Data analysis method, device, computer readable medium, and electronic device

Abstract

The present disclosure relates to a data analysis method, device, computer readable medium and electronic equipment, the method comprising: obtaining historical business data and a preset analysis index; processing the historical business data into binary classification data under different data dimensions; determining the distribution difference between each type of data in each binary classification data, and determining the category interpretation degree of each type of data in each binary classification data in the historical business data according to the preset analysis index; and determining a target data sub-group in the historical business data according to the distribution difference and the category interpretation degree. Through the above technical solution, different data dimensions of each historical business data can be automatically analyzed to accurately determine a target data sub-group meeting the preset analysis index from different data dimensions.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, and more specifically, to a data analysis method, apparatus, computer-readable medium, and electronic device. Background Technology

[0002] In real-world business scenarios, it is often necessary to break down and analyze historical business data by dimension in order to locate subgroups (people / things) that meet certain target indicators, and then apply specific business strategies to those subgroups.

[0003] In related technologies, manual decomposition analysis is generally used to perform dimensional decomposition analysis on historical business data in order to locate subgroups that meet certain target indicators. However, manual decomposition analysis is inefficient, time-consuming, and it is difficult to find the optimal subgroup by exhaustively combining different dimensions. Summary of the Invention

[0004] This summary section is provided to briefly introduce the concepts, which will be described in detail in the detailed description section below. This summary section is not intended to identify key or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0005] In a first aspect, this disclosure provides a data analysis method, the method comprising:

[0006] Acquire historical business data and preset analysis indicators. The preset analysis indicators are used to characterize the relationship between target data and target indicators that are expected to be obtained from the historical business data. The target indicators are obtained by dividing a first data indicator by a second data indicator.

[0007] The historical business data is processed into binary classification data under different data dimensions;

[0008] Determine the distribution difference between each type of data in each binary classification data, and determine the category explanatory power of each type of data in the historical business data according to the preset analysis index. The category explanatory power is used to characterize the data proportion of each type of data in the historical business data.

[0009] Based on the distribution difference and the category explanatory power, a target data subgroup is determined in the historical business data.

[0010] Secondly, this disclosure provides a data analysis apparatus, comprising:

[0011] The acquisition module is used to acquire historical business data and preset analysis indicators. The preset analysis indicators are used to characterize the relationship between target data and target indicators that are expected to be obtained from the historical business data. The target indicators are obtained by dividing a first data indicator by a second data indicator.

[0012] The processing module is used to process the historical business data into binary data under different data dimensions;

[0013] The first determining module is used to determine the distribution difference degree between each type of data in each binary data, and to determine the category explanatory power of each type of data in the historical business data according to the preset analysis index. The category explanatory power is used to characterize the data proportion of each type of data in the historical business data.

[0014] The second determining module is used to determine a target data subgroup in the historical business data based on the distribution difference degree and the category explanatory degree.

[0015] Thirdly, this disclosure provides a computer-readable medium having a computer program stored thereon, which, when executed by a processing device, implements the steps of the method described in the first aspect.

[0016] Fourthly, this disclosure provides an electronic device, comprising:

[0017] A storage device on which computer programs are stored;

[0018] A processing device for executing the computer program in the storage device to implement the steps of the method described in the first aspect.

[0019] The above technical solution can process historical business data under different data dimensions into binary data. Then, the distributional dissimilarity and category explanatory power between each data category can be determined. Based on these dissimilarity and explanatory power, target data subgroups can be identified within the historical business data. This allows for automatic analysis of different data dimensions of each historical business data set, accurately identifying target data subgroups that meet preset analysis indicators and improving data analysis efficiency.

[0020] Other features and advantages of this disclosure will be described in detail in the following detailed description section. Attached Figure Description

[0021] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and the originals and elements are not necessarily drawn to scale. In the drawings:

[0022] Figure 1 This is a flowchart of a data analysis method provided according to an exemplary embodiment of the present disclosure;

[0023] Figure 2 This is a schematic diagram of a target data decision tree provided according to an exemplary embodiment of the present disclosure;

[0024] Figure 3 This is a flowchart of another data analysis method provided according to an exemplary embodiment of the present disclosure;

[0025] Figure 4 This is a block diagram of a data analysis apparatus provided according to an exemplary embodiment of the present disclosure;

[0026] Figure 5 This is a schematic diagram of the structure of an electronic device provided according to an exemplary embodiment of the present disclosure. Detailed Implementation

[0027] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0028] It should be understood that the steps described in the method embodiments of this disclosure may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of this disclosure is not limited in this respect.

[0029] The term "comprising" and its variations as used herein are open-ended inclusions, meaning "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". Definitions of other terms will be given in the description below.

[0030] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0031] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0032] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0033] It is understood that before using the technical solutions disclosed in the various embodiments of this disclosure, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in this disclosure in an appropriate manner in accordance with relevant laws and regulations, and user authorization should be obtained.

[0034] For example, upon receiving a user's active request, a prompt message is sent to the user to explicitly inform them that the requested operation will require the acquisition and use of the user's personal information. This allows the user to independently choose whether to provide personal information to the software or hardware, such as the electronic device, application, server, or storage medium performing the operations of this disclosed technical solution, based on the prompt message.

[0035] As an optional but non-limiting implementation, in response to a user's active request, sending a prompt message to the user can be done via a pop-up window, where the prompt message can be presented in text format. Furthermore, the pop-up window can also include a selection control allowing the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0036] It is understood that the above notification and user authorization process are merely illustrative and do not constitute a limitation on the implementation of this disclosure. Other methods that comply with relevant laws and regulations may also be applied to the implementation of this disclosure.

[0037] Meanwhile, it is understood that the data involved in this technical solution (including but not limited to the data itself, the acquisition or use of the data) shall comply with the requirements of relevant laws, regulations and related provisions.

[0038] As mentioned in the background section, in real-world business scenarios, it's often necessary to break down and analyze historical business data dimensionally to identify subgroups (people / things) that meet certain target metrics, thereby enabling the implementation of specific business strategies for these subgroups. For example, in marketing and promotion, historical business data can be dimensionally broken down to target specific audiences for targeted advertising. Similarly, in product experience optimization, historical business data can be dimensionally broken down and analyzed to identify target audiences and optimize user experience.

[0039] In related technologies, manual decomposition analysis is generally used to perform dimensional decomposition analysis on historical business data to locate subgroups that meet certain target indicators. However, due to the large volume, complex structure, and numerous dimensions of historical business data, manual decomposition analysis for subgroup location is inefficient, time-consuming, and struggles to find the optimal subgroup by exhaustively combining different dimensions. Alternatively, gain model analysis is also used for dimensional decomposition analysis. However, gain model analysis is costly and complex, and is typically used for subgroup location of complex target indicators, thus limiting its application scenarios.

[0040] In view of the above, embodiments of this disclosure provide a data analysis method, apparatus, computer-readable medium, and electronic device to solve the aforementioned technical problems.

[0041] The embodiments of this disclosure will be further explained below with reference to the accompanying drawings.

[0042] Figure 1 This is a flowchart illustrating a data analysis method according to an exemplary embodiment of the present disclosure, with reference to... Figure 1 The method may include the following steps:

[0043] S101: Obtain historical business data and preset analysis indicators. The preset analysis indicators are used to characterize the relationship between the target data and the target indicators that are expected to be obtained from the historical business data. The target indicators are obtained by dividing the first data indicator by the second data indicator.

[0044] It should be understood that, with user authorization, historical business data can be obtained for the execution of the data analysis methods disclosed herein. The method for obtaining user authorization is as described above and will not be repeated here.

[0045] Furthermore, it should be understood that the relationship between target data and target metrics can be determined based on the actual business scenario, and this disclosure does not impose any limitations on this. In possible implementations, the relationship between target data and target metrics can be a positive or negative correlation. Additionally, target metrics can also be determined based on the actual business scenario, and this disclosure does not impose any limitations on this. In possible implementations, target metrics can be average click-through rate or average spending per user.

[0046] S102: Process historical business data into binary data under different data dimensions.

[0047] It should be understood that any piece of business data can include multiple data dimensions, and each data dimension can be used to characterize a feature of the business data or to represent a factor that has an impact on the business data. For example, the data dimensions corresponding to certain historical business data may include age, geographical location, etc.

[0048] It should also be understood that each data dimension can contain multiple dimension elements, and the number of dimension elements contained in different data dimensions may be the same or different. To avoid analysis errors caused by the different number of dimension elements contained in different data dimensions, this embodiment of the disclosure performs binary classification processing on historical business data to ensure that the number of dimension elements under each data dimension is consistent, thereby avoiding analysis errors caused by the different number of dimension elements contained in different data dimensions. Here, dimension elements are used to represent the element category under each data dimension. For example, when the data dimension is age, its corresponding dimension elements can be different ages, such as 20 years old, 30 years old, 40 years old, and 50 years old, etc.

[0049] In a possible implementation, processing historical business data into binary classification data under different data dimensions may include:

[0050] For each data dimension corresponding to historical business data, determine the type of data dimension, and based on the type of data dimension, determine the target partitioning information corresponding to the data dimension; based on the target partitioning information, process the historical business data into binary data from different data dimensions.

[0051] It should be understood that, based on the data type of the data elements within a data dimension, data dimensions can be divided into two types: qualitative dimensions and quantitative dimensions. Specifically, data dimensions where the data element's data type is character or text are qualitative dimensions, such as geographical location. Data dimensions where the data element's data type is numeric are quantitative dimensions, such as age. Therefore, in possible implementations, the qualitative or quantitative nature of a data dimension can be determined by identifying the data type of the data elements within it.

[0052] Once the type of data dimension is determined, the corresponding target partitioning information can be determined based on the type of data dimension, and then binary classification processing can be performed based on the target partitioning information.

[0053] In a possible implementation, determining the target partitioning information corresponding to the data dimension based on the data dimension type may include:

[0054] Based on the type of data dimension, determine the candidate partitioning information corresponding to the data dimension; determine the distribution difference between each type of data after dividing the historical business data according to the candidate partitioning information; and determine the target partitioning information from the candidate partitioning information based on the distribution difference between each type of data after dividing according to the candidate partitioning information.

[0055] The candidate partitioning information can be set according to actual conditions, and this embodiment does not impose any restrictions on it. For example, the candidate partitioning information corresponding to the data dimension can be that each dimension element under the data dimension is used as a partitioning information. Alternatively, the dimension elements under the data dimension can be filtered based on preset selection rules to obtain multiple target dimension elements, and then each target dimension can be used as a partitioning information.

[0056] It should be understood that the greater the distribution difference between each class of data after partitioning according to the candidate partitioning information, the greater the discriminative power between the data, and the more interpretable the determined data subgroup is. Therefore, in this embodiment of the disclosure, the accuracy of the analysis of the target data subgroup can be improved by comparing the distribution difference between each class of data after partitioning according to the candidate partitioning information and determining the candidate partitioning information that maximizes the distribution difference between each class of data as the target partitioning information.

[0057] In a possible implementation, determining the candidate partitioning information corresponding to the data dimension based on the data dimension type may include:

[0058] When the data dimension is a quantitative dimension, each element of the data dimension is used as a partitioning information to obtain the candidate partitioning information corresponding to the data dimension.

[0059] For example, when the data dimension is age, and the included dimension elements are 20, 21, and 22 years old, then 20 years old can be used as the first dividing point to obtain the first candidate dividing information as: less than or equal to 20 years old and greater than 20 years old. Then, 21 years old can be used as the second dividing point to obtain the second candidate dividing information as: less than or equal to 21 years old and greater than 21 years old. Finally, 22 years old can be used as the third dividing point to obtain the third candidate dividing information as: less than or equal to 22 years old and greater than 22 years old. Thus, multiple candidate dividing information can be obtained for the data dimension of age.

[0060] In a possible implementation, determining the candidate partitioning information corresponding to the data dimension based on the data dimension type may include:

[0061] When the data dimension is a qualitative dimension, the explanatory power of each dimension element in the historical business data is determined according to the preset analysis indicators. The dimension elements under the data dimension are then sorted according to the explanatory power to obtain the sorted dimension elements. The explanatory power is used to represent the proportion of the dimension element's data in the historical business data. Then, a partitioning information is determined for every two dimension elements in the sorted dimension elements to obtain the candidate partitioning information corresponding to the data dimension.

[0062] As mentioned earlier, preset analysis indicators are used to characterize the relationship between target data and target indicators expected to be obtained from historical business data. This relationship can be positively or negatively correlated. Therefore, in a possible implementation, when the relationship between target data and target indicators is positively correlated, it means that the target data is historical business data that maximizes the target indicator. Since the target indicator is obtained by dividing the first data indicator by the second data indicator, when the relationship is positively correlated, it is necessary to find a target data subset from the historical business data that has a larger first data indicator value and a larger data volume. In other words, when the relationship between target data and target indicators is positively correlated, the explanatory power of an element can be determined by calculating the proportion of the first data indicator value corresponding to each dimension element in the first data indicator values ​​of the historical business data. That is, the explanatory power of an element can be determined by the proportion of the numerator indicator as follows:

[0063]

[0064] Among them, EP ij a represents the element interpretability of the j-th dimension element in data dimension i. ij S1 represents the first data indicator value corresponding to the j-th dimension element in data dimension i, and S1 represents the first data indicator value corresponding to historical business data.

[0065] For example, if the target metric is click-through rate per user, the first data metric is click volume, and the second data metric is the number of users, and the target data obtained from historical business data is expected to have a positive correlation with the target metric, then according to the above formula, the explanatory power of a certain dimension element can be obtained by dividing the click volume corresponding to that dimension element by the total click volume corresponding to the historical business data.

[0066] When the target data and target metric are negatively correlated, it means the target data is historical business data that minimizes the target metric. Since the target metric is obtained by dividing the first data metric by the second data metric, when the relationship is negative, it's necessary to find a target data subset from the historical business data that has a larger second data metric value and a larger data volume. In other words, when the relationship is negatively correlated, the explanatory power of an element can be determined by calculating the proportion of the second data metric value corresponding to each dimension element in the historical business data. Specifically, the explanatory power of an element can be determined by the proportion of the denominator metric as follows:

[0067]

[0068] Among them, EP ij b represents the element interpretability of the j-th dimension element in data dimension i. ij S1 represents the second data indicator value corresponding to the j-th dimension element in data dimension i, and S2 represents the second data indicator value corresponding to historical business data.

[0069] After determining the explanatory power of each element in the historical business data, the data can be sorted according to the explanatory power to obtain the sorted dimension elements. Then, a partitioning information is determined for every two elements in the sorted dimension to obtain the candidate partitioning information corresponding to the data dimension.

[0070] For example, when the data dimension is geographical location, and the sorting dimension elements are: area A, area D, area B, and area C, a partitioning information is determined sequentially based on every two dimension elements in the sorting dimension. The candidate partitioning information corresponding to the data dimension may include:

[0071] First, using area A as the first dividing point, the first candidate dividing information is: {Area A} and {Areas D, B, C}. Next, using area D as the second dividing point, the second candidate dividing information is: {Area A, D} and {Areas B, C}. Then, using area B as the third dividing point, the third candidate dividing information is: {Area A, D, B} and {Area C}. Thus, dividing information is determined sequentially based on every two elements of the sorting dimensions, resulting in multiple candidate dividing information.

[0072] In possible implementations, to avoid identifying some sparse dimensional elements based on element interpretation degree, the data analysis method may also include:

[0073] Based on the first data indicator, determine the proportion of the numerator indicator of each dimension element in the historical business data, and based on the second data indicator, determine the proportion of the denominator indicator of each dimension element in the historical business data; according to the preset analysis indicators, the proportion of the numerator indicator and the proportion of the denominator indicator, determine the indicator difference value corresponding to each dimension element, and according to the indicator difference value and the element explanatory power, filter the dimension elements under the data dimension to obtain the target dimension element.

[0074] Accordingly, sorting the dimension elements under the data dimension according to their explanatory power can include:

[0075] Sort the target dimension elements according to their interpretability.

[0076] It should be understood that the indicator difference value refers to the difference between two data indicators, which is usually determined by the ratio between the two data indicators. For example, if one data indicator is A and the other data indicator is B, then the indicator difference value of data indicator A relative to data indicator B can be expressed as A / B.

[0077] Furthermore, it should be understood that the preset analytical indicators are used to characterize the relationship between the target data and the target indicators expected to be obtained from historical business data. This relationship can be positive or negative. When the relationship between the target data and the target indicator is positively correlated, it means that the target data is historical business data that maximizes the target indicator. Since the target indicator is obtained by dividing the first data indicator by the second data indicator, the indicator difference value corresponding to each dimension element can be determined by calculating the ratio of the numerator indicator's proportion to the denominator indicator's proportion. That is, the indicator difference value can be determined by the following formula:

[0078]

[0079]

[0080]

[0081] Among them, TGI ij q represents the difference value of the index of the j-th dimension element in data dimension i. ij p represents the proportion of the molecular index of the j-th dimension element in data dimension i. ij b represents the proportion of the denominator index of the j-th dimension element in data dimension j. ij a represents the second data metric value corresponding to the j-th dimension element in data dimension i. ijS1 represents the first data indicator value corresponding to the j-th dimension element in data dimension i, S2 represents the first data indicator value corresponding to historical business data, and S2 represents the second data indicator value corresponding to historical business data.

[0082] Similarly, when the target data and target indicators are negatively correlated, the difference value of the indicator corresponding to each dimension element can be determined by calculating the ratio of the proportion of the indicator in the denominator to the proportion of the indicator in the numerator. That is, the indicator difference value can be determined by the following formula:

[0083]

[0084]

[0085]

[0086] Among them, TGI ij q represents the difference value of the index of the j-th dimension element in data dimension i. ij p represents the proportion of the molecular index of the j-th dimension element in data dimension i. ij b represents the proportion of the denominator index of the j-th dimension element in data dimension i. ij a represents the value of the second data indicator corresponding to the j-th dimension element in data dimension i of the second data indicator. ij S1 represents the first data indicator value corresponding to the j-th dimension element in data dimension i, S2 represents the first data indicator value corresponding to historical business data, and S2 represents the second data indicator value corresponding to historical business data.

[0087] S103: Determine the distribution difference between each category of data in each binary classification data, and determine the category explanatory power of each category of data in the historical business data according to the preset analysis indicators. The category explanatory power is used to characterize the proportion of each category of data in the historical business data.

[0088] In a possible implementation, determining the degree of distributional dissimilarity between each class of data in each binary classification dataset may include:

[0089] Determine the JS divergence between each class of data in each binary classification dataset; based on the JS divergence, determine the distributional dissimilarity between each class of data in each binary classification dataset.

[0090] It should be understood that the JS (Jenson's Shannon) divergence can be used to represent the difference between two probability distributions. The larger the JS divergence, the greater the difference between the two probability distributions, and the higher the corresponding discriminative power. Furthermore, the JS divergence has a symmetric characteristic, which can improve the interpretability of actual business data. Therefore, in the embodiments of this disclosure, the distributional difference between each class of data in each binary classification can be determined based on the JS divergence.

[0091] For example, the degree of distributional dissimilarity between each class of data in each binary classification dataset can be determined by the following formula:

[0092]

[0093] Among them, jS i This represents the degree of distributional difference between each class of data in the binary classification data obtained based on data dimension i, where j represents the dimension element under data dimension i, and n represents the number of dimension elements under data dimension i.

[0094] S104: Determine the target data subgroup in historical business data based on the distribution difference and category explanatory power.

[0095] In a possible implementation, determining the target data subset from historical business data based on distribution dissimilarity and category explanatory power may include:

[0096] Identify the target data dimension with the largest distributional dissimilarity and a corresponding category explanatory power greater than or equal to a preset threshold; use the calculation results of historical business data under the target indicator as the root node, and use each category of data corresponding to the target data dimension as a child node of the root node to generate a target data decision tree; based on the target data decision tree, determine the target data subgroup in the historical business data.

[0097] It should be understood that the greater the distributional difference between data dimensions, the greater the distinguishability between the data dimensions, and the more interpretable the target data subgroup determined thereby. Therefore, based on the distributional difference between each data dimension, the target data dimension with the largest distributional difference can be found, and thus the target data subgroup can be determined from historical business data based on the target data dimension. Meanwhile, to avoid a small number of target data subgroups determined based on data dimensions, this disclosure also introduces category interpretability, combining distributional difference and category interpretability to obtain a target data subgroup that meets the requirements. The preset threshold can be set according to actual conditions, and this disclosure does not impose any restrictions on it.

[0098] It should also be understood that a decision tree represents the results of data classification in a tree structure. It can include a root node and child nodes. The root node can be used to represent the features or attributes of the data, and the child nodes can be used to represent the categories corresponding to the root node. By constructing a decision tree using the features (or attributes) of the data and their corresponding categories, the data classification results can be visually observed. Therefore, this embodiment of the disclosure generates a target data decision tree by using the calculation results of historical business data under the target indicator as the root node and each category of data corresponding to the target data dimension as a child node of the root node. This enables the visualization of obtaining target data subgroups from historical business data.

[0099] In a possible implementation, the calculation results of historical business data under the target metric are used as the root node, and each type of data corresponding to the target data dimension is used as a child node of the root node to generate a target data decision tree, which may include:

[0100] Using the calculation results of historical business data under the target metric as the root node, and each data category corresponding to the target data dimension as a child node of the root node, repeat the following steps to obtain the target data decision tree:

[0101] Using child nodes as parent nodes and the data corresponding to the child nodes as target data, the target data is processed into target binary classification data under different data dimensions. The target distribution difference between each class of data in each target binary classification data is determined. Based on preset analysis indicators, the target category explanatory power of each class of data in each target binary classification data in the target data is determined. A new target data dimension is determined with the largest target distribution difference and the corresponding target category explanatory power greater than or equal to a preset threshold. Each class of data corresponding to the new target data dimension is used as a child node of the parent node, until a preset stopping condition is reached.

[0102] For example, the preset stopping conditions include at least one of the following: reaching a set tree depth, the interpretability of a node being less than a first set threshold, the distribution difference between data of nodes at the same level being less than a second set threshold, and there being no more data dimensions to divide. This embodiment of the disclosure does not limit these conditions. The first set threshold and the second set threshold are different.

[0103] It should be understood that the target data dimension contains different dimensional elements, and each dimensional element corresponds to different historical business data. Furthermore, these different business data are related to new and different dimensional elements. Therefore, in order to facilitate the intuitive representation of the target data subgroups based on the target decision tree, this embodiment can recursively execute the above process when generating the target data decision tree to obtain the target data decision tree.

[0104] In possible implementations, the data analysis method may further include:

[0105] Based on the data subgroup corresponding to each node in the target data decision tree, determine at least one of the following data indicators: the data proportion corresponding to the data subgroup, the distribution difference between the data subgroup and other data subgroups corresponding to the same level node, and the calculation result of the data subgroup under the target indicator; output display the target data decision tree, wherein each node in the target data decision tree is associated with at least one data indicator.

[0106] The data percentage corresponding to the data subgroup can be set according to actual conditions, and this disclosure does not impose any restrictions on this. In possible implementations, the data percentage corresponding to the data subgroup can be a numerator index percentage or a denominator index percentage.

[0107] For example, refer to Figure 2 The root node of the target data decision tree can display the calculation result T of historical business data under the target metric. The target data dimension has dimension elements i1 and i2, so the root node can be divided into two child nodes. One child node contains the historical business data corresponding to dimension element i1. Figure 2 The symbol 'a' represents the element; another child node contains the historical business data corresponding to dimension element i2. Figure 2 The diagram uses 'b' for illustration. Furthermore, following the above method, child node 'a' can be used as the parent node. The business data within child node 'a' can then be further split into two new child nodes according to the new target data dimension. Each new child node contains the historical business data corresponding to dimension element j1 under the new target data dimension. Figure 2 The diagram uses 'c' for illustration; another new child node contains the historical business data corresponding to dimension element j2 under the new target data dimension. Figure 2 The symbol 'd' is used to represent the character.

[0108] Furthermore, to demonstrate the criteria for identifying the target data subgroup, the values ​​of corresponding defining indicators can be displayed at each sub-node, such as the values ​​of analytical indicators, category explanatory power, distribution variability, numerator indicator proportions, and denominator indicator ratios. For example, Figure 2 Each child node displays the following metric values: target_indicator, ep, js, observed_rate, and control_rate. Here, target_indicator represents the target metric value, ep represents the category explanatory power, js represents the distribution variance, observed_rate represents the proportion of the numerator metric, and control_rate represents the proportion of the denominator metric. It should be understood that... Figure 2 The specific values ​​of each indicator are for illustrative purposes only and are not intended to limit this disclosure.

[0109] By employing any of the aforementioned data analysis methods, historical business data across different data dimensions can be processed into binary classification data. This allows for the determination of the distributional dissimilarity and category explanatory power between each category, and further, the identification of target data subgroups within the historical business data based on these factors. Thus, by automatically analyzing different data dimensions of each historical business data set, target data subgroups meeting preset analytical indicators can be accurately identified from various data dimensions. Furthermore, a decision tree can be output and displayed, visualizing the analysis process and results for easy viewing.

[0110] The data analysis method provided in this disclosure will now be described through another exemplary embodiment.

[0111] Reference Figure 3 This data analysis method may include:

[0112] Step 1: Obtain historical business data;

[0113] Step 2: Data cleaning;

[0114] It should be understood that some invalid historical business data may exist. To avoid the invalid historical business data affecting the accuracy of data filtering, data cleaning can be performed on the historical business data to remove invalid data. The data cleaning of historical business data can be carried out according to the actual situation. For example, null values ​​in historical business data can be deleted, or values ​​can be assigned to null values, etc. This disclosure does not impose any restrictions on this. Furthermore, to facilitate the differentiation of different types of input indicators, input indicators of the same type can be uniformly named. The naming method can be set according to the actual situation, and this disclosure does not impose any restrictions on this. For example, the numerator indicator can be named observed_value, and the denominator indicator can be named control_value.

[0115] Step 3: Configure decision tree parameters;

[0116] The decision tree parameters may include the maximum depth of the decision tree, stopping conditions, and analysis metrics. The values ​​of the decision tree parameters can be set according to actual conditions, and this embodiment does not impose any restrictions on them. In a possible implementation, the maximum depth of the decision tree can be set to 4, the stopping condition can be set to an explanatory power of 0.05, and the preset analysis metrics can be set to a positive correlation between the target data and the target metric.

[0117] Step 4: Binary classification;

[0118] It should be understood that any business data may include multiple data dimensions, and each data dimension may contain multiple dimension elements. The number of dimension elements contained in different data dimensions may be the same or different. To avoid analysis errors caused by the different number of dimension elements contained in different data dimensions, historical business data can be binary classified to ensure that the number of dimension elements under each data dimension is consistent, thereby avoiding analysis errors caused by the different number of dimension elements contained in different data dimensions. In addition, according to the data type of the data elements under the data dimension, data dimensions can be divided into two types: qualitative dimensions and quantitative dimensions. Therefore, when performing binary classification, the type of data dimension can be determined first, and then different binary classification methods can be adopted according to different data dimension types. For details on adopting different binary classification methods for different data dimension types, please refer to the relevant description above, which will not be repeated in this embodiment.

[0119] By processing historical business data under different data dimensions into binary data, the distribution difference and category explanatory power between each type of data can be determined based on the binary data. Then, the target data subgroup can be determined in the historical business data based on the distribution difference and category explanatory power between each type of data.

[0120] Step 5: Construct a decision tree;

[0121] It should be understood that a decision tree represents the results of data classification in a tree structure, and the data classification results can be intuitively seen from the decision tree. Therefore, in order to intuitively demonstrate the determination of the target data subgroup from historical business data, this embodiment of the present disclosure can also use the calculation results of historical business data under the target indicator as the root node, and each type of data corresponding to the target data dimension as the child node of the root node, and generate the target data decision tree based on the decision tree parameters configured in step 3 and output it externally.

[0122] The specific implementation methods for each of the above steps have been described in detail above and will not be repeated here. It should also be understood that, for the sake of simplicity, the above method embodiments are described as a series of actions; however, those skilled in the art should understand that this disclosure is not limited to the order of actions described above. Furthermore, those skilled in the art should also understand that the embodiments described above are preferred embodiments, and the steps involved are not necessarily essential to this disclosure.

[0123] By employing the methods described above, different data dimensions of each historical business dataset are automatically analyzed, thereby accurately identifying target data subgroups that meet preset analysis indicators from these dimensions. Furthermore, a decision tree can be output, visualizing the analysis process and results for easy viewing.

[0124] Based on the same concept, embodiments of this disclosure also provide a data analysis apparatus, such as... Figure 4 As shown, the data analysis device 400 may include:

[0125] The acquisition module 401 is used to acquire historical business data and preset analysis indicators. The preset analysis indicators are used to characterize the relationship between target data and target indicators that are expected to be obtained from the historical business data. The target indicators are obtained by dividing a first data indicator by a second data indicator.

[0126] Processing module 402 is used to process the historical business data into binary data under different data dimensions;

[0127] The first determining module 403 is used to determine the distribution difference degree between each type of data in each binary data, and to determine the category explanatory power of each type of data in the historical business data according to the preset analysis index. The category explanatory power is used to characterize the data proportion of each type of data in the historical business data.

[0128] The second determining module 404 is used to determine a target data subgroup from the data in the historical business based on the distribution difference degree and the category explanatory degree.

[0129] In a possible implementation, the processing module 402 may include:

[0130] The first determining submodule is used to determine the type of the data dimension for each data dimension corresponding to the historical business data, and to determine the target partitioning information corresponding to the data dimension based on the type of the data dimension.

[0131] The processing submodule is used to process the historical business data into binary data from different data dimensions based on the target segmentation information.

[0132] In a possible implementation, the first determining submodule may include:

[0133] The first determining unit is used to determine the candidate partitioning information corresponding to the data dimension according to the type of the data dimension;

[0134] The second determining unit is used to determine the distribution difference degree between each type of data after the historical business data is divided according to the candidate division information;

[0135] The third determining unit is used to determine the target partitioning information from the candidate partitioning information based on the distribution difference degree between each type of data after partitioning according to the candidate partitioning information.

[0136] In possible implementations, the first determining unit may include:

[0137] The first determining subunit is used to, when the data dimension is a quantitative dimension type, treat each dimension element under the data dimension as a partitioning information to obtain the candidate partitioning information corresponding to the data dimension.

[0138] In possible implementations, the first determining unit may include:

[0139] The sorting subunit is used to determine the element explanatory power of each dimension element in the historical business data according to the preset analysis indicators when the data dimension type is a qualitative dimension type, and to sort the dimension elements in the data dimension according to the element explanatory power to obtain sorted dimension elements, wherein the element explanatory power is used to characterize the proportion of the data of the dimension element in the historical business data.

[0140] The second determining subunit is used to determine a partitioning information according to every two of the sorting dimension elements in turn, so as to obtain the candidate partitioning information corresponding to the data dimension.

[0141] In a possible implementation, the data analysis device 400 may further include:

[0142] The third determining module is used to determine the proportion of the data of each dimension element in the historical business data based on the first data indicator, and to determine the proportion of the data of each dimension element in the historical business data based on the second data indicator.

[0143] The filtering module is used to determine the index difference value corresponding to each dimension element based on the preset analysis index, the proportion of the numerator index and the proportion of the denominator index, and to filter the dimension elements under the data dimension based on the index difference value and the element explanatory power to obtain the target dimension element.

[0144] Accordingly, the sorting subunit can be used to sort the target dimension elements according to the element interpretation degree.

[0145] In a possible implementation, the first determining module 403 may include:

[0146] The second determining submodule is used to determine the JS divergence between each class of data in each of the binary classification data;

[0147] The third determining submodule is used to determine the distribution difference between each class of data in each binary classification data based on the JS divergence.

[0148] In a possible implementation, the second determining module 404 may include:

[0149] The fourth determination submodule is used to determine the target data dimension with the largest corresponding distribution difference and the corresponding category explanatory power greater than or equal to a preset threshold;

[0150] The decision tree generation submodule is used to take the calculation result of the historical business data under the target indicator as the root node, and take each type of data corresponding to the target data dimension as the child node of the root node to generate the target data decision tree.

[0151] The fifth determination submodule is used to determine the target data subgroup in the historical business data based on the target data decision tree.

[0152] In a possible implementation, the decision tree generation submodule includes:

[0153] The decision tree generation unit is used to take the calculation results of the historical business data under the target indicator as the root node, and take each type of data corresponding to the target data dimension as the child node of the root node, and repeatedly execute the following: take the child node as the parent node, take the data corresponding to the child node as the target data, process the target data into target binary classification data under different data dimensions, determine the target distribution difference degree between each type of data in each target binary classification data, determine the target category explanatory power of each type of data in each target binary classification data in the target data according to the preset analysis indicator, determine the new target data dimension with the largest target distribution difference degree and the corresponding target category explanatory power greater than or equal to a preset threshold, take each type of data corresponding to the new target data dimension as the child node of the parent node, until a preset stopping condition is reached, and obtain the target data decision tree.

[0154] In a possible implementation, the data analysis device 400 may further include:

[0155] The fourth determining module is used to determine at least one of the following data indicators based on the data subgroup corresponding to each node in the target data decision tree: the data proportion corresponding to the data subgroup, the distribution difference degree between the data subgroup and other data subgroups corresponding to the same level node, and the calculation result of the data subgroup under the target indicator.

[0156] An output module is used to output and display the target data decision tree, wherein each node in the target data decision tree is associated with and displayed with at least one data indicator.

[0157] Based on the same concept, embodiments of this disclosure also provide a computer-readable medium having a computer program stored thereon, which, when executed by a processing device, implements the steps of the above-described data analysis method.

[0158] Based on the same concept, embodiments of this disclosure also provide an electronic device, including:

[0159] A storage device on which computer programs are stored;

[0160] A processing device is used to execute the computer program in the storage device to implement the steps of the above-described data analysis method.

[0161] The following is for reference. Figure 5 The diagram illustrates a structural schematic of an electronic device 500 suitable for implementing embodiments of the present disclosure. The terminal devices in the embodiments of the present disclosure may include, but are not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 5 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments disclosed herein.

[0162] like Figure 5 As shown, the electronic device 500 may include a processing unit (e.g., a central processing unit, a graphics processor, etc.) 501, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage device 508 into a random access memory (RAM) 503. The RAM 503 also stores various programs and data required for the operation of the electronic device 500. The processing unit 501, ROM 502, and RAM 503 are interconnected via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0163] Typically, the following devices can be connected to I / O interface 505: input devices 506 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 507 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 508 including, for example, magnetic tapes, hard disks, etc.; and communication devices 509. Communication device 509 allows electronic device 500 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 5 An electronic device 500 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively.

[0164] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 509, or installed from a storage device 508, or installed from a ROM 502. When the computer program is executed by the processing device 501, it performs the functions defined in the methods of embodiments of this disclosure.

[0165] It should be noted that the computer-readable medium described in this disclosure can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this disclosure, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In this disclosure, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0166] In some implementations, communication can be conducted using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol), and can be interconnected with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and end-to-end networks (e.g., ad hoc end-to-end networks), as well as any currently known or future-developed networks.

[0167] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device.

[0168] The aforementioned computer-readable medium carries one or more programs that, when executed by the electronic device, cause the electronic device to: acquire historical business data and preset analysis indicators, wherein the preset analysis indicators are used to characterize the relationship between target data and target indicators expected to be obtained from the historical business data, and the target indicators are obtained by dividing a first data indicator by a second data indicator; process the historical business data into binary data under different data dimensions; determine the distribution difference degree between each type of data in each of the binary data, and determine the category explanatory power of each type of data in the historical business data according to the preset analysis indicators, wherein the category explanatory power is used to characterize the data proportion of each type of data in the historical business data; and determine a target data subgroup in the historical business data according to the distribution difference degree and the category explanatory power.

[0169] Computer program code for performing the operations of this disclosure can be written in one or more programming languages ​​or a combination thereof, including but not limited to object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0170] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0171] The modules described in the embodiments of this disclosure can be implemented in software or hardware. The names of the modules are not, in some cases, intended to limit the functionality of the module itself.

[0172] The functions described above in this document can be performed, at least in part, by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: Field Programmable Gate Arrays (FPGAs), Application-Specific Integrated Circuits (ASICs), Application Standard Products (ASSPs), System-on-Chip (SoCs), Complex Programmable Logic Devices (CPLDs), and so on.

[0173] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0174] The above description is merely a preferred embodiment of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features disclosed in this disclosure that have similar functions.

[0175] Furthermore, while the operations are described in a specific order, this should not be construed as requiring these operations to be performed in the specific order shown or in a sequential order. In certain environments, multitasking and parallel processing may be advantageous. Similarly, while several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of this disclosure. Certain features described in the context of individual embodiments may also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may also be implemented individually or in any suitable sub-combination in multiple embodiments.

[0176] Although the subject matter has been described using language specific to structural features and / or methodological logic, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. Rather, the specific features and actions described above are merely illustrative forms of implementing the claims. Regarding the apparatus in the above embodiments, the specific manner in which the various modules perform their operations has been described in detail in the embodiments relating to the method, and will not be elaborated upon here.

Claims

1. A data analysis method, characterized in that, The method includes: Acquire historical business data and preset analysis indicators. The preset analysis indicators are used to characterize the relationship between the target data and the target indicators expected to be obtained from the historical business data. The target indicators are obtained by dividing a first data indicator by a second data indicator. The historical business data includes historical business data of marketing and promotion business. The target indicators include average click-through rate or average payment per user. The historical business data is processed into binary classification data under different data dimensions; Determine the JS divergence between each class of data in each binary classification data; based on the JS divergence, determine the distribution difference between each class of data in each binary classification data; and based on the preset analysis index, determine the category explanatory power of each class of data in the historical business data, wherein the category explanatory power is used to characterize the data proportion of each class of data in the historical business data. Determine the target data dimension with the largest distribution difference and the corresponding category explanatory power greater than or equal to a preset threshold; use the calculation result of the historical business data under the target indicator as the root node, and use each type of data corresponding to the target data dimension as a child node of the root node to generate a target data decision tree; determine the target data subgroup in the historical business data according to the target data decision tree.

2. The method according to claim 1, characterized in that, The process of dividing the historical business data into binary categories under different data dimensions includes: For each data dimension corresponding to the historical business data, determine the type of the data dimension, and determine the target partitioning information corresponding to the data dimension based on the type of the data dimension; Based on the target segmentation information, the historical business data is processed into binary data from different data dimensions.

3. The method according to claim 2, characterized in that, The step of determining the target partitioning information corresponding to the data dimension based on the data dimension type includes: Based on the type of the data dimension, determine the candidate partitioning information corresponding to the data dimension; Determine the distribution difference between each category of data after dividing the historical business data according to the candidate partitioning information; Based on the distribution difference between each type of data after being divided according to the candidate partitioning information, the target partitioning information is determined from the candidate partitioning information.

4. The method according to claim 3, characterized in that, The step of determining the candidate partitioning information corresponding to the data dimension based on the data dimension type includes: When the data dimension is a quantitative dimension, each element of the data dimension is used as a partitioning information to obtain the candidate partitioning information corresponding to the data dimension.

5. The method according to claim 3, characterized in that, The step of determining the candidate partitioning information corresponding to the data dimension based on the data dimension type includes: When the data dimension is a qualitative dimension, the explanatory power of each dimension element in the historical business data is determined according to the preset analysis indicators, and the dimension elements are sorted according to the explanatory power to obtain sorted dimension elements. The explanatory power is used to characterize the proportion of the data of the dimension element in the historical business data. By sequentially determining a partitioning information based on every two dimensions in the sorting dimension, candidate partitioning information corresponding to the data dimension is obtained.

6. The method according to claim 5, characterized in that, The method further includes: Based on the first data indicator, determine the proportion of the numerator indicator of the data of each dimension element in the historical business data, and based on the second data indicator, determine the proportion of the denominator indicator of the data of each dimension element in the historical business data. Based on the preset analysis indicators, the proportion of the numerator indicators and the proportion of the denominator indicators, the indicator difference value corresponding to each dimension element is determined, and the dimension elements under the data dimension are filtered according to the indicator difference value and the element explanatory power to obtain the target dimension element. The step of sorting the dimension elements under the data dimension according to the element interpretability includes: The target dimension elements are sorted according to their explanatory power.

7. The method according to claim 1, characterized in that, The step of generating a target data decision tree by using the calculation results of the historical business data under the target indicator as the root node and each type of data corresponding to the target data dimension as a child node of the root node includes: Using the calculation results of the historical business data under the target indicator as the root node, and each type of data corresponding to the target data dimension as a child node of the root node, the following steps are repeated to obtain the target data decision tree: Using the child node as the parent node and the data corresponding to the child node as the target data, the target data is processed into target binary classification data under different data dimensions. The target distribution difference degree between each class of data in each target binary classification data is determined. According to the preset analysis index, the target category explanatory power of each class of data in each target binary classification data in the target data is determined. A new target data dimension is determined with the largest target distribution difference degree and the corresponding target category explanatory power greater than or equal to a preset threshold. Each class of data corresponding to the new target data dimension is used as the child node of the parent node until a preset stopping condition is reached.

8. The method according to claim 1, characterized in that, The method further includes: Based on the data subgroup corresponding to each node in the target data decision tree, at least one of the following data indicators is determined: the data proportion corresponding to the data subgroup, the distribution difference between the data subgroup and other data subgroups corresponding to the same level node, and the calculation result of the data subgroup under the target indicator. The output displays the target data decision tree, wherein each node in the target data decision tree is associated with and displays the at least one data indicator.

9. A data analysis device, characterized in that, include: The acquisition module is used to acquire historical business data and preset analysis indicators. The preset analysis indicators are used to characterize the relationship between target data and target indicators expected to be obtained from the historical business data. The target indicators are obtained by dividing a first data indicator by a second data indicator. The historical business data includes historical business data of marketing and promotion business. The target indicators include average click-through rate or average payment per user. The processing module is used to process the historical business data into binary data under different data dimensions; The first determining module is used to determine the JS divergence between each class of data in each binary classification data; based on the JS divergence, determine the distribution difference between each class of data in each binary classification data; and based on the preset analysis index, determine the category explanatory power of each class of data in the historical business data, wherein the category explanatory power is used to characterize the data proportion of each class of data in the historical business data. The second determining module is used to determine the target data dimension with the largest corresponding distribution difference and the corresponding category explanatory power greater than or equal to a preset threshold. The calculation results of the historical business data under the target indicator are used as the root node, and each type of data corresponding to the target data dimension is used as a child node of the root node to generate a target data decision tree; based on the target data decision tree, a target data subgroup is determined in the historical business data.

10. A computer-readable medium having a computer program stored thereon, characterized in that, When the program is executed by the processing device, it implements the steps of the method described in any one of claims 1-8.

11. An electronic device, characterized in that, include: A storage device on which computer programs are stored; A processing device for executing the computer program in the storage device to implement the steps of the method according to any one of claims 1-8.

Images