Data processing method and device, computer device and storage medium

Abstract

This application relates to the field of big data technology, and in particular to a data processing method, apparatus, computer equipment, and storage medium. The method includes: acquiring target resource transfer data of a target object; determining target detection features from multiple detection features based on the target resource transfer data; determining at least one target primary event feature from multiple primary event features corresponding to the target detection feature; determining a target detection dimension from multiple detection dimensions corresponding to any target primary event feature; and obtaining a detection result for the target object based on multiple secondary event features corresponding to the target primary event feature in the target detection dimension. This method can improve the stability of analyzing the influencing factors of resource transfer data.

Description

Technical Field

[0001] This application relates to the field of big data technology, and in particular to a data processing method, apparatus, computer equipment, and storage medium. Background Technology

[0002] With the development of the big data field, in the area of ​​resource transfer data processing for service providers, analyzing and processing the influencing factors of resource transfer data based on the analysis results of big data has become a high-priority demand.

[0003] In traditional technologies, there is no systematic attribution method when diagnosing and analyzing the influencing factors of resource transfer data. Usually, problems are attributed manually based on business experience. This method requires a high level of personal expertise, is time-consuming, has unstable returns, and cannot support automated operations.

[0004] It is evident that current methods for analyzing the influencing factors of resource transfer data suffer from problems such as the need for manual processing, long processing times, and low stability. Summary of the Invention

[0005] Therefore, it is necessary to provide a data processing method, apparatus, computer equipment, and storage medium that can automatically analyze the influencing factors of resource transfer data to address the aforementioned technical problems.

[0006] Firstly, this application provides a data processing method. The method includes:

[0007] Obtain the target resource transfer data of the target object;

[0008] Based on the target resource transfer data, target detection features are determined from multiple detection features, wherein the target resource transfer data is constructed from the detection data corresponding to the multiple detection features;

[0009] From the multiple primary event features corresponding to the target detection feature, at least one target primary event feature is determined, and the detection data corresponding to the target detection feature is constructed from the feature data corresponding to the multiple primary event features;

[0010] For any of the target primary event features, a target detection dimension is determined from multiple detection dimensions corresponding to the target primary event feature;

[0011] Based on the target primary event features and the corresponding secondary event features in the target detection dimension, the detection result for the target object is obtained, and the feature data corresponding to the target primary event features is constructed from the feature data corresponding to the multiple secondary event features.

[0012] In one embodiment, obtaining the detection result for the target object based on the multiple secondary event features corresponding to the primary event features in the target detection dimension includes:

[0013] If the secondary event feature corresponding to the primary event feature of the target in the target detection dimension is a target feature, then the detection result for the target object is determined based on the secondary event feature; or...

[0014] If the secondary event feature corresponding to the primary event feature of the target in the target detection dimension is not the target feature, the secondary event feature of the target is determined from multiple secondary event features, the target feature corresponding to the secondary event feature of the target is determined, and the detection result for the target feature is determined based on the target feature.

[0015] In one embodiment, acquiring the target resource transfer data of the target object includes:

[0016] Obtain resource transfer data of the target object in the current period, and obtain comparative resource transfer data of the target object in the comparison period;

[0017] A first probability value is determined based on the resource transfer data and the comparative resource transfer data;

[0018] If the first probability value is less than the first preset threshold, the resource transfer data of the current period shall be used as the target resource transfer data of the target object.

[0019] In one embodiment, determining the target detection feature from multiple detection features based on the target resource transfer data includes:

[0020] Obtain multiple detection features corresponding to the target resource transfer data;

[0021] For any of the aforementioned detection features, the contribution of the detection feature is determined based on the detection data corresponding to the detection feature, the target resource transfer data, and the comparison resource transfer data of the comparison period.

[0022] The target detection feature is determined from the plurality of detection features based on the contribution of each of the detection features.

[0023] In one embodiment, determining the contribution of the detection feature based on the detection data corresponding to the detection feature, the target resource transfer data, and the comparison resource transfer data of the comparison period includes:

[0024] The target resource transfer data difference is obtained based on the target resource transfer data and the comparison resource transfer data of the comparison period;

[0025] Based on the detection data of the detection feature in the current period and the detection data of the detection feature in the comparison period, the detection data difference corresponding to the detection feature is obtained;

[0026] The contribution of the detection feature is obtained based on the difference in the resource transfer data and the difference in the detection data.

[0027] In one embodiment, determining at least one target first-level event feature from a plurality of first-level event features corresponding to the target detection feature includes:

[0028] For any first-level event feature corresponding to the target detection feature, the contribution of the first-level event feature is determined based on the feature data corresponding to the first-level event feature, the detection data of the target detection feature in the current period, and the detection data of the target detection feature in the comparison period.

[0029] Based on the contribution of each of the primary event features, at least one target primary event feature is determined from the plurality of primary event features.

[0030] In one embodiment, determining the target detection dimension from multiple detection dimensions corresponding to the target primary event feature includes:

[0031] Obtain multiple detection dimensions corresponding to the target primary event features;

[0032] For any of the detection dimensions, obtain multiple secondary event features corresponding to the target primary event features in the detection dimension;

[0033] For any of the detection dimensions, a first probability distribution of the target primary event feature on the detection dimension is determined based on the feature data of the secondary event feature in the current period, and a second probability distribution of the target primary event feature on the detection dimension is determined based on the feature data of the secondary event feature in the comparison period.

[0034] Based on the first probability distribution and the second probability distribution corresponding to each of the detection dimensions, the target detection dimension corresponding to the target first-level event feature is determined from each of the detection dimensions.

[0035] In one embodiment, determining the target detection dimension corresponding to the target primary event feature from each of the detection dimensions based on the first probability distribution and the second probability distribution corresponding to each of the detection dimensions includes:

[0036] For any of the detection dimensions, the target divergence of the target primary event feature on the detection dimension is determined based on the first probability distribution and the second probability distribution corresponding to the detection dimension.

[0037] The feature divergence is determined from the target divergence corresponding to the target primary event feature in each of the detection dimensions, wherein the feature divergence is the maximum value of the target divergence of the target primary event feature in each of the detection dimensions;

[0038] The detection dimension corresponding to the feature divergence is used as the target detection dimension corresponding to the target first-level event feature.

[0039] In one embodiment, obtaining the target resource transfer data of the target object further includes:

[0040] Obtain resource transfer data for the target object in the current period;

[0041] A second probability value is obtained by performing linear regression on the resource transfer data;

[0042] If the second probability value is less than the second preset threshold, the current period is taken as the target period, and the resource transfer data of the current period is taken as the target resource transfer data of the target object.

[0043] Secondly, this application also provides a data processing apparatus. The apparatus includes:

[0044] The data acquisition module is used to acquire target resource transfer data of the target object;

[0045] The detection feature determination module is used to determine target detection features from multiple detection features based on the target resource transfer data, wherein the target resource transfer data is constructed from the detection data corresponding to the multiple detection features;

[0046] The event feature determination module is used to determine at least one target primary event feature from multiple primary event features corresponding to the target detection feature, wherein the detection data corresponding to the target detection feature is constructed from the feature data corresponding to the multiple primary event features;

[0047] The detection dimension determination module is used to determine the target detection dimension from multiple detection dimensions corresponding to any one of the target first-level event features;

[0048] The data detection module is used to obtain the detection result for the target object based on the target primary event features and the corresponding secondary event features in the target detection dimension. The feature data corresponding to the target primary event features is constructed from the feature data corresponding to the multiple secondary event features.

[0049] In one embodiment, the data detection module is further configured to: determine a detection result for the target object based on the secondary event feature when the secondary event feature corresponding to the primary event feature in the target detection dimension is a target feature; or, when the secondary event feature corresponding to the primary event feature in the target detection dimension is not the target feature, determine a target secondary event feature from multiple secondary event features, determine the target feature corresponding to the target secondary event feature, and determine the detection result for the target feature based on the target feature.

[0050] In one embodiment, the data acquisition module is further configured to acquire resource transfer data of the target object in the current period and comparative resource transfer data of the target object in the comparison period; determine a first probability value based on the resource transfer data and the comparative resource transfer data; and, if the first probability value is less than a first preset threshold, use the resource transfer data of the current period as the target resource transfer data of the target object.

[0051] In one embodiment, the detection feature determination module is further configured to acquire multiple detection features corresponding to the target resource transfer data; for any one of the detection features, determine the contribution of the detection feature based on the detection data corresponding to the detection feature, the target resource transfer data, and the comparison resource transfer data of the comparison period; and determine the target detection feature from the multiple detection features based on the contribution of each detection feature.

[0052] In one embodiment, the detection feature determination module is further configured to: obtain a target resource transfer data difference based on the target resource transfer data and the comparison resource transfer data of the comparison period; obtain a detection data difference corresponding to the detection feature based on the detection data of the detection feature in the current period and the detection data of the detection feature in the comparison period; and obtain the contribution of the detection feature based on the target resource transfer data difference and the detection data difference.

[0053] In one embodiment, the event feature determination module is further configured to, for any first-level event feature corresponding to the target detection feature, determine the contribution of the first-level event feature based on the feature data corresponding to the first-level event feature, the detection data of the target detection feature in the current period, and the detection data of the target detection feature in the comparison period; and determine at least one target first-level event feature from the plurality of first-level event features based on the contribution of each of the first-level event features.

[0054] In one embodiment, the detection dimension determination module is further configured to: acquire multiple detection dimensions corresponding to the target primary event feature; acquire multiple secondary event features corresponding to the target primary event feature in the detection dimension for any one of the detection dimensions; determine a first probability distribution of the target primary event feature on the detection dimension based on the feature data of the secondary event feature in the current period for any one of the detection dimensions, and determine a second probability distribution of the target primary event feature on the detection dimension based on the feature data of the secondary event feature in the comparison period; and determine the target detection dimension corresponding to the target primary event feature from each of the detection dimensions based on the first probability distribution and the second probability distribution corresponding to each of the detection dimensions.

[0055] In one embodiment, the detection dimension determination module is further configured to, for any detection dimension, determine the target divergence corresponding to the target primary event feature on the detection dimension based on the first probability distribution and the second probability distribution corresponding to the detection dimension; determine the feature divergence from the target divergences corresponding to the target primary event feature on each detection dimension, wherein the feature divergence is the maximum value of the target divergences of the target primary event feature on each detection dimension; and use the detection dimension corresponding to the feature divergence as the target detection dimension corresponding to the target primary event feature.

[0056] In one embodiment, the data acquisition module is further configured to acquire resource transfer data of the target object in the current period; perform linear regression on the resource transfer data to obtain a second probability value; and if the second probability value is less than a second preset threshold, use the current period as the target period and the resource transfer data of the current period as the target resource transfer data of the target object.

[0057] Thirdly, this application also provides a computer device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0058] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the above-described method embodiments.

[0059] Fifthly, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0060] The aforementioned data processing method, apparatus, computer equipment, storage medium, and computer program product acquire target resource transfer data of a target object; based on the target resource transfer data, determine target detection features from multiple detection features, wherein the target resource transfer data is constructed from the detection data corresponding to the multiple detection features; determine at least one target primary event feature from multiple primary event features corresponding to the target detection feature, wherein the detection data corresponding to the target detection feature is constructed from the feature data corresponding to the multiple primary event features; for any target primary event feature, determine a target detection dimension from multiple detection dimensions corresponding to the target primary event feature; and based on multiple secondary event features corresponding to the target primary event feature in the target detection dimension, obtain a detection result for the target object, wherein the feature data corresponding to the target primary event feature is constructed from the feature data corresponding to the multiple secondary event features. The data processing method, apparatus, computer equipment, storage medium, and computer program product provided in this application determine target detection features from detection features layer by layer, determine target first-level event features from the first-level event features corresponding to the target detection features, and then determine multiple second-level event features corresponding to the target first-level event features in the target detection dimension, thereby obtaining detection results for the target object. This achieves automated attribution analysis of changes in target resource transfer data of the target object and improves the stability of analyzing the influencing factors of resource transfer data. Attached Figure Description

[0061] Figure 1 This is a flowchart illustrating a data processing method in one embodiment;

[0062] Figure 2 This is a flowchart illustrating step 102 in one embodiment;

[0063] Figure 3 This is a flowchart illustrating step 104 in another embodiment;

[0064] Figure 4 This is a flowchart illustrating step 304 in one embodiment;

[0065] Figure 5This is a flowchart illustrating step 106 in one embodiment;

[0066] Figure 6 This is a flowchart illustrating step 108 in one embodiment;

[0067] Figure 7 This is a flowchart illustrating step 608 in one embodiment;

[0068] Figure 8 This is a flowchart illustrating step 102 in another embodiment;

[0069] Figure 9 This is a flowchart illustrating a data processing method in one embodiment;

[0070] Figure 10 This is a structural block diagram of a data processing device in one embodiment;

[0071] Figure 11 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0072] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0073] Offline shopping mall revenue can be divided into two main parts: rent and GMV (Gross Merchandise Volume) commission. Increasing store GMV can bring additional commission revenue to the mall. To improve a store's GMV, it is first necessary to determine whether the current GMV is good or bad. Therefore, the ability to accurately and efficiently diagnose and analyze the GMV situation, identify problems, and provide rectification suggestions becomes a high priority for the business.

[0074] Traditional technologies have two major drawbacks in diagnosing store revenue and performing revenue attribution analysis:

[0075] First, offline stores that have joined the customer flow system lack systematic diagnostic indicators, diagnostic methods, and early warning methods during operation, and their operating methods are relatively traditional and extensive.

[0076] Secondly, there are many dimensions that can be attributed to the indicators that have been diagnosed as problems. Currently, there is no systematic method for attribution or to provide suggestions, and the manpower and time costs of each analysis are very high.

[0077] Based on this, embodiments of this application provide a data processing method to solve the above problems and improve the stability of the influencing factors of resource transfer data analysis.

[0078] In one embodiment, such as Figure 1 As shown, a data processing method is provided. This embodiment illustrates the method applied to a server, but it is understood that the method can also be applied to a terminal, or to a system including both a terminal and a server, and implemented through interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0079] Step 102: Obtain the target resource transfer data of the target object.

[0080] The target object can be any service provider to be analyzed, such as a store in a shopping mall. The target resource transfer data is the data of the influencing factors to be analyzed, such as the store's GMV.

[0081] Step 104: Based on the target resource transfer data, determine the target detection features from multiple detection features. The target resource transfer data is constructed from the detection data corresponding to the multiple detection features.

[0082] The detection features can be derived from the calculation formula of the target resource transfer data. In other words, the target resource transfer data can be obtained by calculating the detection data corresponding to multiple detection features. The target detection feature is the one that has the greatest impact on changes in the target resource transfer data among multiple detection features. Taking GMV as the target resource transfer data as an example, GMV = number of deep-browsing users × conversion rate × average order value. Multiple detection features can include the number of deep-browsing users, conversion rate, and average order value. The number of deep-browsing users is the number of people who stay on the target object for more than a certain threshold, and the conversion rate refers to the conversion rate from deep-browsing users to the number of completed transactions. It should be noted that this application embodiment does not specifically limit the method of obtaining the detection data corresponding to the detection features. For example, the number of deep-browsing users can be directly retrieved from the server, and the conversion rate and average order value can be directly obtained from the mall or store.

[0083] Specifically, the contribution of each detection feature can be determined based on the target resource transfer data, the detection data of each detection feature in the current period and the comparison period, and the detection feature with the largest contribution can be used as the target detection feature.

[0084] Step 106: Determine at least one target first-level event feature from multiple first-level event features corresponding to the target detection feature. The detection data corresponding to the target detection feature is constructed from the feature data corresponding to multiple first-level event features.

[0085] The primary event features can be obtained by decomposing them according to the calculation formula of the target detection features. That is, the detection data corresponding to the target detection features can be obtained by calculating the feature data corresponding to multiple primary event features. The target primary event feature is the primary event feature that has the greatest impact on the change of the detection data corresponding to the target detection feature among multiple primary event features. Taking the target detection feature as the number of people who browse deeply as an example, the number of people who browse deeply = the number of people entering the mall × the floor climbing rate × the store passing rate × the store entry rate × the deep browsing rate. Multiple primary event features can include the number of people entering the mall, the floor climbing rate, the store passing rate, the store entry rate, the deep browsing rate, etc. The number of people entering the mall is the number of people entering the mall, the floor climbing rate is the ratio of people entering the floor where the target object is located, the store passing rate is the ratio of people passing by the target object, the store entry rate is the ratio of people entering the target object, and the deep browsing rate is the ratio of people browsing deeply in the target object. It should be noted that the embodiments of this application do not specifically limit the method of obtaining the feature data corresponding to the primary event features. For example, the number of people entering the mall, the floor climbing rate, the store passing rate, the store entry rate, the deep browsing rate, etc. can all be directly retrieved from the server.

[0086] Specifically, the contribution of each primary event feature can be determined based on the detection data corresponding to the target detection feature, the feature data corresponding to each primary event feature in the current period and the comparison period, and the contribution can be arranged from largest to smallest. At least one primary event feature with a large contribution is selected as the target primary event feature according to the preset number of target primary event features.

[0087] Step 108: For any target first-level event feature, determine the target detection dimension from multiple detection dimensions corresponding to the target first-level event feature.

[0088] For any target primary event feature, it can be divided into multiple secondary event features according to various detection dimensions. The detection dimension represents the classification of multiple secondary event features obtained by attributing the target primary event feature to the next layer of influencing factors. When a target primary event feature is divided into multiple secondary event features according to a single detection dimension, the difference in feature data among the multiple secondary event features is greatest between the current period and the comparison period; in this case, the detection dimension is the target detection dimension.

[0089] For example, taking the number of attendees as the primary event feature, the detection dimensions could be gender (male / female), age (less than 15 / 16-22 / 23-28 / 29-36 / 37-45 / 46-55 / greater than 55), and whether they are in the same group (single / family / couple / friends), etc. The Jensen-Shannon divergence (JS divergence) of the primary event feature across each detection dimension can be calculated, and the detection dimension with the largest JS divergence can be selected as the target detection dimension.

[0090] Step 110: Based on the target's primary event features and the corresponding secondary event features in the target detection dimension, the detection results for the target object are obtained. The feature data corresponding to the target's primary event features are constructed from the feature data corresponding to the secondary event features.

[0091] The secondary event features can be derived from the calculation formula of the primary event features of the target. In other words, the feature data corresponding to the primary event features of the target can be obtained by calculating the feature data corresponding to multiple secondary event features. The detection results of the target object can include the attribution results of the target object's resource transfer data, i.e., the target features that have the greatest impact on changes in the target resource transfer data.

[0092] The above data processing method determines the target detection features from the detection features layer by layer, determines the target first-level event features from the first-level event features corresponding to the target detection features, and then determines the multiple second-level event features corresponding to the target first-level event features in the target detection dimension, thereby obtaining the detection results for the target object. This achieves automated attribution analysis of changes in the target resource transfer data of the target object and improves the stability of the analysis of the influencing factors of resource transfer data.

[0093] In one embodiment, in step 110, obtaining the detection result for the target object based on the multiple secondary event features corresponding to the primary event features of the target in the target detection dimension may include:

[0094] If the secondary event feature corresponding to the primary event feature of the target in the target detection dimension is the target feature, then the detection result for the target object is determined based on the secondary event feature; or...

[0095] If the secondary event feature corresponding to the primary event feature of the target is not the target feature in the target detection dimension, the target secondary event feature is determined from multiple secondary event features, the target feature corresponding to the target secondary event feature is determined, and the detection result for the target feature is determined based on the target feature.

[0096] Specifically, the target feature can be a pre-defined feature that can be manually intervened. When the secondary event feature corresponding to the primary event feature in the target detection dimension is the target feature, the secondary event feature can be output as the detection result for the target object. Alternatively, the contribution of multiple secondary event features can be calculated, and the secondary event feature with the highest contribution can be output as the detection result for the target object. For example, if the primary event feature is the number of people entering and the target detection dimension is the peer dimension, the secondary event features can include the number of single people entering, the number of families entering, the number of couples entering, and the number of friends entering. In this case, the number of single people entering, the number of families entering, the number of couples entering, and the number of friends entering are pre-defined target features. The detection result for the target object can include the target features such as the number of single people entering, the number of families entering, the number of couples entering, and the number of friends entering, or it can include the feature data changes of the target features such as the number of single people entering, the number of families entering, the number of couples entering, and the number of friends entering in the current period and the comparison period. Alternatively, a unique target feature can be obtained as the detection result by calculating the contribution of the target features such as the number of single people entering, the number of families entering, the number of couples entering, and the number of friends entering.

[0097] When the secondary event feature corresponding to the primary event feature in the target detection dimension is not the target feature, the target secondary event feature can be determined from multiple secondary event features. The target secondary event feature is the secondary event feature that has the greatest impact on the change of the feature data corresponding to the primary event feature among multiple secondary event features. The target secondary event feature can be determined by calculating the contribution of multiple secondary event features and selecting the secondary event feature with the highest contribution. The corresponding target feature can be obtained according to the calculation formula of the target secondary event feature. Taking the primary event feature as store entry rate and the target detection dimension as gender as an example, the breakdown of store entry rate satisfies the following formula (I).

[0098]

[0099] In this context, RATE represents rate, CNT represents number of people, and RATIO represents percentage. The subscript of RATE, CNT, or RATIO indicates an event, such as enter or pass, representing an entry or exit event. The superscript indicates a value for a specific dimension, such as male or female, representing values ​​segmented by gender. enterThe store entry rate is calculated as follows: Store Entry Rate = Male Store Entry Rate × Male Store Visit Rate + Female Store Entry Rate × Female Store Visit Rate. This means that secondary event features can include both male store entry rate × male store visit rate and female store entry rate × female store visit rate. However, these two rates are not the preset target features. Instead, the contribution of each secondary event feature can be calculated, and the secondary event feature with the highest contribution is selected as the target secondary event feature. For example, if the target secondary event feature is female store entry rate × female store visit rate, we can further break it down into two features: female store entry rate and female store visit rate. Then, by calculating the contribution, we can further break down the female store entry rate into two target features: the number of female store visitors and the number of female store visitors. The detection results for the target object can include target features such as the number of female store visitors and the number of female store visitors, or they can include the changes in feature data between the current period and the comparison period for these target features. Alternatively, a unique target feature can be obtained as the detection result by calculating the relative change rate of target features such as the number of female customers entering the store and the number of female customers passing through the store.

[0100] In this embodiment of the disclosure, the target resource transfer data of the target object is logically split into preset target features, and the detection results for the target features are determined based on the target features. This realizes automated attribution analysis of changes in the target resource transfer data of the target object and improves the stability of the analysis of the influencing factors of resource transfer data.

[0101] In one embodiment, such as Figure 2 As shown, in step 102, obtaining the target resource transfer data of the target object may include:

[0102] Step 202: Obtain the resource transfer data of the target object in the current period, and obtain the comparative resource transfer data of the target object in the comparison period.

[0103] In this embodiment, the duration of the current period and the comparison period are not specifically limited, and can be set according to actual needs. For example, the store's GMV in the current period can be obtained as resource transfer data, and the store's GMV in the comparison period can be obtained as comparative resource transfer data.

[0104] Step 204: Determine the first probability value based on the resource transfer data and by comparing the resource transfer data.

[0105] Among them, a t-test (Student's t-test) can be performed on the resource transfer data and the comparative resource transfer data, and the p-value obtained is used as the first probability value.

[0106] Step 206: If the first probability value is less than the first preset threshold, the resource transfer data of the current period is used as the target resource transfer data of the target object.

[0107] In this embodiment, the first preset threshold is not specifically limited; for example, it can be 0.05. Taking resource transfer data as GMV as an example, if the first probability value is less than the first preset threshold, it indicates that the target object's GMV in the current period and the comparison period have a significant difference; otherwise, it does not. If there is a significant difference, the difference between the target object's average GMV in the current period and the average GMV in the comparison period is calculated. If the result is positive, it indicates that the GMV in the current period is significantly better than the GMV in the comparison period; if the result is negative, it indicates that the GMV in the current period is significantly worse than the GMV in the comparison period. Only when the target object shows a significant difference between the current period and the comparison period is it necessary to perform the next step of attributing the changes in the target resource transfer data. Therefore, the resource transfer data in the current period is used as the target resource transfer data of the target object.

[0108] In this embodiment of the disclosure, the resource transfer data of the target object in the current period and the comparative resource transfer data of the target object in the comparison period are used to determine whether there is a significant difference between the two periods. The resource transfer data of the current period with significant differences is used as the target resource transfer data for attribution processing, thereby improving the efficiency of analyzing the influencing factors of resource transfer data.

[0109] In one embodiment, such as Figure 3 As shown, in step 104, determining the target detection features from multiple detection features based on the target resource transfer data may include:

[0110] Step 302: Obtain multiple detection features corresponding to the target resource transfer data.

[0111] Among them, the detection features can be obtained by splitting the target resource transfer data according to the calculation formula. Taking GMV as the target resource transfer data as an example, GMV = number of people who browsed deeply × conversion rate × average order value. Multiple detection features can include the number of people who browsed deeply, conversion rate, and average order value.

[0112] Step 304: For any detection feature, determine the contribution of the detection feature based on the detection data corresponding to the detection feature, the target resource transfer data, and the comparison resource transfer data of the comparison period.

[0113] The contribution calculation method can include two approaches: additive contribution and multiplicative contribution. The calculation method can be determined based on the calculation formula of the preceding feature. For example, taking GMV as the target resource transfer data, the preceding feature of the detection feature is the target resource transfer data. The formula for GMV is: GMV = Number of Deep Browsers × Conversion Rate × Average Transaction Value. In this case, the GMV calculation formula is a multiplicative formula, and the contribution of the detection feature can be calculated using the multiplicative contribution method. The multiplicative contribution is obtained by taking the logarithm of both sides of the GMV calculation formula to obtain an additive formula, and then calculating it using the additive contribution method. The additive contribution can satisfy the following formula (I).

[0114] Contribution of factor a = diff of factor a / diff of the overall market (Formula 1)

[0115] Factor a is the detection feature, diff is the difference between the detection data of the detection feature in the current period and the detection data in the comparison period, and the diff of the overall market is the difference between the target resource transfer data and the comparison resource transfer data in the comparison period.

[0116] Step 306: Determine the target detection feature from multiple detection features based on the contribution of each detection feature.

[0117] The contribution degree characterizes the extent to which changes in each detection feature affect changes in the target resource transfer data. After determining the contribution degree of each detection feature, the detection feature with the highest contribution degree can be used as the target detection feature.

[0118] In this embodiment of the disclosure, target detection features are selected from detection features based on contribution calculation. This allows for the selection of detection features that have the greatest impact on changes in target resource transfer data, thereby improving the stability of the analysis of influencing factors in resource transfer data.

[0119] In one embodiment, such as Figure 4 As shown, in step 304, the contribution of the detection feature is determined based on the detection data corresponding to the detection feature, the target resource transfer data, and the comparison resource transfer data of the comparison period. This may include:

[0120] Step 402: Based on the target resource transfer data and the comparative resource transfer data of the comparison period, obtain the target resource transfer data difference.

[0121] For example, the target resource transfer data GMV = number of deep browsing visitors × conversion rate × average order value. In this case, the GMV calculation formula is a multiplicative formula, and the contribution of the detection feature can be calculated using the multiplicative contribution method. The multiplicative contribution is calculated by taking the logarithm of both sides of the GMV calculation formula and then calculating it using the additive contribution method. The difference between the logarithm of the store's GMV in the current period and the logarithm of the store's GMV in the comparison period is obtained.

[0122] Step 404: Based on the detection data of the detection feature in the current period and the detection data of the detection feature in the comparison period, obtain the detection data difference corresponding to the detection feature.

[0123] For example, taking the number of deep-browsing users as a detection feature, the difference between the logarithmic value of the number of deep-browsing users in the current period and the logarithmic value of the number of deep-browsing users in the comparison period can be obtained.

[0124] Step 406: Based on the difference in resource transfer data and the difference in detection data, obtain the contribution of the detection features.

[0125] For example, the contribution of a detection feature is positively correlated with the difference in detection data. The contribution of a detection feature can be obtained by dividing the difference in detection data by the difference in resource transfer data.

[0126] In this embodiment of the disclosure, target detection features are selected from detection features based on contribution calculation. This allows for the selection of detection features that have the greatest impact on changes in target resource transfer data, thereby improving the stability of the analysis of influencing factors in resource transfer data.

[0127] In one embodiment, such as Figure 5 As shown, in step 106, determining at least one target-level event feature from multiple first-level event features corresponding to the target detection feature may include:

[0128] Step 502: For any first-level event feature corresponding to the target detection feature, determine the contribution of the first-level event feature based on the feature data corresponding to the first-level event feature, the detection data of the target detection feature in the current period, and the detection data of the target detection feature in the comparison period.

[0129] The contribution calculation method can include two approaches: additive contribution and multiplicative contribution. The calculation method can be determined based on the calculation formula of the preceding feature. Taking the number of visitors as the primary event feature as an example, the preceding target detection feature is the number of deep-browsing visitors. Deep-browsing visitors = number of visitors × floor climbing rate × store pass rate × store entry rate × deep-browsing rate. In this case, the calculation formula for deep-browsing visitors is a multiplicative formula, and the contribution of the primary event feature can be calculated using the multiplicative contribution method. The multiplicative contribution is calculated by taking the logarithm of both sides of the GMV calculation formula and then applying the additive contribution method. The difference between the logarithmic value of deep-browsing visitors in the current period and the logarithmic value of deep-browsing visitors in the comparison period is obtained. Similarly, the difference between the logarithmic value of visitors in the current period and the visitor count in the comparison period is obtained. The ratio of the visitor count difference to the deep-browsing visitor difference yields the contribution of the number of visitors.

[0130] Step 504: Based on the contribution of each primary event feature, determine at least one target primary event feature from multiple primary event features.

[0131] For example, the top two primary event features in terms of contribution can be selected as target primary event features. For instance, the target primary event features could be the number of people entering the venue and the store entry rate.

[0132] In this embodiment of the disclosure, target primary event features are selected from primary event features based on contribution calculation. This allows for the selection of target primary event features that have the greatest impact on changes in target resource transfer data, thereby improving the stability of the analysis of influencing factors on resource transfer data.

[0133] In one embodiment, such as Figure 6 As shown, in step 108, determining the target detection dimension from multiple detection dimensions corresponding to the target primary event features may include:

[0134] Step 602: Obtain multiple detection dimensions corresponding to the target primary event features.

[0135] For example, the detection dimension is a classification of multiple secondary event features obtained by attributing the primary event features of the target to the next level of influencing factors. Detection dimensions can be dimensions such as gender (male / female), age (less than 15 / 16-22 / 23-28 / 29-36 / 37-45 / 46-55 / greater than 55), and peers (single / family / couple / friends).

[0136] Step 604: For any detection dimension, obtain multiple secondary event features corresponding to the target primary event features in the detection dimension.

[0137] For example, taking the number of attendees as the primary event feature and peers as the detection dimension, multiple secondary event features can include the number of single attendees, the number of family attendees, the number of couple attendees, and the number of friends attendees.

[0138] Step 606: For any detection dimension, determine the first probability distribution of the target first-level event feature on the detection dimension based on the feature data of the second-level event feature in the current period, and determine the second probability distribution of the target first-level event feature on the detection dimension based on the feature data of the second-level event feature in the comparison period.

[0139] For example, taking the number of attendees as the primary event feature, the probability distribution is based on the proportion of feature data of different secondary event features under that detection dimension for any detection dimension. Taking gender as the detection dimension, the secondary event features are the number of male attendees and the number of female attendees. The first probability distribution is determined based on the feature data of male attendees and female attendees in the current period. The first probability distribution is also determined based on the feature data of male attendees and female attendees in the comparison period. For example, the first probability distribution of male and female genders in the current period is [0.3, 0.7], and the probability distribution of genders in the comparison period is [0.6, 0.4].

[0140] Step 608: Based on the first probability distribution and the second probability distribution corresponding to each detection dimension, determine the target detection dimension corresponding to the target first-level event feature from each detection dimension.

[0141] The JS divergence between the current period and the comparison period in the detection dimension can be calculated based on the first and second probability distributions. JS divergence measures the difference between the two probability distributions. The JS divergences for each detection dimension are sorted from largest to smallest; the detection dimension corresponding to the largest JS divergence is the target detection dimension.

[0142] In this embodiment of the disclosure, the first probability distribution and the second probability distribution of each detection dimension are calculated to obtain the corresponding JS divergence. The target detection dimension is determined based on the JS divergence, which improves the stability of the influencing factors of resource transfer data analysis.

[0143] In one embodiment, such as Figure 7 As shown, in step 608, determining the target detection dimension corresponding to the target primary event feature from each detection dimension based on the first probability distribution and the second probability distribution corresponding to each detection dimension may include:

[0144] Step 702: For any detection dimension, determine the target divergence corresponding to the target primary event feature on the detection dimension based on the first probability distribution and the second probability distribution corresponding to the detection dimension.

[0145] The target divergence is the JS divergence, which measures the difference between two first probability distributions and a second probability distribution. It should be noted that this embodiment does not impose a specific formula for calculating the JS divergence; it is sufficient that the JS divergence can be calculated from the first and second probability distributions.

[0146] Step 704: Determine the feature divergence from the target divergence corresponding to the target primary event features in each detection dimension. The feature divergence is the maximum value of the target divergence of the target primary event features in each detection dimension.

[0147] For example, taking the number of attendees as the primary event feature, after obtaining the JS divergence of the number of attendees on each detection dimension, the JS divergence of the number of attendees on the peer dimension is the largest, that is, the JS divergence on the peer dimension is taken as the feature divergence.

[0148] Step 706: Use the detection dimension corresponding to the feature divergence as the target detection dimension corresponding to the target first-level event feature.

[0149] Among them, the feature divergence is the maximum value of the target divergence of the target first-level event features in each detection dimension. The feature divergence is used as the target detection divergence, and the target first-level event features are split according to the target detection divergence to obtain multiple second-level event features.

[0150] In this embodiment of the disclosure, the first probability distribution and the second probability distribution of each detection dimension are calculated to obtain the corresponding JS divergence. The target detection dimension is determined based on the JS divergence, which improves the stability of the influencing factors of resource transfer data analysis.

[0151] In one embodiment, such as Figure 8 As shown, in step 102, obtaining the target resource transfer data of the target object may include:

[0152] Step 802: Obtain the resource transfer data of the target object in the current period.

[0153] For example, the store's GMV in the current period can be obtained as resource transfer data.

[0154] Step 804: Perform linear regression on the resource transfer data to obtain the second probability value.

[0155] In this process, linear regression is performed on the resource transfer data for the current period to determine whether there is a significant increase or decrease in the resource transfer data for the current period itself. The p-value obtained from the linear regression on the resource transfer data for the current period is used as the second probability value. It should be noted that the specific method of linear regression is not described in detail in the embodiments of this application. For example, its null hypothesis is that the slope is zero, and the Wald test with a t-distribution and a test statistic is used.

[0156] Step 806: If the second probability value is less than the second preset threshold, the current period is taken as the target period, and the resource transfer data of the current period is taken as the target resource transfer data of the target object.

[0157] In this embodiment, the second preset threshold is not specifically limited; for example, it can be 0.05. If the second probability value is less than the second preset threshold, it indicates that the resource transfer data of the target object has a significant trend change in the current period; otherwise, it does not. If there is a significant trend change, if the slope of the linear regression is greater than 0, it indicates that the resource transfer data in the current period has a significant upward trend; if the slope of the linear regression is less than 0, it indicates that the resource transfer data in the current period has a significant downward trend.

[0158] In this embodiment of the disclosure, the resource transfer data of the target object in the current period is used to determine whether there are significant changes in the target object in the current period. The resource transfer data of the current period with significant changes is used as the target resource transfer data for attribution processing, thereby improving the efficiency of analyzing the influencing factors of resource transfer data.

[0159] To facilitate a further understanding of the embodiments of this application, this application provides a most complete embodiment. Taking the application of the data processing method to store GMV analysis as an example, current offline store operation methods have the following shortcomings: First, when diagnosing store revenue, the business side usually judges the revenue performance by observing year-on-year and month-on-month comparisons, without deeper evaluation, such as whether there are significant differences in revenue between the diagnosis period and the comparison period, or whether there is a significant upward or downward trend in the diagnosis period. Second, when attributing the causes of problem indicators, there is no systematic attribution method. There are many factors affecting offline store revenue, and currently, most cases rely on business experience to attribute problems and provide rectification suggestions, which may not be supported by data or theory. Relying on manual diagnosis and analysis to find the causes of indicator changes is very costly, requires a high level of personal expertise, is time-consuming, and has unstable returns, and also cannot support automated operation. Third, there is currently no scientific method to estimate the potential benefits of rectification; estimation can only be done through post-event review. Therefore, the data processing method provided in this application, when applied to store GMV analysis, can solve the above problems.

[0160] The data processing method provided in this application can be applied to an intelligent diagnostic analysis system, which may include an automated diagnostic module and an automated attribution module. The automated diagnostic module is used to establish an indicator system, diagnostic methods, and early warning methods to help the business side automatically identify problems or opportunities in the business. The automated attribution module is used to calculate the contribution of changes in each factor x to changes in the target value y by splitting the logic tree, thereby quantitatively and intuitively giving the importance of each factor. It can also estimate how much change in the target value y is caused by changes in each factor x, thus completing the attribution. The automated diagnostic module can, based on the resource transfer data of the target object in the current period (diagnosis period) and the comparative resource transfer data in another period (comparison period), perform t-tests or linear regressions on the resource transfer data of the two periods to determine whether there are significant differences or significant trend changes, in order to determine whether to proceed with further attribution. When there are significant differences or significant trend changes in the resource transfer data between the two periods, the automated attribution module can use the calculation of contribution and JS divergence to break down the detection features corresponding to the target resource transfer data step by step into target features, thereby obtaining the detection results for the target object. Figure 9As shown, in the attribution process from target resource transfer data to the second-layer target detection features, changes in GMV are attributed to changes in the number of deep-browsing users, conversion rate, and average order value. This attribution step uses a multiplicative contribution method, obtaining a maximum contribution of 80% for the number of deep-browsing users, thus using them as the target detection feature. In the attribution process from the second-layer target detection features to the third-layer target primary event features, changes in the number of deep-browsing users are attributed to changes in the number of visitors, floor climbing rate, store visit rate, store entry rate, and deep-browsing rate. Using the multiplicative contribution method, the contribution of visitors is 27%, and the contribution of store entry rate is 50%. The top two contributors, store entry rate and visitors, are selected as the target primary event features. In the attribution process from the third-level target primary event feature to the fourth-level secondary event feature, when attributing the store entry rate or number of visitors downwards, multiple detection dimensions can be used. For example, the store entry rate can be attributed downwards based on dimensions such as gender, age, and peers. The detection dimension with the greatest difference can be calculated using JS divergence as the target detection dimension. For instance, if the store entry rate is attributed downwards based on gender, the number of visitors can be attributed downwards based on peers. The secondary event feature corresponding to the number of visitors is the preset target feature, and each secondary event feature and its contribution can be directly used as the detection result. The secondary event feature corresponding to the store entry rate is not the preset target feature. In the attribution process from the fifth-level secondary event feature to the target feature, the above process is repeated, selecting the secondary event feature with the greatest contribution as the target secondary event feature and attributing downwards until the preset target features of female store entry rate and female store visit percentage are obtained. The attribution of the female store entry rate can be obtained by calculating the relative change rate between the number of female store visitors and the number of female store visitors; the attribution of the female store visit percentage can be obtained by calculating the relative change rate between the number of female store visitors and the number of male store visitors. The data processing method provided in this application allows for the attribution of GMV down the hierarchy, all the way down to factors that can be manipulated by humans, such as the number of employees.

[0161] Intelligent diagnostic analysis systems can also include an automated suggestion module. This module takes pre-set suggestions for each detection result and outputs target suggestions based on the detection results of the actual target audience. For example, if the number of people browsing deeper into the store is the main contributor, the system can attribute the cause layer by layer to target characteristics, that is, which dimension of the customer flow funnel has the main contribution to the change in the number of people, and suggest corresponding actions to attract specific groups of people. If there is a significant difference in GMV between the current diagnostic period and the comparison period, it is because the conversion rate or average order value is the main contributor, and the system can suggest corresponding actions to adjust the conversion rate or average order value.

[0162] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0163] Based on the same inventive concept, this application also provides a data processing apparatus for implementing the data processing method described above. The solution provided by this apparatus is similar to the implementation scheme described in the above method; therefore, the specific limitations in one or more data processing apparatus embodiments provided below can be found in the limitations of the data processing method described above, and will not be repeated here.

[0164] In one embodiment, such as Figure 10 As shown, a data processing device 1000 is provided, including: a data acquisition module 1002, a detection feature determination module 1004, an event feature determination module 1006, a detection dimension determination module 1008, and a data detection module 1010, wherein:

[0165] The data acquisition module 1002 is used to acquire the target resource transfer data of the target object.

[0166] The detection feature determination module 1004 is used to determine the target detection feature from multiple detection features based on the target resource transfer data, wherein the target resource transfer data is constructed from the detection data corresponding to the multiple detection features.

[0167] The event feature determination module 1006 is used to determine at least one target first-level event feature from multiple first-level event features corresponding to the target detection feature. The detection data corresponding to the target detection feature is constructed from the feature data corresponding to multiple first-level event features.

[0168] The detection dimension determination module 1008 is used to determine the target detection dimension from multiple detection dimensions corresponding to any target first-level event feature.

[0169] The data detection module 1010 is used to obtain the detection results for the target object based on the target primary event features and the corresponding secondary event features in the target detection dimension. The feature data corresponding to the target primary event features is constructed from the feature data corresponding to the multiple secondary event features.

[0170] The data processing device 1000 provided in this application determines the target detection features from the detection features layer by layer, determines the target first-level event features from the first-level event features corresponding to the target detection features, and then determines multiple second-level event features corresponding to the target first-level event features in the target detection dimension, thereby obtaining the detection results for the target object. This realizes automated attribution analysis of changes in the target resource transfer data of the target object and improves the stability of the analysis of the influencing factors of resource transfer data.

[0171] In one embodiment, the data detection module 1010 is further configured to determine the detection result for the target object based on the secondary event feature when the secondary event feature corresponding to the primary event feature in the target detection dimension is the target feature; or, when the secondary event feature corresponding to the primary event feature in the target detection dimension is not the target feature, determine the target secondary event feature from multiple secondary event features, determine the target feature corresponding to the target secondary event feature, and determine the detection result for the target feature based on the target feature.

[0172] In one embodiment, the data acquisition module 1002 is further configured to acquire resource transfer data of the target object in the current period and comparative resource transfer data of the target object in the comparison period; determine a first probability value based on the resource transfer data and the comparative resource transfer data; and, if the first probability value is less than a first preset threshold, use the resource transfer data of the current period as the target resource transfer data of the target object.

[0173] In one embodiment, the detection feature determination module 1004 is further configured to acquire multiple detection features corresponding to the target resource transfer data; for any detection feature, determine the contribution of the detection feature based on the detection data corresponding to the detection feature, the target resource transfer data, and the comparison resource transfer data of the comparison period; and determine the target detection feature from the multiple detection features based on the contribution of each detection feature.

[0174] In one embodiment, the detection feature determination module 1004 is further configured to obtain the target resource transfer data difference based on the target resource transfer data and the comparison resource transfer data of the comparison period; obtain the detection data difference corresponding to the detection feature based on the detection data of the detection feature in the current period and the detection data of the detection feature in the comparison period; and obtain the contribution of the detection feature based on the target resource transfer data difference and the detection data difference.

[0175] In one embodiment, the event feature determination module 1006 is further configured to determine the contribution of any first-level event feature corresponding to the target detection feature based on the feature data corresponding to the first-level event feature, the detection data of the target detection feature in the current period, and the detection data of the target detection feature in the comparison period; and to determine at least one target first-level event feature from multiple first-level event features based on the contribution of each first-level event feature.

[0176] In one embodiment, the detection dimension determination module 1008 is further configured to acquire multiple detection dimensions corresponding to the target primary event features; for any detection dimension, acquire multiple secondary event features corresponding to the target primary event features in the detection dimension; for any detection dimension, determine a first probability distribution of the target primary event features on the detection dimension based on the feature data of the secondary event features in the current period, and determine a second probability distribution of the target primary event features on the detection dimension based on the feature data of the secondary event features in the comparison period; and determine the target detection dimension corresponding to the target primary event features from each detection dimension based on the first probability distribution and the second probability distribution corresponding to each detection dimension.

[0177] In one embodiment, the detection dimension determination module 1008 is further configured to, for any detection dimension, determine the target divergence corresponding to the target primary event feature on the detection dimension based on the first probability distribution and the second probability distribution corresponding to the detection dimension; determine the feature divergence from the target divergences corresponding to the target primary event feature on each detection dimension, wherein the feature divergence is the maximum value among the target divergences of the target primary event feature on each detection dimension; and use the detection dimension corresponding to the feature divergence as the target detection dimension corresponding to the target primary event feature.

[0178] In one embodiment, the data acquisition module 1002 is further configured to acquire resource transfer data of the target object in the current period; perform linear regression on the resource transfer data to obtain a second probability value; and if the second probability value is less than a second preset threshold, use the current period as the target period and the resource transfer data of the current period as the target resource transfer data of the target object.

[0179] Each module in the aforementioned data processing device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the operations corresponding to each module.

[0180] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 11As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database stores target resource transfer data and data corresponding to various characteristics of the target object. The network interface communicates with external terminals via a network connection. The computer program, when executed by the processor, performs data processing.

[0181] Those skilled in the art will understand that Figure 11 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0182] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0183] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0184] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0185] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0186] 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, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0187] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0188] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A data processing method, characterized by, The method includes: Obtain the target resource transfer data of the target object; Based on the target resource transfer data, target detection features are determined from multiple detection features, wherein the target resource transfer data is constructed from the detection data corresponding to the multiple detection features; From the multiple primary event features corresponding to the target detection feature, at least one target primary event feature is determined, and the detection data corresponding to the target detection feature is constructed from the feature data corresponding to the multiple primary event features; For any of the target primary event features, a target detection dimension is determined from multiple detection dimensions corresponding to the target primary event feature; Based on the target primary event features and the multiple secondary event features corresponding to the target detection dimension, a detection result for the target object is obtained, wherein the feature data corresponding to the target primary event features is constructed from the feature data corresponding to the multiple secondary event features. The acquisition of target resource transfer data of the target object includes: Obtain resource transfer data of the target object in the current period, and obtain comparative resource transfer data of the target object in the comparison period; A first probability value is determined based on the resource transfer data and the comparative resource transfer data; If the first probability value is less than the first preset threshold, the resource transfer data of the current period shall be used as the target resource transfer data of the target object.

2. The method of claim 1, wherein, The step of obtaining the detection result for the target object based on the multiple secondary event features corresponding to the primary event features in the target detection dimension includes: If the secondary event feature corresponding to the primary event feature of the target in the target detection dimension is a target feature, then the detection result for the target object is determined based on the secondary event feature; or... If the secondary event feature corresponding to the primary event feature of the target in the target detection dimension is not the target feature, the secondary event feature of the target is determined from multiple secondary event features, the target feature corresponding to the secondary event feature of the target is determined, and the detection result for the target feature is determined based on the target feature.

3. The method of claim 1, wherein, The step of determining target detection features from multiple detection features based on the target resource transfer data includes: Obtain multiple detection features corresponding to the target resource transfer data; For any of the aforementioned detection features, the contribution of the detection feature is determined based on the detection data corresponding to the detection feature, the target resource transfer data, and the comparison resource transfer data of the comparison period. The target detection feature is determined from the plurality of detection features based on the contribution of each of the detection features.

4. The method of claim 3, wherein, The step of determining the contribution of the detection feature based on the detection data corresponding to the detection feature, the target resource transfer data, and the comparison resource transfer data of the comparison period includes: The target resource transfer data difference is obtained based on the target resource transfer data and the comparison resource transfer data of the comparison period; Based on the detection data of the detection feature in the current period and the detection data of the detection feature in the comparison period, the detection data difference corresponding to the detection feature is obtained; The contribution of the detection feature is obtained based on the difference in the resource transfer data and the difference in the detection data.

5. The method of claim 1, wherein, Determining at least one target primary event feature from multiple primary event features corresponding to the target detection feature includes: For any first-level event feature corresponding to the target detection feature, the contribution of the first-level event feature is determined based on the feature data corresponding to the first-level event feature, the detection data of the target detection feature in the current period, and the detection data of the target detection feature in the comparison period. Based on the contribution of each of the primary event features, at least one target primary event feature is determined from the plurality of primary event features.

6. The method of claim 1, wherein, Determining the target detection dimension from multiple detection dimensions corresponding to the target primary event features includes: Obtain multiple detection dimensions corresponding to the target primary event features; For any of the detection dimensions, obtain multiple secondary event features corresponding to the target primary event features in the detection dimension; For any of the detection dimensions, a first probability distribution of the target primary event feature on the detection dimension is determined based on the feature data of the secondary event feature in the current period, and a second probability distribution of the target primary event feature on the detection dimension is determined based on the feature data of the secondary event feature in the comparison period. Based on the first probability distribution and the second probability distribution corresponding to each of the detection dimensions, the target detection dimension corresponding to the target first-level event feature is determined from each of the detection dimensions.

7. The method of claim 6, wherein, The step of determining the target detection dimension corresponding to the target primary event feature from each of the detection dimensions based on the first probability distribution and the second probability distribution corresponding to each of the detection dimensions includes: For any of the detection dimensions, the target divergence of the target primary event feature on the detection dimension is determined based on the first probability distribution and the second probability distribution corresponding to the detection dimension. The feature divergence is determined from the target divergence corresponding to the target primary event feature in each of the detection dimensions, wherein the feature divergence is the maximum value of the target divergence of the target primary event feature in each of the detection dimensions; The detection dimension corresponding to the feature divergence is used as the target detection dimension corresponding to the target first-level event feature.

8. The method according to claim 1, characterized in that, The acquisition of target resource transfer data of the target object also includes: Obtain resource transfer data for the target object in the current period; A second probability value is obtained by performing linear regression on the resource transfer data; If the second probability value is less than the second preset threshold, the current period is taken as the target period, and the resource transfer data of the current period is taken as the target resource transfer data of the target object.

9. A data processing apparatus, characterized by, The device includes: The data acquisition module is used to acquire target resource transfer data of the target object; The detection feature determination module is used to determine target detection features from multiple detection features based on the target resource transfer data, wherein the target resource transfer data is constructed from the detection data corresponding to the multiple detection features; The event feature determination module is used to determine at least one target primary event feature from multiple primary event features corresponding to the target detection feature, wherein the detection data corresponding to the target detection feature is constructed from the feature data corresponding to the multiple primary event features; The detection dimension determination module is used to determine the target detection dimension from multiple detection dimensions corresponding to any one of the target first-level event features; The data detection module is used to obtain the detection result for the target object based on the multiple secondary event features corresponding to the primary event features of the target in the target detection dimension. The feature data corresponding to the primary event features of the target is constructed from the feature data corresponding to the multiple secondary event features. Specifically, the data acquisition module is used to acquire the resource transfer data of the target object in the current period, and to acquire the comparative resource transfer data of the target object in the comparison period. A first probability value is determined based on the resource transfer data and the comparative resource transfer data; If the first probability value is less than the first preset threshold, the resource transfer data of the current period shall be used as the target resource transfer data of the target object.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 8.

11. A computer readable storage medium having stored thereon a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 8.

12. A computer program product comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 8.

Images