Financial marketing modeling method and related equipment

Abstract

The invention discloses a financial marketing modeling method and device, a server, a computer readable storage medium and a computer program product. The method provides a two-stage sampling modeling mode and a sampling interval construction mode. Wherein in the first stage, a feature screening mode is adopted to construct a marketing model, the marketing model serves as a rough arrangement model, and the model has the good distinguishing capacity for samples in a head interval relative to the sampling scale; in the second stage, a rough arrangement model is used for scoring full-amount modeling samples, segmented sampling intervals with increasing sample numbers are constructed according to scores from low to high, the higher the sample scores are, the higher the sampling proportion of the sampling intervals is, and therefore difficult negative samples are reserved to the maximum extent, and the sampling efficiency is improved. The training of the marketing model is carried out on the sample set constructed based on the sampling mode, and the modeling effect of the fine arrangement model can be obviously improved.

Description

Technical Field

[0001] The present disclosure generally relates to the field of financial technology, and in particular to a financial marketing modeling method, device, server, computer-readable storage medium, and computer program product. Background Art

[0002] A common marketing modeling method in the financial field is to take all users at a fixed time point as samples, observe whether they apply for credit / initiate credit use in the future (such as a week or a month), and build a binary classification model with users who actually use credit as positive samples and users who do not use credit as negative samples. However, in reality, since the number of users is usually very large and the number of features used in modeling is also large, it is difficult to directly use all samples and all features for modeling due to the limitations of the CPU / GPU and memory resources of the training platform.

[0003] Normally, the ratio of positive and negative samples in the full sample is quite different. In order to solve the problem of not being able to directly use the full sample and full features to train the marketing model, common practices include random negative sampling of samples and feature screening. The first method is to retain all positive samples and use a fixed positive-negative sample ratio (such as 1:100) to randomly sample negative samples to reduce the number of samples and reduce the demand for training resources; the second method retains all modeling samples, but first uses a small part of the sample set (such as 1 / 10) to train a model with a smaller sample size with smaller resource requirements, and then selects a part of the features with higher importance according to the feature importance of the trained model before using the full sample for modeling.

[0004] However, the random negative sampling method will cause the distribution of positive and negative samples in the modeling to be inconsistent with the actual distribution. The trained model can usually only have a certain degree of discrimination under coarse-grained binning. It is not applicable in some scenarios where the model only needs to have high discrimination for a very small number of user groups with the highest scores (such as precise manual telemarketing reach). The feature screening method will screen out most of the features that are not important enough, and the modeling effect will be compromised. Summary of the invention

[0005] In view of the above-mentioned defects or deficiencies in the prior art, it is desired to provide a financial marketing modeling method, device, server, computer-readable storage medium and computer program product, which not only overcome the problem of low discrimination of the interval with the highest model score caused by random sampling of negative samples, but also solve the problem of removing features with low importance resulting in deterioration of modeling effect, thereby effectively improving the modeling effect of the financial scenario marketing model.

[0006] In a first aspect, an embodiment of the present application provides a financial marketing modeling method, including:

[0007] Extract the full amount of data corresponding to important features from the full sample set;

[0008] Training a first marketing model according to the full amount of data as a rough ranking model;

[0009] Calling the coarse ranking model to evaluate each sample in the full sample set, obtaining a score corresponding to each sample, and ranking each sample according to the score;

[0010] Construct segmented sampling intervals with increasing sample numbers according to the scores from low to high;

[0011] According to the sampling ratio, a specified number of samples are extracted from each of the segmented sampling intervals as a modeling sample set for the refined ranking model;

[0012] A second marketing model is trained based on the modeling sample set as a refined financial marketing model.

[0013] In one embodiment, constructing segmented sampling intervals with increasing sample numbers from low to high scores includes:

[0014] Construct segmented sampling intervals with increasing sample numbers in geometric proportion from low to high scores.

[0015] In one embodiment, constructing segmented sampling intervals with increasing sample numbers in geometric proportion from low to high scores includes:

[0016] Evaluate the polynomial equation q c -1=ck(q-1) positive real number solution q c ; where c is the number of sampling intervals, q c is the sample quantity growth ratio for each sampling interval, q is the initial ratio value, and k is the sampling ratio;

[0017] According to the scores from low to high, the number of structures are There are c sampling intervals; where N is the number of samples required to be sampled.

[0018] In one embodiment, it also includes:

[0019] Evaluate the available resources of the training platform and determine the number of samples N that need to be sampled.

[0020] In one embodiment, extracting the full amount of data corresponding to the important features from the full amount of sample sets includes:

[0021] Extract a small sample set from the full sample set, and train the third marketing model based on the small sample set as the small sample model;

[0022] Calling the small sample model to evaluate the weight of each feature in the small sample set;

[0023] Select features with weights greater than the threshold as the target feature set;

[0024] The full amount of data corresponding to the target feature set is extracted from the full amount of sample set.

[0025] In one embodiment, the first marketing model and the second marketing model are XGBOOST models.

[0026] In a second aspect, an embodiment of the present application provides a financial marketing modeling device, including:

[0027] A data extraction unit is used to extract the full amount of data corresponding to important features from the full amount of sample sets;

[0028] A first model training unit, used for training a first marketing model according to the full amount of data as a rough ranking model;

[0029] A sample evaluation unit, used to call the coarse ranking model to evaluate each sample in the full sample set, obtain a score corresponding to each sample, and sort each sample according to the score;

[0030] An interval construction unit is used to construct segmented sampling intervals with increasing sample quantities from low to high scores;

[0031] A sample extraction unit, used to extract a specified number of samples from each of the segmented sampling intervals according to a sampling ratio as a modeling sample set for a refined ranking model;

[0032] The second model training unit is used to train a second marketing model according to the modeling sample set as a refined financial marketing model.

[0033] In a third aspect, an embodiment of the present application provides a server, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method described in the embodiment of the present application when executing the program.

[0034] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in the embodiment of the present application.

[0035] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the method described in the embodiment of the present application.

[0036] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings:

[0038] Figure 1 A schematic diagram showing a flow chart of a financial marketing modeling method provided in an embodiment of the present application is shown;

[0039] Figure 2 A schematic diagram of data flow provided by an embodiment of the present application is shown;

[0040] Figure 3 An exemplary structural block diagram of a financial marketing modeling device provided in an embodiment of the present application is shown;

[0041] Figure 4 A schematic diagram of the structure of a computer system suitable for implementing a server of an embodiment of the present application is shown. DETAILED DESCRIPTION

[0042] The present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the invention are shown in the accompanying drawings.

[0043] It should be noted that, in the absence of conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. The present application will be described in detail below with reference to the drawings and in combination with the embodiments. Although the embodiments of the present application provide the method operation instruction steps shown in the following embodiments or drawings, more or fewer operation instruction steps may be included in the method based on routine or no creative labor. In the steps where there is no necessary causal relationship logically, the execution order of these steps is not limited to the execution order provided by the embodiments of the present application. The method may be executed in the order of the methods shown in the embodiments or drawings or in parallel during the actual processing process or when the device is executed.

[0044] It should be noted that the acquisition or use of data in the embodiments of the present application requires the user's consent. The relevant data can only be obtained after the user's authorization and permission, and the acquisition or use of the data complies with relevant legal provisions.

[0045] Please refer to Figure 1 , Figure 1 FIG. 1 is a flow chart of a financial marketing modeling method provided by an embodiment of the present application. Figure 1 As shown, the method includes:

[0046] S101, extracting the full amount of data corresponding to important features from the full amount of sample sets;

[0047] A portion of features with higher importance are extracted from the entire sample set and entered into the model, that is, the entire data corresponding to the important features, to train the first marketing model. Concentrating important feature resources on training the first marketing model can help it concentrate resources to identify key marketing factors.

[0048] Among them, the selection of important features can be customized by the user. For example, the user pre-selects features with greater information value in use and marks them as important features; or the feature weights or values ​​are evaluated after model learning, and features with results higher than a threshold are screened out from the evaluation results as important features. In this embodiment, the selection process of important features is not limited and can be set according to actual application needs.

[0049] To deepen the understanding, a specific screening method of important features is provided here, such as Figure 2 The following is a schematic diagram of data flow using this feature screening method. The implementation method is as follows:

[0050] (1) extracting a small sample set from the full sample set, and training a third marketing model based on the small sample set as the small sample model;

[0051] A small portion of the sample set is used to train a marketing model with a smaller sample size, namely the third marketing model, which is taken as the small sample model.

[0052] (2) Call the small sample model to evaluate the weight of each feature in the small sample set;

[0053] Call the trained small sample model to evaluate the importance of each feature in the small sample set and determine the weight corresponding to each feature. The specific small sample model can use the XGBOOST (Extreme Gradient Boosting) model.

[0054] (3) Select features whose weights are greater than the threshold as the target feature set;

[0055] A weight threshold is predefined, and features with weights greater than the threshold are taken as important features for data extraction. If the weight is not greater than the threshold, this part of the data is discarded to reduce the amount of data analysis.

[0056] (4) Extract the full data corresponding to the target feature set from the full sample set.

[0057] It should be noted that Figure 2 The data flow process after obtaining the top feature set can be compared with Figure 1 Steps S102 to S106 and related introductions are compared with each other. In this embodiment, only the screening method of the above important features is taken as an example. Other methods can refer to the introduction of this embodiment and will not be repeated here.

[0058] S102, training a first marketing model based on the full amount of data as a rough ranking model;

[0059] This method uses feature screening to construct a coarse-ranking model. Although the model effect is somewhat compromised compared to the ideal model that uses all samples and all features for modeling, it still has good discrimination ability for samples in the head interval under the sampling scale. At the same time, the full amount of data corresponding to important features is used to train the marketing model, which can significantly improve the accuracy, generalization ability and stability of the model while reducing resource consumption, and can also reduce the risk of bias and overfitting.

[0060] S103, calling the coarse ranking model to evaluate each sample in the full sample set, obtaining a score corresponding to each sample, and ranking each sample according to the score;

[0061] Use the coarse sorting model to score all the modeling samples and sort them by score. To facilitate the subsequent sample extraction, the scores of all the samples after sorting can be replaced by the sorting sequence number. The higher the sample score, the later the ranking is when sorting from high to low by score, and the larger the sequence number. Of course, you can also directly determine the important features by looking up the score, which is not limited here, and only takes the example of determining the score by sequence number.

[0062] S104, constructing segmented sampling intervals with increasing sample numbers from low to high scores;

[0063] The coarse ranking model is used to score all the modeling samples. The higher the sample score, the more difficult it is to distinguish the negative samples, and the more they should be added to the fine ranking model modeling sample set. To this end, this method constructs a segmented sampling interval with increasing sample numbers, ensuring that the higher the sample score, the higher the sampling ratio of the sampling interval, thereby retaining the difficult negative samples to the greatest extent and improving the modeling effect of the fine ranking model.

[0064] The specific construction method of the specific segmented sampling interval is not limited in this embodiment. The segmented sampling interval with increasing sample quantity can be constructed according to the needs of the actual sampling ratio.

[0065] S105, extracting a specified number of samples from each segmented sampling interval according to the sampling ratio as a modeling sample set for the refined ranking model;

[0066] For all sampling intervals, a specified number of samples are extracted from each sampling interval, where the number of samples extracted from each sampling interval can be determined by the user directly or based on system performance and available resources. For example, if a total of 100 samples are collected and there are 5 segmented sampling intervals, 20 samples can be randomly sampled from each sampling interval without replacement to construct a set of 100 sample sets for fine-tuning the model.

[0067] S106. Train a second marketing model based on the modeling sample set as a refined financial marketing model.

[0068] Based on the full feature set of the modeling samples, the second marketing model is trained to obtain the final refined ranking model, and the trained refined ranking model is used as the final financial marketing model for evaluating user demand for financial products.

[0069] It should be noted that all the marketing models in this embodiment, including the first marketing model, the second marketing model, and the third marketing model, are only used for distinction and have no other practical meaning. The type of model and the specific model structure parameters are not limited here. For example, other gradient boosting decision tree models such as lightgbm, catboost, or deep models such as MaskNet can be selected. Specifically, the gradient boosting decision tree XGBOOST model can be selected. Of course, other types of models can also be selected. Please refer to the introduction of this embodiment and will not be repeated here.

[0070] It should be noted that although the operations of the method of the present invention are described in a particular order in the drawings, this does not require or imply that the operations must be performed in this particular order or that all illustrated operations must be performed to achieve desired results.

[0071] Based on the above introduction, the method provided in this embodiment provides a two-stage sampling modeling method and a sampling interval construction method. Among them, the first stage adopts the feature screening method to construct the marketing model. As a coarse ranking model, the model has a better ability to distinguish the samples in the head interval under the sampling scale; the second stage uses the coarse ranking model to score the full amount of modeling samples. The higher the sample score, the more difficult the negative samples are to distinguish. For this reason, this method proposes to construct a segmented sampling interval with increasing sample numbers from low to high according to the score. The higher the sample score, the higher the sampling ratio of the sampling interval in which it is located, thereby retaining the difficult negative samples to the greatest extent. The sample set constructed based on this sampling method is used to train the marketing model, which can significantly improve the modeling effect of the fine ranking model while consuming less resources. Therefore, this method not only overcomes the problem that the random sampling of negative samples leads to low discrimination of the interval with the highest model score, but also solves the problem that the modeling effect is compromised due to the removal of features with low importance, and can effectively improve the modeling effect of the financial scenario marketing model.

[0072] The above embodiments do not limit the specific construction method of the specific segmented sampling intervals. In this regard, this embodiment proposes a specific construction method of the segmented sampling intervals, which can be achieved by constructing segmented sampling intervals with a geometrically increasing number of samples from low to high scores.

[0073] Geometric increase means that the amount of each increase is a fixed proportion of the previous value. Constructing a geometrically increasing segmented sampling interval can better adapt to the distribution of financial sample data, especially when the data changes exponentially, and the growth rate can be controlled by adjusting the proportional factor, so that it can be more flexible to adapt to different needs.

[0074] Specifically, the implementation steps of constructing a segmented sampling interval with a geometric increase in the number of samples from low to high scores are as follows:

[0075] (1) Calculate the polynomial equation q c -1=ck(q-1) positive real number solution q c ; where c is the number of sampling intervals, q c is the sample quantity growth ratio for each sampling interval, q is the initial ratio value, and k is the sampling ratio;

[0076] Calculate the polynomial equation q based on the number of sampling intervals c (typical value is 5, which can be modified according to actual conditions) c -1=ck(q-1) positive real number solution q c , where q c represents the growth ratio of the number of samples in each sampling interval, q represents the initial ratio value, M is the total number of samples, N is the number of samples required to be sampled, then the sampling ratio k = M / N.

[0077] (2) According to the scores from low to high, the number of There are c sampling intervals.

[0078] According to the order of the full sample numbers from 1 to M, the number of samples constructed is There are c sampling intervals.

[0079] That is, the sum of the number of samples in all sampling intervals is equal to the number of total samples.

[0080] The number of samples N required to be sampled can be determined by evaluating the available resources of the training platform (such as CPU / GPU / memory resources). This determination method is automated and can adapt to the needs of different training platforms to avoid affecting other processes on the platform. Of course, other determination methods can also be adopted, such as user customization, which is not limited here.

[0081] Further references Figure 3 , which shows an exemplary structural block diagram of a financial marketing modeling device according to an embodiment of the present application, which mainly includes:

[0082] The data extraction unit 101 is used to extract the full amount of data corresponding to the important features from the full amount of sample sets;

[0083] A first model training unit 102, used to train a first marketing model according to the full amount of data as a rough ranking model;

[0084] The sample evaluation unit 103 is used to call the coarse ranking model to evaluate each sample in the full sample set, obtain the score corresponding to each sample, and sort each sample according to the score;

[0085] An interval construction unit 104 is used to construct a segmented sampling interval with increasing sample quantity from low to high scores;

[0086] The sample extraction unit 105 is used to extract a specified number of samples from each segmented sampling interval according to the sampling ratio as a modeling sample set of the refined ranking model;

[0087] The second model training unit 106 is used to train a second marketing model according to the modeling sample set as a refined financial marketing model.

[0088] It should be understood that the units described in the above device are similar to those in the reference Figure 1The steps in the method described above correspond to each other. Therefore, the operations and features described above for the method are also applicable to the device and the units contained therein, and will not be repeated here. The device can be pre-implemented in the browser or other security application of the server, or loaded into the browser or its security application of the server by downloading or the like. The corresponding units in the device can cooperate with the units in the server to implement the solution of the embodiment of the present application.

[0089] For the several units mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more units described above can be embodied in one unit. On the contrary, the features and functions of one unit described above can be further divided into multiple units to be embodied.

[0090] It should be noted that for details not disclosed in the financial marketing modeling device of the embodiment of the present application, please refer to the details disclosed in the above embodiments of the present application, which will not be repeated here.

[0091] Reference below Figure 4 , Figure 4 A schematic diagram of the structure of a computer system suitable for implementing a server of an embodiment of the present application is shown.

[0092] like Figure 4 As shown, the computer system includes a central processing unit (CPU) 401, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 402 or the program loaded from the storage part 408 into the random access memory (RAM) 403. In the RAM 403, various programs and data required for the operation instructions of the system are also stored. The CPU 401, the ROM 402, and the RAM 403 are connected to each other through a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0093] The following components are connected to the I / O interface 405: an input section 406 including a keyboard, a mouse, etc.; an output section 407 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN card, a modem, etc. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to the I / O interface 405 as needed. A removable medium 411, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 410 as needed, so that a computer program read therefrom is installed into the storage section 408 as needed.

[0094] In particular, according to an embodiment of the present application, the above reference flow chart Figure 1The described process can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer readable medium, and the computer program includes a program code for executing the method shown in the flow chart. In such an embodiment, the computer program includes a program code for executing the method shown in the flow chart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 409, and / or installed from the removable medium 411. When the computer program is executed by the central processing unit (CPU) 401, the above-mentioned functions defined in the system of the present application are executed.

[0095] It should be noted that the computer-readable medium shown in the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal can take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium such as a computer-readable storage medium that can send, propagate or transmit a program for use by or in conjunction with an instruction execution system, device or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0096] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, functions and operating instructions of the system, method and computer program product according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the aforementioned module, program segment or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the box can also occur in a sequence different from that marked in the accompanying drawings. For example, the boxes represented by two connections can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operating instruction, or can be implemented with a combination of dedicated hardware and computer instructions.

[0097] The units or modules involved in the embodiments described in the present application may be implemented by software or hardware. The units or modules described may also be arranged in a processor. The names of these units or modules do not, in some cases, constitute limitations on the units or modules themselves.

[0098] As another aspect, the present application further provides a computer-readable storage medium, which may be included in the server described in the above embodiment, or may exist independently without being assembled into the server. The above computer-readable storage medium stores one or more programs, and when the above programs are used by one or more processors to execute the financial marketing modeling method described in the present application.

[0099] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the aforementioned disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in this application (but not limited to) by each other to form a technical solution.

Claims

1. A financial marketing modeling method, characterized in that: include: Extract the full amount of data corresponding to important features from the full sample set; Training a first marketing model according to the full amount of data as a rough ranking model; Calling the coarse ranking model to evaluate each sample in the full sample set, obtaining a score corresponding to each sample, and ranking each sample according to the score; Construct segmented sampling intervals with increasing sample numbers according to the scores from low to high; According to the sampling ratio, a specified number of samples are extracted from each of the segmented sampling intervals as a modeling sample set for the refined ranking model; A second marketing model is trained based on the modeling sample set as a refined financial marketing model.

2. The method according to claim 1, characterized in that The stepwise sampling intervals with increasing sample numbers are constructed from low to high scores, including: Construct segmented sampling intervals with increasing sample numbers in geometric proportion from low to high scores.

3. The method according to claim 2, characterized in that The stepwise sampling intervals with increasing sample numbers from low to high scores include: Evaluate the polynomial equation q c -1=ck(q-1) positive real number solution q c ; where c is the number of sampling intervals, q c is the sample quantity growth ratio for each sampling interval, q is the initial ratio value, and k is the sampling ratio; According to the scores from low to high, the number of constructions are N / c, q c N / c, There are c sampling intervals; where N is the number of samples required to be sampled.

4. The method according to claim 3, characterized in that Also includes: Evaluate the available resources of the training platform and determine the number of samples N that need to be sampled.

5. The method according to claim 1, characterized in that The step of extracting the full amount of data corresponding to the important features from the full amount of sample sets includes: Extract a small sample set from the full sample set, and train the third marketing model based on the small sample set as the small sample model; Calling the small sample model to evaluate the weight of each feature in the small sample set; Select features with weights greater than the threshold as the target feature set; The full amount of data corresponding to the target feature set is extracted from the full amount of sample set.

6. The method according to claim 1, characterized in that The first marketing model and the second marketing model use the XGBOOST model.

7. A financial marketing modeling device, characterized in that: include: A data extraction unit is used to extract the full amount of data corresponding to important features from the full amount of sample sets; A first model training unit, used for training a first marketing model according to the full amount of data as a rough ranking model; A sample evaluation unit, used to call the coarse ranking model to evaluate each sample in the full sample set, obtain a score corresponding to each sample, and sort each sample according to the score; An interval construction unit is used to construct segmented sampling intervals with increasing sample quantities from low to high scores; A sample extraction unit, used to extract a specified number of samples from each of the segmented sampling intervals according to a sampling ratio as a modeling sample set for a refined ranking model; The second model training unit is used to train a second marketing model according to the modeling sample set as a refined financial marketing model.

8. A server comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images