Unified identity generation method and device and storage medium

Abstract

The invention discloses a unified identity generation method. The method comprises the following steps: acquiring a plurality of advertisement request events, wherein each advertisement request event comprises at least two mutually associated feature identifiers; according to the plurality of advertisement request events, determining a plurality of feature identifiers and association relationships among the plurality of feature identifiers, each association relationship having a corresponding event time; a connected subgraph is generated by taking a group of feature identifiers with the incidence relation in the plurality of feature identifiers as nodes and taking the incidence relation between the feature identifiers as edges, each edge of the connected subgraph has a corresponding weight factor, and the weight factor is attenuated along with the event time of the corresponding incidence relation; and generating a corresponding unified identity label based on the connected sub-graph.

Description

Technical Field

[0001] This application relates to Internet communication technology, and more specifically, to a method, apparatus and storage medium for generating a unified identity identifier. Background Technology

[0002] In existing internet service businesses, user behavior, device IDs, and account data are scattered across databases on different platforms, making direct correlation difficult. For example, a user might browse ads on a short video app and then order the promoted product on an e-commerce platform. Because the short video app's backend might only record the user's device ID when browsing ads, and only the user's phone number when placing an order on the e-commerce platform, these two actions would be identified as being performed by two different users. This difficulty in correlating such data is known as the data silo problem. To solve this problem and integrate data from different platforms, OneID (Unified Identity) technology was proposed. OneID technology maps various IDs, such as device identifiers (i.e., IDs like MAC address, IMEI, IMSI, etc.), platform accounts, and phone numbers, to a unified OneID, thus defining a user's unique identity. By relying on OneID to identify potential advertising users, advertisers can achieve precise marketing.

[0003] Existing technologies can use rule engines to associate heterogeneous data from multiple platforms, such as structured and unstructured data (e.g., logs, images, and text), to generate OneIDs. However, this method relies on manually defined association rules, which is inefficient and prone to incomplete coverage due to human oversight. Furthermore, if new types of data are received, they cannot be processed because the rule engine does not predefine them, resulting in poor scalability for rule-engine-based OneID technology. In addition, rule-engine-based OneID generation methods cannot handle real-time changes in user behavior such as device switching and multi-account operations, and are susceptible to noise data interference, potentially leading to misidentification of users.

[0004] Existing technologies can use clustering algorithms such as K-means to generate OneIDs by associating user identifiers with data similarity. However, data generated from user behavior across platforms is high-dimensional and sparse. For example, data generated when browsing an e-commerce platform without logging in and logging in on a social media platform may misclassify the same user's behavior into different clusters due to the lack of dense connections that clustering depends on, leading to misidentification of the user.

[0005] Existing technologies also utilize graph algorithms to generate OneIDs. By representing scattered IDs with vertices and the relationships between IDs with edges, graph algorithms can automatically map multiple IDs of the same user to a unique OneID by identifying connected subgraphs, without requiring pre-defined rules. Compared to clustering algorithms, graph algorithms can also handle multi-source heterogeneous data well. Furthermore, graph algorithms exhibit strong dynamic reasoning capabilities when generating OneIDs, adapting well to newly added relationships generated by dynamic updates; for example, representing data from different periods with a unified OneID. However, existing graph algorithms generate OneIDs that remain valid indefinitely, leading to an excessive number of OneIDs and a large amount of redundant and invalid data. Once the total number of OneIDs reaches its limit, it becomes impossible to assign OneIDs to new IDs.

[0006] Therefore, an improved method for generating OneID is needed. Summary of the Invention

[0007] One objective of this application is to provide a unified identity generation method, apparatus, and storage medium to solve the problem of a large amount of redundant data in existing graph-based OneID generation methods.

[0008] In one aspect of this application, a unified identity generation method is provided. The method includes: acquiring multiple advertising request events, wherein each advertising request event includes at least two related feature identifiers; determining multiple feature identifiers and their relationships based on the multiple advertising request events, wherein each relationship has a corresponding event time; generating a connected subgraph using a group of related feature identifiers as nodes and the relationships between the feature identifiers as edges, wherein each edge of the connected subgraph has a corresponding weight factor, the weight factor decaying with the event time of the corresponding relationship; and generating a corresponding unified identity identifier based on the connected subgraph.

[0009] Another aspect of this application provides a unified identity generation apparatus based on a connected graph, the unified identity generation apparatus including a non-transitory computer storage medium storing one or more executable instructions, the one or more executable instructions being executed by a processor to implement a unified identity generation method.

[0010] In another aspect, this application provides a non-transitory computer storage medium storing one or more executable instructions, which, when executed by a processor, implement a unified identity generation method.

[0011] The above is an overview of this application, and there may be simplifications, generalizations, and omissions of details. Therefore, those skilled in the art should recognize that this section is merely illustrative and not intended to limit the scope of this application in any way. This overview section is neither intended to identify the key or essential features of the claimed subject matter nor to serve as an aid in determining the scope of the claimed subject matter. Attached Figure Description

[0012] The above and other features of this application will become more fully clear through the following description and appended claims, in conjunction with the accompanying drawings. It is understood that these drawings depict only a few embodiments of the application and should not be construed as limiting the scope of the application. The application will be described more clearly and in more detail through the use of the drawings.

[0013] Figure 1 This application illustrates a unified identity generation method for all scenarios according to one embodiment of the present application;

[0014] Figure 2a An example of multiple connected subgraphs generated based on the ad request events in Table 1 is shown;

[0015] Figure 2b It shows in Figure 2a An example of multiple connected subgraphs obtained by updating the weight factors and removing edges whose weight factors are less than a predetermined weight threshold;

[0016] Figure 3 This application illustrates a unified identity generation method for incremental scenarios according to one embodiment of the present application;

[0017] Figure 4 It shows in Figure 2b Based on this, an example of generating multiple connected subgraphs using the ad request events in Table 2 as incremental ad request events. Detailed Implementation

[0018] In the following detailed description, reference is made to the accompanying drawings, which form a part thereof. In the drawings, similar symbols generally denote similar components unless the context otherwise requires. The illustrative embodiments described in the detailed description, drawings, and claims are not intended to be limiting. Other embodiments and variations may be employed without departing from the spirit or scope of the subject matter of this application. It will be understood that various different configurations, substitutions, combinations, and designs can be made to the various aspects of the general description and illustrated in the drawings of this application, all of which explicitly form part of the subject matter of this application.

[0019] To address the data silo problem and achieve unified identification and data integration of user identities across platforms, this application provides a method for generating unified identity identifiers. This method uses a graph algorithm to generate OneIDs, requiring no pre-defined rules and effectively handling multi-source heterogeneous data. Specifically, the method sets a weight factor that decays over time as the ad request event occurs. This reduces the impact of ad request events with excessively small weight factors on OneID calculation and generation, and in some cases, can even be disregarded. In this way, the raw data of ad request events can be flexibly processed, avoiding the generation of excessive OneID data that could negatively impact the efficiency and reliability of the database system.

[0020] Figure 1 This illustration shows a unified identity generation method 100 according to an embodiment of this application, targeting a full-scale scenario, i.e., all existing advertising request events. For example... Figure 1 As shown, the method 100 includes steps S101-S104.

[0021] In step S101, multiple ad request events are acquired.

[0022] An ad request event refers to the process by which a user triggers ad display conditions, such as opening a software application or browsing a webpage, and the relevant application or software initiates an ad request to an advertising platform. Data in an ad request event can include ad behavior logs, text, images, videos, etc., and is generally unstructured data, i.e., data without a predefined model or not organized in a predefined way. In some cases, ad request events may also include structured data. Structured data refers to a data representation in which the data structure of each record, composed of data elements, is consistent and can be effectively described using a relational model. In some embodiments, the method of this application can process both structured and unstructured data simultaneously, or only one type of data.

[0023] Feature identification refers to information used in advertising or other similar businesses to identify users, typically including user identifiers and device identifiers. User identifiers uniquely identify a user and can be represented by a user's account on a specific platform or database, ID number, biometric identifier, etc. Device identifiers identify physical devices (such as mobile phones, tablets, personal computers, etc.) and are usually bound to hardware or software environments. Device identifiers can include device "fingerprints," which are generated by collecting and analyzing the device's hardware and software characteristics. For example, a device fingerprint can be generated by combining parameters such as the device's hardware identifiers (e.g., MEI, MAC address) and software configurations (operating system version, etc.). Generally, device fingerprints are unique and cannot be reset, but they are susceptible to forgery and impersonation. Therefore, to meet privacy compliance requirements and avoid directly exposing user identities, device identifiers can also be anonymized, forming anonymized device identifiers such as IDFA, CAID, IDFV, AAID, OAID, etc. Among them, IDFA is a dedicated advertising identifier designed by Apple for iOS devices; CAID is an alternative launched by the China Advertising Association when IDFA cannot be obtained on iOS devices; IDFV is also an identifier for iOS devices, and applications from the same developer share the same IDFV; AAID is a dedicated advertising identifier for Android devices; OAID is designed as a dedicated advertising identifier for Android devices in China.

[0024] Feature identifiers are typically included in advertising behavior logs. After obtaining an advertising request event, the advertising behavior log can be parsed using an existing language model to obtain the feature identifiers. In some embodiments of this application, each advertising request event may include at least two related feature identifiers. When two or more feature identifiers exist simultaneously in the same advertising request event, these two or more feature identifiers can be considered related. In other cases, an advertising request event may also include only a single feature identifier. The relationship between feature identifiers cannot be obtained through a single feature identifier, but for this type of feature identifier, a unified identity identifier can be directly assigned to it instead of using a connected graph. When obtaining an advertising request event, the time when the advertising request occurred can also be obtained simultaneously as the event time.

[0025] Table 1 provides examples of multiple ad request events across the entire scenario. Ad request event 1 includes only IDFA identifier A, and occurred on January 1, 2025. Ad request event 2 includes both IDFA identifier A and CAID identifier B, and occurred on January 1, 2025. Ad request event 3 includes both CAID identifier B and IDFV identifier C, and occurred on January 1, 2025. Ad request event 4 includes only IDFA identifier D, and occurred on January 1, 2025. Ad request event 5 includes only CAID identifier E, and occurred on January 1, 2025. Ad request event 6 includes both IDFA identifier D and CAID identifier E, and occurred on January 2, 2025. Ad Request Event 7 includes only the IDFV identifier F, and Ad Request Event 7 occurred on January 2, 2025.

[0026] Table 1. Multiple ad request events in the full scenario.

[0027]

[0028] It is understood that the multiple ad request events shown in Table 1 are only examples. In actual applications, each event may include three or more feature identifiers.

[0029] In step S102, multiple feature identifiers and their associations are determined based on multiple advertising request events.

[0030] Based on the feature identifiers obtained in each ad request event in step S101, for identical feature identifiers, only one is retained, thereby determining the multiple feature identifiers contained in all ad request events. When an ad request event contains two or more feature identifiers, it can be determined that there is a correlation between the two or more feature identifiers, and each correlation can have a corresponding event time. In some embodiments, when at least two ad request events contain a correlation between the same two feature identifiers, the event time closest to the current time can be used as the event time corresponding to the correlation between the two feature identifiers. In this way, the amount of data for ad request events in subsequent processing can be reduced.

[0031] The following uses the ad request events listed in Table 1 as an example to further illustrate how to obtain multiple feature identifiers and the relationships between them. Based on ad requests 1 to 7, it can be determined that the ad request events in Table 1 include feature identifiers A, B, C, D, E, and F. Specifically, based on ad request events 2 and 3, it can be determined that: there is a relationship between IDFA identifier A, CAID identifier B, and IDFV identifier C, and the event time corresponding to the relationship between IDFA identifier A and CAID identifier B is the event time of event 2, i.e., 2025.1.1; the event time corresponding to the relationship between CAID identifier B and IDFV identifier C is the event time of event 3, i.e., 2025.1.2. Based on ad request event 6, it can be determined that there is a relationship between IDFA identifier D and CAID identifier E, and the event time corresponding to this relationship is the event time of event 6, i.e., 2025.1.2. Based on ad request event 7, it can be determined that IDFV identifier F is an independent feature identifier and has no relationship with the other feature identifiers in Table 1.

[0032] In step S103, a connected subgraph is generated by using a group of feature identifiers that have a relationship among multiple feature identifiers as nodes and the relationships between feature identifiers as edges. It can be understood that when there are multiple relationships, each relationship can generate a connected subgraph, and the overall structure of these connected subgraphs represents a connected graph representing the relationships between all feature identifiers.

[0033] Based on graph computing theory, an undirected graph is a graph structure composed of nodes and edges, where edges are used only to represent bidirectional relationships between vertices and have no directionality. If there is at least one path from one node to another in an undirected graph, then the two nodes are said to be connected. If there are at least one pair of nodes in an undirected graph that are not connected by a path, i.e., it contains multiple independent parts (also called a connected subgraph), then the undirected graph is a disconnected graph. In a connected subgraph, there is a path between any two nodes. A connected subgraph is a special type of connected subgraph where all nodes are interconnected and cannot be kept connected by adding other nodes or edges. To save on the number of unified identity identifiers and associate as many special identifiers as possible with each unified identity identifier to provide precise advertising and marketing, the connected subgraphs in some embodiments of this application are preferably maximal connected subgraphs.

[0034] In some embodiments of this application, each edge of the connected subgraph is assigned a corresponding weight factor, and the weight factor decays with the event time of the corresponding association. The magnitude of the weight factor reflects the strength of the association between the feature identifiers connected by the edge; a larger weight factor indicates a stronger association between the feature identifiers connected by the edge, and a smaller weight factor indicates a weaker association between the feature identifiers connected by the edge. The weight factor can be determined as follows: first, determine the current time, and then determine the decay of the weight factor based on the time difference between the current time and the event time of the association corresponding to the weight factor. Generally speaking, the larger the time difference, the further away the event time that establishes the association is from the current time, and the weaker the association between the feature identifiers connected by the edge, thus the smaller the determined weight factor; the smaller the time difference, the closer the event time that establishes the association is to the current time, and the stronger the association between the feature identifiers connected by the edge, thus the larger the determined weight factor.

[0035] In some embodiments of this application, the attenuated weighting factor W(t) can be determined according to the following equation (1):

[0036] W(t) = W0 * e -λt (1)

[0037] Wherein, W0 is the preset initial weight factor, λ is the preset decay rate, and t is the time difference between the current time and the event time corresponding to the weight factor. In some embodiments of this application, the preset initial weight factor for different sides can be set to the same value, for example, uniformly set to 1. In other embodiments of this application, the preset initial weight factor can be set to different values ​​according to the source of the advertising request event. For example, for services with higher stability, the value of the preset initial weight factor can be set larger, and for services with lower stability, the value of the preset initial weight factor can be set smaller. To save computing resources, unified identity identifiers are usually generated on a daily basis, so the preset decay rate can be set to change by a certain value every day, with λ ranging from 0 to 1 / day, and its specific value determined by the business during actual execution. In some embodiments of this application, the preset decay rate for different sides can be set to the same value, for example, uniformly set to 0.1 / day. In other embodiments of this application, the preset decay rate for different sides can be set to different values. Of course, the preset decay rate can also be set to other values ​​according to the actual business needs, and this embodiment does not impose specific limitations.

[0038] Using the data in Table 1 as a basis, and assuming the current time is January 11, 2025, the preset decay rate is 0.1 / day, and the preset initial weight factor for all edges is 1, this will be explained as an example. Based on the association relationships obtained in step S102, a disconnected graph can be generated, which includes multiple connected subgraphs. Figure 2aAn example of multiple connected subgraphs generated based on the ad request events in Table 1 is shown. For example... Figure 2a As shown, using the set of feature identifiers IDFA identifier A, CAID identifier B, and IDFV identifier C as nodes, a connected component is generated. Figure 1 In the connected sub Figure 1 In the process, the first edge is established between IDFA identifier A and CAID identifier B. The time difference between the current time and the event time of event 2, which is associated with the first edge, is 10. The weight factor corresponding to the first edge is calculated as 1*e. -0.1*10 =0.37. Similarly, a second edge is established between CAID identifier B and IDFV identifier C, and the weight factor for the second edge is calculated to be 0.41. Similarly, using the feature identifiers IDFA identifier D and CAID identifier E as nodes, a connected subgraph 2 is generated. In connected subgraph 2, a third edge is established between IDFA identifier D and CAID identifier E, and similarly, the weight factor for the third edge is calculated to be 0.41. Using the isolated node IDFA identifier F as a connected component... Figure 3 However, since it is an isolated node, this connected component... Figure 3 There are no edges in it.

[0039] In step S104, a corresponding unified identity identifier is generated based on the connected subgraph.

[0040] In some embodiments of this application, when a connected subgraph includes multiple feature identifiers, one feature identifier can be selected, and a corresponding unified identity identifier can be generated based on that feature identifier. Specifically, one of the multiple feature identifiers can be randomly selected, or certain feature identifiers or a certain type of feature identifiers can be selected preferentially according to a preset rule and a certain order, and a corresponding unified identity identifier can be generated based on the selected feature identifier. After selecting the feature identifiers in the connected subgraph, a predetermined function can be used to generate the corresponding unified identity identifier with the feature identifier as the independent variable.

[0041] In other embodiments of this application, a unique numerical identifier can be assigned to each feature identifier contained in the connected subgraph. Specifically, a pre-defined function can be used to generate a corresponding unique numerical identifier for each feature identifier. For example, multiple feature identifiers included in the connected subgraph can be stored in list form or a similar data indexing method, for example, storing one feature identifier per row. Then, the row_number function can be used to generate a corresponding unique numerical identifier for each feature identifier. The row_number function is a window function that can assign a unique number to each row in the query result set. Taking the data in Table 1 as an example, the list is arranged according to connected subgraphs... Figure 1-3 The sequential storage stores the feature identifiers contained in all connected subgraphs. `row_number` represents the number of connected subgraphs. Figure 1-3 Assign integers starting from 1 as numerical identifiers. Assuming the list is assigned numerical identifiers in the order of feature identifiers A, B, C, D, E, and F, then the numerical identifiers corresponding to feature identifiers A, B, C, D, E, and F are 1, 2, 3, 4, 5, and 6, respectively. Finally, one of the numerical identifiers in the connected subgraph can be selected according to a preset rule, such as selecting the minimum or maximum value as the unified identifier of the connected subgraph. In this application, the minimum value among all numerical identifiers in the connected subgraph is used as the unified identifier for the connected subgraph as an example for explanation. Therefore, the connected subgraph... Figure 1 Contains the numbers 1, 2, and 3, and connects the substrings. Figure 1 The unified identity identifier is 1. Connected subgraph 2 contains the numeric identifiers 4 and 5; the unified identity identifier of connected subgraph 2 is 4. Connected subgraph... Figure 3 It contains the number 6, and its unified identity identifier is 6.

[0042] In this application, the weight factors of the edges in the connected subgraph are updated by triggering events, and the weight factors are set to decay with the event time of the corresponding association, thus achieving a balance between association reasoning and historical behavior decay. Therefore, the unified identity identifier generated in the end has timeliness, which is conducive to the accurate identification and association of user identity information.

[0043] Since edge weights decay over time, when two feature identifiers connected by an edge have not appeared in the same event for a long period, the edge weight will decay to a very small value. This indicates that the probability of the two feature identifiers appearing simultaneously is low, less than the probability of each feature identifier appearing alone. Therefore, it can be assumed that there is no direct correlation between the two feature identifiers, or the probability of a direct correlation recurring is low. Consequently, these two feature identifiers can correspond to their respective unified identity identifiers, rather than sharing the same unified identity identifier. For example, if a user discards an old mobile phone, the device identifier associated with that phone will no longer appear in the future. Clearing the association data associated with that phone and the edges in the connected subgraph can save database storage space and improve subsequent data processing efficiency. Therefore, in some embodiments of this application, before step S104, edges with weight factors less than a predetermined weight threshold can be removed from the connected subgraph. After removing some edges, the connectivity between nodes in the connected subgraph may change. Therefore, the connected subgraph can be further updated to generate an updated connected subgraph, and a corresponding unified identity identifier can be generated based on the updated connected subgraph.

[0044] In some embodiments of this application, changes in the connectivity between nodes in a connected subgraph may cause an original connected subgraph to be split into multiple connected subgraphs. This step may specifically include: determining whether the connected subgraph includes multiple groups of nodes that are not connected by any edges; and in response to determining that the connected subgraph includes multiple groups of nodes that are not connected by any edges, generating multiple updated connected subgraphs based on each group of nodes and the edges connecting that group of nodes.

[0045] The following explanation uses the data in Table 1 and the connected subgraph updated every 10 days as an example (refer to the calculation method of equation (1)). Of course, the connected subgraph can also be updated at other intervals, such as every day, every 2 days, every 5 days, etc. In this application, the time interval for updating is not restricted. Assuming the current time is January 21, 2025, and no new advertising requests are received from January 11, 2025 to January 20, 2025, only the weights of each edge are updated: the weight factor of the first edge is updated to 0.14, and the weight factors of the second and third edges are updated to 0.15. Assuming the predetermined weight threshold is 0.15, since the weight factor of the first edge 0.14 is less than the predetermined weight threshold of 0.15, and the weight factors of the second and third edges 10.15 are both greater than or equal to the predetermined weight threshold of 0.15, the first edge is removed from the connected subgraph, and the second and third edges are retained. Figure 2b It shows in Figure 2a This is an example of multiple connected subgraphs obtained by updating the weight factors and removing edges whose weight factors are less than a predetermined weight threshold. For example... Figure 2b As shown, after removing the first edge, the connected sub-element Figure 1 Split into connected components Figure 1 a and 1b, connected subgraph 2 and connected subgraph Figure 3 It remains unchanged. Among them, the connected component Figure 1 a includes the IDFA identifier A of the isolated node. Connected sub-nodes Figure 1 b includes the CAID identifier B and the IDFV identifier C. Let's continue with the example of selecting the minimum value among all numerical identifiers in a connected subgraph as the unified identifier for that subgraph. In this case, the connected subgraph... Figure 1 'a' contains the numeric identifier 1, and the connected sub-sub- Figure 1 The unified identity identifier of 'a' is 1; connected sub Figure 1 b contains the digits 2 and 3, and is a connected component. Figure 1 The unified identity of b is 2. The unified identity of the connected subgraph 2 remains 4. Figure 3 The unified identity identifier remains at 6.

[0046] In real-world business scenarios, fraudulent order placement (or fake advertising requests) may occur. In such cases, a single device identifier may be associated with many user identifiers, resulting in a very large number of edges in the connected subgraph. Similarly, in scenarios where multiple feature identifiers are artificially bound to the same feature identifier—for example, assigning a default user identifier to unlogged-in users—edges may form between unrelated device identifiers and the default user identifier, creating a false connected subgraph with a very large number of edges.

[0047] To address the data issues arising from abnormal advertising request events caused by fraudulent order placement, some embodiments of this application provide an exception handling mechanism. Specifically, before generating the corresponding unified identity identifier based on the connected subgraph, the number of edges included in the connected subgraph is determined. If the number of edges exceeds a predetermined edge count threshold, at least a portion of the edges and nodes unrelated to the remaining edges are removed from the connected subgraph, ensuring that the number of edges does not exceed the predetermined edge count threshold. This allows nodes and edges that might affect real-world business scenarios to be removed, thereby improving the robustness of the unified identity identifier generation method against abnormal scenarios.

[0048] In some embodiments of this application, a predetermined edge number threshold can be determined based on the characteristics of normal business. For example, the predetermined edge number threshold can be set as the upper limit of the number of edges in normal business. Alternatively, the predetermined edge number threshold can be dynamically adjusted based on an algorithm, such as dynamically generating the predetermined edge number threshold based on business time periods and / or historical data.

[0049] In practice, edges and nodes to be removed can be determined based on various rules. For example, a single device identifier typically has no more than three accounts associated with the same application. Based on this, with some tolerance, edges and nodes with more than five accounts associated with the same application for a single device identifier can be removed. Additionally, edges and nodes can be removed based on other conflicting node attributes, such as those with excessively large geographical spans within a short period.

[0050] Figure 3 This application illustrates a unified identity generation method 300 according to an embodiment of the present application for incremental scenarios (i.e., based on an existing unified identity, for the receipt of an incremental advertising request event). Figure 3 As shown, the method 300 includes steps S301-S304.

[0051] In step S301, one or more incremental ad request events are acquired.

[0052] For cases where no new ad request event acquisition channels are added, the event time of an incremental ad request event is after the generation of the existing unified identity. For cases where new ad request event acquisition channels are added, all ad request events through that channel can be classified as incremental ad request events. This includes ad request events whose event time is after the generation of the existing unified identity, as well as ad request events whose event time is before the generation of the existing unified identity, but whose acquisition time is after the generation of the existing unified identity. A description of incremental ad request events can be found in [reference needed]. Figure 1 The description of ad request events in all scenarios will not be repeated in this embodiment.

[0053] Table 2 provides examples of multiple incremental ad request events acquired on January 31, 2025. After January 21, 2025, new channels were added for acquiring ad request events from Android devices. Therefore, the new ad request events can include both ad request event 8 and ad request event 10 from these new channels. Ad request event 8 includes IDFV identifier C and OAID identifier G, and its occurrence date is January 13, 2025. Ad request event 10 includes IDFV identifier I and OAID identifier J, and its occurrence date is January 25, 2025. Furthermore, the new ad request events also include ad request events 9 and 11 acquired from existing channels. Ad request event 9 includes IDFA identifier H and CAID identifier B, and its occurrence date is January 22, 2025. Ad request event 11 includes CAID identifier E and IDFV identifier F, and the date of ad request event 11 is January 30, 2025.

[0054] Table 2. Multiple ad request events in incremental scenarios.

[0055]

[0056] In step S302, based on multiple incremental advertising request events, multiple incremental feature identifiers and incremental association relationships between the multiple incremental feature identifiers are determined. Each incremental association relationship has a corresponding event time.

[0057] In some embodiments of this application, taking the acquisition of multiple incremental ad request events listed in Table 2 as an example, we illustrate how to determine multiple incremental feature identifiers and the association relationships between them. Based on ad request events 8-11, the newly added feature identifiers can be determined to be G, H, I, and J. Based on ad request event 8, it can be determined that there is an association relationship between IDFV identifier C and OAID identifier G, and the event time corresponding to the association relationship between IDFV identifier C and OAID identifier G is the event time of event 8. Based on ad request event 9, it can be determined that there is an association relationship between IDFA identifier H and CAID identifier B, and the event time corresponding to the association relationship between IDFA identifier H and CAID identifier B is the event time of event 9. Based on ad request event 10, it can be determined that there is an association relationship between IDFV identifier I and OAID identifier J, and the event time corresponding to the association relationship between IDFV identifier I and OAID identifier J is the event time of event 10. Based on the ad request event 11, it can be determined that there is an association between CAID identifier E and IDFV identifier F, and the event time corresponding to the association between CAID identifier E and IDFV identifier F is the event time of event 11.

[0058] In step S303, a set of incremental feature identifiers that are the same as or related to a set of feature identifiers that are nodes in the existing connected subgraph are selected from multiple incremental feature identifiers, and the existing connected subgraph is updated to generate an updated connected subgraph by using the association relationship between the set of incremental feature identifiers as edges.

[0059] In step S304, a corresponding unified identity identifier is generated based on the updated connected subgraph.

[0060] In this application, updating a connected subgraph to generate an updated connected subgraph includes the following three methods.

[0061] Method 1: Dynamically add nodes and edges to the existing connected subgraph to obtain the updated connected subgraph.

[0062] Method 2: A connected subgraph is connected to at least one or more other connected subgraphs to form an updated connected subgraph.

[0063] Method 3: Incremental feature identification generates new connected subgraphs. This can be referenced from [the previous method]. Figure 1 The method for generating connected subgraphs will not be described in detail in this application.

[0064] by Figure 2bUsing the existing connected subgraph as an example, this illustrates how to generate an updated connected subgraph and generate corresponding unified identity identifiers based on the updated connected subgraph. First, using the aforementioned method, a unique numerical identifier 7, 8, 9, and 10 are assigned to each of the newly added feature identifiers G, H, I, and J. Second, feature identifiers B, C, E, and F are... Figure 2b The existing feature identifiers, feature identifiers G and H are related to... Figure 2b The existing feature identifiers are associated with the feature identifiers, and feature identifiers I and J do not belong to the feature identifiers. Figure 2b The existing feature identifiers are not the same as those in the text. Figure 2b The existing feature identifiers are associated with each other. Therefore, feature identifiers B, C, E, F, G, and H are incremental feature identifiers. Figure 4 It shows in Figure 2b Based on this, using the ad request events in Table 2 as incremental ad request events, examples of multiple connected subgraphs are generated. For example... Figure 4 As shown, connected sub Figure 1 a remains unchanged. Without change Figure 2b Connecting elements Figure 1 Based on the existing edges of b, the new nodes G and H, along with their associated edges, are dynamically added to the connected component. Figure 1 b, thus obtaining the updated connected component Figure 1 b. Updated connected components Figure 1 b contains feature identifiers B, C, G, and H, where: the edge between B and C remains unchanged, but its weight factor changes to 0.05; a fourth edge is added between B and H, with a weight factor of 0.41; and a fifth edge is added between C and G, with a weight factor of 0.17. Since event 11 associated feature identifiers E and F, the connected subgraph 2 and the connected subgraph... Figure 3 The connection is transformed into an updated connected subgraph 2, which contains feature identifiers D, E, and F. Specifically, the edge between D and E remains unchanged, but its weight factor changes to 0.06; a fifth edge is added between E and F, with a weight factor of 0.9. Then, a corresponding unified identity identifier can be generated based on the updated connected subgraph. The unified identity identifier is generated using the same method described above. The connected subgraph... Figure 1 The unified identity identifier of 'a' remains at 1; connected sub-identity Figure 1 If subgraph b contains the digits 2, 3, 7, and 8, then the minimum value 2 is selected as the unified identity identifier; if the connected subgraph 2 contains the digits 4, 5, and 6, then the minimum value 4 is selected as the unified identity identifier; connected subgraph Figure 3 If the data contains the digits 9 and 10, the minimum value 9 is selected as the unified identity identifier. Furthermore, to prevent confusion caused by duplicate digits in incremental data, the numerical identity identifier can be the sum of the calculated result of the `row_number` function and the maximum value of existing digits.

[0065] Of course, before generating the corresponding unified identity based on the updated connected subgraph, edges with weight factors less than a predetermined weight threshold can be removed from the updated connected subgraph. Abnormal edges and nodes can also be removed. For details, please refer to [link / reference]. Figure 1 The relevant descriptions in the corresponding embodiments will not be repeated here.

[0066] In addition, after generating the corresponding unified identity based on the updated connected subgraph, when updating the connected subgraph again, the connected subgraph and the unified identity can still be updated based on the incremental ad request event.

[0067] In some embodiments, this application also provides computer program products including a non-transitory computer-readable storage medium. The non-transitory computer-readable storage medium includes computer-executable code for performing... Figure 1 and Figure 3 The steps in the illustrated method embodiments. In some embodiments, the computer program product may be stored in a hardware device.

[0068] Embodiments of the present invention can be implemented in hardware, software, or a combination of both. The hardware portion can be implemented using dedicated logic; the software portion can be stored in memory and executed by a suitable instruction execution system, such as a microprocessor or dedicated-design hardware. Those skilled in the art will understand that the above-described devices and methods can be implemented using computer-executable instructions and / or included in processor control code, for example, on a carrier medium such as a disk, CD, or DVD-ROM, a programmable memory such as read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The devices and modules of the present invention can be implemented using hardware circuitry such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field-programmable gate arrays, programmable logic devices, etc., or using software executed by various types of processors, or using a combination of the above-described hardware circuitry and software, such as firmware.

[0069] It should be noted that although several steps or modules of the unified identity generation method, apparatus, and storage medium are mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of this application, the features and functions of two or more modules described above can be embodied in one module. Conversely, the features and functions of one module described above can be further divided and embodied by multiple modules.

[0070] Those skilled in the art can understand and implement other modifications to the disclosed embodiments by studying the specification, the disclosure, the drawings, and the appended claims. In the claims, the word "comprising" does not exclude other elements and steps, and the words "a" or "an" do not exclude a plurality. In practical applications of this application, a single part may perform the function of multiple technical features referenced in the claims. Any reference numerals in the claims should not be construed as limiting the scope.

Claims

1. A method for generating a unified identity identifier, characterized in that, The method includes: Acquire multiple ad request events, wherein each ad request event includes at least two related feature identifiers; Based on the multiple advertising request events, multiple feature identifiers and the association relationships between the multiple feature identifiers are determined, wherein each association relationship has a corresponding event time; Using a group of feature identifiers that are related among the plurality of feature identifiers as nodes, and the relationships between the feature identifiers as edges, a connected subgraph is generated. Each edge of the connected subgraph has a corresponding weight factor, which decays over time with the event duration of the corresponding relationship. A corresponding unified identity identifier is generated based on the connected subgraph.

2. The method according to claim 1, characterized in that, Before the step of generating the corresponding unified identity based on the connected subgraph, the method further includes: Remove edges with weight factors less than a predetermined weight threshold from the connected subgraph; and The connected subgraph is updated to generate an updated connected subgraph.

3. The method according to claim 2, characterized in that, The weighting factor is determined in the following manner: Determine the current time; and The decay of the weighting factor is determined based on the time difference between the current time and the event time.

4. The method according to claim 3, characterized in that, Determining the decay of the weighting factor based on the time difference between the current time and the event time further includes determining the decayed weighting factor W(t) using the following equation: W(t)=W0·e -λt Where W0 is a preset initial weighting factor, λ is a preset decay rate, and t is the time difference between the current time and the event time.

5. The method according to claim 2, characterized in that, The step of updating the connected subgraph to generate an updated connected subgraph further includes: Determine whether the connected subgraph includes multiple sets of nodes that are not connected by any edges; and In response to determining that the connected subgraph includes multiple groups of nodes that are not connected by any edges, multiple updated connected subgraphs are generated based on each group of nodes and the edges connecting that group of nodes.

6. The method according to claim 1, characterized in that, After generating the corresponding unified identity identifier based on the connected subgraph, the method further includes: Acquire one or more incremental ad request events, wherein each incremental ad request event includes at least two related feature identifiers; Based on the multiple incremental ad request events, multiple incremental feature identifiers and incremental association relationships between the multiple incremental feature identifiers are determined, wherein each incremental association relationship has a corresponding event time; Select a set of incremental feature identifiers that are the same as or related to a set of feature identifiers that are nodes in the connected subgraph from the plurality of incremental feature identifiers, and update the connected subgraph with the association relationship between the set of incremental feature identifiers as edges to generate an updated connected subgraph. A corresponding unified identity identifier is generated based on the updated connected subgraph.

7. The method according to claim 6, characterized in that, The step of updating the connected subgraph to generate an updated connected subgraph further includes: Connect the connected subgraph with at least one or more other connected subgraphs to form an updated connected subgraph.

8. The method according to claim 1, characterized in that, Before the step of generating the corresponding unified identity based on the connected subgraph, the method further includes: Determine the number of edges included in the connected subgraph; and In response to the number of edges exceeding a predetermined edge number threshold, at least a portion of the edges and nodes unrelated to the remaining edges are removed from the connected subgraph, so that the number of edges does not exceed the predetermined edge number threshold.

9. The method according to claim 1, characterized in that, The step of generating a corresponding unified identity identifier based on the connected subgraph further includes: Select a feature identifier from the connected subgraph and generate a corresponding unified identity identifier based on that feature identifier.

10. The method according to claim 1, characterized in that, The step of generating a corresponding unified identity identifier based on the connected subgraph further includes: Assign a unique numerical identifier to each feature identifier contained in the connected subgraph; The minimum value among all numerical identifiers is selected as the unified identity identifier of the connected subgraph.

11. A unified identity generation device, characterized in that, The unified identity generation device includes a non-transitory computer storage medium storing one or more executable instructions, which are executed by a processor to implement the method of any one of claims 1-10.

12. A non-transitory computer storage medium having stored thereon one or more executable instructions, which, when executed by a processor, implement the method of any one of claims 1-10.

Images