Index determination method, service processing method and device and electronic equipment

Abstract

The embodiment of the invention provides an index determination method, a business processing method and device and electronic equipment, and relates to the field of financial science and technology or other related fields. The method comprises the steps that category labels of a plurality of objects are determined, and the category labels are used for indicating that the corresponding objects have handled a first service or indicating that the corresponding objects do not handle the first service; obtaining respective index values of the plurality of objects under the plurality of indexes; according to respective category labels of the plurality of objects and respective index values of the plurality of objects under the plurality of indexes, at least one target index having an association relationship with the object category is determined in the plurality of indexes, and the at least one target index is used for determining the object category of the target object; the object category of the target object is used for indicating whether the target object belongs to an associated object of the first business. According to the scheme, the method for determining the target index in the multiple indexes is achieved.

Description

Technical Field

[0001] The present application relates to the field of financial technology or other related fields, and in particular to an indicator determination method, a business processing method, a device and an electronic device. Background Art

[0002] In some business scenarios, it is necessary to determine the object category of an object, where the object category refers to whether the object belongs to an associated object of the business.

[0003] At present, the object category of an object can be determined by obtaining the indicator value of the object under at least one evaluation indicator. However, for any business scenario, the object has multiple indicators under the business scenario, and it is necessary to determine at least one evaluation indicator among the multiple indicators, and then obtain the indicator value of the object under the at least one evaluation indicator to determine the object category of the object. Therefore, how to determine the evaluation indicator for object category classification among multiple indicators is a technical problem that needs to be solved urgently. Summary of the invention

[0004] The embodiments of the present application provide an indicator determination method, a business processing method, an apparatus and an electronic device to implement a method for determining an evaluation indicator among multiple indicators.

[0005] In a first aspect, an embodiment of the present application provides an indicator determination method, the method comprising:

[0006] Determine a category label of each of the plurality of objects, the category label being used to indicate that the corresponding object has handled the first service, or to indicate that the corresponding object has not handled the first service;

[0007] Get the indicator values ​​of multiple objects under multiple indicators;

[0008] According to respective category labels of the multiple objects and respective indicator values ​​of the multiple objects under multiple indicators, at least one target indicator associated with the object category is determined among the multiple indicators, the at least one target indicator is used to determine the object category of the target object, and the object category of the target object is used to indicate whether the target object belongs to an associated object of the first business.

[0009] In a second aspect, an embodiment of the present application provides a service processing method, the method comprising:

[0010] Obtain the indicator value of each target object under at least one target indicator;

[0011] Determining an object category of the target object according to respective indicator values ​​of the target object under at least one target indicator; the object category is used to indicate whether the target object belongs to an associated object of the first business;

[0012] Determining a first push frequency according to an object category of the target object;

[0013] The service information of the first service is sent to the client according to the first push frequency.

[0014] In a third aspect, an embodiment of the present application provides an indicator determination device, the device comprising a first determination module, a first acquisition module and a first processing module, wherein

[0015] A first determination module is used to determine a category label of each of the plurality of objects, where the category label is used to indicate that the corresponding object has handled the first service, or indicates that the corresponding object has not handled the first service;

[0016] A first acquisition module is used to acquire the index value of each of the multiple objects under multiple indicators;

[0017] The first processing module is used to determine at least one target indicator having an association relationship with the object category among the multiple indicators based on the category labels of the multiple objects and the indicator values ​​of the multiple objects under the multiple indicators, wherein the at least one target indicator is used to determine the object category of the target object, and the object category of the target object is used to indicate whether the target object belongs to an associated object of the first business.

[0018] In a fourth aspect, an embodiment of the present application provides a service processing device, the device comprising a second acquisition module, a second processing module, a second determination module and a sending module, wherein:

[0019] A second acquisition module is used to acquire the respective indicator values ​​of the target objects under at least one target indicator;

[0020] A second processing module, configured to determine an object category of the target object according to respective indicator values ​​of the target object under at least one target indicator; the object category is used to indicate whether the target object belongs to an associated object of the first business;

[0021] A second determination module, configured to determine a first push frequency according to an object category of a target object;

[0022] The sending module is used to send the service information of the first service to the client according to the first push frequency.

[0023] In a fifth aspect, an embodiment of the present application provides an electronic device, including: a processor, and a memory communicatively connected to the processor;

[0024] Memory stores computer-executable instructions;

[0025] The processor executes the computer-executable instructions stored in the memory to implement the indicator determination method involved in the first aspect and any possible implementation method, and the business processing method involved in the second aspect.

[0026] In a sixth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer execution instructions, which, when executed by a processor, are used to implement the indicator determination method involved in the first aspect and any possible implementation method, as well as the business processing method involved in the second aspect.

[0027] In the seventh aspect, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the indicator determination method involved in the first aspect and any possible implementation method, and the business processing method involved in the second aspect.

[0028] The indicator determination method, business processing method, device and electronic device provided in the embodiments of the present application can determine at least one target indicator that has an association relationship with the object category among multiple indicators by analyzing the category labels of each of the multiple objects and the indicator values ​​of each of the multiple objects under multiple indicators. In the scheme of the present application, the indicator values ​​of each of the multiple objects under multiple indicators are analyzed in combination with the category labels of each of the multiple objects, and at least one target indicator is determined among the multiple indicators, and any one of the at least one target indicator determined has an association relationship with the object category. Therefore, the scheme of the present application realizes a method of determining a target indicator among multiple indicators, so that the object category of the target object can be determined according to at least one target indicator. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0030] Figure 1 A schematic diagram of applicable scenarios provided for embodiments of the present application;

[0031] Figure 2 A flowchart of an indicator determination method provided in an embodiment of the present application;

[0032] Figure 3 A schematic diagram of a process for determining at least one target object provided in an embodiment of the present application;

[0033] Figure 4 A schematic diagram of a cause-effect diagram provided in an embodiment of the present application;

[0034] Figure 5 A schematic diagram of a causal graph in which a node satisfies a front-door criterion provided in an embodiment of the present application;

[0035] Figure 6 A schematic diagram of a cause-and-effect diagram that satisfies front door adjustment provided in an embodiment of the present application;

[0036] Figure 7 A flowchart of a business processing method provided in an embodiment of the present application;

[0037] Figure 8 A schematic diagram of the structure of an indicator determination device provided in an embodiment of the present application;

[0038] Fig. 9 A schematic diagram of the structure of a service processing device provided in an embodiment of the present application;

[0039] Fig.10 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0040] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0041] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0042] 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, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of the relevant data comply with the relevant laws, regulations and standards of the relevant countries and regions, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0043] In addition, this application involves big data analysis of user information (including but not limited to personal biometrics, identity data, consumption data, asset data, electronic terminal operation data, etc.), and the use of artificial intelligence technology for automated decision-making, and a technical solution for making decisions that have a significant impact on personal rights and interests based on the results of automated decision-making. The application provides users with corresponding operation entrances for them to choose to agree or reject the results of automated decision-making; if the user chooses to reject, the expert decision-making process will be entered.

[0044] It should be noted that the indicator determination method, business processing method, device and electronic equipment provided in this application can be used in the field of financial technology or other related fields, and can also be used in any field other than the field of financial technology or other related fields. The application field of the indicator determination method, business processing method, device and electronic equipment in this application is not limited.

[0045] It is crucial to analyze whether the object is an associated object of the first service according to the indicator values ​​of the object under multiple indicators. For example, if the object is an associated object of the first service, it means that the object is willing to handle the first service, or the object has a need to use or handle the first service, etc.; if the object is not an associated object of the first service, it means that the object is not willing to handle the first service, or the object has no need to use or handle the first service, etc. For the two different situations of the object being an associated object of the first service and the object not being an associated object of the first service, the first service can be pushed to the object in different ways, thereby improving the user stickiness and activity of the object.

[0046] In the related art, the object category of the target object can be determined by statistical analysis methods such as mean, median, standard deviation, etc. However, the statistical analysis method focuses on the analysis of the central tendency and dispersion degree between the indicator values ​​of the object under multiple indicators, ignoring the complex correlation and potential correlation between the indicator values ​​of multiple objects under multiple indicators and the category labels of multiple objects, limiting the ability to explain the behavior of the object through the indicator values ​​of each object under multiple indicators, and to determine at least one target indicator that has an association relationship with the category label of the object. In addition, the object category of the target object can also be determined by a logistic regression model. However, the logistic regression model has certain limitations when processing data with nonlinear relationships or high-dimensional data, and the generalization ability of the logistic regression model is low. In summary, in order to improve the accuracy of determining the object category, the target indicator can be determined among multiple indicators, and the object category can be determined based on the target indicator.

[0047] In response to the above technical problems, in an embodiment of the present application, by analyzing the indicator values ​​of multiple objects under multiple indicators and the association relationship between the category labels of multiple objects, at least one target indicator that has an association relationship with the object category of the object is determined from the multiple indicators, thereby achieving determination of the target indicator among the multiple indicators.

[0048] For ease of understanding, the following Figure 1 , briefly describe the application scenarios to which the embodiments of the present application are applicable.

[0049] Figure 1 A schematic diagram of an applicable scenario provided for an embodiment of the present application, such as Figure 1As shown, it includes a client 11 and a server 12. The client 11 is installed with an indicator determination system. The object can log in to the platform of the indicator determination system through the client 11. The client 11 can be a terminal device such as a mobile phone, a computer, or a tablet. The server 12 can be a background server of the indicator determination system. The indicator determination system can be a system for determining at least one target indicator having an association relationship with the object category from among multiple indicators.

[0050] In actual application, information exchange may be performed between the client 11 and the server 12. For example, the client 11 may send the index values ​​of multiple objects under multiple indexes to the server 12.

[0051] It should be noted that Figure 1 It is just an example to illustrate a system architecture diagram, and it is not a limitation of the system architecture diagram.

[0052] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0053] Figure 2 A flowchart of an indicator determination method provided in an embodiment of the present application. Figure 2 , the method may include the following steps:

[0054] S21. Determine a category label of each of the plurality of objects, where the category label is used to indicate that the corresponding object has handled the first service, or indicates that the corresponding object has not handled the first service.

[0055] The execution subject of the embodiment of the present application can be a background server of the indicator determination system, or can be an indicator determination device set in the server. The indicator determination device can be implemented by software, or by a combination of software and hardware.

[0056] The category labels of multiple objects can be represented by numbers. For example, the object category label can be represented as 1 or 2. When the object category label of the object is 1, it indicates that the object has handled the first business; when the object category label of the object is 2, it indicates that the object has not handled the first business, or when the object category label of the object is 1, it indicates that the object has not handled the first business, and when the object category label of the object is 2, it indicates that the object has handled the first business. The first business can be various types of activities or projects. For example, the first business can be an application for a financial product. In the embodiment of the present application, the first business is to apply for a financial product as an example for explanation.

[0057] S22. Obtain indicator values ​​of each of the multiple objects under multiple indicators.

[0058] For any object, the multiple indicators may be indicators of the object related to the first business. For example, if the first business is applying for a financial product, the multiple indicators may be at least one of gender, region, monthly income, monthly expenditure, age, and credit score. For each indicator, the indicator value of the object under the indicator is used to represent the state or feature of the object under the indicator, and the indicator values ​​of the object under the multiple indicators may be expressed in numerical form.

[0059] In some embodiments, for the indicator of gender, the indicator value of the object under this indicator can be 1 or 2, where 1 represents female and 2 represents male, or 1 represents male and 2 represents female; for the indicator of region, the indicator value of the object under this indicator can be the code of different regions. For example, assuming that the code of region A is 10001, when the object is located in region A, the indicator value of the object under this indicator can be 10001; for the indicator of monthly income, the indicator value of the object under this indicator can be the object's average monthly income; for the indicator of monthly expenditure, the indicator value of the object under this indicator can be the object's average monthly expenditure; for the indicator of credit score, the indicator value of the object under this indicator can be the object's current credit rating, credit points and other values.

[0060] Optionally, the index values ​​of each of the multiple objects under the multiple indicators can be obtained after data preprocessing, and the data preprocessing method can be: descriptive statistics through special values, abnormal values, missing values, minimum values, maximum values, modes, missing rates, information value (IV) values, variances, correlations, etc. Through the above data preprocessing methods, the index values ​​of each of the multiple objects under the multiple indicators can more comprehensively and accurately reflect the correlation between the multiple indicators, as well as the correlation between each indicator and the object category.

[0061] S23. According to respective category labels of the multiple objects and respective indicator values ​​of the multiple objects under multiple indicators, determine at least one target indicator associated with the object category from among the multiple indicators, wherein the at least one target indicator is used to determine the object category of the target object, and the object category of the target object is used to indicate whether the target object belongs to an associated object of the first business.

[0062] The target object is an object of a to-be-determined object category, and there are two types of object categories: one is an associated object belonging to the first business, and the other is an associated object not belonging to the first business. Exemplarily, if the target object is an associated object of the first business, it means that the target object is willing to handle the first business, or the target object has a need to handle the first business; if the target object is not an associated object of the first business, it means that the target object is not willing to handle the first business, or the target object has no need to handle the first business.

[0063] According to the category labels of each of the multiple objects and the indicator values ​​of each of the multiple objects under multiple indicators, at least one target indicator that has an association relationship with the object category is determined among the multiple indicators. Each target indicator in the at least one target indicator has an association relationship with the object category. Therefore, the object category of the target object can be determined according to the at least one target indicator, that is, whether the target object belongs to an associated object of the first business.

[0064] exist Figure 2 In the illustrated embodiment, at least one target indicator associated with the object category is determined from among the multiple indicators by analyzing the category labels of each of the multiple objects and the indicator values ​​of each of the multiple objects under multiple indicators. In the scheme of the present application, the indicator values ​​of each of the multiple objects under multiple indicators are analyzed in combination with the category labels of each of the multiple objects, and at least one target indicator is determined from among the multiple indicators, and any one of the at least one target indicator determined has an associated relationship with the object category, so the scheme of the present application implements a method of determining a target indicator from among multiple indicators, so that the object category of the target object can be determined based on at least one target indicator.

[0065] exist Figure 2 Based on the embodiment shown below, combined with Figure 3 A detailed process of determining at least one target object having an association relationship with an object category among multiple indicators is further introduced.

[0066] Figure 3 For a flow chart of determining at least one target object provided in an embodiment of the present application, see Figure 3 , the method may include the following steps:

[0067] S31. Performing initial analysis and processing on the indicator values ​​of each of the multiple objects under multiple indicators according to the respective category labels of the multiple objects, and determining the initial association relationship between the object category and the multiple indicators.

[0068] The method of determining the initial association relationship between the object category and multiple indicators can be as follows: according to the category labels of each of the multiple objects, and the indicator values ​​of each of the multiple objects under the multiple indicators, determine the indicator category of each of the multiple indicators, the indicator category is the first indicator category or the second indicator category, the first indicator category is an indicator category that has an association relationship with the object category, and the second indicator category is an indicator category that has an association relationship with both the indicators of the first indicator category and the object category; according to the category labels of each of the multiple objects, the indicator values ​​of each of the multiple objects under the multiple indicators, and the indicator category of each of the multiple indicators, determine the initial association weights between each of the multiple indicators and the object category; wherein the initial association relationship includes the indicator category of each of the multiple indicators, and the initial association weights between each of the multiple indicators and the object category.

[0069] For any indicator, the initial association weight between the indicator and the object category is used to indicate the degree of association between the indicator and the object category, and the initial association weight between the indicator and the object category can be expressed in the form of percentage, decimal, etc. Taking the first business as applying for financial products as an example, assuming that multiple indicators include monthly income and monthly expenditure, where the initial association weight between monthly income and the object category is 30%, and the initial association weight between monthly expenditure and the object category is 60%, it means that the degree of association between monthly income and the object category is small, and the degree of association between the monthly expenditure indicator and the object category is large.

[0070] Optionally, the indicator values ​​of multiple objects under multiple indicators can be processed by methods such as random forest model and eXtreme Gradient Boosting (XGBoost) algorithm to determine the indicator categories of the multiple indicators and the initial association weights between the multiple indicators and the object categories.

[0071] Taking the XGBoost algorithm as an example to process the index values ​​of multiple objects under multiple indicators, the Gradient Boosting Decision Tree (GBDT) algorithm mainly consists of three parts: loss function, weak evaluator (decision tree) and weak evaluator integration method. On this basis, the XGBoost algorithm redefines the loss function and weak evaluator to improve the algorithm's operational efficiency. In addition to the loss function, the XGBoost algorithm also adds the algorithm complexity as a measurement indicator when measuring the algorithm's operational efficiency. Therefore, the calculation formula of the XGBoost algorithm objective function obj can be shown as formula (1):

[0072]

[0073] Among them, obj is used to represent the calculation result of the objective function. represents the loss function of the algorithm, which is used to measure the difference between the true value of the object category and the predicted value of the object category. i represents the i-th sample, which includes the index value of at least one object under at least one index. m represents the total amount of data imported into the k-th weak evaluator. y i represents the true value of the object category, represents the predicted value of the object category; Indicates the complexity of the algorithm, using a transformation Ω of a certain weak evaluator model. K represents the sum of all weak evaluators currently established, and f k Represents the output of the kth weak evaluator. In each iteration of the weak evaluator, the optimal prediction value of the object category is sought by minimizing the objective function, thereby reducing the error rate and complexity of the algorithm at the same time.

[0074] For the loss function, the XGBoost algorithm adds a regularization part to the loss function, which can effectively prevent the occurrence of overfitting, thereby making the algorithm more generalizable. In addition, the XGBoost algorithm uses the second-order Taylor method to process the error part in the calculation of the loss function. Compared with the first-order Taylor method used in the GBDT algorithm, the calculation of the loss function in the XGBoost algorithm is more accurate. At the same time, the XGBoost algorithm makes parallel selections in the weak evaluator learning establishment stage, finds the appropriate child node and splits the features and eigenvalues ​​at the node, thereby improving the algorithm operation efficiency.

[0075] The evaluation indicators for evaluating the performance of the XGBoost algorithm include accuracy, receiver operating characteristic (ROC) curve and area under the curve (AUC) value, and Kolmogorov-Smirnov (KS) curve.

[0076] The accuracy rate refers to the proportion of all correctly predicted samples to the total number of samples. The calculation formula of the accuracy rate can be shown as formula (2):

[0077]

[0078] Among them, T (True) represents the number of correct predictions, F (False) represents the number of incorrect predictions, P (Positives) represents the number of associated objects whose object category is that the object belongs to the first business, and N (Negatives) represents the number of associated objects whose object category is that the object does not belong to the first business. Therefore, TP (True Positives) refers to the number of correctly predicted associated objects whose object category is that the object belongs to the first business, TN (True Negatives) refers to the number of correctly predicted associated objects whose object category is that the object does not belong to the first business, FP (False Positives) refers to the number of incorrectly predicted associated objects whose object category is that the object belongs to the first business, and FN (True Negatives) refers to the number of incorrectly predicted associated objects whose object category is that the object does not belong to the first business.

[0079] The ROC curve is a comprehensive indicator used to show the relationship between sensitivity and specificity. By constructing the ROC curve, the continuous changes in the sensitivity and specificity of the XGBoost algorithm can be revealed. In the ROC curve, the horizontal axis represents the false positive rate (FPR), that is, the number of objects whose predicted object category is the associated objects of the first business and whose predictions are wrong, accounts for the number of associated objects whose object category is the associated objects of the first business. The vertical axis represents the true positive rate (TPR), that is, the number of objects whose predicted object category is the associated objects of the first business and whose predictions are correct, accounts for the number of associated objects whose object category is the associated objects of the first business. The calculation formulas for FPR and TPR can be shown as formula (3):

[0080]

[0081] In the ROC curve, ideally, the TPR should be close to 1 and the FPR should be close to 0, that is, the point (0, 1) on the ROC curve. The closer the ROC curve is to the point (0, 1) and the farther it is from the 45-degree diagonal line, the better the performance of the XGBoost algorithm. The area under the ROC curve is called the AUC value, which usually ranges from 0.5 to 1. When AUC = 1, it means that the XGBoost algorithm is a perfect classifier. When using this XGBoost algorithm, there is at least one threshold that can obtain perfect prediction results. When 0.5 < AUC < 1, the results of the model are better than random guessing. When the threshold is set, the XGBoost algorithm has a certain predictive value. When AUC = 0.5, it means that the prediction results of the XGBoost algorithm are no different from random guessing, and the XGBoost algorithm has no predictive value. When AUC < 0.5, it means that the prediction effect of the XGBoost algorithm is worse than random guessing. To sum up, the larger the AUC value of the classifier, the higher the accuracy rate of the XGBoost algorithm.

[0082] The KS curve, also known as the Lorenz curve, evaluates the risk discrimination ability of the XGBoost algorithm by comparing the differences between the cumulative distributions of good and bad samples. In the KS curve, the vertical axis represents the true positive rate (TPR) and the false positive rate (FPR), and the horizontal axis represents the threshold. By plotting these two curves, the KS curve can be obtained. The KS curve is the difference between the two curves at each threshold, that is, the maximum value of the difference between TPR and FPR. The KS value calculation formula can be expressed as: KS = MAX{TPR, FPR}. Among them, the larger the KS value, the better the prediction accuracy of the XGBoost algorithm. The value range of the KS value is [0, 1]. Generally, when KS is less than 0.2, it can be considered that the performance of the XGBoost algorithm is poor, and at this time the XGBoost algorithm has no discrimination ability; when KS is between 0.2 and 0.3, it can be considered that the performance of the XGBoost algorithm is average; when KS is between 0.3 and 0.4, it can be considered that the performance of the XGBoost algorithm is good and has a certain discrimination ability; when KS is between 0.41 and 0.5, it can be considered that the XGBoost algorithm has discrimination ability; when the KS value is between 0.51 and 0.6, it can be considered that the XGBoost algorithm has good discrimination ability; when the KS value is between 0.61 and 0.75, it can be considered that the XGBoost algorithm has very good discrimination ability; when KS is greater than 0.75, it can be considered that the XGBoost algorithm model is abnormal.

[0083] Taking the first business as applying for financial products as an example, assuming that the multiple indicators of object 1 are gender, age, monthly income, monthly expenditure, and credit score, the indicator value of gender is 1, indicating that object 1 is female, the indicator value of age is 20, the indicator value of monthly income is 5000, the indicator value of monthly expenditure is 3000, and the credit score is 60. The category label of object 1 is used to indicate that object 1 has handled the first business; the multiple indicators of object 2 are gender, age, monthly income, monthly expenditure, and credit score, the indicator value of gender is 1, indicating that object 2 is female, the indicator value of age is 40, the indicator value of monthly income is 8000, the indicator value of monthly expenditure is 5000, and the credit score is 70. The category label of object 2 is used to indicate that object 2 has not handled the first business.

[0084] The category label of object 1, the indicator value of object 1 under multiple indicators, the category label of object 2, and the indicator value of object 2 under multiple indicators are processed by random forest model or XGBoost algorithm, and it is determined that the three indicators of monthly income, monthly expenditure and credit score have an association relationship with the object category, that is, the indicator category of the three indicators of monthly income, monthly expenditure and credit score is the first indicator category; it is determined that the indicator of age has an association relationship with the object category and also has an association relationship with the indicator of monthly income, that is, the indicator category of the indicator of age is the second indicator category. At the same time, it is determined that the initial association weight between gender and object category in multiple indicators is 0, the initial association weight between age and object category is 5%, the initial association weight between monthly income and object category is 30%, the initial association weight between monthly expenditure and object category is 40%, and the initial association weight between credit score and object category is 5%.

[0085] S32. Determine at least one first candidate indicator from a plurality of indicators according to the initial association relationship.

[0086] The method for determining the first candidate indicator can be as follows: for any indicator, the initial association weight between the indicator and the object category is compared with the initial association weight threshold. If the initial association weight between the indicator and the object category is greater than or equal to the initial association weight threshold, the indicator is determined as the first candidate indicator.

[0087] Taking the first business as applying for financial products as an example, assuming that multiple indicators include gender, age, monthly income, and monthly expenditure, the initial association weight between gender and object category is 0, the initial association weight between age and object category is 10%, and the initial association weight between monthly income and object category is 40%. The initial association weight between monthly expenditure and object category is 50%, and the initial association weight threshold is 30%, then monthly income and monthly expenditure are determined as the first candidate indicators.

[0088] S33. Determine, according to the index value of each of the multiple objects under at least one first candidate index, an associated index value between each of the at least one first candidate index and the object category.

[0089] For any one indicator of the at least one first candidate indicator, the association index value between the indicator and the object category is used to indicate the degree of association between the indicator and the object category.

[0090] The method of determining the associated index value between each of at least one first candidate indicator and an object category can be as follows: generating a causal graph corresponding to at least one first candidate indicator according to the indicator category of each of the at least one first candidate indicator and the indicator value of each of the multiple objects under at least one first candidate indicator; processing the causal graph to obtain the associated index value between each of the at least one first candidate indicator and the object category.

[0091] The causal graph corresponding to at least one first candidate indicator is a directed acyclic graph. To further understand the causal graph, Figure 4 Let's take an example to explain. Figure 4 For a schematic diagram of a cause-effect diagram provided in an embodiment of the present application, see Figure 4 The causal graph includes at least one node and a directed line segment, wherein at least one node corresponds to a first candidate indicator or object category respectively, and the directed line segment between two nodes is used to indicate that there is an association relationship between the two nodes, and the node pointed by the arrow of the directed line segment is the "offspring", and the node without the arrow pointing is the "father", wherein the "father" node has a certain influence on the "offspring" node; the length of the directed line segment indicates the degree of association between the two nodes, the longer the length of the directed line segment, the smaller the degree of association between the two nodes; the longer the length of the directed line segment, the greater the degree of association between the two nodes.

[0092] exist Figure 4In the causal graph, the nodes are respectively node X, node Y and node Z, among which there are directed line segments between node Z and node X and node Y respectively, indicating that there is an association relationship between node Z and node X and node Y, and the arrows of the directed line segments point to node X and node Y, indicating that node Z will have a certain influence on node X and node Y respectively; there is a directed line segment between node X and node Y, indicating that there is an association relationship between node X and node Y, and the arrow of the directed line segment points to node Y, indicating that node X will have a certain influence on node Y; the length of the directed line segment between node Z and node X is shorter, and the length of the directed line segment between node Z and node Y is longer, indicating that the degree of association between node Z and node X is greater, and the degree of association between node Z and node Y is smaller, and the length of the directed line segment between node X and node Y is shorter, indicating that the degree of association between X and node Y is greater.

[0093] Generally speaking, correlation can be divided into three levels. The first refers to pure statistical correlation, which is to observe the distribution law of data from the data. This relationship is mainly a summary of historical data, and its mathematical expression is P(y|x); the second is based on the correlation between the two, and is a speculation for the future, and its mathematical expression is P(y|do(x),z); the third is equivalent to considering the cause in combination with the result, which is a reflection on the past, and its mathematical expression is P(y x |x',y').

[0094] In the causal graph, the causal effect between nodes can be defined according to the do operator. Taking node X and node Y as an example, the formula for calculating the average causal effect of node X corresponding to node Y using the do operator can be shown as formula (4):

[0095] ACE(X→Y)=E(Y|do(X=1)-E(Y|do(X=0) (4)

[0096] Among them, ACE(X→Y) represents the average influence of variable X on variable Y. In the context of the do operator, do(X=1) means setting the value of variable X to 1 (i.e., intervention or treatment), while do(X=0) means setting the value of variable X to 0 (i.e., no intervention or as a control). These operations simulate the situation in which the value of X is changed by external means in an actual experiment. E(Y|do(X=1)) and E(Y|do(X=0)) represent the expected value of variable Y when the value of variable X is intervened to 1 or 0, respectively. These expected values ​​are calculated from the mean of the probability distribution of variable Y under given intervention conditions, reflecting the average level of Y under different intervention conditions.

[0097] The method of generating a causal graph corresponding to at least one first candidate indicator may be as follows: input the indicator values ​​of each of the multiple objects under the multiple indicators and the category labels of each of the multiple objects into the causal graph model, so that the causal graph model identifies the association between each indicator and the object category based on the least squares estimation method. Then, input the indicator values ​​of each of the multiple objects under the at least one first candidate indicator and the indicator category of the at least one first candidate indicator into the causal graph model to generate a causal graph corresponding to the at least one first candidate indicator.

[0098] Among them, the causal graph model can be a model for depicting a causal graph, for example, the causal graph model can be a Pearl causal graph model, and the Pearl causal graph model uses a directed graph to describe the association between variables. First, the indicator values ​​of multiple objects under multiple indicators are input into the causal graph model, which can clearly show the association between the indicators and the association between the indicators and the object categories. For example, the indicator whose indicator category is the second indicator category, the association between the first indicator category, and the association between the indicator and the object category can eliminate the false relationship caused by the indicator whose indicator category is the second indicator category, so that the causal graph model can more accurately understand the real association between the indicators and the object category, so that the generated causal graph can more accurately reflect the association between the indicators in at least one first candidate indicator and the object category.

[0099] For the causal graph corresponding to at least one first candidate indicator, the causal graph is processed by backdoor criteria, frontdoor criteria, backdoor adjustment, frontdoor adjustment and IV identification to identify the causal effect of each indicator on the object category in at least one first candidate indicator.

[0100] When identifying the causal effect of each indicator in at least one first candidate indicator on the object category, it is necessary to solve the endogeneity problem in the causal diagram. The endogeneity problem is mainly caused by omitted variables, sample selection bias, two-way causality and measurement error. The endogeneity problem can be solved by natural experiment method or double difference method (Difference-in-Differences, DID).

[0101] The natural experiment method is a random experiment. For any one of the at least one first candidate indicator, Y1(x) is used to indicate that the object category of object i is affected by the indicator, and Y0(x) is used to indicate that the object category of object i is not affected by the indicator. For object i, the object category of object i is either affected by the indicator or not affected by the indicator, that is, one of Y1(x) and Y0(x) must be missing. The average treatment effect (ATE) between the indicator and the object category can be expressed as: ATE = E(Y1(x) - Y0(x)). Under the premise of T randomization, the calculation formula of ATE can be shown as formula (5):

[0102] ATE=E(Y1(x)-Y0(x))=E(Y1(x)|T i =1)-E(Y0(x)|T i =0) = E(Y i |T i =1)-E(Y i |T i =0) (5)

[0103] Among them, E(Y1(x)-Y0(x)) reflects the linear relationship of the expected operator, but the correlation between the nonlinear operators is difficult to identify. i =1)-E(Y0(x)|T i =0) is the embodiment of randomization, that is, T is independent of (Y(1), Y(0)), E(Y i |T i =1)-E(Y i |T i =0), it can be seen that the value of ATE at can be estimated from the observed data.

[0104] In the above natural experiment method, ATE=E(Y1(x)|T i =1)-E(Y0(x)|T i =0) is that T is independent of (Y(1), Y(0)). When this premise is not met, the double difference method can be used. This method requires T0-T i The changing trends between them are consistent (that is, the lines of Di=1 and Di=0 are kept balanced).

[0105] In some embodiments, each node in the causal graph may satisfy backdoor criteria and / or frontdoor criteria, e.g. Figure 4Node Z in the causal graph shown in the figure satisfies the following two conditions: the first condition is that node Z is not a descendant of node X, and the other condition is that node Z blocks the backdoor path from node X to node Y (that is, node X cannot reach node Y through node Z). In this case, node Z is said to satisfy the backdoor criterion.

[0106] For the front door criteria, Figure 5 Take this as an example to illustrate: Figure 5 For a schematic diagram of a causal graph in which a node satisfies the front-door criterion provided in an embodiment of the present application, see Figure 5 , the causal graph includes nodes X, Y, Z and U, among which there are directed line segments between node U and nodes X and Y respectively, and the arrows of the directed line segments point to nodes X and Y respectively; there is a directed line segment between node X and node Z, and the arrow of the directed line segment points to node Z; there is a directed line segment between node Z and node Y, and the arrow of the directed line segment points to node Y. Figure 5 In the causal diagram shown, node Y can be the object category in the embodiment of the present application, node U can be the indicator of the second indicator type in the embodiment of the present application, for example, it can be an indicator such as the age of the object; node X and node Z can be the indicator of the first indicator type in the embodiment of the present application, for example, it can be an indicator such as the monthly expenditure of the object.

[0107] For node Z, node Z satisfies the following three conditions: the first condition is that node Z blocks the direct path from node X to node Y; the second condition is that node X has no backdoor path to node Z; under the three conditions, the backdoor from node Z to node Y is blocked, and node Z satisfies the front door criterion. At this time, if in the joint probability distribution of node X and node Z, the probability that node X and node Z simultaneously take a certain value is greater than 0, that is, P(x,z)>0, then it means that the association effect from node X to node Y can be identified, and the formula for calculating the association effect from node X to node Y through the front door criterion can be shown as formula (6):

[0108] P(y|do(X) = x) = ∑ z P(z|x)∑ x' P(y|x',z)P(x') (6)

[0109] Among them, P(y|do(X)=x) represents the conditional probability distribution of the value y of node Y after the value of node X is intervened by the do operator and the value of node X is set to x; P(z|x) represents the conditional probability distribution of the value z of node Z after the value of node X is intervened by the do operator and the value of node X is set to x; P(y|x',z) represents the conditional probability distribution of the value y of node Y when the value of node X is set to x' and the value of node Z is set to z; P(x') represents the probability that the value of node X is x'.

[0110] For backdoor adjustment, based on the definition of the backdoor criterion, that is, in the path from node X to node Y, node Z is not a descendant of node X, and node Z will block the backdoor path from node X to node Y. It has been proved that if there is a node Z that satisfies the backdoor criterion, then the formula for calculating the causal effect of node X on node Y through backdoor adjustment can be shown as formula (7):

[0111] P(y|do(X)= x) =∑ z P(y|x,z)P(z) (7)

[0112] Among them, P(y|do(X)=x) represents the conditional probability distribution of the value y of node Y after the value of node X is intervened by the do operator and the value of node X is set to x; P(y|x,z) represents the conditional probability distribution of the value y of node Y when the value of node X is set to x and the value of node Z is set to z; P(z) represents the probability that the value of node Z is z.

[0113] Adjust for the front door. Figure 6 Take this as an example to illustrate: Figure 6 For a schematic diagram of a cause-effect diagram that satisfies front door adjustment provided in an embodiment of the present application, see Figure 6 , the causal graph includes nodes X, Y, Z and U, among which there are directed line segments between node U and nodes X and Y respectively, and the arrows of the directed line segments point to nodes X and Y respectively; there is a directed line segment between node X and node Z, and the arrow of the directed line segment points to node Z; there is a directed line segment between node Z and node Y, and the arrow of the directed line segment points to node Y. Figure 6 In the causal graph shown, node Y can be an object category in the embodiment of the present application, node U can be an indicator of the second indicator type in the embodiment of the present application, for example, an indicator such as the age of the object; node X and node Z can be indicators of the first indicator type in the embodiment of the present application, for example, an indicator such as the monthly expenditure of the object. The causal relationship between node X and node Z and the causal relationship between node Z and node Y can be analyzed through a forward process, where the calculation formula of the causal effect of node X on node Z can be shown as formula (8):

[0114]

[0115] The calculation formula of the causal effect of node Z on node Y can be shown as formula (9):

[0116]

[0117] Therefore, the calculation formula for the causal effect of node X on node Y through front door adjustment can be shown as formula (10):

[0118]

[0119] According to formula (10), the causal effect of node X on node Y can be regarded as the product of the causal effect of node X on node Z and the causal effect of node Z on node Y.

[0120] The above method can be used to identify the causal effect of any one of the at least one first candidate indicators on the object category in the causal graph. Then, the association index value between each of the at least one first candidate indicator and the object category is calculated according to the Rubin Causal Model (RCM) or Propensity Score Matching (PSM) and other methods. The association index value is used to indicate the degree of influence of each indicator in the at least one first candidate indicator on the object category. The association index value can be, for example, ATE, Conditional Average Treatment Effect (CATE), etc.

[0121] RCM can estimate the possible results of each node in the causal graph or the overall average, and then obtain the intervention effect. The core of RCM is to find a suitable control group. Usually, RCM is used to observe the degree of influence between two nodes in the control group, the degree of influence between the two nodes in the experimental group, and the difference in the degree of influence between the two nodes in the control group and the experimental group. For the experimental group, only a state of being affected or not affected between the two nodes in the experimental group can be observed. Therefore, it is necessary to find a suitable control group for processing. In an embodiment of the present application, the index values ​​of multiple objects under at least one first candidate indicator are used as the experimental group, and the index values ​​of multiple objects under multiple indicators are used as the control group for processing.

[0122] PSM is a matching method using propensity scores. Propensity scores refer to the degree to which an object tends to belong to the experimental group. For objects with the same propensity score, the treatment and characteristics should be independent, and the treatment and potential results should also be independent to meet the random requirements in the experiment. There are three main methods for propensity scores: stratification, matching, and weighting. PSM belongs to the matching method. By matching samples with similar propensity scores in the experimental group and the control group, a matching group is obtained. Then, the association index value between each of the at least one first candidate indicator and the object category is calculated in the matching group. The calculation formula of the propensity score can be shown as formula (11):

[0123] e(x)=Pr(T=1|X=x) (11)

[0124] Among them, e(x) is the function representation of the propensity score, wherein x represents the indicator value of the indicator for the indicator category being the first indicator category, and Pr(T=1|X=x) represents the probability that the object category is the object having handled the first business (T=1) given that the indicator value of the indicator X is x for any indicator for the indicator category being the first indicator category.

[0125] In the process of matching samples with similar propensity scores in the experimental group and the control group, the most commonly used matching method is the nearest neighbor matching method. Therefore, after the propensity score is calculated, matching can be performed based on the nearest neighbor matching method and the propensity score.

[0126] In addition, there are other ways to determine the association index value between each of the at least one first candidate indicator and the object category, such as Propensity Score Stratification (PSS), Propensity Score Weighting (PSW) and breakpoint regression. Among them, the PSS method is to divide at least one candidate indicator into multiple different levels, and reduce the influence of indicators whose indicator category is the second indicator category according to the similarity of the propensity scores; the PSW method is to calculate the propensity score of each indicator for at least one first candidate indicator, and weight the propensity score of each indicator, so as to reduce the influence of indicators whose indicator category is the second indicator category; the breakpoint regression method determines the association index value between each of the at least one first candidate indicator and the object category by identifying the situation where the influence state of any one of the at least one first candidate indicator on the object category changes at a certain specific threshold.

[0127] Individual Treatment Effect (ITE) refers to the causal effect between each indicator in at least one first candidate indicator and the object category of any object, which can be defined by comparing the potential results of the object category of the object under the influence of each indicator and under the influence of each indicator. Specifically, for any one of the at least one first candidate indicator, Y1(X) represents the potential result when the object category of object i is affected by the indicator, and Y0(X) represents the potential result when the object category of object i is not affected by the indicator. Then the individual treatment effect of object i can be expressed as TE=Y1(X)-Y0(X). It should be noted that for any one of the at least one first candidate indicator, it is impossible for the object category of an object to be affected by the indicator and not affected by the indicator at the same time. Therefore, Y1(X) and Y0(X) are usually not observable at the same time, and these potential results are usually estimated through randomized experiments or statistical methods.

[0128] When a single object is extended to multiple objects, ATE can be defined, which represents the difference in potential results between the object category under the influence of any one of the at least one first candidate indicators and under the influence of the indicator in the multiple objects. ATE can be expressed as ATE=E(Y1(X)-Y0(X)), where E() represents the expected or average value. Since ATE is the average value of the difference in potential results based on multiple objects, the ATE between each of the at least one first candidate indicator and the object category provides an overall assessment of the degree to which the object category is affected by each of the at least one first candidate indicator. It should be noted that the average treatment effect of the experimental group can be called the average treatment effect on the intervention group (Average Treatment Effects on Treated, ATT), and the average treatment effect of the control group can be called the average treatment effect on the control group (Average Treatment Effect on the Control, ATC), where the calculation formula of ATT is ATT=E(Y(W=1)|W=1)-E(Y(W=0)|W=1), Y(W=1) represents the result obtained when the object is in the experimental group, Y(W=0) represents the result obtained when the object is in the control group, E(Y(W=1)|W=1) represents the average result of the object category in the experimental group affected by the corresponding indicator, and E(Y(W=0)|W=1) represents the average counterfactual result of the object category in the control group not affected by the corresponding indicator. In summary, ATT is used to represent the average result of the object category in the experimental group due to the influence of the corresponding indicator.

[0129] Heterogeneous Treatment Effects (HTE) refers to the phenomenon that in the same experiment or study, for any indicator of at least one first candidate indicator, there are significant differences in the degree to which the object category in different objects is affected by the indicator. To study this heterogeneity, a common method is to group the indicator values ​​of multiple objects under at least one first candidate indicator, and compare the differences in the degree to which the object category in each subgroup is affected by the indicator. The degree to which the object in each subgroup is affected by the indicator is called the CATE between the indicator and the object category.

[0130] In addition to determining the association index value between each of at least one first candidate indicator and an object category according to the above method, an embodiment of the present application also provides a method for determining the association index value between each of at least one first candidate indicator and an object category. The method can be: inputting the index category of each of at least one first candidate indicator, the index value of each of multiple objects under at least one first candidate indicator, and the initial association weight between each of at least one first candidate indicator and the object category into a first model to obtain the association index value between each of at least one first candidate indicator and the object category; wherein the first model is trained based on training data of sample objects, and the training data includes the index category of each of at least one first candidate indicator, the sample index value of the sample object under at least one first candidate indicator, and the initial association weight label between each of at least one first candidate indicator and the object category.

[0131] In a process of training the first model, the server can input the training data of the sample object into the first model, process the training data of the sample object through the first model, and obtain the first association index value between at least one first candidate indicator and the object category output by the first model, and the first association index value between at least one first candidate indicator and the object category is the association index value between at least one first candidate indicator and the object category predicted by the first model. Then, the server calculates the loss value according to the first association index value between at least one first candidate indicator and the object category, and the initial association weight label between at least one first candidate indicator and the object category, and adjusts the model parameters of the first model according to the calculated loss value, thereby completing a round of training process.

[0132] The server may perform one or more rounds of training on the first model until the training termination condition of the first model is reached, and the training process is stopped to obtain the trained first model. The training termination condition may be set according to actual needs, for example, the training termination condition may be set to a loss value less than or equal to a preset loss value, for example, the training termination condition may be set to a training number reaching a preset number of times, and so on. The trained first model has the ability to determine the associated index value between each of the at least one first candidate index and the object category.

[0133] S34. Determine at least one target indicator from at least one first candidate indicator according to the associated indicator value between each of the at least one first candidate indicators and the object category.

[0134] At least one target indicator may be an indicator in at least one first candidate indicator, and at least one target indicator may be a main indicator affecting the object category.

[0135] The method of determining at least one target indicator in at least one first candidate indicator can be as follows: determine at least one second candidate indicator in at least one first candidate indicator whose association indicator value with the object category is greater than or equal to a preset association indicator value; and determine an indicator of the first indicator category in at least one second candidate indicator as at least one target indicator.

[0136] Taking the first business of applying for financial products as an example, assuming that at least one candidate indicator includes age, monthly income, monthly expenditure and credit score, wherein the indicator category of age is the second indicator category, the indicator category of monthly income, monthly expenditure and credit score is the first indicator category, and the correlation indicator value between age and object category is 40%, the correlation indicator value between monthly income and object category is 50%, the correlation indicator value between monthly expenditure and object category is 60%, the correlation indicator value between credit score and object category is 20%, and the preset correlation indicator value is 30%, then at least one second candidate indicator is determined to include age, monthly income and monthly expenditure. Since only the indicator categories of the two indicators of monthly income and monthly expenditure in at least one second candidate indicator category are the first indicator category, therefore, at least one target indicator is determined to be monthly income and monthly expenditure.

[0137] exist Figure 3In the embodiment shown, the index values ​​of multiple objects under multiple indicators are initially analyzed and processed by random forest model or XGBoost algorithm, the initial association relationship between the object category and multiple indicators is determined, and at least one first candidate indicator is determined from the multiple indicators according to the initial association relationship, and then the association index value between at least one first candidate indicator and the object category is determined according to the causal graph or the first model by constructing a causal graph of at least one first candidate indicator or training a first model, and finally, at least one target indicator is determined according to the association index value. In the above method, multiple indicators are first screened to determine at least one first candidate indicator, so that when the causal graph is subsequently constructed and the first model is trained, the influence of indicators that have no association relationship or a small association relationship with the object category is reduced, and computing resources are saved. Since the causal graph or the first model can more accurately determine the association index value between at least one first candidate indicator and the object category, at least one target indicator that mainly affects the object category can be more accurately determined according to the association index value.

[0138] exist Figure 3 In the embodiment shown, it is introduced to determine at least one target indicator among multiple indicators. The embodiment of the present application also provides a method for performing business processing according to at least one target indicator. Figure 7 , introduces a method for performing business processing based on at least one target indicator.

[0139] Figure 7 For a flow chart of a business processing method provided in an embodiment of the present application, please refer to Figure 7 , the method may include the following steps:

[0140] S71. Obtain respective indicator values ​​of target objects under at least one target indicator.

[0141] The target object may be an object of an object category to be determined, and at least one target indicator may be an indicator of the target object that has an association with the object category among multiple indicators related to the first business, and the respective indicator values ​​of the target object under at least one target indicator are used to represent the indicator value of the target object under any one of the at least one target indicator.

[0142] The first business may be various types of activities or projects, and the first business may be, for example, applying for a financial product. Taking the first business as applying for a financial product as an example, assuming that the target object is object 1, at least one target indicator includes monthly income and monthly expenditure, and the monthly income of object 1 is 4000 and the monthly expenditure is 2000.

[0143] S72. Determine an object category of the target object according to the respective indicator values ​​of the target object under at least one target indicator; the object category is used to indicate whether the target object is an associated object of the first business.

[0144] The method of determining the object category of the target object according to the respective indicator values ​​of the target object under at least one target indicator can be as follows: determine the object category of the target object under each target indicator of at least one target indicator according to the respective indicator values ​​of the target object under at least one target indicator, and then determine the object category of the target object as an associated object of the first business if the number of associated objects of the target object belonging to the first business under each target indicator is greater than a preset threshold; otherwise, determine the object category of the target object as an associated object that the target object does not belong to the first business.

[0145] Taking the first business as applying for a financial product as an example, assuming that at least one target object includes monthly income, monthly expenditure and credit score, the target object's monthly income is 5000, the monthly expenditure is 4500, and the credit score is 70% (full score 100%). According to the index value of the target object's monthly income, the object category of the target object is determined to be that the target object does not belong to the associated object of the first business; according to the index value of the target object's monthly expenditure, the object category of the target object is determined to be that the target object belongs to the first business; according to the index value of the target object's credit score, the object category of the target object is determined to be that the target object belongs to the first business. The preset threshold is 50% of the number of at least one target object, that is, the preset threshold is 1.5, and the number of associated objects whose object category under each target indicator is that the target object belongs to the first business is 2. Therefore, the object category of the target object is determined to be that the target object belongs to the associated object of the first business.

[0146] S73: Determine a first push frequency according to the object category of the target object.

[0147] The first push frequency may be a frequency of pushing the business information of the first business to the target object, and the first push frequency may be daily, weekly, monthly, etc. The business information of the first business may be, for example, information such as activities and products of the first business.

[0148] When the target object is an associated object of the first business, the first push frequency is determined to be a more frequent frequency such as daily or weekly; when the target object is not an associated object of the first business, the first push frequency is determined to be a frequency with a longer time interval such as monthly or every two months.

[0149] S74. Send service information of the first service to the client according to the first push frequency.

[0150] The client may be a client of the target object, for example, a mobile phone, a computer or other device of the target object. The service information of the first service may be advertisements, activities or other related content of the first service.

[0151] For example, assuming that the first push frequency is daily, then the relevant information such as the activities of the first business on that day will be sent to the client of the target object every day; assuming that the first push frequency is weekly, then the relevant information such as the activities of the first business within that week will be sent to the client of the target object every week; assuming that the first push frequency is monthly, then the relevant information such as the activities of the first business within that month will be sent to the client of the target object every month.

[0152] exist Figure 7 In the illustrated embodiment, according to obtaining the respective index values ​​of the target object under at least one target index, and according to the respective index values ​​of the target object under at least one target index, the object category of the target object is determined, and then according to the object category of the target object, when the object category is that the target object belongs to an associated object of the first business, the first push frequency of the target object is determined to be a relatively frequent frequency, and when the object category is that the target object does not belong to an associated object of the first business, the first push frequency of the target object is determined to be a frequency with a longer time interval, and finally, according to the first push frequency, the business information of the first business is sent to the client of the target object. In the above method, since at least one target index is an index with an associated relationship with the object category, when the object category of the target object is determined according to the respective index values ​​of the target object under at least one target index, the accuracy of determining the object category is improved. Further, according to the object category of the target object, the first push frequency of the target object is determined, and the business information of the first business is sent to the client of the target object according to the first push frequency of the target object, which can improve the use experience of the target object and increase the user stickiness and activity of the target object.

[0153] Figure 8 For a schematic diagram of a structure of an indicator determination device provided in an embodiment of the present application, see Figure 8 The indicator determination device 80 includes: a first determination module 81, a first acquisition module 82 and a first processing module 83, wherein:

[0154] A first determination module 81 is used to determine a category label of each of the plurality of objects, where the category label is used to indicate that the corresponding object has handled the first service, or indicates that the corresponding object has not handled the first service;

[0155] A first acquisition module 82 is used to acquire the index value of each of the multiple objects under multiple indicators;

[0156] The first processing module 83 is used to determine at least one target indicator that has an association relationship with the object category among the multiple indicators based on the category labels of the multiple objects and the indicator values ​​of the multiple objects under the multiple indicators, the at least one target indicator is used to determine the object category of the target object, and the object category of the target object is used to indicate whether the target object belongs to an associated object of the first business.

[0157] In a possible implementation manner, the first processing module 83 is specifically configured to:

[0158] According to the category labels of the multiple objects, initial analysis and processing are performed on the indicator values ​​of the multiple objects under the multiple indicators to determine the initial correlation relationship between the object category and the multiple indicators;

[0159] Determine at least one first candidate indicator from the multiple indicators according to the initial association relationship;

[0160] Determine, according to the index value of each of the multiple objects under at least one first candidate index, an associated index value between each of the at least one first candidate index and the object category;

[0161] At least one target indicator is determined from the at least one first candidate indicator according to the associated indicator value between each of the at least one first candidate indicators and the object category.

[0162] In a possible implementation manner, the first processing module 83 is specifically configured to:

[0163] Determine the indicator category of each of the multiple indicators according to the category labels of each of the multiple objects and the indicator values ​​of each of the multiple objects under the multiple indicators, where the indicator category is a first indicator category or a second indicator category, the first indicator category is an indicator category associated with the object category, and the second indicator category is an indicator category associated with both the indicators of the first indicator category and the object category;

[0164] Determine initial association weights between each of the multiple indicators and the object category according to the category labels of each of the multiple objects, the indicator values ​​of each of the multiple objects under the multiple indicators, and the indicator category of each of the multiple indicators;

[0165] The initial association relationship includes the indicator category of each of the multiple indicators and the initial association weight between each of the multiple indicators and the object category.

[0166] In a possible implementation manner, the first processing module 83 is specifically configured to:

[0167] Generate a causal graph corresponding to at least one first candidate indicator according to the indicator category of each at least one first candidate indicator, the initial association weight between each at least one first candidate indicator and the object category, and the indicator value of each of the multiple objects under the at least one first candidate indicator;

[0168] The causal graph is processed to obtain an association index value between each of the at least one first candidate index and the object category.

[0169] In a possible implementation manner, the first processing module 83 is specifically configured to:

[0170] Determine at least one second candidate indicator from the at least one first candidate indicator, the association index value between which the indicator and the object category is greater than or equal to a preset association index value;

[0171] An indicator of the first indicator category among at least one second candidate indicator is determined as at least one target indicator.

[0172] The indicator determination device provided in the embodiment of the present application can execute the technical solution of the indicator determination method in the above method embodiment, and its implementation principle and beneficial effects are similar, which will not be repeated here.

[0173] Fig. 9 A schematic diagram of the structure of a service processing device provided in an embodiment of the present application. Fig. 9 As shown, the service processing device 90 includes a second acquisition module 91, a second processing module 92, a second determination module 93 and a sending module 94, wherein:

[0174] The second acquisition module 91 is used to obtain the indicator value of each target object under at least one target indicator;

[0175] The second processing module 92 is used to determine the object category of the target object according to the respective indicator values ​​of the target object under at least one target indicator; the object category is used to indicate whether the target object belongs to an associated object of the first business;

[0176] A second determination module 93, configured to determine a first push frequency according to an object category of a target object;

[0177] The sending module 94 is used to send the service information of the first service to the client according to the first push frequency.

[0178] The business processing device provided in the embodiment of the present application can execute the technical solution of the business processing method in the above method embodiment, and its implementation principle and beneficial effects are similar, which will not be repeated here.

[0179] Fig.10 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Fig.10As shown, the electronic device 100 may include: a processor 102, and a memory 101 communicatively connected to the processor 102;

[0180] The memory 101 stores computer executable instructions;

[0181] The processor 102 executes the computer-executable instructions stored in the memory 101 to implement the methods involved in the above method embodiments.

[0182] Optionally, the processor may be a central processing unit (CPU), a graphics processing unit (GPU), other general-purpose processors, a digital signal processor (DSP), or an application specific integrated circuit (ASIC). A general-purpose processor may be a microprocessor or any conventional processor. The steps of the method disclosed in the embodiments of the present application may be directly implemented as being executed by a hardware processor, or may be implemented by a combination of hardware and software modules in the processor.

[0183] The electronic device 100 provided in the embodiment of the present application can execute the technical solution in the above method embodiment, and its implementation principles and beneficial effects are similar, which will not be repeated here.

[0184] An embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, they are used to implement the method involved in the above method embodiment.

[0185] An embodiment of the present application provides a computer program product, including a computer program, which implements the method involved in the above method embodiment when executed by a processor.

[0186] An embodiment of the present application provides a chip having a computer program stored thereon. When the computer program is executed by the chip, the method involved in the above method embodiment is implemented.

[0187] An embodiment of the present application provides a chip module, on which a computer program is stored. When the computer program is executed by the chip module, the method involved in the above method embodiment is implemented.

[0188] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by the present application.

[0189] It should be further noted that, although the various steps in the flowchart are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless there is a clear description in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowchart may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these sub-steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

[0190] It should be understood that the above-mentioned device embodiments are only illustrative, and the device of the present application can also be implemented in other ways. For example, the division of units / modules in the above-mentioned embodiments is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units, modules or components can be combined, or can be integrated into another system, or some features can be ignored or not executed.

[0191] In addition, unless otherwise specified, each functional unit / module in each embodiment of the present application may be integrated into one unit / module, each unit / module may exist physically separately, or two or more units / modules may be integrated together. The above-mentioned integrated unit / module may be implemented in the form of hardware or in the form of a software program module.

[0192] If the integrated unit / module is implemented in the form of hardware, the hardware may be a digital circuit, an analog circuit, etc. The physical implementation of the hardware structure includes but is not limited to transistors, memristors, etc. If not specifically stated, the processor may be any appropriate hardware processor, such as a CPU, a GPU, a Field Programmable Gate Array (FPGA), a DSP, and an ASIC, etc. If not specifically stated, the storage unit may be any appropriate magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory RRAM (Resistive Random Access Memory), a dynamic random access memory DRAM (Dynamic Random Access Memory), a static random access memory SRAM (Static Random-Access Memory), an enhanced dynamic random access memory EDRAM (Enhanced Dynamic Random Access Memory), a high-bandwidth memory HBM (High-Bandwidth Memory), a hybrid memory cube HMC (Hybrid Memory Cube), etc.

[0193] If the integrated unit / module is implemented in the form of a software program module and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or all or part of the technical solution, can be embodied in the form of a software product, which is stored in a memory and includes several instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the various embodiments of the present application. The aforementioned memory includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk and other media that can store program codes.

[0194] In the above embodiments, the description of each embodiment has its own emphasis. For the part not described in detail in a certain embodiment, please refer to the relevant description of other embodiments. The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not 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.

[0195] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary techniques in the art that are not disclosed in the present application. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present application are indicated by the following claims.

[0196] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A method for determining an indicator, characterized in that: The method comprises: Determine a category label of each of the plurality of objects, wherein the category label is used to indicate that the corresponding object has handled the first service, or indicates that the corresponding object has not handled the first service; Obtaining the indicator value of each of the multiple objects under multiple indicators; According to the category labels of each of the multiple objects and the indicator values ​​of each of the multiple objects under the multiple indicators, at least one target indicator associated with the object category is determined among the multiple indicators, and the at least one target indicator is used to determine the object category of the target object, and the object category of the target object is used to indicate whether the target object belongs to an associated object of the first business.

2. The method according to claim 1, characterized in that The determining, according to the respective category labels of the plurality of objects and the respective indicator values ​​of the plurality of objects under the plurality of indicators, at least one target indicator associated with the object category from the plurality of indicators comprises: According to the category labels of the multiple objects, initial analysis and processing are performed on the indicator values ​​of the multiple objects under the multiple indicators to determine the initial association relationship between the object category and the multiple indicators; Determining at least one first candidate indicator from the multiple indicators according to the initial association relationship; Determine, according to the index value of each of the plurality of objects under the at least one first candidate index, an associated index value between each of the at least one first candidate index and the object category; The at least one target indicator is determined from the at least one first candidate indicator according to the associated indicator value between each of the at least one first candidate indicator and the object category.

3. The method according to claim 2, characterized in that The performing initial analysis and processing on the indicator values ​​of each of the multiple objects under the multiple indicators according to the category labels of each of the multiple objects to determine the initial association relationship between the object category and the multiple indicators includes: Determine, according to the respective category labels of the plurality of objects and the respective indicator values ​​of the plurality of objects under the plurality of indicators, the indicator category of each of the plurality of indicators being a first indicator category or a second indicator category, the first indicator category being an indicator category having an association relationship with the object category, and the second indicator category being an indicator category having an association relationship with both the indicators of the first indicator category and the object category; Determine, according to the respective category labels of the plurality of objects, the respective indicator values ​​of the plurality of objects under the plurality of indicators, and the respective indicator categories of the plurality of indicators, an initial association weight between each of the plurality of indicators and the object category; The initial association relationship includes the indicator category of each of the multiple indicators and the initial association weight between each of the multiple indicators and the object category.

4. The method according to claim 2 or 3, characterized in that: The determining, according to the index value of each of the plurality of objects under the at least one first candidate index, the associated index value between each of the at least one first candidate index and the object category comprises: Generate a cause-and-effect graph corresponding to the at least one first candidate indicator according to the indicator category of each of the at least one first candidate indicator and the indicator values ​​of each of the multiple objects under the at least one first candidate indicator; The cause-effect diagram is processed to obtain an association index value between each of the at least one first candidate index and the object category.

5. The method according to claim 2 or 3, characterized in that: The determining, according to the index value of each of the plurality of objects under the at least one first candidate index, the associated index value between each of the at least one first candidate index and the object category comprises: Inputting the indicator category of each of the at least one first candidate indicator, the indicator value of each of the multiple objects under the at least one first candidate indicator, and the initial association weight between each of the at least one first candidate indicator and the object category into the first model to obtain the association indicator value between each of the at least one first candidate indicator and the object category; Among them, the first model is trained according to the training data of the sample object, and the training data includes the indicator category of each of the at least one first candidate indicators, the sample indicator value of each of the sample objects under the at least one first candidate indicator, and the initial association weight label between each of the at least one first candidate indicator and the object category.

6. The method according to claim 3, characterized in that The determining the at least one target indicator from the at least one first candidate indicator according to the associated indicator value between each of the at least one first candidate indicator and the object category comprises: Determine, from the at least one first candidate indicator, at least one second candidate indicator whose association index value with the object category is greater than or equal to a preset association index value; An indicator of the first indicator category among the at least one second candidate indicator is determined as the at least one target indicator.

7. A business processing method, characterized in that: The method comprises: Obtain the indicator value of each target object under at least one target indicator; Determining an object category of the target object according to respective indicator values ​​of the target object under the at least one target indicator; the object category is used to indicate whether the target object belongs to an associated object of the first business; Determining a first push frequency according to the object category of the target object; Sending the service information of the first service to the client according to the first push frequency.

8. An indicator determination device, characterized in that: The device comprises: A first determination module is used to determine a category label of each of the plurality of objects, wherein the category label is used to indicate that the corresponding object has handled the first service, or indicates that the corresponding object has not handled the first service; A first acquisition module is used to acquire the index value of each of the multiple objects under multiple indicators; A first processing module is used to determine at least one target indicator associated with the object category among the multiple indicators based on the category labels of each of the multiple objects and the indicator values ​​of each of the multiple objects under the multiple indicators, wherein the at least one target indicator is used to determine the object category of the target object, and the object category of the target object is used to indicate whether the target object belongs to an associated object of the first business.

9. A service processing device, characterized in that: The device comprises: A second acquisition module is used to acquire the respective indicator values ​​of the target objects under at least one target indicator; A second processing module, configured to determine an object category of the target object according to respective indicator values ​​of the target object under the at least one target indicator; the object category is used to indicate whether the target object belongs to an associated object of the first business; A second determining module, configured to determine a first push frequency according to an object category of the target object; A sending module is used to send the service information of the first service to the client according to the first push frequency.

10. An electronic device, characterized in that: include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the method according to any one of claims 1 to 7.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 7 when executed by a processor.

12. A computer program product, characterized in that The invention comprises a computer program, which implements the method according to any one of claims 1 to 7 when being executed by a processor.

Images