Multi-scene data construction method and device, electronic equipment and storage medium

Abstract

The invention provides a multi-scene data construction method and device, electronic equipment and a storage medium. The invention relates to the technical field of computers, in particular to the technical fields of big data, big models, deep learning, intelligent recommendation and the like. According to the specific implementation scheme, an operation summary graph is constructed according to operation data of a first object in multiple scenes; performing multi-dimensional migration based on the operation summary graph, and generating a first migration path in each single scene and a second migration path in a mixed scene; and generating co-occurrence relation data between the first object and the second object based on the first walk path and the second walk path.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, specifically to technical fields such as big data, big models, deep learning, and intelligent recommendation, and specifically to a multi-scenario data construction method, device, electronic device, and storage medium. Background Art

[0002] In information flow recommendation systems, there are often multiple scenarios for users to consume. The user's behavior in multiple scenarios constitutes the user's complete behavior trajectory. Unified modeling of user behavior in multiple scenarios is conducive to improving recommendation accuracy. Graph models are a popular user behavior modeling technology in recent years, but there is little related work on multi-scenario behavior modeling. Therefore, how to construct multi-scenario data based on graph models has become a technical problem that needs to be solved urgently. Summary of the invention

[0003] The present invention provides a multi-scenario data construction method, device, electronic device and storage medium.

[0004] According to one aspect of the present disclosure, a multi-scenario data construction method is provided, the method comprising:

[0005] Constructing an operation summary graph according to operation data of a first object in multiple scenarios, the operation summary graph comprising a first node representing the first object, a second node representing a second object, and an edge connecting the first node and the second node, wherein the second object is an object for publishing resources;

[0006] Perform multi-dimensional walking based on the operation summary graph to generate a first walking path in each single scene and a second walking path in the mixed scene;

[0007] Co-occurrence relationship data between the first object and the second object is generated based on the first travel path and the second travel path.

[0008] According to another aspect of the present disclosure, a multi-scenario data construction device is provided, the device comprising:

[0009] A first construction module is used to construct an operation summary graph according to operation data of a first object in multiple scenarios, wherein the operation summary graph includes a first node representing the first object, a second node representing a second object, and an edge connecting the first node and the second node, wherein the second object is an object for publishing resources;

[0010] A first generating module, configured to perform multi-dimensional walking based on the operation summary graph, and generate a first walking path in each single scenario and a second walking path in a mixed scenario;

[0011] The second generating module is used to generate co-occurrence relationship data between the first object and the second object based on the first wandering path and the second wandering path.

[0012] According to another aspect of the present disclosure, there is provided an electronic device, comprising:

[0013] at least one processor; and

[0014] a memory communicatively connected to the at least one processor; wherein,

[0015] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute any method in the embodiments of the present disclosure.

[0016] According to another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable the computer to execute any method according to the embodiments of the present disclosure.

[0017] According to another aspect of the present disclosure, a computer program product is provided, including a computer program, which implements any method according to the embodiments of the present disclosure when executed by a processor.

[0018] By adopting the solution disclosed in the present invention, by simultaneously modeling the user's behavior in a single scene and a mixed scene in a graph model, it is possible to learn the commonalities of mixed scene behaviors and the characteristics of single scene behaviors, thereby improving the learning accuracy of graph representation.

[0019] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0021] Figure 1 is a flowchart of a method for constructing multi-scenario data according to an embodiment of the present disclosure;

[0022] Figure 2 is a schematic diagram of an operation summary diagram of a recommendation platform according to an embodiment of the present disclosure;

[0023] Figure 3 is a schematic diagram of walking paths of a single scene and a mixed scene according to an embodiment of the present disclosure;

[0024] Figure 4 is a schematic diagram of the structure of a multi-scenario data construction device according to an embodiment of the present disclosure;

[0025] Figure 5 is a schematic diagram of a scene constructed according to multi-scene data of an embodiment of the present disclosure;

[0026] Figure 6 It is a structural diagram of an electronic device used to implement the multi-scenario data construction method of an embodiment of the present disclosure. DETAILED DESCRIPTION

[0027] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope of the present disclosure. Similarly, for the sake of clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0028] The term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there may be three relationships. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. The term "at least one" in this article means any combination of at least two of any one or more of a plurality of. For example, including at least one of A, B, and C can mean including any one or more elements selected from the set consisting of A, B, and C. The terms "first" and "second" in this article refer to multiple similar technical terms and distinguish them. They do not mean to limit the order or to limit them to only two. For example, the first feature and the second feature refer to two types / two features. The first feature can be one or more, and the second feature can also be one or more.

[0029] In addition, in order to better illustrate the present disclosure, numerous specific details are given in the following specific embodiments. It should be understood by those skilled in the art that the present disclosure can also be implemented without certain specific details. In some examples, methods, means, components and circuits well known to those skilled in the art are not described in detail in order to highlight the main purpose of the present disclosure.

[0030] Before introducing the technical solutions of the embodiments of the present disclosure, the technical terms that may be used in the present disclosure are further explained:

[0031] Operation summary graph: A graph structure used to represent the interaction relationship between multiple objects. It contains nodes and connecting edges, where nodes represent objects and connecting edges represent the interaction relationship between objects.

[0032] Multi-dimensional walk: The process of traversing a graph structure according to certain rules and strategies. Multi-dimensional walk takes into account the different attributes and dimensions of the graph and can generate rich walk paths.

[0033] Co-occurrence relationship data: data indicating the co-occurrence of two objects in different scenarios. This data can reflect the associations and patterns between objects.

[0034] In the related technologies, multi-scenario modeling technologies are divided into the following categories:

[0035] 1. Unified walking learning: Ignoring scene information, it is impossible to learn the user's preferences in each single scene.

[0036] 2. Single walk learning: It can only learn the user's preferences in a single scenario, and cannot transfer the scenario behavior, and thus cannot learn the user's overall preferences.

[0037] In order to at least partially solve one or more of the above problems and other potential problems, the present disclosure proposes a multi-scenario data construction method based on a graph model, which simultaneously models the behavior of objects in single and mixed scenes in the graph model, and can learn the commonalities of mixed scene behaviors as well as the characteristics of single scene behaviors, thereby improving the learning accuracy of graph representation. This solves the problem of the single learning dimension of multi-scenario behavior data in existing graph model technology, improves the model learning effect, and is conducive to improving downstream task indicators.

[0038] The present disclosure provides a method for constructing multi-scenario data. Figure 1 It is a flowchart of a multi-scenario data construction method according to an embodiment of the present disclosure. The multi-scenario data construction method can be applied to a multi-scenario data construction device. The multi-scenario data construction device is located in an electronic device. The electronic device includes but is not limited to a fixed device and / or a mobile device. For example, a fixed device includes but is not limited to a server, and the server can be a cloud server or an ordinary server. For example, a mobile device includes but is not limited to: a mobile phone, a tablet computer, and a vehicle-mounted terminal. In some possible implementations, the multi-scenario data construction method can also be implemented by a processor calling a computer-readable instruction stored in a memory. Figure 1 As shown, the multi-scenario data construction method includes:

[0039] S101, constructing an operation summary graph according to operation data of a first object in multiple scenarios; the operation summary graph includes a first node representing the first object, a second node representing the second object, and an edge connecting the first node and the second node, where the second object is an object that publishes resources;

[0040] S102, performing multi-dimensional walking based on the operation summary graph to generate a first walking path in each single scenario and a second walking path in the mixed scenario;

[0041] S103: Generate co-occurrence relationship data between the first object and the second object based on the first wandering path and the second wandering path.

[0042] In the disclosed embodiment, operation data refers to the interactive behavior data generated by the first object and the second object in a specific scenario, such as clicks, browsing, purchases, comments, etc. The resources published by the second object include but are not limited to articles, long videos, short videos, pictures, product links, etc.

[0043] In the embodiments of the present disclosure, the user is aware of and agrees to the acquisition and use of operation data, which complies with the provisions of relevant laws and regulations and does not violate public order and good morals.

[0044] In the embodiment of the present disclosure, in the operation summary graph, the first node represents the first object (such as a user), and the second node represents the second object (such as an object that publishes resources, such as a merchant, a content creator, etc.).

[0045] In the embodiment of the present disclosure, each single scene includes 1 scene; each mixed scene includes at least 2 scenes.

[0046] In the disclosed embodiment, in the operation summary graph, an edge represents an interactive relationship between a first object and a second object, such as a user browsing a certain merchant's products or clicking on a certain author's work.

[0047] In some embodiments, constructing an operation summary graph includes: first, collecting operation data of a first object (e.g., a user) in multiple scenarios. The operation data may include user behavior records such as clicks, browsing, purchases, following, forwarding, sharing, and liking. Then, using a graph database or a graph processing framework, the collected operation data is constructed into an operation summary graph.

[0048] In the disclosed embodiment, the operation summary graph can display the behavior characteristics of the first object in the entire scene, and can also display the behavior characteristics of the first object in a single scene (also called a separate scene or a single scene). In this way, the operation summary graph can integrate data from multiple scenes to form a behavior portrait from a global perspective. Through this global perspective behavior portrait, the electronic device can quickly and accurately identify the commonalities and differences in the behavior of the object in different scenes. The operation summary graph can also focus on a specific scene, and analyze the specific behavior of the object in a specific scene based on the operation summary graph, which can improve the accuracy of identifying the behavior preference of the object in a specific environment.

[0049] In some implementations, the operation summary graph may be a directed graph or an undirected graph, wherein both nodes and edges have attributes to represent different objects and interaction relationships.

[0050] Figure 2 A schematic diagram showing an operational summary diagram of the recommendation platform is shown, Figure 2As shown in the figure, double coils represent user nodes, and single coils represent author nodes; for example, u1, u2, u3, u4, u5, and u6 represent user nodes; a1, a2, a3, a4, a5, a6, and a7 represent author nodes. When connecting edges, different attributes are assigned to the edges according to different scenarios. For example, recommendation scenarios can be divided into one-hop single-column, one-hop double-column, and two-hop scenarios. Figure 2 In the figure, the short dashed line represents satisfactory consumption in a one-hop single-column scenario, the thick solid line represents satisfactory consumption in a one-hop double-column scenario, and the dotted line represents satisfactory consumption in a two-hop scenario. Figure 2 It shows the satisfied consumption objects of users in the mixed scenarios supported by the recommendation platform, and it can also show the satisfied consumption objects of users in a single scenario. It can be understood that the diagram is schematic and not restrictive, and the number of nodes and scenarios in the diagram can be increased or decreased according to needs, and the node representation form and line type can also be set or changed according to needs.

[0051] In the disclosed embodiment, multi-dimensional walking includes single-scene walking and multi-scene walking.

[0052] In some implementations, in a single scene, walking is performed according to the operation summary graph to generate a first walking path. The first walking path represents an interaction sequence between the first object and the second object in the single scene.

[0053] In some implementations, in a mixed scene, the interaction of at least two single scenes is considered to generate a second walking path, which spans across different single scenes and reflects the interaction relationship between the first object and the second object in different single scenes.

[0054] In some implementations, a graph traversal algorithm such as Depth-First Search (DFS), Breadth-First Search (BFS), or a random walk-based algorithm may be used to generate a walk path.

[0055] In the disclosed embodiment, the co-occurrence relationship can be used to find the correlation between different features or items in the data set. For example, in a shopping data set, if two commodities are often purchased by the same user, there is a co-occurrence relationship between them, which can be used to build a recommendation system. For example, in a video data set, the co-occurrence relationship can be used to find the correlation between users and authors. For example, the co-occurrence of user viewing behavior and author works. When a video of an author is watched by a large number of users, it can be considered that there is a co-occurrence relationship between the author's video and the user. This relationship can be used to recommend other works of the author to users, or to recommend the author's videos to other users. For another example, the co-occurrence relationship between users. In a video data set, the co-occurrence relationship between users can be analyzed. When two users watch a large number of the same videos, it can be considered that there is a co-occurrence relationship between the two users. By analyzing the co-occurrence relationship between users, user groups with similar interests or preferences can be found, thereby providing them with more personalized recommendation services. For another example, the co-occurrence relationship between authors. In a video data set, the co-occurrence relationship between authors can be analyzed. For example, when the videos of two authors are often watched by users at the same time, it can be considered that there is a co-occurrence relationship between the two authors. This relationship may be due to the similarity of their creative styles, the proximity of their topics, or their common influence in certain specific fields. By analyzing the co-occurrence relationship between authors, we can further understand the diversity and relevance of video content, thereby improving the accuracy and relevance of recommendations.

[0056] In some embodiments, the co-occurrence relationship data includes the identifier of the first object, the identifier of the second object, and the identifier of the co-occurrence scene. Optionally, the co-occurrence relationship data may further include data such as the number of co-occurrences (how many times the co-occurrences occurred within a set time period), the co-occurrence time period (the time period when the co-occurrences occurred), and the co-occurrence location (the location when the co-occurrences occurred). For example, the co-occurrence relationship data may be represented as the interaction frequency, interaction depth, interaction time, interaction location, etc. between users and merchants, users and content creators, etc.

[0057] For example, suppose there is an e-commerce platform where the first object is the user and the second object is the merchant. By collecting the user's browsing, purchasing, commenting and other operation data on the platform, an operation summary graph can be constructed. Then, based on the graph, multi-dimensional walks are performed to generate the interaction paths between users and merchants in different scenarios (such as search, browsing, purchasing, etc.). Finally, based on these paths, the co-occurrence relationship data between users and merchants is generated, such as the number of times a user purchases a merchant's products and the number of times the user browses. These data can be used to optimize the accuracy of recommended products, the satisfaction of recommended products, etc.

[0058] The technical solution of the disclosed embodiment constructs an operation summary graph based on the behavior data of the first object in multiple scenarios, and can integrate the behavior data in multiple scenarios to generate a more comprehensive operation summary graph, thereby more accurately reflecting the interactive relationship between the first object and the second object. Through multi-dimensional wandering, rich wandering paths can be generated to provide data support for subsequent co-occurrence relationship analysis, which is helpful to discover potential associations and patterns. The generated co-occurrence relationship data can be applied to multiple fields, such as product recommendations, content recommendations, service recommendations, etc. By deeply exploring the co-occurrence relationships between different fields, the accuracy and breadth of the recommended content can be improved.

[0059] In some embodiments, an operation summary graph is constructed based on the operation data of the first object in multiple scenarios, including: for each single scene, constructing a first node, a second node and an edge, wherein the edge identifiers corresponding to the edges in different single scenes are different; and constructing an operation summary graph based on the first node, the second node and the edge in each single scene.

[0060] In the disclosed embodiment, a node represents an entity or object in a graph, and specifically may be a user, a resource, an author, and the like.

[0061] In the disclosed embodiments, an edge represents a connection or relationship between nodes, and may specifically represent an interactive relationship between a user and a resource, such as browsing, purchasing, forwarding, commenting, etc. It may also represent an interactive relationship between a user and the author of a published resource, such as following, unfollowing, recommending, commenting, and replying, etc.

[0062] In the embodiment of the present disclosure, the edge identifier is an identifier or label used to distinguish the interaction relationship between nodes in different scenarios. The edge identifier at least includes a scene identifier used to distinguish scenes.

[0063] In some embodiments, the operation data of the first object (such as a user) in different scenarios are collected and sorted. These data may include interactive behaviors such as clicks, browsing, purchases, and comments of users in various scenarios, as well as the time, frequency, results, and other attributes of these behaviors. Data collection can be carried out in a variety of ways, such as software development kits (SDKs), log analysis, questionnaires, third-party data platforms, etc. In each single scenario, the first object (such as a user) and the second object (such as goods, services, other users, authors, etc.) with which it interacts are respectively used as nodes. Two nodes with an interactive relationship are connected to form an edge, and an edge identifier is assigned to the edge to indicate the type or nature of the interaction. For example, if user E purchases product F, an edge can be connected between user E and product F, and the edge identifier "purchase" can be assigned. In order to distinguish the interactive relationships in different scenarios, different edge identifiers need to be assigned to the edges in different scenarios. For example, in an e-commerce scenario, the edge identifier can be "purchase", "browse", "comment", etc.; in a social scenario, the edge identifier can be "follow", "like", "forward", etc. This ensures that when integrating data from multiple scenarios, the interaction relationships between different scenarios can be clearly identified.

[0064] In some implementations, the operation summary graph can be flexibly adjusted and optimized according to business needs and analysis objectives. At the same time, after the operation summary graph is constructed, new scenarios and data sources can be expanded on the constructed operation summary graph.

[0065] For example, suppose you need to analyze the user operation data of a search website to understand the user's preferences. You can follow the steps below to build an operation summary graph. Clearly define the analysis goal: understand the user's preferences on the search website. Data collection and organization: Collect the user's browsing, searching, commenting, liking, forwarding and other behavior data on the search website, and organize them into a structured format. Define nodes and edges: Use node e to represent user E and node g to represent author G. When a user browses the works published by author G on the recommendation page, an edge is formed between node e and node g, and the edge label is given as "recommendation"; when user E watches the live broadcast of author G, another edge is formed between node e and node g, and the edge label is given as "live broadcast". Distinguish edge labels in different scenarios: In the search scenario, the edge label can be a variety of search display forms supported by the search website, such as "recommendation", "discovery", "hot search", "live broadcast", etc. Integrate nodes and edges into the graph: Integrate all user nodes, author nodes and the edges between them into the same graph to form an operation summary graph.

[0066] The technical solution of the disclosed embodiment can construct a comprehensive and accurate operation summary diagram by integrating behavioral data in multiple scenarios, providing strong support for recommendation analysis and decision-making. Through visualization processing, complex behavioral data can be converted into an intuitive and easy-to-understand visual representation, significantly reducing the complexity of operations in the data processing process and effectively improving the overall data processing efficiency.

[0067] In some embodiments, in the operation summary graph, an object identifier representing a first object is recorded on a first node, an object identifier representing a second object is recorded on a second node, and an edge identifier representing a scene is recorded on an edge.

[0068] In the disclosed embodiment, an object identifier (OID) is a string or number used to uniquely identify a node in a graph, such as a user identifier (ID), an author ID, a product ID, and the like.

[0069] In the disclosed embodiment, the edge identifier can be used to uniquely identify the interaction type or property of the edge in the graph. The edge identifier includes at least a scene identifier (Scene Identifier, SID). The scene identifier can be used to uniquely identify the scene category in the graph.

[0070] In some embodiments, when constructing an operation summary graph, a unique object identifier is assigned to each node. The object identifier is used to distinguish different nodes in the graph, ensuring that each node can be uniquely identified. When creating a node, its object identifier is recorded as one of the attributes of the node. In this way, in the subsequent analysis and query process, a specific node can be quickly located through the object identifier. Similar to the node, a unique scene identifier is assigned to each edge. The scene identifier is used to distinguish the interactive relationship under different scenes in the graph, ensuring that each edge can be accurately classified into the scene to which it belongs. The scene identifier can be a string describing the type or nature of the interaction, such as "purchase", "browse", "comment", etc., and the scene identifier can also be a string describing the scene category, such as "search website", "e-commerce platform", "video publishing platform", etc. When creating an edge, the scene identifier is recorded as one of the attributes of the edge. In this way, in the subsequent analysis and query process, the interactive relationship under a specific scene can be quickly identified through the scene identifier. The nodes with object identifiers and the edges with scene identifiers are integrated into the same graph to form a complete operation summary graph. This graph not only contains the information of nodes and edges, but also their respective object identifiers and scene identifiers, providing rich data support for subsequent analysis and query.

[0071] For example, there is an operation summary graph of an e-commerce website, where nodes represent users and products, and edges represent the interaction between users and products. Object identifiers and scenario identifiers can be assigned to nodes and edges in the following steps. First, assign object identifiers to nodes. For user nodes, assign a unique user ID to each user as the object identifier. For product nodes, assign a unique stock keeping unit (SKU) code to each product as the object identifier. Then, assign scenario identifiers to edges. For users browsing products, the scenario identifier of the edge is "browse". For users purchasing products, the scenario identifier of the edge is "purchase". For users commenting on products, the scenario identifier of the edge is "comment". The scenario identifier of an edge can contain a series of diverse scenario identifier sequences, such as a combination of specific scenario identifiers such as browsing, purchasing, and commenting. In this way, each edge can be accurately labeled, reflecting the interaction between users and products or services in different scenarios. Through these scenario identifiers, users' preferences and behavior patterns can be more deeply understood, so as to provide users with more personalized recommendations and services. Finally, user nodes with user IDs and product nodes with SKU codes are integrated into the graph. Connect the edges with scene identifiers (such as "browse", "purchase", "comment") to the corresponding nodes. In this way, an operation summary graph containing object identifiers and scene identifiers is constructed, which provides strong support for subsequent analysis and query.

[0072] The technical solution described in this embodiment improves the accuracy and consistency of data representation by assigning unique object identifiers and scene identifiers to nodes and edges. Through object identifiers and scene identifiers, specific nodes and edges can be quickly located, improving the efficiency of data query and processing.

[0073] In some embodiments, multi-dimensional walking is performed based on the operation summary graph to generate a first walking path in each single scenario and a second walking path in the mixed scenario, including: distinguishing different single scenarios according to edge identifiers in the operation summary graph; in each single scenario, walking according to the connection relationship between the first node and the second node to obtain a first walking path; in the mixed scenario, walking according to the connection relationship between the first node and the second node to obtain a second walking path.

[0074] In the embodiment of the present disclosure, the wandering path is a path generated by wandering in a graph according to the connection relationship between nodes.

[0075] In the disclosed embodiment, a single scene is a single specific scene in multiple scenes, and each scene in the multiple scenes can be used as a single scene.

[0076] In the disclosed embodiment, a mixed scene includes a mixture of at least two single scenes. Multiple scenes can form multiple mixed scenes. For example, if multiple scenes include scene 1, scene 2, and scene 3, then the first mixed scene includes scene 1 and scene 2; the second mixed scene includes scene 2 and scene 3; the third mixed scene includes scene 1 and scene 3; and the fourth mixed scene includes scene 1, scene 2, and scene 3.

[0077] In some implementations, in the operation summary graph, different scenarios are distinguished according to edge identifiers. For example, all edges with an edge identifier of "purchase" represent purchase scenarios, and all edges with an edge identifier of "browse" represent browse scenarios.

[0078] In some embodiments, generating a first walking path in a single scenario includes: selecting a specific scenario from the operation summary graph, such as a purchase scenario. Selecting a starting node in the scenario, such as a user node. Performing DFS or BFS from the starting node based on the connection relationship between the nodes to generate a walking path in the scenario. The path will contain a series of connected nodes, which represent objects that the user may visit or interact with in the scenario. Record the generated walking path as the first walking path.

[0079] In some embodiments, generating a second wandering path in a mixed scene includes: selecting at least two single scenes from the operation summary diagram to combine into a mixed scene. Selecting a starting node and setting the conditions for scene switching, for example, the condition for switching from a browsing scene to a purchase scene may be that the user clicks a purchase button. According to the connection relationship between the nodes and the scene switching conditions, wandering is started from the starting node. During the wandering process, when the scene switching conditions are met, the wandering path is switched to a new scene, and wandering continues in the scene. In this way, a mixed wandering path containing multiple scenes can be generated.

[0080] by Figure 2 Based on the operation summary diagram shown, Figure 3 The schematic diagram of the walking path of a single scene and a mixed scene is shown as follows: Figure 3As shown, in the single-column behavior scenario, the roaming path is: author a2-user u2-author a6-user u5; in the double-column behavior scenario, the roaming path is: author a5-user u3-author a2-user u1; in the two-hop behavior scenario, the roaming path is: author a4-user u3-author a7-user u6; in the mixed behavior scenario (including single-column behavior, double-column behavior and two-hop behavior), the roaming path is: author a1-user u3-author a4-user u4. It can be understood that the figure is schematic and not restrictive. The number of nodes in each roaming path in the single scenario in the figure can be increased or decreased according to demand, and the number of nodes and the number of scenarios in the roaming path under the mixed scenario in the figure can be increased or decreased according to demand.

[0081] For example, suppose there is an operation summary graph of a search website, which contains user nodes, author nodes, and the interactive relationships between them (such as browsing, following, commenting, and collecting). The following steps can be followed to generate a walk path: Walk path generation in a single scenario: Select the follow scenario. Select a user node as the starting node. Based on the interactive relationship between the user and the author (such as clicking the follow button), perform a depth-first search or a breadth-first search from the starting node to generate a walk path in the follow scenario. Walk path generation in a mixed scenario: Select the browse scenario and the follow scenario as the mixed scenario. Select a user node as the starting node. Set the scene switching condition, such as the user clicking the follow button while browsing. Based on the interactive relationship between the user and the author and the scene switching condition, start walking from the starting node. For example, a user may browse from the homepage to multiple work pages, and then click the follow button when browsing to a certain work.

[0082] The technical solution of the disclosed embodiment can generate a walking path containing multiple single scenes and mixed scenes through multi-dimensional walking, thereby improving the comprehensiveness and accuracy of the walking path. Furthermore, the generated walking path can be used for visual analysis to help users more intuitively understand the user's behavior path and interaction relationship in different scenes, thereby improving the user experience. Based on the generated walking path, the recommendation system can identify the user's behavior pattern and preference trend, thereby optimizing the personalized recommendation algorithm and improving the accuracy of the recommended content and user satisfaction.

[0083] In some embodiments, generating co-occurrence relationship data between a first object and a second object based on a first wandering path and a second wandering path includes: selecting at least two single scenes from multiple scenes as target scenes; determining a target first wandering path and a target second wandering path from the first wandering path and the second wandering path based on edge identifiers corresponding to the target scenes; and traversing the target first wandering path and the target second wandering path to obtain co-occurrence relationship data.

[0084] In the embodiment of the present disclosure, the target scene is a specific scene selected from multiple scenes for analyzing the co-occurrence relationship of objects.

[0085] In the disclosed embodiment, the co-occurrence relationship data can reflect the co-occurrence of objects in different scenarios, and such data is usually used to analyze the correlation, similarity or potential connection between objects. For example, the co-occurrence relationship data reflects the co-occurrence of a first object and a second object. The first object and the second object are two objects that need to be paid attention to when analyzing the co-occurrence relationship of objects, such as user D and author H.

[0086] In some implementations, at least two single scenes are selected from multiple scenes as target scenes. These target scenes may be determined based on business analysis requirements. For example, if it is desired to analyze the object co-occurrence relationship of users in browsing and purchasing scenes, these two scenes may be selected as target scenes.

[0087] In some embodiments, based on the edge identifier corresponding to the target scene, the target first wandering path and the target second wandering path related to the target scene are screened out from the previously generated first wandering path and the second wandering path, and these wandering paths should only contain nodes and edges under the target scene. For example, single scene one and single scene two are selected as the target scenes. If the mixed scene includes a mixed scene composed of single scene one and single scene two, the second wandering path with the characteristics of the mixed scene is screened out from multiple second wandering paths as the target second wandering path. If the mixed scene does not include such a mixed scene composed of single scene one and single scene two, the target second wandering path is generated based on the first wandering path of single scene one, the second wandering path of single scene two, and the switching form of single scene one and single scene two.

[0088] In some embodiments, the target first wandering path and the target second wandering path are traversed, and the number and position of the first object and the second object co-occurring in each wandering path are recorded. The co-occurrence relationship data can be a two-dimensional matrix, in which rows and columns represent different objects, and the values ​​in the matrix represent the number of times the two objects co-occur. In addition, the location information of the co-occurrence can also be recorded, such as in which scene or near which node the co-occurrence occurs.

[0089] In some implementations, the generated co-occurrence relationship data is sorted and optimized. The data is deduplicated, invalid data is filtered, and other operations are performed to ensure the accuracy and reliability of the data. At the same time, the data can also be further analyzed and mined as needed to extract more valuable information.

[0090] In some embodiments, the method further includes: in response to detecting a preset trigger operation, outputting the co-occurrence relationship data in a visual display form. In this way, it can help users more intuitively understand the co-occurrence of objects in different scenarios and provide an intuitive basis for business decision-making.

[0091] For example, suppose there is user operation data on an e-commerce website, and the travel paths in the browsing scenario and the purchase scenario have been generated. If you want to analyze the co-occurrence relationship between users' preferences for product F and product S in the browsing and purchase scenarios. Select the browsing scenario and the purchase scenario as the target scenarios. Filter out the target travel paths related to the browsing scenario and the purchase scenario from the previously generated travel paths. Traverse the target travel paths and record the number of co-occurrences and positions of product F and product S in each travel path. For example, in the browsing scenario, product F and product S may often appear on the same page or adjacent pages; in the purchase scenario, users may often purchase product F and product S at the same time.

[0092] The technical solution described in the embodiment of the present disclosure can accurately reflect the co-occurrence relationship between the first object and the second object in different scenarios by traversing the walking path and recording the co-occurrence of objects, thereby providing strong support for accurate recommendation. The method can adapt to different scenarios and objects, and only needs to select the corresponding target scene and object to generate co-occurrence relationship data, thereby improving the efficiency of generating co-occurrence relationship data. At the same time, the data can be further analyzed and mined as needed to meet the recommendation needs in different scenarios, thereby improving the accuracy of recommendations.

[0093] In some embodiments, co-occurrence relationship data of the first object and the second object is generated based on the first wandering path and the second wandering path, including: traversing all nodes in the first wandering path and the second wandering path, recording the co-occurrence relationships between all other nodes; traversing all recorded co-occurrence relationships, and filtering the data according to at least one of a preset co-occurrence number threshold and scene importance information to obtain co-occurrence relationship data.

[0094] In the embodiment of the present disclosure, the wandering path is a path generated by wandering in a graph according to the connection relationship between nodes, and is used to represent the behavior pattern of an object in different scenarios.

[0095] In the embodiment of the present disclosure, the co-occurrence relationship is an association relationship in which two objects appear simultaneously in the same scene or path.

[0096] In the embodiment of the present disclosure, the preset co-occurrence number threshold is a number threshold for screening meaningful co-occurrence relationships. Only co-occurrence relationships whose co-occurrence number exceeds the threshold will be retained.

[0097] In the disclosed embodiment, the scene importance information is an indicator for evaluating the importance of a scene. For example, only those co-occurrence relationships that occur in important scenes will be given priority or retained. For another example, co-occurrence relationships that occur in unimportant scenes will be given priority or retained.

[0098] In some implementations, traversing the walk path and recording the co-occurrence relationship includes:

[0099] Initialize the data structure. Specifically, create a data structure (such as a dictionary or database) to store co-occurrence relationship data. This data structure should be able to record the number of co-occurrences of each node with other nodes and the scene information of the co-occurrence.

[0100] Traverse each node in the first and second wandering paths in turn. For each node in the path, check its connection relationship with other nodes.

[0101] For each node, record its co-occurrence relationship with adjacent nodes. For example, record the number of co-occurrences (i.e., the number of times two nodes appear adjacent to each other in the path) and the scene of co-occurrence (i.e., in which walking path or scene the co-occurrence relationship occurs).

[0102] In some embodiments, preset co-occurrence thresholds and / or scene importance information are set according to business needs. These thresholds and indicators will be used to filter meaningful co-occurrence relationships. Traverse all recorded co-occurrence relationships and filter data according to the set thresholds and importance indicators. Only those co-occurrence relationships with co-occurrence times exceeding the threshold or frequently appearing in important scenes are retained. The filtered data is sorted for further analysis and visualization. This may include converting the data into a specific format, removing redundant information, or merging similar co-occurrence relationships.

[0103] In some embodiments, the sorted co-occurrence relationship data is output to a format that is easy to analyze and visualize, such as a comma-separated values ​​(CSV) file, a table, or a database.

[0104] For example, suppose there is operational data on a social media platform, and the user's travel paths in different scenarios have been generated. Now, if you want to analyze the co-occurrence relationship of users on specific topics (such as "travel" and "food") in different scenarios. Then set the co-occurrence threshold (such as 5 times) and the scene importance index (such as "hot topic discussion area" and "user personal homepage"); traverse the user's travel path in different scenarios, record each user's mention and co-occurrence of the topic; filter out the co-occurrence relationships that occur more than 5 times or frequently appear in important scenarios such as "hot topic discussion area" and "user personal homepage". Use visualization tools to analyze the co-occurrence of the topics of "travel" and "food" in different scenarios, and find out the user's preferences and associations for different topics.

[0105] The technical solution described in this embodiment can accurately capture the co-occurrence of objects in different scenes by traversing the wandering path and recording the co-occurrence relationship. By adjusting the preset co-occurrence threshold or scene importance information, co-occurrence relationship data that meets specific requirements can be generated more flexibly, thereby improving the generation efficiency and diversity of co-occurrence relationship data.

[0106] In some embodiments, the multiple scenes include scenes corresponding to different applications.

[0107] In some implementations, multiple scenarios are defined as specific environments or contexts corresponding to different applications. Each application has its own unique user interface, functions, and services, which together constitute the scenario of the application.

[0108] In some embodiments, collecting the walking path data includes: for each application, collecting the user's behavior data in its scene through a built-in behavior tracking system (such as logging, event tracking, etc.). Based on the collected behavior data, generating the user's walking path in each application scene. These paths reflect the user's navigation and interaction patterns between different pages, functions or services.

[0109] In some embodiments, generating co-occurrence relationship data includes: defining pages, functions, or services as nodes in the scenario of each application. These nodes are the basic units of user behavior. Traversing the user's travel path in each application scenario, recording the co-occurrence relationship between different nodes. The co-occurrence relationship can be defined as the number of times or frequency that two nodes appear adjacent to each other in the same travel path. If the analysis target is the co-occurrence relationship across applications, it is also necessary to integrate and associate the scenario data of different applications, which can be achieved through user ID, timestamp or other unique identifiers.

[0110] In some implementations, meaningful co-occurrence relationship data is screened based on a preset co-occurrence threshold or scenario importance index, and co-occurrence relationship data in different application scenarios are displayed using visualization tools (such as network diagrams, heat maps, etc.) to help discover user behavior patterns and associations between different applications.

[0111] For example, suppose there are two applications, one is a social media platform and the other is an e-commerce platform. Users switch between the two applications and perform different actions in each application. Users browse posts, like, comment, and share content on social media platforms, and these behaviors constitute the wandering path of social media platforms. Users search for products, view details, add to carts, and purchase on e-commerce platforms, and these behaviors constitute the wandering path of e-commerce websites. By integrating the scenario data of these two applications, it is found that after users see an advertisement for a certain product on a social media platform, they may jump to the e-commerce platform to purchase it. This co-occurrence relationship across applications provides valuable insights for personalized recommendations and market strategies. Understanding the behavioral patterns and associations of users between different applications can help companies develop more effective market strategies and marketing plans.

[0112] The technical solution described in the embodiment of the present disclosure, because multiple scenarios include scenarios corresponding to different applications, can obtain a comprehensive understanding of the user's behavior on multiple applications by integrating and analyzing the co-occurrence relationship data in different application scenarios, and thus more comprehensively understand the user's preferences and needs. This cross-application data integration not only promotes the consistency and personalization of the service experience, but also significantly improves the accuracy of recommendations and user satisfaction.

[0113] In some embodiments, the multiple scenes include scenes corresponding to different projects within the same application.

[0114] In some implementations, in a single application, the scenarios corresponding to different projects or functional modules are clearly defined. These scenarios can be divided based on multiple dimensions such as user behavior, content type, functional module, etc. For example, a social media application may include news browsing scenarios, friend interaction scenarios, video viewing scenarios, live broadcast scenarios, etc.

[0115] In some implementations, the user's behavior data in different scenarios is collected using the logging system, event tracking tools, etc. in the application. The collected data is pre-processed by cleaning, deduplication, formatting, etc. to ensure the quality and consistency of the data.

[0116] In some implementations, scene feature extraction includes: analyzing user behavior data in different scenarios and extracting key features, such as visit frequency, dwell time, number of interactions, etc. For scenes involving content (such as articles, pictures, videos, etc.), extracting content features, such as categories, tags, keywords, etc. For certain specific scenarios, it may be necessary to extract technology-related features, such as network speed, device type, etc.

[0117] In some embodiments, multi-scenario data analysis includes: discovering user behavior patterns and preferences by analyzing user behavior data in different scenarios. Exploring the association and mutual influence between different scenarios, such as the probability of a user jumping from a news browsing scenario to a video watching scenario. Based on the results of multi-scenario data analysis, providing users with personalized content recommendations, function optimization, etc.

[0118] In some implementations, the project layout, function design, content recommendation, etc. within the application are adjusted based on the analysis results. The results of multi-scenario data analysis are used to formulate more accurate marketing strategies, such as targeted advertising push, event planning, etc. A continuous data monitoring and feedback mechanism is established to promptly discover and handle problems in the application, and continuously optimize the application effect of multi-scenario data analysis.

[0119] For example, suppose there is a social media application that includes news browsing scenarios, friend interaction scenarios, and video viewing scenarios. In the news browsing scenario, users can browse the latest news information and like, comment, or share the content they are interested in. In the friend interaction scenario, users can chat, like, comment, and interact with friends to share their daily lives and interests. In the video viewing scenario, users can watch short videos or live broadcasts, like, comment, share, or purchase related products. Through multi-scenario data analysis, users' behavior patterns and preferences in different scenarios can be discovered. For example, some users prefer to get the latest news information in the news browsing scenario, while others prefer to watch entertainment content in the video viewing scenario. Based on these analysis results, personalized recommendations and optimizations can be made to the application, such as pushing more news information to users who like news and providing more video content recommendations to users who like videos.

[0120] The technical solution described in the embodiment of the present disclosure, because multiple scenarios include scenarios corresponding to different projects within the same application, can integrate and optimize co-occurrence relationship data from different projects through multi-scenario data analysis, thereby revealing the potential connections and rules between data and improving data utilization. By sharing co-occurrence relationship data, the application can have a deeper understanding of the user's preferences and behavior patterns, which enables the application to provide users with more personalized content and services, such as customized recommendations, personalized interfaces, etc., thereby improving the accuracy of recommendations and improving user satisfaction with recommendations.

[0121] In some embodiments, the multi-scenario data construction method also includes: based on the co-occurrence relationship data, constructing a first representation vector of the first object and a second representation vector of the second object in the operation summary graph; based on the similarity between the first representation vector and the second representation vector, determining the recommended resources that match the first object from the resources published by the second object.

[0122] In the embodiments of the present disclosure, co-occurrence relationship data refers to the association data between objects in different scenarios. These objects can be users, pages, commodities, etc., depending on the application scenario. Co-occurrence relationship data can be used to analyze the association and similarity between objects.

[0123] In the embodiment of the present disclosure, the representation vector is a vector used to describe the characteristics of an object. In the resource recommendation method, the representation vector is usually constructed based on co-occurrence relationship data to represent the similarity and correlation between objects.

[0124] In the embodiment of the present disclosure, similarity is an indicator for measuring the similarity between two objects. In resource recommendation methods, similarity is usually used to evaluate the proximity between the representation vectors of two objects, thereby determining the degree of matching between them.

[0125] In the embodiment of the present disclosure, the resources to be recommended are resources that meet the needs of the user and are screened out from the resource library.

[0126] In some embodiments, based on the co-occurrence relationship data, features of each object are extracted. These features may include the frequency of occurrence of the object, the number of co-occurrences with other objects, the context of co-occurrence, etc. Using the extracted features, a representation vector is constructed for each object. This vector can be a high-dimensional sparse vector, in which each dimension represents a feature, and the value of the vector represents the importance or strength of the feature.

[0127] In some embodiments, for a given first object and a second object, the similarity of their representation vectors is calculated. There are many methods for calculating similarity, such as cosine similarity, Euclidean distance, Manhattan distance, etc. The specific method to be selected depends on the application scenario and data characteristics.

[0128] In some implementations, determining the resource to be recommended includes: based on the similarity between the first representation vector and the second representation vector, selecting resources matching the first object from the resources published by the second object as the resource to be recommended. The higher the similarity, the more likely it is that the resource meets the needs and preferences of the first object.

[0129] In some implementations, the selected resources to be recommended are recommended to the first object, and feedback data of the first object is collected. The feedback data can be used to evaluate the recommendation effect and further optimize the recommendation algorithm.

[0130] For example, suppose there is an online shopping platform where user K browses multiple products and buys some of them. At the same time, user Q posts some product information on the platform. By analyzing user K's browsing and purchase records, we can build co-occurrence relationship data between user K and different products. Based on the co-occurrence relationship data, build a representation vector for user K and each product. Calculate the similarity between user K's representation vector and the representation vector of each product. Filter out products with high similarity to user K from the products posted by user Q as resources to be recommended. Recommend the selected products to user K, and collect user feedback data to optimize the recommendation algorithm.

[0131] The technical solution described in the embodiment of the present disclosure can provide personalized recommendation services for each user by constructing a representation vector and calculating similarity, thereby improving the accuracy of recommendations. By constructing a representation vector using co-occurrence relationship data, resources that match user needs can be quickly screened out, thereby improving recommendation efficiency.

[0132] Below, we take the user-author graph in the recommendation system scenario as an example to illustrate the construction of multi-scenario data.

[0133] 1. Multi-scenario user behavior graphing: Extract the user's behavior in all scenarios. When a user has a satisfactory consumption of a resource, a <user, author> pair is formed. An edge is connected between the user and the author of the user's satisfactory resource. All satisfactory consumption behaviors are summarized on a graph, and all satisfactory consumption behaviors of all users in the used scenarios can be obtained. When connecting edges, different attributes are given to the edges according to different scenarios. For example, recommendation scenarios can be divided into one-hop single column, one-hop double column, two-hop scenarios, etc. For example, user u2 consumes articles of authors a2, a4, and a6 in one-hop single column, and consumes articles of author a3 in one-hop double column. User u3 consumes articles of authors a2, a1, and a5 in one-hop double column, and consumes articles of authors a4 and a7 in two-hop scenarios. When building a graph for multiple scenarios, user nodes (u2, u3) and author nodes (a1, a2, a3, a4, a5, a6, a7) are connected by consumption behavior to form an operation summary graph, and different attributes (colors) are used to distinguish consumption behaviors in each scenario.

[0134] 2. Multi-dimensional walking: Multi-dimensional walking is performed based on the operation summary graph, and single-scene and mixed-scene walking is performed simultaneously to achieve comprehensive co-occurrence learning of single-scene and mixed-scene behaviors. According to the different attributes of the edge, walking on the same edge is single-scene walking; without limiting the edge attributes, it is mixed-scene walking. For example, in the recommendation scenario, one-hop single-column behavior, one-hop double-column behavior, two-hop behavior walking and other mixed behavior walking in single scenes and mixed scenes are distinguished. Multi-dimensional walking is performed on the operation summary graph to produce multiple walking paths, and the walking method is not limited. The walking path of a single scene (for example, a2-u2-a6-u5 for single-column behavior) and the walking path of mixed multi-scenes (for example, a1-u3-a4-u4) are produced.

[0135] 3. Semantic vector transfer: Based on multi-dimensional walking paths, co-occurrence pairs of users in single and mixed scenarios are extracted for graph representation learning and training. During the training process, all nodes use a common embedding vector under co-occurrence pairs in different scenarios. On the basis of ensuring the unique information of each scenario, the collaboration and knowledge transfer capabilities between multiple scenarios are improved, thereby more comprehensively describing user behavior and improving the expression ability of the graph model.

[0136] The graph model is trained on the multi-dimensional walking path. First, the co-occurrence pairs are generated, and then the node vector representation is optimized to maximize the pair similarity. The similarity can be measured using cosine similarity.

[0137]

[0138] Or use the inner product directly:

[0139] sim=A·B

[0140] Under the multi-dimensional wandering path, we can produce both the co-occurrence behavior of users in the same scenario and the co-occurrence behavior of users in a single scenario, so that the node vector representation can not only characterize the user's universal behavior in all scenarios, but also retain the unique preferences in a single scenario.

[0141] During training, input a positive sample pair (a pair with co-occurrence relationship, such as a2, u2, u2, u5) and N (N is a positive integer not less than 1) negative sample pairs (pairs without co-occurrence relationship, such as u1, a7, u1, u6) at the same time, and calculate the contrastive learning loss:

[0142]

[0143] where e i represents the vector of each node, Represents the vector of positive sample pair nodes that have a co-occurrence relationship with node i, e jA vector of negative sample pair nodes representing co-occurrence relationships.

[0144] 4. Recommendation based on node vector representation: In a recommendation process, the vector representation of the target user is obtained u and the vector representation of all candidate resources e a , calculate e u and all e a The similarity of the images will be recommended to the users.

[0145] In this way, through multi-scenario user behavior mapping, multi-dimensional walking and semantic vector migration, the behavior sequences of objects in single and mixed scenarios are modeled in the graph model at the same time, which can learn the commonalities of mixed scenario behaviors and the characteristics of single scenario behaviors, and improve the learning accuracy of graph representation. This multi-scenario data construction method not only solves the problem of the single learning dimension of multi-scenario behavior data in existing graph model technology, but also improves the model learning effect, which is conducive to the improvement of downstream task indicators.

[0146] The present disclosure provides a multi-scenario data construction device, such as Figure 4 As shown, the multi-scenario data construction device may include: a first construction module 401, used to construct an operation summary graph according to the operation data of the first object in multiple scenarios, the operation summary graph including a first node representing the first object, a second node representing the second object, and an edge connecting the first node and the second node, and the second object is an object for publishing resources; a first generation module 402, used to perform multi-dimensional wandering based on the operation summary graph, and generate a first wandering path in each single scenario and a second wandering path in the mixed scenario; a second generation module 403, used to generate co-occurrence relationship data between the first object and the second object based on the first wandering path and the second wandering path.

[0147] In some embodiments, the first construction module 401 includes: a production sub-module, used to construct a first node, a second node and an edge for each single scene, wherein the edge identifiers corresponding to the edges in different single scenes are different; an integration sub-module, used to construct an operation summary graph according to the first node, the second node and the edge in each single scene.

[0148] In some embodiments, in the operation summary graph, an object identifier representing a first object is recorded on a first node, an object identifier representing a second object is recorded on a second node, and an edge identifier representing a scene is recorded on an edge.

[0149] In some embodiments, the first generation module 402 includes: an identification submodule, which is used to distinguish different single scenarios according to edge identifiers in the operation summary graph; a first wandering submodule, which is used to wander in each single scenario according to the connection relationship between the first node and the second node to obtain a first wandering path; and a second wandering submodule, which is used to wander in a mixed scenario according to the connection relationship between the nodes to obtain a second wandering path.

[0150] In some embodiments, the second generation module 403 includes: a selection submodule, used to select at least two scenes from multiple scenes as target scenes; a determination submodule, used to determine the target first wandering path and the target second wandering path from the first wandering path and the second wandering path based on the edge identifiers corresponding to the target scenes; and a first generation submodule, used to traverse the target first wandering path and the target second wandering path to obtain co-occurrence relationship data.

[0151] In some embodiments, the second generation module 403 includes: a recording submodule, which is used to record the co-occurrence relationship between all nodes by traversing all nodes in the first wandering path and the second wandering path; a second generation submodule, which is used to traverse all recorded co-occurrence relationships, and perform data screening according to at least one of a preset co-occurrence number threshold and scene importance information to obtain co-occurrence relationship data.

[0152] In some embodiments, the multiple scenes include scenes corresponding to different applications.

[0153] In some embodiments, the multiple scenes include scenes corresponding to different projects within the same application.

[0154] In some embodiments, the multi-scenario data construction device further includes: a second construction module ( Figure 4 (not shown), for constructing a first characterization vector of the first object and a second characterization vector of the second object in the operation summary graph based on the co-occurrence relationship data; determining module ( Figure 4 (not shown) for determining, based on the similarity between the first characterization vector and the second characterization vector, resources to be recommended that match the first object from the resources published by the second object.

[0155] For the description of specific functions and examples of each module and submodule of the device in the embodiment of the present disclosure, reference can be made to the relevant description of the corresponding steps in the above method embodiment, which will not be repeated here.

[0156] The multi-scenario data construction device of the disclosed embodiment can learn the commonalities of mixed-scenario behaviors and the characteristics of single-scenario behaviors by simultaneously modeling the user's behaviors in single scenarios and mixed scenarios in a graph model, thereby improving the learning accuracy of graph representation.

[0157] The present disclosure provides a scenario diagram of a method for constructing multi-scenario data. Figure 5 shown.

[0158] As mentioned above, the multi-scenario data construction method provided by the embodiment of the present disclosure is applied to electronic devices. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices, and other similar computing devices.

[0159] Specifically, the electronic device can perform the following operations:

[0160] Constructing an operation summary graph according to the operation data of the first object in multiple scenarios;

[0161] Perform multi-dimensional walking based on the operation summary graph to generate the first walking path in each single scenario and the second walking path in the mixed scenario;

[0162] Co-occurrence relationship data between the first object and the second object is generated based on the first wandering path and the second wandering path.

[0163] It should be understood that Figure 5 The scene diagram shown is only illustrative and not restrictive. Those skilled in the art can Figure 5 Various obvious changes and / or substitutions can be made to the examples, and the resulting technical solutions still fall within the scope of the disclosure of the embodiments of the present disclosure.

[0164] In the technical solution disclosed herein, the acquisition, storage and application of user personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0165] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0166] Figure 6 A schematic block diagram of an example electronic device 600 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0167] like Figure 6 As shown, the device 600 includes a computing unit 601, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 602 or a computer program loaded from a storage unit 608 to a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the device 600 can also be stored. The computing unit 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0168] A number of components in the device 600 are connected to the I / O interface 605, including: an input unit 606, such as a keyboard, a mouse, etc.; an output unit 607, such as various types of displays, speakers, etc.; a storage unit 608, such as a disk, an optical disk, etc.; and a communication unit 609, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 609 allows the device 600 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0169] The computing unit 601 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSP), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 601 performs the various methods and processes described above, such as a multi-scenario data construction method. For example, in some embodiments, the multi-scenario data construction method may be implemented as a computer software program, which is tangibly included in a machine-readable medium, such as a storage unit 608. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 600 via the ROM 602 and / or the communication unit 609. When the computer program is loaded into the RAM 603 and executed by the computing unit 601, one or more steps of the multi-scenario data construction method described above may be performed. Alternatively, in other embodiments, the computing unit 601 may be configured to execute the multi-scenario data construction method in any other appropriate manner (eg, by means of firmware).

[0170] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0171] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0172] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0173] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a cathode ray tube (CRT) or a liquid crystal display (LCD) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0174] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: Local Area Networks (LANs), Wide Area Networks (WANs), and the Internet.

[0175] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0176] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0177] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the principles of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A multi-scenario data construction method, comprising: Constructing an operation summary graph according to operation data of a first object in multiple scenarios, the operation summary graph comprising a first node representing the first object, a second node representing a second object, and an edge connecting the first node and the second node, wherein the second object is an object for publishing resources; Perform multi-dimensional walking based on the operation summary graph to generate a first walking path in each single scene and a second walking path in a mixed scene, wherein the mixed scene includes at least two of the single scenes; Co-occurrence relationship data between the first object and the second object is generated based on the first travel path and the second travel path.

2. The method according to claim 1, wherein: The step of constructing an operation summary graph according to the operation data of the first object in multiple scenarios includes: For each of the single scenarios, construct the first node, the second node, and the edge, wherein the edge identifiers corresponding to the edges in different single scenarios are different; The operation summary graph is constructed according to the first node, the second node, and the edge in each of the single scenarios.

3. The method according to claim 2, wherein: In the operation summary graph, the first node records an object identifier representing the first object, the second node records an object identifier representing the second object, and the edge records the edge identifier representing the scene.

4. The method according to claim 1, wherein: The performing multi-dimensional walking based on the operation summary graph to generate a first walking path in each single scene and a second walking path in the mixed scene includes: Differentiate different single scenarios according to edge identifiers in the operation summary graph; In each of the single scenarios, walking is performed according to the connection relationship between the first node and the second node to obtain the first walking path; In the hybrid scenario, the second wandering path is obtained by wandering according to the connection relationship between the first node and the second node.

5. The method according to claim 1, wherein: The generating co-occurrence relationship data between the first object and the second object based on the first wandering path and the second wandering path includes: Selecting at least two of the single scenes from the multiple scenes as target scenes; Determining a target first wandering path and a target second wandering path from the first wandering path and the second wandering path based on the edge identifier corresponding to the target scene; The target first wandering path and the target second wandering path are traversed to obtain the co-occurrence relationship data.

6. The method according to claim 1, wherein: The generating co-occurrence relationship data between the first object and the second object based on the first wandering path and the second wandering path includes: By traversing all nodes in the first wandering path and the second wandering path, recording the co-occurrence relationship between all the nodes; The co-occurrence relationships of all records are traversed, and data is screened according to at least one of a preset co-occurrence number threshold and scene importance information to obtain the co-occurrence relationship data.

7. The method according to claim 1, wherein: The multiple scenes include scenes corresponding to different application programs, or scenes corresponding to different projects within the same application program.

8. The method according to claim 1, wherein: The method further comprises: Based on the co-occurrence relationship data, construct a first representation vector of the first object and a second representation vector of the second object in the operation summary graph; Based on the similarity between the first characterization vector and the second characterization vector, a resource to be recommended that matches the first object is determined from the resources published by the second object.

9. A multi-scenario data construction device, comprising: A first construction module is used to construct an operation summary graph according to operation data of a first object in multiple scenarios, wherein the operation summary graph includes a first node representing the first object, a second node representing a second object, and an edge connecting the first node and the second node, wherein the second object is an object for publishing resources; A first generating module, configured to perform multi-dimensional walking based on the operation summary graph, and generate a first walking path in each single scene and a second walking path in a mixed scene, wherein the mixed scene includes at least two of the single scenes; The second generating module is used to generate co-occurrence relationship data between the first object and the second object based on the first wandering path and the second wandering path.

10. The device according to claim 9, wherein: The first building block comprises: A making submodule, used for constructing the first node, the second node and the edge for each single scene, wherein the edge identifiers corresponding to the edges in different single scenes are different; An integration submodule is used to construct the operation summary graph according to the first node, the second node and the edge in each of the single scenarios.

11. The device according to claim 10, wherein: In the operation summary graph, the first node records an object identifier representing the first object, the second node records an object identifier representing the second object, and the edge records the edge identifier representing the scene.

12. The device according to claim 9, wherein: The first generating module comprises: an identification submodule, used to distinguish different single scenarios according to edge identifiers in the operation summary graph; A first walking submodule, configured to walk according to a connection relationship between the first node and the second node in each of the single scenarios to obtain the first walking path; The second wandering submodule is used to wander according to the connection relationship between the first node and the second node in the hybrid scenario to obtain the second wandering path.

13. The device according to claim 9, wherein: The second generating module comprises: A selection submodule, used for selecting at least two of the single scenes from the multiple scenes as target scenes; a determination submodule, configured to determine a target first wandering path and a target second wandering path from the first wandering path and the second wandering path based on an edge identifier corresponding to the target scene; The first generating submodule is used to traverse the first target wandering path and the second target wandering path to obtain the co-occurrence relationship data.

14. The device according to claim 9, wherein: The second generating module comprises: A recording submodule, configured to record the co-occurrence relationship between all nodes by traversing all nodes in the first wandering path and the second wandering path; The second generation submodule is used to traverse the co-occurrence relationship of all records, and perform data screening according to at least one of a preset co-occurrence number threshold and scene importance information to obtain the co-occurrence relationship data.

15. The device according to claim 9, wherein: The multiple scenes include scenes corresponding to different application programs, or scenes corresponding to different projects within the same application program.

16. The device according to claim 9, wherein: The device also includes: A second construction module, configured to construct a first representation vector of the first object and a second representation vector of the second object in the operation summary graph based on the co-occurrence relationship data; A determination module is configured to determine, based on the similarity between the first characterization vector and the second characterization vector, a resource to be recommended that matches the first object from the resources published by the second object.

17. An electronic device comprising: at least one processor; as well as a memory communicatively connected to at least one processor; wherein, The memory stores instructions that can be executed by at least one processor, and the instructions are executed by at least one processor to enable the at least one processor to perform the method of any one of claims 1 to 8.

18. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are for causing a computer to execute a method according to any one of claims 1-8.

19. A computer program product comprising a computer program stored on a storage medium, the computer program implementing the method according to any one of claims 1 to 8 when executed by a processor.

Images