A target user retrieval method, apparatus, medium and device

Abstract

This invention relates to the field of data retrieval technology, and in particular to a method, apparatus, medium, and device for retrieving target user data. By obtaining the number of aggregates corresponding to each aggregated data in the current level, the method acquires the number of aggregated data corresponding to each aggregated data in the next level and the number of aggregates corresponding to each aggregated data, thereby progressively constructing a multi-level aggregated data structure. A suitable target level identifier is determined through preset conditions to avoid over-segmentation or insufficient aggregation in the hierarchical structure, balancing data aggregation degree and retrieval efficiency. Then, the method progressively compares and searches downwards within the hierarchical structure to ultimately determine whether the target user data has been retrieved. This approach effectively utilizes the hierarchical structure to rationally organize massive amounts of data, improving retrieval accuracy, and leverages the characteristics of Bloom filters to achieve efficient data retrieval.

Description

Technical Field

[0001] This invention relates to the field of data retrieval technology, and in particular to a method, apparatus, medium and device for retrieving data from a target user. Background Technology

[0002] In today's digital age, various applications and systems have accumulated massive amounts of user data. For businesses and developers, efficiently and accurately retrieving target user information from this vast amount of data is crucial for achieving precision marketing, personalized services, and user behavior analysis.

[0003] Currently, common user retrieval methods include linear search and hash table retrieval. Linear search compares user data with target user data one by one according to the data storage order. While simple to implement, this method has extremely high time complexity and low retrieval efficiency when dealing with large datasets, severely impacting system response speed and user experience. To improve retrieval efficiency, hash table retrieval is further used, mapping user data to specific hash values. With moderate datasets, this allows for rapid location of target data and higher retrieval efficiency. However, hash tables suffer from hash collisions, where different data may map to the same hash value, leading to lower accuracy of search results and poor performance with complex data structures.

[0004] Therefore, how to balance search accuracy and search efficiency when searching for target users in massive amounts of data has become an urgent problem to be solved. Summary of the Invention

[0005] To address the aforementioned technical problems, the present invention provides a target user retrieval method, which includes the following steps:

[0006] S1. Obtain the initial user data set corresponding to the preset label, as well as the preset bit array length and N preset hash functions. The initial user data set includes M initial user data, where M and N are both integers greater than 0.

[0007] S2, initialize the level identifier i=1, the data identifier within the level j=1, and determine the first aggregated data A in the first level. 11 The corresponding number of aggregations is M, j is an integer in the range [1, J], and J is the total number of aggregations corresponding to the i-th level. The value of J for the first level is determined to be 1, and the first aggregation data A in the first level is obtained. 11 .

[0008] S3, based on the number of aggregations corresponding to the j-th aggregation data in the i-th level, obtain the K corresponding to the j-th aggregation data in the i+1-th level. ij Individual aggregated data and Kij The number of aggregations corresponding to each aggregation data.

[0009] S4, if K ij If the maximum number of aggregations among the aggregations corresponding to each aggregation data satisfies the first preset condition, then the current value of i is determined to be the target level identifier R, and step S7 is executed; otherwise, step S5 is executed.

[0010] S5, update i=i+1, update J=∑ for the i-th level. j=1 J K (i-1)j .

[0011] S6, return to step S3, until the maximum value of the number of aggregates corresponding to all aggregate data in the (i-1)th level satisfies the second preset condition, and determine the current i value as the target level identifier R.

[0012] S7. When the target user data corresponding to the target user under the preset tag is obtained, the search results are obtained based on the target user data, the preset bit array length, N preset hash functions and all aggregated data.

[0013] The present invention also provides a target user retrieval device, the target user retrieval device comprising:

[0014] The data acquisition module is used to acquire the initial user data set corresponding to the preset tag, as well as the preset bit array length and N preset hash functions. The initial user data set includes M initial user data, where M and N are both integers greater than 0.

[0015] The first-level data acquisition module is used to initialize the level identifier i=1, the data identifier within the level j=1, and determine the first aggregated data A in the first level. 11 The corresponding number of aggregations is M, j is an integer in the range [1, J], and J is the total number of aggregations corresponding to the i-th level. The value of J for the first level is determined to be 1, and the first aggregation data A in the first level is obtained. 11 .

[0016] The second-level data acquisition module is used to obtain the K corresponding to the j-th aggregate data in the i-th level in the (i+1)-th level, based on the number of aggregates corresponding to the j-th aggregate data in the i-th level. ij Individual aggregated data and K ij The number of aggregations corresponding to each aggregation data.

[0017] The first identifier determination module is used if K ijIf the maximum number of aggregates among the aggregates corresponding to each aggregated data satisfies the first preset condition, then the current value of i is determined to be the target level identifier R, and the target bit array acquisition module is executed; otherwise, the identifier update module is executed.

[0018] The identifier update module is used to update i=i+1 and update J=∑ for the i-th level. j=1 J K (i-1)j .

[0019] The second identifier determination module is used to return to the execution of the second-level data acquisition module until the maximum value of the number of aggregates corresponding to all aggregate data in the (i-1)th level satisfies the second preset condition, and determine the current i value as the target level identifier R.

[0020] The search result acquisition module is used to obtain search results based on the target user data, the preset bit array length, N preset hash functions, and all aggregated data when the target user data corresponding to the preset tag is obtained.

[0021] The present invention also provides a non-transitory computer-readable storage medium storing at least one instruction or at least one program, wherein the at least one instruction or at least one program is loaded and executed by a processor to implement the above-described target user retrieval method.

[0022] The present invention also provides an electronic device, including a processor and the aforementioned non-transitory computer-readable storage medium.

[0023] This invention has at least the following beneficial effects: By obtaining an initial user data set, a preset bit array length, and N preset hash functions, and based on the number of aggregations corresponding to each aggregated data in the current level, the number of aggregated data corresponding to each aggregated data in the next level is obtained, as well as the number of aggregations corresponding to each aggregated data. This allows for further subdivision of the aggregated data in the current level, constructing a multi-level aggregated data structure. A suitable target level identifier is determined through a first preset condition and a second preset condition, avoiding over-subdivision or insufficient aggregation in the hierarchical structure, ensuring the rationality of the hierarchical structure, balancing data aggregation degree and retrieval efficiency. Then, the hierarchical structure is progressively compared and searched downwards to ultimately determine whether the target user data has been retrieved. Through the combination of hierarchical aggregation and Bloom filters, in large-scale user data retrieval scenarios, the hierarchical structure can be effectively utilized to rationally organize massive amounts of data, improving retrieval accuracy. Furthermore, the characteristics of Bloom filters can be leveraged to achieve efficient data retrieval with a small memory footprint, meeting the needs of large-scale user data retrieval in practical applications. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 This is a flowchart of a target user retrieval method provided in Embodiment 1 of the present invention;

[0026] Figure 2 This is a schematic diagram of the structure of a target user retrieval device provided in Embodiment 2 of the present invention. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It is understood that, where appropriate, the terms used to distinguish similar objects can be interchanged so that the invention can also be implemented in other embodiments besides the illustrated or described embodiments. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products, or devices.

[0029] Example 1

[0030] This embodiment provides a target user retrieval method, which includes the following steps: Figure 1 As shown:

[0031] S1. Obtain the initial user data set corresponding to the preset label, as well as the preset bit array length and N preset hash functions. The initial user data set includes M initial user data, where M and N are both integers greater than 0.

[0032] S2, initialize the level identifier i=1, the data identifier within the level j=1, and determine the first aggregated data A in the first level. 11The corresponding number of aggregations is M, j is an integer in the range [1, J], and J is the total number of aggregations corresponding to the i-th level. The value of J for the first level is determined to be 1, and the first aggregation data A in the first level is obtained. 11 .

[0033] S3, based on the number of aggregations corresponding to the j-th aggregation data in the i-th level, obtain the K corresponding to the j-th aggregation data in the i+1-th level. ij Individual aggregated data and K ij The number of aggregations corresponding to each aggregation data.

[0034] S4, if K ij If the maximum number of aggregations among the aggregations corresponding to each aggregation data satisfies the first preset condition, then the current value of i is determined to be the target level identifier R, and step S7 is executed; otherwise, step S5 is executed.

[0035] S5, update i=i+1, update J=∑ for the i-th level. j=1 J K (i-1)j .

[0036] S6, return to step S3, until the maximum value of the number of aggregates corresponding to all aggregate data in the (i-1)th level satisfies the second preset condition, and determine the current i value as the target level identifier R.

[0037] S7. When the target user data corresponding to the target user under the preset tag is obtained, the search results are obtained based on the target user data, the preset bit array length, N preset hash functions and all aggregated data.

[0038] Among them, the preset labels are used to identify specific types of user data, and the preset bit array length and N preset hash functions are the basis for building the Bloom filter.

[0039] The hierarchical identifier and the data identifier within the hierarchical level are initialized to 1. The M initial user data in the initial user data set are used as the basis for the Bloom filter. By performing hash value calculation, index value calculation and bit array update on the M initial user data as a whole, the unique aggregate data corresponding to the first level is obtained, and the initial state is set for the construction of hierarchical aggregate data.

[0040] The aggregation quantity is the initial user data quantity that is used as a data group for hash value calculation, index value calculation, and bit array update. The aggregated data is the updated bit array corresponding to a data group after hash value calculation, index value calculation, and bit array update. The sum of the aggregation quantities corresponding to all aggregated data in each level is the total quantity M of the initial user data.

[0041] Based on the number of aggregates corresponding to each aggregated data in the current level, the number of aggregated data corresponding to each aggregated data in the next level is determined, as well as the number of aggregates corresponding to each aggregated data. This allows for further subdivision of the aggregated data in the current level, constructing the aggregated data structure for the next level. Through this hierarchical aggregation method, large-scale initial user data can be gradually organized into a more detailed hierarchical structure, facilitating subsequent retrieval operations.

[0042] The decision to stop the current hierarchical aggregation process is made by determining whether the maximum number of aggregations meets a first preset condition. If the condition is met, it means the current hierarchical structure is suitable for retrieval operations, and no further subdivision is needed. The current value of i is then set as the target hierarchical identifier R, and subsequent retrieval operations are executed. Otherwise, hierarchical aggregation needs to continue.

[0043] Specifically, while continuing hierarchical aggregation, the range of the hierarchical identifier and the data identifier within the layer is updated to provide the necessary parameters for the aggregation operation of the next level, until the second preset condition is met, that is, a suitable target hierarchical identifier R is determined, thereby determining a hierarchical structure that is neither too coarse (high data aggregation degree but low retrieval accuracy) nor too fine (low data aggregation degree but reduced retrieval efficiency), so as to balance retrieval efficiency and accuracy.

[0044] When performing subsequent operations to retrieve target users based on hierarchical data structures, the system searches for aggregated data that may contain target user data at different levels, based on the target user data, the preset bit array length, N preset hash functions, and all aggregated data, until the target user data is retrieved or all levels have been traversed to obtain the corresponding retrieval results.

[0045] In one specific implementation, the preset tag belongs to the preset tag set, which includes several preset tags of different types. The types of preset tags in the preset tag set include Wi-Fi, APP, device ID, and identity ID.

[0046] Among them, Wi-Fi tags can be used to mark wireless network connection information related to users, such as specific Wi-Fi hotspots that users often connect to, which helps to understand information such as the user's activity area, living or working location.

[0047] App tags can reflect the various applications a user uses, helping to understand the user's interests, habits, and other behaviors.

[0048] Device ID tags are used to identify the devices used by users, which helps track user activity on different devices and analyze device usage preferences.

[0049] Identity ID tags can be directly linked to a user's identity information, such as ID card number or membership card number, which helps to accurately identify individual users, provide personalized services, and manage users.

[0050] When obtaining target user data corresponding to preset tags in step S7, the target user may have multiple preset tags of different types. For example, it may be necessary to find relevant data based on the user's device ID and also analyze it in conjunction with APP usage. During the construction of the target bit array and subsequent comparison and retrieval process, the target bit array and retrieval logic constructed for different types of tags may need to be adjusted and optimized accordingly. For example, the selection of hash functions for different types of data in the Bloom filter may need to be customized based on the uniqueness of the data to improve the accuracy and efficiency of the retrieval.

[0051] In one specific implementation, S2 further includes the following steps:

[0052] S21, based on the initial user data set and the first aggregated data A in the first level. 11 Given the corresponding number of aggregations M, obtain the initial user data subset corresponding to the first aggregate data in the first level.

[0053] S22, based on the initial user data subset corresponding to the first aggregated data in the first level, the preset bit array length, and N preset hash functions, obtain the reference bit array corresponding to the first aggregated data in the first level.

[0054] S23, determine the reference bit array corresponding to the first aggregated data in the first level as aggregated data A. 11 .

[0055] Since the first-level aggregated data in the hierarchical structure aggregates all M initial user data, the subset of initial user data is actually the entire initial user data set.

[0056] As described above, by applying the principle of Bloom filters, all data in the initial user data subset is mapped to an array of a preset bit array length through N preset hash functions to generate a reference bit array. This transforms the initial user data subset into a more compact data representation that facilitates subsequent retrieval and comparison. It allows for quick determination of whether a target user might exist among all the initial users corresponding to the initial user data subset while using less memory. This provides an efficient data structure for subsequent hierarchical aggregated data retrieval, reducing the complexity of data storage and processing while ensuring retrieval efficiency.

[0057] In one specific implementation, S3 further includes the following steps:

[0058] S31, based on the number of aggregates corresponding to the j-th aggregate data in the i-th level, obtain the number K of aggregate data corresponding to the j-th aggregate data in the i+1-th level. ij .

[0059] S32, based on the total number M of the initial user data and the number K of the aggregate data corresponding to the j-th aggregate data in the i+1-th level. ij Obtain the K corresponding to the j-th aggregated data in the i+1-th level. ij The number of aggregations corresponding to each aggregation data.

[0060] S33, based on the initial user data set and the K corresponding to the j-th aggregate data in the i+1th level. ij The number of aggregations corresponding to each aggregation data is obtained, and the K corresponding to the j-th aggregation data in the i-th level in the (i+1)-th level is obtained. ij Each aggregated data point corresponds to a subset of the initial user data.

[0061] S34, based on K corresponding to the j-th aggregated data in the (i+1)-th level. ij Given the initial user data subset corresponding to each aggregated data point, the preset bit array length, and N preset hash functions, obtain the K corresponding to the j-th aggregated data point in the i-th level of the (i+1)-th level. ij Each aggregated data point corresponds to a reference bit array.

[0062] S35, in the (i+1)th level, the K corresponding to the j-th aggregated data in the i-th level. ij The reference bit arrays corresponding to each aggregated data are respectively determined as K corresponding to the j-th aggregated data in the i-th level of the (i+1)-th level. ij Aggregated data.

[0063] In this process, based on the number of aggregates corresponding to each aggregate data in the current level, it is determined how many aggregate data should be further subdivided into in the next level. This helps to build a more reasonable hierarchical structure, so that the data can be effectively organized and distributed in different levels, which facilitates subsequent retrieval operations.

[0064] Based on the average allocation method, assign M initial user data to K. ij Given aggregated data, obtain the K corresponding to the j-th aggregated data in the i-th level of the (i+1)-th level. ij The number of aggregations corresponding to each aggregated data is determined to clarify the amount of data contained in each subdivided aggregated data in the next level, so as to accurately divide the initial user data subset in the future.

[0065] For each initial user data subset, the principle of the Bloom filter is applied, and the data in the initial user data subset is mapped to an array of a preset bit array length through N preset hash functions to generate a corresponding reference bit array. This transforms the initial user data subset into a more compact reference bit array form that is easier to retrieve and compare, thereby improving the efficiency of subsequent retrieval.

[0066] The above-mentioned process involves continuous subdivision and construction of hierarchical levels, gradually forming a complete hierarchical structure that provides a foundation for subsequent target user data retrieval.

[0067] In one specific embodiment, S31 includes the following steps:

[0068] S311, Substitute the number of aggregated data corresponding to the j-th aggregated data in the i-th level into the preset retrieval quantity calculation model to obtain the target retrieval quantity calculation model corresponding to the j-th aggregated data in the i-th level, wherein the preset retrieval quantity calculation model takes the number of aggregated data as the independent variable and the retrieval quantity as the dependent variable.

[0069] S312, based on the target retrieval quantity calculation model corresponding to the j-th aggregated data in the i-th level, the quantity of aggregated data corresponding to the minimum retrieval quantity is determined as the reference quantity value corresponding to the j-th aggregated data in the i-th level.

[0070] S313, round up the reference quantity value corresponding to the j-th aggregated data in the i-th level to obtain the first intermediate quantity value corresponding to the j-th aggregated data in the i-th level.

[0071] S314, round down the reference quantity value corresponding to the j-th aggregated data in the i-th level to obtain the second intermediate quantity value corresponding to the j-th aggregated data in the i-th level.

[0072] S315, substitute the first intermediate quantity value corresponding to the j-th aggregated data in the i-th level into the corresponding target retrieval quantity calculation model to obtain the first intermediate retrieval quantity corresponding to the j-th aggregated data in the i-th level.

[0073] S316, Substitute the second intermediate quantity value corresponding to the j-th aggregated data in the i-th level into the corresponding target retrieval quantity calculation model to obtain the second intermediate retrieval quantity corresponding to the j-th aggregated data in the i-th level.

[0074] S317, if the first intermediate retrieval quantity corresponding to the j-th aggregated data in the i-th level is less than the corresponding second intermediate retrieval quantity, then the first intermediate quantity value corresponding to the j-th aggregated data in the i-th level is determined as the quantity K of the aggregated data corresponding to the j-th aggregated data in the i-th level in the (i+1)-th level. ij.

[0075] Here, the number of aggregations corresponding to the j-th aggregate data in the i-th level is set to V. ij The model Q is used to calculate the target retrieval quantity corresponding to the j-th aggregated data in the i-th level. ij =k×log k (V ij ), where Q ij refers to the number of searches corresponding to the j-th aggregated data in the i-th level, and k refers to the number of aggregated data corresponding to the j-th aggregated data in the i-th level.

[0076] The above describes how, using a pre-defined retrieval quantity calculation model, a reference quantity value that minimizes the retrieval quantity is first found. Then, this reference quantity value is rounded up and down to obtain two intermediate quantity values. The retrieval quantity corresponding to each of these two intermediate quantity values ​​is calculated. Finally, the two retrieval quantities are compared, and the intermediate quantity value corresponding to the smaller retrieval quantity is selected as the aggregated data quantity for the next level. By optimizing the aggregated data quantity, the retrieval quantity is reduced, thereby improving retrieval efficiency.

[0077] In one specific implementation, S4 includes the following steps:

[0078] S41, obtain the preset false positive rate threshold.

[0079] S42, according to K ij Given the maximum number of aggregates for each aggregated data, the preset bit array length, and the total number N of preset hash functions, obtain K. ij The minimum false positive rate corresponding to each aggregated data point.

[0080] S43, if K ij If the minimum false positive rate corresponding to each aggregated data point is less than the preset false positive rate threshold, then K is determined. ij The maximum number of aggregates among the aggregates corresponding to each aggregated data satisfies the first preset condition.

[0081] The specific value of the preset false positive rate threshold can be set by the implementer according to the actual situation.

[0082] As mentioned above, under the premise of ensuring a certain misjudgment rate, determining an appropriate target level identifier can optimize the accuracy and efficiency of the entire data retrieval process.

[0083] In one specific implementation, S6 includes the following steps:

[0084] S61, obtain the preset aggregation quantity threshold.

[0085] S62, if the maximum value of the number of aggregates corresponding to all aggregate data in the (i-1)th level is less than or equal to the preset number of aggregates threshold, then the maximum value of the number of aggregates corresponding to all aggregate data in the (i-1)th level satisfies the second preset condition.

[0086] The specific value of the preset aggregation quantity threshold can be set by the implementer according to the actual situation.

[0087] The above can control the depth of hierarchical aggregation, avoiding excessive subdivision or insufficient aggregation, so as to ensure that the entire data processing and retrieval process is carried out under an appropriate hierarchical structure, thereby improving retrieval efficiency and accuracy.

[0088] In one specific implementation, S7 includes the following steps:

[0089] S71, when the target user data corresponding to the target user under the preset tag is obtained, the target bit array corresponding to the target user data is obtained according to the target user data, the preset bit array length and N preset hash functions.

[0090] S72, initialize the reference level identifier u=1.

[0091] S73, compare the target bit array with each aggregated data in the u-th level that meets the third preset condition. If the comparison result with any aggregated data meets the fourth preset condition, then determine that the aggregated data in the (u+1)-th level corresponding to the aggregated data that meets the fourth preset condition meets the third preset condition. Otherwise, determine that the search result is that no target user data was found. Among them, the first aggregated data in the first level meets the third preset condition.

[0092] S74, update u=u+1, return to step S73, until u=R+1, confirm that the search result is that the target user data has been retrieved.

[0093] Specifically, when performing subsequent operations to retrieve target users based on hierarchical data structures, the target user data is mapped to a target bit array based on the principle of Bloom filters. Bloom filters can quickly determine whether an element might exist in a set while using relatively little memory, providing an efficient data representation for subsequent retrieval operations.

[0094] The initial reference level identifier is set to 1, which sets the starting level for retrieving target user data from the hierarchical structure.

[0095] Specifically, the system searches for aggregated data that may contain the target user data within the current level. By comparing the target bit array with aggregated data that meets the third preset condition, it determines whether to continue searching in the next level corresponding to that aggregated data based on the fourth preset condition. This process is repeated hierarchically down the hierarchy until the target user data is found or all levels have been traversed to obtain the corresponding search results.

[0096] In one specific implementation, S71 includes the following steps:

[0097] S711, Obtain an initial bit array that meets the preset bit array length, wherein the value of each position in the initial bit array is initialized to 0.

[0098] S712 calculates the target user data according to N preset hash functions to obtain N target hash values.

[0099] S713: Obtain N target index values ​​by taking the modulo of the preset bit array length according to the N target hash values.

[0100] S714, update the value at the corresponding position of each target index in the initial bit array to 1, and obtain the target bit array.

[0101] As described above, by utilizing the principle of Bloom filters, the target user data is mapped to multiple positions in the initial bit array through a hash function, and the value of the corresponding position is marked as 1, thereby obtaining a target bit array that can be used for fast retrieval and comparison.

[0102] In one specific implementation, S73 includes the following steps:

[0103] S731, for any aggregated data in the u-th level that satisfies the third preset condition, if the values ​​of each position with a value of 1 in the target bit array are all 1 in the corresponding position in the current aggregated data, then it is determined that the comparison result between the target bit array and the current aggregated data satisfies the fourth preset condition, and it is determined that the aggregated data in the (u+1)-th level corresponding to the current aggregated data satisfies the third preset condition.

[0104] S732, if the values ​​of each position with a value of 1 in the target bit array are not all 1 in the corresponding position in the current aggregated data, then it is determined that the comparison result between the target bit array and the current aggregated data does not meet the fourth preset condition.

[0105] S733, traverse all aggregated data in the u-th level that meet the third preset condition. If the comparison result with all aggregated data in the u-th level that meet the third preset condition does not meet the fourth preset condition, then determine that the search result is that no target user data was found.

[0106] The above describes a method that, by acquiring an initial user data set, a preset bit array length, and N preset hash functions, obtains the number of aggregated data corresponding to each aggregated data in the current level in the next level, as well as the number of aggregated data corresponding to each aggregated data, to further subdivide the aggregated data in the current level and construct a multi-level aggregated data structure. A suitable target level identifier is determined through the first and second preset conditions to avoid over-subdivision or insufficient aggregation in the hierarchical structure, ensuring its rationality and balancing data aggregation degree and retrieval efficiency. Then, the system progressively compares and searches downwards within the hierarchical structure to ultimately determine whether the target user data has been retrieved. Through the combination of hierarchical aggregation and Bloom filters, in large-scale user data retrieval scenarios, it effectively utilizes the hierarchical structure to rationally organize massive amounts of data, improving retrieval accuracy, while leveraging the characteristics of Bloom filters to achieve efficient data retrieval with minimal memory usage, thus meeting the needs of large-scale user data retrieval in practical applications.

[0107] Example 2

[0108] This second embodiment provides a target user retrieval device, which includes, for example: Figure 2 As shown:

[0109] The data acquisition module 201 is used to acquire the initial user data set corresponding to the preset tag, as well as the preset bit array length and N preset hash functions. The initial user data set includes M initial user data, where M and N are both integers greater than 0.

[0110] The first-level data acquisition module 202 is used to initialize the level identifier i=1, the data identifier within the level j=1, and determine the first aggregated data A in the first level. 11 The corresponding number of aggregations is M, j is an integer in the range [1, J], and J is the total number of aggregations corresponding to the i-th level. The value of J for the first level is determined to be 1, and the first aggregation data A in the first level is obtained. 11 .

[0111] The second-level data acquisition module 203 is used to obtain the K corresponding to the j-th aggregate data in the i-th level in the (i+1)-th level, based on the number of aggregates corresponding to the j-th aggregate data in the i-th level. ij Individual aggregated data and K ij The number of aggregations corresponding to each aggregation data.

[0112] The first identifier determination module 204 is used if K ijIf the maximum number of aggregates among the aggregates corresponding to each aggregated data satisfies the first preset condition, then the current value of i is determined to be the target level identifier R, and the target bit array acquisition module 207 is executed; otherwise, the identifier update module 205 is executed.

[0113] The identifier update module 205 is used to update i=i+1 and update J=∑ corresponding to the i-th level. j=1 J K (i-1)j .

[0114] The second identifier determination module 206 is used to return to the execution of the second-level data acquisition module 203 until the maximum value of the number of aggregates corresponding to all aggregate data in the (i-1)th level satisfies the second preset condition, and determine the current i value as the target level identifier R.

[0115] The retrieval result acquisition module 207 is used to obtain retrieval results based on the target user data, the preset bit array length, N preset hash functions, and all aggregated data when the target user data corresponding to the target user under the preset tag is obtained.

[0116] In one specific implementation, the first-level data acquisition module 202 further includes:

[0117] The first data subset acquisition submodule is used to obtain the initial user data set and the first aggregate data A in the first level. 11 Given the corresponding number of aggregations M, obtain the initial user data subset corresponding to the first aggregate data in the first level.

[0118] The reference bit array acquisition submodule is used to obtain the reference bit array corresponding to the first aggregated data in the first level based on the initial user data subset corresponding to the first aggregated data in the first level, the preset bit array length, and N preset hash functions.

[0119] The first aggregated data acquisition submodule is used to determine the reference bit array corresponding to the first aggregated data in the first level as aggregated data A. 11 .

[0120] In one specific implementation, S3 further includes the following steps:

[0121] The first quantity acquisition submodule is used to input the aggregation quantity corresponding to the j-th aggregation data in the i-th level into a preset quantity determination model, and obtain the quantity K of the aggregation data corresponding to the j-th aggregation data in the i+1-th level. ij .

[0122] The second acquisition submodule is used to determine the total number M of initial user data and the number K of aggregate data corresponding to the j-th aggregate data in the i-th level of the (i+1)-th level. ij Obtain the K corresponding to the j-th aggregated data in the i+1-th level. ij The number of aggregations corresponding to each aggregation data.

[0123] The second data subset acquisition submodule is used to obtain the K data based on the initial user data set and the K data corresponding to the j-th aggregated data in the i-th level of the (i+1)-th level. ij The number of aggregations corresponding to each aggregation data is obtained, and the K corresponding to the j-th aggregation data in the i-th level in the (i+1)-th level is obtained. ij Each aggregated data point corresponds to a subset of the initial user data.

[0124] The reference bit array retrieval submodule is used to obtain K based on the K corresponding to the j-th aggregate data in the i+1-th level. ij Given the initial user data subset corresponding to each aggregated data point, the preset bit array length, and N preset hash functions, obtain the K corresponding to the j-th aggregated data point in the i-th level of the (i+1)-th level. ij Each aggregated data point corresponds to a reference bit array.

[0125] The second aggregation data acquisition submodule is used to obtain the K data corresponding to the j-th aggregation data in the i-th level of the (i+1)-th level. ij The reference bit arrays corresponding to each aggregated data are respectively determined as K corresponding to the j-th aggregated data in the i-th level of the (i+1)-th level. ij Aggregated data.

[0126] In one specific implementation, the first quantity acquisition submodule includes:

[0127] The target model acquisition unit is used to substitute the number of aggregates corresponding to the j-th aggregated data in the i-th level into the preset retrieval quantity calculation model to obtain the target retrieval quantity calculation model corresponding to the j-th aggregated data in the i-th level. The preset retrieval quantity calculation model uses the number of aggregated data as the independent variable and the retrieval quantity as the dependent variable.

[0128] The reference quantity value acquisition unit is used to determine the number of aggregated data corresponding to the minimum retrieval quantity of the j-th aggregated data in the i-th level based on the target retrieval quantity calculation model.

[0129] The first intermediate quantity value acquisition unit is used to round up the reference quantity value corresponding to the j-th aggregated data in the i-th level to obtain the first intermediate quantity value corresponding to the j-th aggregated data in the i-th level.

[0130] The second intermediate quantity value acquisition unit is used to round down the reference quantity value corresponding to the j-th aggregated data in the i-th level to obtain the second intermediate quantity value corresponding to the j-th aggregated data in the i-th level.

[0131] The first intermediate retrieval quantity acquisition unit is used to substitute the first intermediate quantity value corresponding to the j-th aggregated data in the i-th level into the corresponding target retrieval quantity calculation model to obtain the first intermediate retrieval quantity corresponding to the j-th aggregated data in the i-th level.

[0132] The second intermediate retrieval quantity acquisition unit is used to substitute the second intermediate quantity value corresponding to the j-th aggregated data in the i-th level into the corresponding target retrieval quantity calculation model to obtain the second intermediate retrieval quantity corresponding to the j-th aggregated data in the i-th level.

[0133] The quantity acquisition unit is configured to determine the first intermediate retrieval value corresponding to the j-th aggregated data in the i-th level as the quantity K of the aggregated data corresponding to the j-th aggregated data in the i-th level if the first intermediate retrieval quantity corresponding to the j-th aggregated data in the i-th level is less than the corresponding second intermediate retrieval quantity. ij .

[0134] In one specific implementation, the first identifier determination module 204 includes:

[0135] The first threshold acquisition submodule is used to obtain the preset false positive rate threshold.

[0136] The minimum false positive rate acquisition submodule is used to determine the minimum false positive rate based on K. ij Given the maximum number of aggregates for each aggregated data, the preset bit array length, and the total number N of preset hash functions, obtain K. ij The minimum false positive rate corresponding to each aggregated data point.

[0137] The first preset condition judgment submodule is used if K ij If the minimum false positive rate corresponding to each aggregated data point is less than the preset false positive rate threshold, then K is determined. ij The maximum number of aggregates among the aggregates corresponding to each aggregated data satisfies the first preset condition.

[0138] In one specific embodiment, the second identifier determination module 206 includes:

[0139] The second threshold acquisition submodule is used to obtain the preset aggregation quantity threshold.

[0140] The second preset condition judgment submodule is used to determine that the maximum value of the aggregation quantity corresponding to all aggregate data in the (i-1)th level satisfies the second preset condition if the maximum value of the aggregation quantity corresponding to all aggregate data in the (i-1)th level is less than or equal to the preset aggregation quantity threshold.

[0141] In one specific implementation, the target bit array acquisition module 207 includes:

[0142] The target bit array acquisition submodule is used to obtain the target bit array corresponding to the target user data when the target user data corresponding to the target user under the preset tag is obtained, based on the target user data, the preset bit array length and N preset hash functions.

[0143] The identifier initialization submodule is used to initialize the reference level identifier u=1.

[0144] The first search result acquisition submodule is used to compare the target bit array with each aggregated data in the u-th level that meets the third preset condition. If the comparison result with any aggregated data meets the fourth preset condition, it is determined that the aggregated data in the (u+1)-th level corresponding to the aggregated data that meets the fourth preset condition meets the third preset condition. Otherwise, it is determined that the search result is that no target user data was found. The first aggregated data in the first level meets the third preset condition.

[0145] The second search result acquisition submodule is used to update u=u+1, return to execute the first search result acquisition submodule, until u=R+1, and determine that the search result is that the target user data has been retrieved.

[0146] In one specific implementation, the target bit array acquisition submodule includes:

[0147] The initial bit array acquisition unit is used to obtain an initial bit array that conforms to the preset bit array length, wherein the value of each position in the initial bit array is initialized to 0.

[0148] The target hash value acquisition unit is used to calculate the target user data according to N preset hash functions to obtain N target hash values.

[0149] The target index value acquisition unit is used to obtain N target index values ​​by taking the modulo of the preset bit array length according to the N target hash values.

[0150] The target bit array acquisition unit is used to update the value at the corresponding position of each target index value in the initial bit array to 1, thereby obtaining the target bit array.

[0151] In one specific implementation, the first search result acquisition submodule includes:

[0152] The first preset condition judgment unit is used to determine that, for any aggregated data in the u-th level that satisfies the third preset condition, if the values ​​of each position with a value of 1 in the target bit array are all 1 in the corresponding position in the current aggregated data, then the comparison result between the target bit array and the current aggregated data satisfies the fourth preset condition, and the aggregated data in the (u+1)-th level corresponding to the current aggregated data satisfies the third preset condition.

[0153] The second preset condition judgment unit is used to determine that if the values ​​of each position with a value of 1 in the target position array are not all 1 in the corresponding position in the current aggregated data, then the comparison result between the target position array and the current aggregated data does not meet the fourth preset condition.

[0154] The first search result acquisition unit is used to traverse all aggregated data in the u-th level that meet the third preset condition. If the comparison result with all aggregated data in the u-th level that meet the third preset condition does not meet the fourth preset condition, then the search result is determined to be that no target user data was found.

[0155] It should be noted that the information interaction and execution process between the above modules are based on the same concept as the method embodiments of the present invention. For details on their specific functions and technical effects, please refer to the method embodiments section, which will not be repeated here.

[0156] Example 3

[0157] Embodiment 3 of the present invention provides a non-transitory computer-readable storage medium, which stores at least one instruction or at least one program segment, wherein the at least one instruction or at least one program segment is loaded and executed by a processor to implement the following steps:

[0158] S1. Obtain the initial user data set corresponding to the preset label, as well as the preset bit array length and N preset hash functions. The initial user data set includes M initial user data, where M and N are both integers greater than 0.

[0159] S2, initialize the level identifier i=1, the data identifier within the level j=1, and determine the first aggregated data A in the first level. 11 The corresponding number of aggregations is M, j is an integer in the range [1, J], and J is the total number of aggregations corresponding to the i-th level. The value of J for the first level is determined to be 1, and the first aggregation data A in the first level is obtained. 11 .

[0160] S3, based on the number of aggregations corresponding to the j-th aggregation data in the i-th level, obtain the K corresponding to the j-th aggregation data in the i+1-th level. ij Individual aggregated data and Kij The number of aggregations corresponding to each aggregation data.

[0161] S4, if K ij If the maximum number of aggregations among the aggregations corresponding to each aggregation data satisfies the first preset condition, then the current value of i is determined to be the target level identifier R, and step S7 is executed; otherwise, step S5 is executed.

[0162] S5, update i=i+1, update J=∑ for the i-th level. j=1 J K (i-1)j .

[0163] S6, return to step S3, until the maximum value of the number of aggregates corresponding to all aggregate data in the (i-1)th level satisfies the second preset condition, and determine the current i value as the target level identifier R.

[0164] S7. When the target user data corresponding to the target user under the preset tag is obtained, the search results are obtained based on the target user data, the preset bit array length, N preset hash functions and all aggregated data.

[0165] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0166] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0167] Example 4

[0168] Embodiment 4 of the present invention provides an electronic device, which includes a processor and a non-transitory computer-readable storage medium as described in Embodiment 3 of the present invention.

[0169] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for retrieving target users, characterized in that, The target user retrieval method includes the following steps: S1, obtain the initial user data set corresponding to the preset tag, as well as the length of the preset bit array and N preset hash functions, wherein the initial user data set includes M initial user data, and M and N are both integers greater than 0; S2, initialize the level identifier i=1, the intra-layer data identifier j=1, determine the first aggregation data A in the first level 11 The corresponding aggregation quantity is M, j is an integer in the range of [1, J i ] and J i is the total number of aggregation data corresponding to the ith level, determine the value of J1 corresponding to the first level is 1, and obtain the first aggregation data A in the first level 11 ; S3, according to the aggregation quantity corresponding to the jth aggregated data in the ith level, obtaining K ij aggregated data corresponding to the jth aggregated data in the ith level in the i+1th level and the aggregation quantity corresponding to the K ij aggregated data respectively; S4, according to K ij Given the maximum number of aggregates for each aggregated data point, the length of the preset bit array, and the total number N of preset hash functions, obtain K. ij The minimum false positive rate corresponding to aggregated data, if the K ij If the minimum false positive rate corresponding to each aggregated data point is less than the preset false positive rate threshold, then K is determined. ij If the maximum number of aggregations among the aggregations corresponding to each aggregation data satisfies the first preset condition, and the current value of i is determined to be the target level identifier R, then proceed to step S7; otherwise, proceed to step S5. S5, update i=i+1, update the corresponding level i. , where J i-1 This represents the total number of aggregated data corresponding to the (i-1)th level. S6, return to step S3, until the maximum value of the number of aggregates corresponding to all aggregate data in the (i-1)th level is less than or equal to the preset number of aggregates threshold, then determine that the maximum value of the number of aggregates corresponding to all aggregate data in the (i-1)th level satisfies the second preset condition, and determine that the current i value is the target level identifier R; S7, when the target user data corresponding to the target user under the preset tag is obtained, the retrieval result is obtained based on the target user data, the length of the preset bit array, N preset hash functions, and all aggregated data. S7 includes the following steps: S71, when the target user data corresponding to the target user under the preset tag is obtained, the target bit array corresponding to the target user data is obtained according to the target user data, the length of the preset bit array and N preset hash functions; S72, Initialize the reference level identifier u=1; S73, compare the target bit array with each aggregated data in the u-th level that satisfies the third preset condition. If the comparison result with any aggregated data satisfies the fourth preset condition, then determine that the aggregated data in the (u+1)-th level corresponding to the aggregated data that satisfies the fourth preset condition satisfies the third preset condition; otherwise, determine that the search result is that the target user data was not found. The first aggregated data in the first level satisfies the third preset condition. S73 includes the following steps: S731, for any aggregated data in the u-th level that satisfies the third preset condition, if the values ​​of each position with a value of 1 in the target bit array are all 1 in the corresponding position in the current aggregated data, then it is determined that the comparison result between the target bit array and the current aggregated data satisfies the fourth preset condition, and it is determined that the aggregated data in the (u+1)-th level corresponding to the current aggregated data satisfies the third preset condition. S732, if the values ​​of each position with a value of 1 in the target bit array are not all 1 in the corresponding position in the current aggregated data, then it is determined that the comparison result between the target bit array and the current aggregated data does not meet the fourth preset condition. S733, traverse all aggregated data in the u-th level that meet the third preset condition. If the comparison result with all aggregated data in the u-th level that meet the third preset condition does not meet the fourth preset condition, then determine that the search result is that no target user data was found. S74, update u=u+1, return to step S73, until u=R+1, and determine that the search result is that the target user data has been retrieved.

2. The target user retrieval method according to claim 1, characterized in that, S2 also includes the following steps: S21, according to the initial user data set and the first aggregation data A in the first level 11 The corresponding aggregation quantity M, the initial user data sub-set corresponding to the first aggregation data in the first level is obtained; S22, based on the initial user data subset corresponding to the first aggregated data in the first level, the length of the preset bit array, and N preset hash functions, obtain the reference bit array corresponding to the first aggregated data in the first level; S23, determine the reference bit array corresponding to the first aggregated data in the first hierarchy as aggregated data A 11 .

3. The target user retrieval method according to claim 1, characterized in that, S3 also includes the following steps: S31, according to the aggregation quantity corresponding to the jth aggregated data in the ith level, obtaining the quantity K of the aggregated data corresponding to the jth aggregated data in the ith level in the i+1th level ij ; S32, according to the total number M of initial user data and the number K of aggregation data corresponding to the jth aggregation data in the ith level in the i+1th level ij , obtaining the aggregation number corresponding to K ij aggregation data corresponding to the jth aggregation data in the ith level in the i+1th level respectively S33, according to the initial user data set and the K ij aggregated data corresponding to the jth aggregated data in the ith level in the i+1th level respectively correspond to the aggregated number, to obtain the K ij aggregated data corresponding to the jth aggregated data in the ith level in the i+1th level respectively correspond to the initial user data sub-set; S34, according to the K ij initial user data subsets corresponding to the K ij reference bit arrays corresponding to the K S35, the K ij reference bit arrays corresponding to the K ij aggregated data in the i+1 hierarchy level respectively.

4. The target user retrieval method according to claim 1, characterized in that, S71 includes the following steps: S711, Obtain an initial bit array that conforms to the preset length of the bit array, wherein the value of each position in the initial bit array is initialized to 0; S712, calculate the target user data according to the N preset hash functions respectively to obtain N target hash values; S713, take the modulo of the length of the preset bit array according to the N target hash values ​​to obtain N target index values; S714, update the value at the position corresponding to each target index value in the initial bit array to 1, and obtain the target bit array.

5. A target user retrieval device, characterized in that, The target user retrieval device includes: The data acquisition module is used to acquire the initial user data set corresponding to the preset tag, as well as the length of the preset bit array and N preset hash functions, wherein the initial user data set includes M initial user data, and M and N are both integers greater than 0; A first hierarchical data acquisition module is configured to initialize a hierarchical identifier i=1 and an intra-hierarchical data identifier j=1, determine a first aggregated data A in a first hierarchical level 11 , wherein M is a corresponding aggregated number, j is an integer in a range of [1, J i ], J i is a total number of aggregated data corresponding to the i-th hierarchical level, J i =1, and acquire the first aggregated data A 11 in the first hierarchical level. The second-level data acquisition module is used to obtain the K corresponding to the j-th aggregate data in the i-th level in the (i+1)-th level, based on the number of aggregates corresponding to the j-th aggregate data in the i-th level. ij Individual aggregated data and K ij The number of aggregates corresponding to each aggregated data; The first identifier determination module is used to determine the identifier based on K. ij Given the maximum number of aggregates for each aggregated data point, the length of the preset bit array, and the total number N of preset hash functions, obtain K. ij The minimum false positive rate corresponding to aggregated data, if the K ij If the minimum false positive rate corresponding to each aggregated data point is less than the preset false positive rate threshold, then K is determined. ij If the maximum number of aggregates in each aggregated data set meets the first preset condition, and the current value of i is determined to be the target level identifier R, the target bit array acquisition module is executed; otherwise, the identifier update module is executed. The identifier update module is used to update i=i+1, and update the identifier corresponding to the i-th level. , where J i-1 This represents the total number of aggregated data corresponding to the (i-1)th level. The second identifier determination module is used to return to the execution of the second level data acquisition module until the maximum value of the number of aggregates corresponding to all aggregate data in the (i-1)th level is less than or equal to the preset number of aggregates threshold. Then, it is determined that the maximum value of the number of aggregates corresponding to all aggregate data in the (i-1)th level satisfies the second preset condition, and the current i value is determined to be the target level identifier R. The search result acquisition module is used to obtain search results based on the target user data, the length of the preset bit array, N preset hash functions, and all aggregated data when the target user's data corresponding to the preset tag is obtained. The search result acquisition module includes: The target bit array acquisition submodule is used to obtain the target bit array corresponding to the target user data when the target user data corresponding to the target user under the preset tag is obtained, based on the target user data, the length of the preset bit array and N preset hash functions; The identifier initialization submodule is used to initialize the reference level identifier u=1; The first search result acquisition submodule is used to compare the target bit array with each aggregated data in the u-th level that satisfies the third preset condition. If the comparison result with any aggregated data satisfies the fourth preset condition, then it is determined that the aggregated data in the (u+1)-th level corresponding to the aggregated data that satisfies the fourth preset condition satisfies the third preset condition; otherwise, it is determined that the search result is that the target user data was not found. The first aggregated data in the first level satisfies the third preset condition. The first search result acquisition submodule includes: The first preset condition judgment unit is used to determine that, for any aggregated data in the u-th level that satisfies the third preset condition, if the values ​​of each position with a value of 1 in the target bit array are all 1 in the corresponding position in the current aggregated data, the comparison result between the target bit array and the current aggregated data satisfies the fourth preset condition, and the aggregated data in the (u+1)-th level corresponding to the current aggregated data satisfies the third preset condition. The second preset condition judgment unit is used to determine that if the values ​​of each position with a value of 1 in the target position array are not all 1 in the corresponding position in the current aggregated data, then the comparison result between the target position array and the current aggregated data does not meet the fourth preset condition. The first search result acquisition unit is used to traverse all aggregated data in the u-th level that meet the third preset condition. If the comparison result with all aggregated data in the u-th level that meet the third preset condition does not meet the fourth preset condition, then the search result is determined to be that no target user data was found. The second search result acquisition submodule is used to update u=u+1, return to execution step S73, until u=R+1, and determine that the search result is that the target user data has been retrieved.

6. A non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores at least one instruction or at least one program, characterized in that, The at least one instruction or the at least one program segment is loaded and executed by the processor to implement the target user retrieval method as described in any one of claims 1-4.

7. An electronic device, characterized in that, Includes a processor and the non-transitory computer-readable storage medium as described in claim 6.

Images