topic, and is more representative of the object's preferences.

[0100] S311: Based on the feature information of each item in the graph network of the first topic and the global feature information of the first group, determine the matching information between each item and the first group.

[0101] In this embodiment of the application, based on the feature information of each item in the graph network of the first topic and the global feature information of the first group, the server can determine the matching information of each object with the first topic.

[0102] In one embodiment, the feature representation of each item in the graph network of the first topic can be matched with the global feature representation of the first group using the following formulas (1-6) to obtain the matching information between each item and the first group:

[0103] Z ic =Z i *Z c ,i∈V h (1-6);

[0104] Among them, Z i The feature representation of each item in the graph network of the first theme can be obtained by formula (1-2), Z. c This is the global feature representation of the first group, which can be obtained through formula (1-3), where * represents the multiplication of corresponding bits, and V h It is the collection of items and tags included in the graph network of the first theme.

[0105] S312: Determine the object matching information of the first group based on the matching information of each object with the first topic.

[0106] In this embodiment of the application, the object matching information of the first group can be determined based on the matching information of each object with the first topic.

[0107] In one embodiment, the object matching information for the first group can be obtained using the following formula (1-7):

[0108]

[0109] Where ∑ represents addition, Z uh The object matching information of a group can be obtained by adding the matching information of each object in the graph network of the first group to the matching information of the first topic, as shown by formula (1-5).

[0110] S313: Determine the project matching information for the first topic based on the matching information of each project with the first group.

[0111] In this embodiment of the application, the matching information of the first topic can be determined based on the matching information of each item and the first group.

[0112] In one embodiment, the project matching information for the first topic can be obtained using the following formula (1-8):

[0113]

[0114] Where ∑ represents addition, Z icThe matching information of the first topic can be obtained by adding the matching information of each item in the graph network of the first topic to the matching information of the first group, which can be obtained by formula (1-6).

[0115] S314: Obtain the first tag set corresponding to the first topic based on the graph network of the first topic.

[0116] In this embodiment of the application, the server can obtain a first tag set corresponding to the first topic based on the graph network of the first topic. The first tag set includes the tags of each item in the first topic.

[0117] S315: Obtain the second label set corresponding to the first group based on the graph network of the first group.

[0118] In this embodiment of the application, the server can obtain a second tag set corresponding to the first group based on the graph network of the first group. The second tag set includes the tags of the items browsed by each object in the first group.

[0119] S316: Determine the label matching information of the first group based on each label included in the first label set and the global feature information of the first group.

[0120] In this embodiment of the application, based on each tag included in the first tag set and the global feature information of the first group, the server can determine the tag matching information of the first group.

[0121] In one embodiment, the feature representation of each label included in the first label set can be matched with the global feature representation of the first group using the following formulas (1-9) to determine the label matching information of the first group:

[0122]

[0123] Where ∑ represents addition, Z tag Z is the feature representation of each label in the first label set corresponding to the first topic. c This is the global feature representation of the first group, * represents the bitwise multiplication, V h This is the set of items and labels included in the graph network of the first topic. By multiplying the feature representation of each label in the first label set corresponding to the first topic with the global feature representation of the first group bit by bit and then adding them together, we can obtain the label matching information of the first group.

[0124] S317: Based on each tag included in the second tag set and the global feature information of the first topic, determine the tag matching information of the first topic.

[0125] In this embodiment of the application, based on each tag included in the second tag set and the global feature information of the first topic, the server can determine the tag matching information of the first topic.

[0126] In one embodiment, the feature representation of each tag included in the second tag set can be matched with the global feature representation of the first topic using the following formula (1-10) to determine the tag matching information of the first topic:

[0127]

[0128] Where ∑ represents addition, Z tag Z is the feature representation of each label in the second label set corresponding to the first group. h This is the global feature representation of the first topic, * represents the bitwise multiplication, V c This is the set of objects and labels included in the graph network of the first group. By multiplying the feature representation of each label in the second label set corresponding to the first group with the global feature representation of the first topic bit by bit and then adding them together, we can obtain the label matching information of the first topic.

[0129] S318: Obtain the first average popularity feature information corresponding to each topic that is related to the first group, and the first target popularity feature information of the first topic.

[0130] In this embodiment of the application, the server can obtain the first average popularity feature information corresponding to each topic that is associated with the first group and the first target popularity feature information of the first topic based on the original data graph. The first average popularity feature information is obtained based on the number of visits to each topic and the first target popularity feature information is obtained based on the number of visits to the first topic.

[0131] In one embodiment, such as Figure 3d As shown, the first average popularity feature information corresponding to each topic that is related to the first group, and the first target popularity feature information of the first topic can be obtained through steps S35-S38.

[0132] S35: Obtain the first association matrix between the first group and the topic set.

[0133] In this embodiment, the server can obtain a first association matrix between the first group and the topic set based on the original data graph. This first association matrix includes the association between each object in the first group and each item in each topic set. The elements in the first association matrix can take values ​​of 0 or 1. If there is an association between an object and an item, the corresponding element in the first association matrix takes the value 1; if there is no association between an object and an item, the corresponding element in the first association matrix takes the value 0.

[0134] S36: Obtain the first popularity feature information matrix corresponding to the topic set.

[0135] In this embodiment of the application, the server can obtain a first popularity feature information matrix corresponding to the topic set based on the original data graph. The first popularity feature information matrix includes the popularity feature information of each topic in the topic set, and the popularity feature information of each topic can be obtained based on the number of visits to each topic.

[0136] In one embodiment, after obtaining the number of visits for each topic, the number of visits for each topic can be converted into a feature representation of the number of visits. Accordingly, the first popularity feature information matrix can be obtained by concatenating the feature representations of the number of visits for each topic.

[0137] S37: Based on the first correlation matrix and the first popularity feature information matrix, determine the first average popularity feature information corresponding to each topic that is related to the first group.

[0138] In this embodiment, the server can determine the first average popularity feature information corresponding to each topic associated with the first group based on the obtained first association matrix and first popularity feature information matrix. This first average popularity feature information can be a feature representation of the first average popularity obtained by averaging the first association matrix and the first popularity feature information matrix.

[0139] S38: Determine the first target popularity feature information of the first topic based on the number of visits to the first topic.

[0140] In this embodiment of the application, the server can determine the first target popularity feature information of the first topic based on the number of accesses to the first topic. The first target popularity feature information can be a feature representation of the first target popularity.

[0141] In one embodiment, after obtaining the number of visits to the first topic, the number of visits to the first topic can be transformed to obtain a feature representation of the number of visits to the first topic. This feature representation of the number of visits to the first topic can be determined as a feature representation of the first target popularity of the first topic.

[0142] S319: Obtain the second average popularity feature information corresponding to each group that is related to the first topic, and the second target popularity feature information of the first group.

[0143] In this embodiment of the application, the server can obtain the second average popularity feature information corresponding to each group that is related to the first topic, and the second target popularity feature information of the first group based on the original data graph. The second average popularity feature information is obtained based on the number of visits to the items in each group, and the second target popularity feature information is obtained based on the number of visits to the items in the first group.

[0144] In one embodiment, such as Figure 3e As shown, the second average popularity feature information corresponding to each group that is related to the first topic, and the second target popularity feature information of the first group, can be obtained through steps S39-S41.

[0145] S39: Obtain the second association moment between the first topic and the group set.

[0146] In this embodiment, the server can obtain a second association matrix between the first topic and the group set based on the original data graph. This second association matrix includes the association relationships between each item in the first topic and the objects in each group of the group set. The values ​​of the elements in the second association matrix can be 0 or 1. If there is an association between an item and an object, the corresponding element in the second association matrix is ​​1; if there is no association between an item and an object, the corresponding element in the second association matrix is ​​0.

[0147] S40: Obtain the second popularity feature information matrix corresponding to the group set.

[0148] In this embodiment of the application, the server can obtain a second popularity feature information matrix corresponding to the group set based on the original data graph. The second popularity feature information matrix includes the popularity feature information of each group in the group set. The popularity feature information of each group is obtained based on the number of visits to the items of the objects included in each group.

[0149] In one embodiment, after obtaining the number of visits to the items of each group, the number of visits to the items of each group can be converted into a feature representation of the number of visits. The second popularity feature information matrix can be obtained by concatenating the feature representations of the number of visits to the items of each group.

[0150] S41: Based on the second correlation matrix and the second popularity feature information matrix, determine the second average popularity feature information corresponding to each group that is related to the first topic.

[0151] In this embodiment, the server can determine the second average popularity feature information corresponding to each group that is associated with the first topic based on the obtained second association matrix and second popularity feature information matrix. This second average popularity feature information can be a feature representation of the second average popularity obtained by averaging the second association matrix and the second popularity feature information matrix.

[0152] S42: Determine the second target popularity feature information of the first group based on the number of times the objects included in the first group are accessed.

[0153] In this embodiment of the application, the server can determine the second target popularity feature information of the first group based on the number of accesses to the items of the objects included in the first group. The second target popularity feature information can be a feature representation of the second target popularity.

[0154] In one embodiment, after obtaining the number of visits to the items of the objects included in the first group, the number of visits to the items of the objects included in the first group can be transformed to obtain a feature representation of the number of visits to the items of the objects included in the first group. This feature representation of the number of visits to the items of the objects included in the first group can be determined as a feature representation of the second target popularity of the first group.

[0155] S320: Determine the structural matching information of the first group based on the first average heat feature information and the first target heat feature information.

[0156] In this embodiment of the application, the server can determine the structural matching information of the first group based on the first average heat feature information and the first target heat feature information.

[0157] In one embodiment, the feature representation of the first average heat and the feature representation of the first target heat can be matched using the following formula (1-11) to determine the structural matching information of the first group:

[0158] m d =mean(A CH d H )*d h (1-11);

[0159] Where, mean(A) CH d H A is a characteristic representation of the first average heat. CH It is the first correlation matrix, d H It is the feature information matrix of the first popularity, dh This represents the characteristic of the first target's heat level, and * indicates the multiplication of corresponding bits.

[0160] S321: Determine the structural matching information of the first topic based on the second average heat feature information and the second target heat feature information.

[0161] In this embodiment of the application, the server can determine the structural matching information of the first topic based on the second average heat feature information and the second target heat feature information.

[0162] In one embodiment, the feature representation of the second average heat and the feature representation of the second target heat can be matched using the following formula (1-12) to determine the structural matching information of the first topic:

[0163] m d' =mean(A HC d C )*d c (1-12);

[0164] Where, mean(A) HC d C A is a characteristic representation of the second average heat. HC It is the second correlation matrix, d C It is the feature information matrix of the second most popular trend, d c This represents the characteristic of the first target's heat level, and * indicates the multiplication of corresponding bits.

[0165] S322: The global feature information, object matching information, label matching information and structure matching information of the first group are fused to obtain the first feature information of the graph network of the first group.

[0166] In this embodiment of the application, the server can perform fusion processing on the global feature information, object matching information, label matching information and structure matching information of the first group to finally obtain the first feature information of the graph network of the first group.

[0167] In one embodiment, the global feature information, object matching information, label matching information, and structure matching information of the first group can be fused using the following formula (1-13):

[0168] Z c '=f fuse (Z c ,m uh ,m ch ,m d (1-13);

[0169] Among them, f fuse It is the fusion function, Z cIt is the global feature information of the first group, m uh This is the object matching information for the first group, m ch This is the tag matching information for the first group, m d This is the structural matching information for the first group.

[0170] S323: The global feature information, item matching information, label matching information and structure matching information of the first theme are fused to obtain the second feature information of the graph network of the first theme.

[0171] In this embodiment of the application, the server can perform fusion processing on the global feature information, item matching information, tag matching information and structure matching information of the first topic to finally obtain the second feature information of the graph network of the first topic.

[0172] In one embodiment, the global feature information of the first topic, item matching information, tag matching information, and structure matching information can be fused using the following formula (1-14):

[0173] Z h '=f fuse (Z h ,m ic ,m hc ,m d' (1-14);

[0174] Among them, f fuse It is the fusion function, Z h It is the global feature information of the first topic, m ic This is the project matching information for the first topic, m hc This is the tag matching information for the primary topic, m d' This is the structural matching information for the first topic.

[0175] S324: Determine the matching degree between the first group and the first topic based on the first feature information and the second feature information, and send the items included in the first topic to the objects included in the first group based on the matching degree.

[0176] In this embodiment of the application, the server can determine the matching degree between the first group and the first topic based on the first feature information and the second feature information. For example, the server can perform a dot product operation on the first feature information and the second feature information to obtain the matching degree between the first group and the first topic, and send the items included in the first topic to the objects included in the first group based on the matching degree.

[0177] By implementing the embodiments of this application, global feature information, object matching information, tag matching information, and structure matching information of a first group, and global feature information, item matching information, tag matching information, and structure matching information of a first topic can be obtained respectively. The global feature information, object matching information, tag matching information, and structure matching information of the first group are fused to obtain the first feature information of the first group. The global feature information, item matching information, tag matching information, and structure matching information of the first topic are fused to obtain the second feature information of the first topic. Based on the first feature information of the first group and the second feature information of the first topic, the matching degree between the first group and the first topic can be determined, and the items included in the first topic are sent to the objects included in the first group based on the matching degree. This makes the feature information used to describe the first group and the first topic richer, and the first feature information of the first group and the second feature information of the first topic more accurate, thereby improving the accuracy of information matching.

[0178] For further details, please see Figure 4 , Figure 4 This is a flowchart illustrating another information matching method provided in this application. This information matching method can be executed by a computer device, which can be a terminal device or a server. For ease of understanding, this application embodiment uses the execution of the method by a server as an example for illustration.

[0179] S401: The server can build based on the raw data. Figure 4 The original data is shown in Figure 43 and the tag library is shown in Figure 44.

[0180] S402: Server based on Figure 4 The original data graph 43 and the tag library 44 shown can be used to construct the graph network 45 of the first group and the graph network 46 of the first topic.

[0181] S403: Based on the original data graph, the server can retrieve 47 items associated with each object in the first group. Specific retrieval steps are described above. Figure 3c The illustrated embodiment.

[0182] S404: The server can obtain global feature information, object matching information, label matching information, and structure matching information of the first group.

[0183] S405: The server can obtain global feature information, item matching information, tag matching information, and structure matching information of the first topic.

[0184] S406: The server can fuse the global feature information, object matching information, label matching information and structure matching information of the first group to obtain the first feature information of the first group.

[0185] S407: The server can fuse the global feature information, item matching information, tag matching information and structure matching information of the first topic to obtain the second feature information of the first topic.

[0186] S408: The server can determine the matching degree between the first group and the first topic based on the first feature information and the second feature information.

[0187] Furthermore, the server can send items included in the first topic to the objects included in the first group based on the matching degree.

[0188] The information matching method provided in this application embodiment can be used in, for example... Figure 5a Information matching verification was performed on the three datasets shown. The three datasets have different data sizes, which improves the reliability of the verification results. The results obtained by comparing the information matching method provided in this application with other models can be found in [reference needed]. Figure 5b , Figure 5b The comparison results of this application with five other different models are shown. The final comparison results include three evaluation metrics: ACC (Accuracy), AUC (Area under the Curve), and F1 (F1 Score). ACC is a metric used to evaluate the accuracy of a model, AUC is another metric used to evaluate the accuracy of a model, and F1 is a metric used to evaluate the precision of a model. It can be understood that the higher the value of any of the three metrics, the better the accuracy of the corresponding model. Therefore, it is evident that the information matching accuracy obtained by the method provided in this application is significantly better than that of the other five models, effectively improving the accuracy of information matching.

[0189] For further details, please see Figure 6 , Figure 6 This is a schematic diagram of the structure of an information matching device provided in an embodiment of this application. In one embodiment, such as... Figure 6 As shown, the information matching device 60 may include:

[0190] The acquisition unit 601 is used to acquire the raw data graph and the tag library. The raw data graph includes the relationship between each object in each group and each item in each topic.

[0191] Processing unit 602 is configured to construct a graph network for each group in the group set and a graph network for each topic in the topic set based on the original data graph and the tag library; obtain first feature information of the graph network of the first group and second feature information of the graph network of the first topic, wherein the first feature information includes one or more of global feature information of the first group, object matching information, tag matching information and structure matching information, and the second feature information includes one or more of global feature information of the first topic, item matching information, tag matching information and structure matching information; the first group is any one of the groups included in the group set, and the first topic is any one of the topics included in the topic set; and determine the matching degree between the first group and the first topic based on the first feature information and the second feature information.

[0192] The sending unit 603 is used to send items included in the first topic to objects included in the first group based on the matching degree.

[0193] In one embodiment, the processing unit 602 can also be used to obtain global feature information, object matching information, label matching information, and structure matching information of the first group, as well as global feature information, item matching information, label matching information, and structure matching information of the first topic, based on the graph network of the first group and the graph network of the first topic; perform fusion processing on the global feature information, object matching information, label matching information, and structure matching information of the first group to obtain the first feature information of the graph network of the first group; and perform fusion processing on the global feature information, item matching information, label matching information, and structure matching information of the first topic to obtain the second feature information of the graph network of the first topic.

[0194] In one embodiment, the processing unit 602 can also be used to obtain feature information of each object included in the graph network of the first group and feature information of each item included in the graph network of the first topic, wherein the feature information of each object is obtained based on the labels of each item associated with each object, and the feature information of each item is obtained based on the labels of each item; determine the global feature information of the first group based on the feature information of each object; and determine the global feature information of the first topic based on the feature information of each item.

[0195] In one embodiment, the processing unit 602 can also be used to obtain a first set of items browsed by each object in the graph network of the first group; determine the relevance between each item in a second set of items that does not belong to the first topic and the first topic; determine a target item from the second set of items based on the relevance; and treat the target item and the items in the first topic as items associated with each object.

[0196] In one embodiment, the processing unit 602 can also be used to determine the matching information between each object and the first topic based on the feature information of each object included in the graph network of the first group and the global feature information of the first topic; determine the object matching information of the first group based on the matching information between each object and the first topic; determine the matching information between each item and the first group based on the feature information of each item included in the graph network of the first topic and the global feature information of the first group; and determine the item matching information of the first topic based on the matching information between each item and the first group.

[0197] In one embodiment, the acquisition unit 601 can also be used to acquire raw data, which includes the access relationship between each group in the group set and each topic in the topic set; the processing unit 602 can also be used to construct a raw data graph based on the access relationship between each group in the group set and each topic in the topic set; acquire the tags of each item accessed by each object in each group; and construct a tag library based on the tags of each item accessed by each object in each group.

[0198] In one embodiment, the processing unit 602 can also be used to obtain a first tag set corresponding to the first topic based on a graph network of the first topic; determine the tag matching information of the first group based on each tag included in the first tag set and the global feature information of the first group; obtain a second tag set corresponding to the first group based on a graph network of the first group; and determine the tag matching information of the first topic based on each tag included in the second tag set and the global feature information of the first topic.

[0199] In one embodiment, the processing unit 602 can also be used to obtain first average popularity feature information corresponding to each topic associated with the first group, and first target popularity feature information of the first topic. The first average popularity feature information is obtained based on the number of visits to each topic, and the first target popularity feature information is obtained based on the number of visits to the first topic. Based on the first average popularity feature information and the first target popularity feature information, the processing unit 602 can also be used to determine the structural matching information of the first group. Based on the first average popularity feature information and the first target popularity feature information, the processing unit 602 can also be used to obtain second average popularity feature information corresponding to each group associated with the first topic, and second target popularity feature information of the first group. The second average popularity feature information is obtained based on the number of visits to items of objects included in each group, and the second target popularity feature information is obtained based on the number of visits to items of objects included in the first group. Based on the second average popularity feature information and the second target popularity feature information, the processing unit 602 can also be used to determine the structural matching information of the first topic.

[0200] In one embodiment, the processing unit 602 can also be used to obtain a first association matrix between the first group and the topic set, the first association matrix including the association between the first group and each topic in the topic set; obtain a first popularity feature information matrix corresponding to the topic set, the first popularity feature information matrix including the popularity feature information of each topic in the topic set, the popularity feature information of each topic being obtained based on the number of visits to each topic; determine the first average popularity feature information corresponding to each topic that has an association with the first group based on the first association matrix and the first popularity feature information matrix; and determine the first target popularity feature information of the first topic based on the number of visits to the first topic.

[0201] In one embodiment, the processing unit 602 can also be used to obtain a second association matrix between the first topic and the group set, the second association matrix including the association between the first topic and each group in the group set; obtain a second popularity feature information matrix corresponding to the group set, the second popularity feature information matrix including the popularity feature information of each group in the group set, the popularity feature information of each group being obtained based on the number of visits to the items of the objects included in each group; determine the second average popularity feature information corresponding to each group that has an association with the first topic based on the second association matrix and the second popularity feature information matrix; and determine the second target popularity feature information of the first group based on the number of visits to the items of the objects included in the first group.

[0202] It should be noted that, Figure 6 For details not mentioned in the corresponding embodiments and the specific implementation methods of each step, please refer to [link to relevant documentation]. Figures 2-5b The illustrated embodiments and the foregoing content will not be repeated here.

[0203] For further details, please see Figure 7 , Figure 7This is a schematic diagram of a computer device provided in an embodiment of this application. The computer device may include a network interface 701, a memory 702, and a processor 703. The network interface 701, memory 702, and processor 703 are connected via one or more communication buses, which are used to enable communication between these components. The network interface 701 may include a standard wired interface or a wireless interface (such as a Wi-Fi interface). The memory 702 may include volatile memory, such as random-access memory (RAM); the memory 702 may also include non-volatile memory, such as flash memory, solid-state drive (SSD), etc.; the memory 702 may also include a combination of the above types of memory. The processor 703 may be a central processing unit (CPU). The processor 703 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), etc. The aforementioned PLD can be a field-programmable gate array (FPGA), a generic array logic (GAL), etc.

[0204] Optionally, the memory 702 is also used to store program instructions, which the processor 703 can also call to implement:

[0205] Based on the original data graph and the tag library, construct the graph network for each group in the group set and the graph network for each topic in the topic set;

[0206] Obtain the first feature information of the graph network of the first group and the second feature information of the graph network of the first topic. The first feature information includes one or more of the global feature information, object matching information, label matching information and structure matching information of the first group. The second feature information includes one or more of the global feature information, item matching information, label matching information and structure matching information of the first topic. The first group is any one of the groups included in the group set, and the first topic is any one of the topics included in the topic set.

[0207] Based on the first feature information and the second feature information, the matching degree between the first group and the first topic is determined, and the items included in the first topic are sent to the objects included in the first group based on the matching degree.

[0208] In one embodiment, the memory 702 is further used to store program instructions, which the processor 703 can also call to implement: based on the graph network of the first group and the graph network of the first topic, obtaining global feature information, object matching information, label matching information, and structure matching information of the first group, as well as global feature information, item matching information, label matching information, and structure matching information of the first topic; performing fusion processing on the global feature information, object matching information, label matching information, and structure matching information of the first group to obtain first feature information of the graph network of the first group; and performing fusion processing on the global feature information, item matching information, label matching information, and structure matching information of the first topic to obtain second feature information of the graph network of the first topic.

[0209] In one embodiment, the memory 702 is further used to store program instructions, which the processor 703 can also invoke to perform: obtaining feature information of each object included in the graph network of the first group and feature information of each item included in the graph network of the first topic, wherein the feature information of each object is obtained based on the labels of each item associated with each object, and the feature information of each item is obtained based on the labels of each item; determining global feature information of the first group based on the feature information of each object; and determining global feature information of the first topic based on the feature information of each item.

[0210] In one embodiment, the memory 702 is further used to store program instructions, which the processor 703 can also invoke to perform: obtaining a first set of items browsed by each object in the graph network of the first group; determining the relevance between each item in a second set of items in the first set that does not belong to the first topic and the first topic; determining a target item from the second set of items based on the relevance; and treating the target item and the items in the first topic as items associated with each object.

[0211] In one embodiment, the memory 702 is further used to store program instructions, which the processor 703 can also invoke to perform: determining matching information between each object and the first topic based on feature information of each object included in the graph network of the first group and global feature information of the first topic; determining object matching information of the first group based on the matching information between each object and the first topic; determining matching information between each item and the first group based on feature information of each item included in the graph network of the first topic and global feature information of the first group; and determining item matching information of the first topic based on the matching information between each item and the first group.

[0212] In one embodiment, the memory 702 is further used to store program instructions, which the processor 703 can also call to perform the following: acquiring raw data, the raw data including access relationships between each group in the group set and each topic in the topic set; constructing a raw data graph based on the access relationships between each group in the group set and each topic in the topic set; acquiring the tags of each item accessed by each object in each group; and constructing a tag library based on the tags of each item accessed by each object in each group.

[0213] In one embodiment, the memory 702 is further used to store program instructions, which the processor 703 can also call to perform the following: obtaining a first tag set corresponding to the first topic based on a graph network of the first topic; determining tag matching information of the first group based on each tag included in the first tag set and global feature information of the first group; obtaining a second tag set corresponding to the first group based on a graph network of the first group; and determining tag matching information of the first topic based on each tag included in the second tag set and global feature information of the first topic.

[0214] In one embodiment, the memory 702 is further used to store program instructions, which the processor 703 can also invoke to perform the following: obtaining first average popularity feature information corresponding to each topic associated with the first group, and first target popularity feature information of the first topic, wherein the first average popularity feature information is obtained based on the number of visits to each topic, and the first target popularity feature information is obtained based on the number of visits to the first topic; determining structural matching information of the first group based on the first average popularity feature information and the first target popularity feature information; obtaining second average popularity feature information corresponding to each group associated with the first topic, and second target popularity feature information of the first group, wherein the second average popularity feature information is obtained based on the number of visits to items of objects included in each group, and the second target popularity feature information is obtained based on the number of visits to items of objects included in the first group; and determining structural matching information of the first topic based on the second average popularity feature information and the second target popularity feature information.

[0215] In one embodiment, the memory 702 is further used to store program instructions, and the processor 703 can also call the program instructions to: obtain a first association matrix between the first group and the topic set, the first association matrix including the association between the first group and each topic in the topic set; obtain a first popularity feature information matrix corresponding to the topic set, the first popularity feature information matrix including the popularity feature information of each topic in the topic set, the popularity feature information of each topic being obtained based on the number of visits to each topic; determine the first average popularity feature information corresponding to each topic that has an association with the first group based on the first association matrix and the first popularity feature information matrix; and determine the first target popularity feature information of the first topic based on the number of visits to the first topic.

[0216] In one embodiment, the memory 702 is further used to store program instructions, which the processor 703 can also call to perform the following: obtaining a second association matrix between a first topic and a set of groups, the second association matrix including the association between the first topic and each group in the set of groups; obtaining a second popularity feature information matrix corresponding to the set of groups, the second popularity feature information matrix including the popularity feature information of each group in the set of groups, the popularity feature information of each group being obtained based on the number of times each group includes the items accessed; determining the second average popularity feature information corresponding to each group that is associated with the first topic based on the second association matrix and the second popularity feature information matrix; and determining the second target popularity feature information of the first group based on the number of times each group includes the items accessed.

[0217] It should be understood that the principles and beneficial effects of the computer device 70 in solving the problem described in the embodiments of this application are the same as those in this application. Figures 2-5b The embodiments shown and the foregoing content have similar principles and beneficial effects in solving the problem, and will not be repeated here for the sake of brevity.

[0218] In addition, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method provided in the foregoing embodiments.

[0219] This application also provides a computer program product or computer program that includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the method provided in the foregoing embodiments.

[0220] The steps in the method of this application embodiment can be adjusted, combined, or deleted according to actual needs.

[0221] The units in the device of this application embodiment can be merged, divided, and deleted according to actual needs.

[0222] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0223] The above-disclosed embodiments are merely some of the embodiments of this application, and should not be construed as limiting the scope of this application. Those skilled in the art can understand that all or part of the processes for implementing the above embodiments, and equivalent changes made in accordance with the claims of this application, still fall within the scope of this application.

Claims

1. An information matching method, characterized in that, The method includes: Based on the original data graph and the tag library, a graph network is constructed for each group in the group set and a graph network for each topic in the topic set. The original data graph includes the relationships between each object in each group and each item in each topic. The graph network of each group includes the relationships between each object in each group and the tags in the tag library, as well as the connections between each group and each object included in that group. The graph network of each topic includes the relationships between each item in each topic and the tags in the tag library, as well as the connections between each topic and each item included in that topic. Obtain first feature information of the graph network of the first group and second feature information of the graph network of the first topic. The first feature information includes one or more of the global feature information, object matching information, label matching information, and structure matching information of the first group. The second feature information includes one or more of the global feature information, item matching information, label matching information, and structure matching information of the first topic. The first group is any one of the groups included in the group set, and the first topic is any one of the topics included in the topic set. Based on the first feature information and the second feature information, the matching degree between the first group and the first topic is determined, and the items included in the first topic are sent to the objects included in the first group based on the matching degree.

2. The method as described in claim 1, characterized in that, The step of obtaining the first feature information of the graph network of the first group and the second feature information of the graph network of the first topic includes: Based on the graph network of the first group and the graph network of the first topic, obtain the global feature information, object matching information, label matching information, and structure matching information of the first group, as well as the global feature information, item matching information, label matching information, and structure matching information of the first topic; The global feature information, object matching information, label matching information and structure matching information of the first group are fused to obtain the first feature information of the graph network of the first group. The global feature information, item matching information, label matching information, and structure matching information of the first topic are fused to obtain the second feature information of the graph network of the first topic.

3. The method as described in claim 2, characterized in that, The graph network based on the first group and the graph network based on the first topic obtains global feature information, object matching information, label matching information, and structure matching information of the first group, as well as global feature information, object matching information, label matching information, and structure matching information of the first topic, including: The feature information of each object included in the graph network of the first group and the feature information of each item included in the graph network of the first topic are obtained. The feature information of each object is obtained based on the labels of each item associated with each object, and the feature information of each item is obtained based on the labels of each item. The global feature information of the first group is determined based on the feature information of each object; The global feature information of the first topic is determined based on the feature information of each project.

4. The method as described in claim 3, characterized in that, The method further includes: Obtain the first set of items browsed by each object in the graph network of the first group; Determine the relevance between each item in the second set of items that does not belong to the first topic and the first topic; Target projects are determined from the second set of projects based on the relevance. The target project and the projects included in the first topic are considered as projects associated with each of the objects.

5. The method as described in claim 2, characterized in that, The method further includes: Based on the feature information of each object included in the graph network of the first group and the global feature information of the first topic, the matching information between each object and the first topic is determined; Based on the matching information of each object with the first topic, determine the object matching information of the first group; Based on the feature information of each item in the graph network of the first topic and the global feature information of the first group, the matching information between each item and the first group is determined; The matching information of the first topic is determined based on the matching information of each item with the first group.

6. The method as described in claim 1, characterized in that, Before constructing the graph network for each group in the group set and the graph network for each topic in the topic set based on the original data graph and the tag library, the method further includes: Obtain raw data, which includes the access relationships between each group in the group set and each topic in the topic set; The original data graph is constructed based on the access relationships between each group in the group set and each topic in the topic set; Obtain the tags of each item accessed by each object in each group; The tag library is constructed based on the tags of each item accessed by each object in each group.

7. The method as described in claim 2, characterized in that, The method further includes: The first set of tags corresponding to the first topic is obtained based on the graph network of the first topic; Based on each tag included in the first tag set and the global feature information of the first group, the tag matching information of the first group is determined; The second label set corresponding to the first group is obtained based on the graph network of the first group; Based on each tag included in the second tag set and the global feature information of the first topic, the tag matching information of the first topic is determined.

8. The method as described in claim 2, characterized in that, The method further includes: Obtain first average popularity feature information corresponding to each topic that is associated with the first group, and first target popularity feature information of the first topic. The first average popularity feature information is obtained based on the number of visits to each topic, and the first target popularity feature information is obtained based on the number of visits to the first topic. The structural matching information of the first group is determined based on the first average heat feature information and the first target heat feature information; Obtain second average popularity feature information corresponding to each group that is related to the first topic, and second target popularity feature information of the first group. The second average popularity feature information is obtained based on the number of visits to the items of the objects included in each group, and the second target popularity feature information is obtained based on the number of visits to the items of the objects included in the first group. The structural matching information of the first topic is determined based on the second average heat feature information and the second target heat feature information.

9. The method as described in claim 8, characterized in that, The step of obtaining the first average popularity feature information corresponding to each topic associated with the first group, and the first target popularity feature information of the first topic, includes: Obtain a first association matrix between the first group and the topic set, wherein the first association matrix includes the association between the first group and each topic in the topic set; Obtain the first popularity feature information matrix corresponding to the topic set. The first popularity feature information matrix includes the popularity feature information of each topic in the topic set. The popularity feature information of each topic is obtained based on the number of visits to each topic. Based on the first association matrix and the first popularity feature information matrix, determine the first average popularity feature information corresponding to each topic that is associated with the first group; The first target popularity feature information of the first topic is determined based on the number of visits to the first topic.

10. The method as described in claim 8, characterized in that, The step of obtaining the second average popularity feature information corresponding to each group that is related to the first topic, and the second target popularity feature information of the first group, includes: Obtain a second association matrix between the first topic and the group set, the second association matrix including the association between the first topic and each group in the group set; Obtain the second popularity feature information matrix corresponding to the group set. The second popularity feature information matrix includes the popularity feature information of each group in the group set. The popularity feature information of each group is obtained based on the number of visits to the items of the objects included in each group. Based on the second association matrix and the second popularity feature information matrix, determine the second average popularity feature information corresponding to each group that is associated with the first topic; The second target popularity feature information of the first group is determined based on the number of times the objects included in the first group are accessed.

11. An information matching device, characterized in that, The device includes: The acquisition unit is used to acquire the raw data graph and the tag library. The raw data graph includes the relationship between each object in each group and each item in each topic. The processing unit is configured to construct a graph network for each group in the group set and a graph network for each topic in the topic set based on the original data graph and the tag library; obtain first feature information of the graph network of the first group and second feature information of the graph network of the first topic, wherein the first feature information includes one or more of global feature information, object matching information, tag matching information, and structure matching information of the first group, and the second feature information includes one or more of global feature information, item matching information, tag matching information, and structure matching information of the first topic; the first group is any one of the groups included in the group set, and the first topic is any one of the topics included in the topic set; determine the matching degree between the first group and the first topic based on the first feature information and the second feature information; the graph network of each group includes the association relationship between each object in each group and the tag in the tag library, and the connection relationship between each group and each object included in the group; the graph network of each topic includes the association relationship between each item in each topic and the tag in the tag library, and the connection relationship between each topic and each item included in the topic; A sending unit is configured to send the items included in the first topic to the objects included in the first group based on the matching degree.

12. A computer device, characterized in that, The device includes a memory, a processor, and a network interface. The processor is connected to the memory and the network interface. The network interface is used to provide network communication functions. The memory is used to store program code. The processor is used to call the program code to execute the method according to any one of claims 1-10.

13. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method described in any one of claims 1-10.

14. A computer program product, characterized in that, The computer program product includes a computer program or computer instructions that, when executed by a processor, implement the method of any one of claims 1-10.

Images