Implementation method and system for calculating core deposit proportion of active deposit
Through the multivariate linear regression model, the misleading problem of ignoring statistical characteristics and correlation coefficients in the existing technology is solved, and a more accurate calculation of core deposit ratios is achieved, and a more objective impact of market interest rates is provided. analyze.
Patent Information
- Application Number
- CN202411023528.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-29
- Publication Date
- 2024-12-06
AI Technical Summary
The existing deposit stability analysis technology ignores the statistical characteristics of real data in practical applications, and quantifies the impact of market interest rates on deposits by calculating the correlation coefficient between execution interest rates and market interest rates, especially in the variable relationship. When linear, the result is distorted.
Using a multivariate linear regression model, the correlation between feature factors is calculated by determining the stable churn rate and feature factors, cleaning and normalizing the processing data, verifying feature factors using a collinear diagnostic tool, and fitting feature factors using a multivariate linear regression model. The relationship with stable deposits can then calculate the core deposit ratio to provide a more accurate impact analysis.
Achieve more objectively quantifying the impact of different factors on stable deposits, especially the impact of market interest rates on stable deposits, providing a more accurate and objective calculation of core deposit ratios, solving the problem of the degree of closeness instead of the real impact, and making full use of the Statistics of the true churn information in the sequence data.
Smart Images

Figure CN119090608A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of computer technology, and in particular to a method and system for calculating the core deposit ratio of demand deposits, a computer-readable storage medium, an electronic device and a program product. Background Art
[0002] In the financial field, deposits are one of the core resources for banks and other financial institutions. For banks, being able to accurately assess and predict the stability of deposits is crucial to the formulation of their business strategies and risk management. To this end, the industry often uses various mathematical models and technical means to analyze deposit data in order to find a more scientific way to quantify and manage the stability of deposits.
[0003] Trend fitting technology is a widely used analytical method that simulates the trend of data changes over time through mathematical and statistical models. These models include but are not limited to the minimum balance method, time series models (such as STL, ARIMA, etc.), and machine learning models that have emerged in recent years (such as multivariate linear regression, tree model, etc.). Through these models, analysts can conduct in-depth mining of deposit data and extract the laws and trends hidden behind the data.
[0004] Residual analysis is an important part of evaluating the effect of trend fitting. When we subtract the original sequence value from the fitted trend item obtained by the trend fitting method, the result is called residual. Ideally, if the fitting effect is good, the residual should show randomness, that is, there is no obvious trend or pattern. Then, by analyzing these residuals, the confidence of the fitted trend item can be further evaluated, so as to obtain a more accurate stable deposit ratio.
[0005] Market interest rates are an important factor affecting deposit stability. Fluctuations in market interest rates may lead to deposit flows, which in turn affect the bank's funding situation. Therefore, banks usually consider the impact of market interest rates when analyzing deposit stability. The execution rate is the interest rate that the bank actually pays to customers, and the difference between it and the market rate can reflect the impact of market conditions on bank deposits to a certain extent. By calculating the Pearson correlation coefficient between the execution rate and the market rate, banks can try to quantify the extent of this impact, although this method has certain limitations.
[0006] However, the existing analysis techniques have some obvious deficiencies in practical applications. First, the method of calculating the stable deposit ratio by fitting the confidence distribution of the residual ignores the statistical characteristics of the real data. Although this method seems feasible in theory, it may lead to distorted results in practical applications due to the non-normality of data distribution or other complex factors.
[0007] Secondly, there are also problems in quantifying the impact of market interest rates on stable deposits by calculating the correlation coefficient between the execution rate and the market interest rate. Although the correlation coefficient can describe the closeness between two variables, it cannot accurately quantify the true impact of one variable on another. In particular, when the relationship between the two variables is nonlinear, simply using the correlation coefficient may lead to misleading conclusions. Summary of the invention
[0008] In order to address the deficiencies of the prior art, the present invention introduces a multivariate linear regression model, which can more objectively quantify and analyze the impact of different factors on stable deposits, especially the impact of market interest rates on stable deposits. Compared with the traditional correlation coefficient calculation method, this solution provides a more accurate method for calculating the core deposit ratio, solving the problem of replacing the real impact with the closeness. In view of the assumption that the calculation of stable deposits relies on the residual as a stationary sequence, it is proposed to calculate the stable deposits at a point in time based on the real churn rate of the balance sequence, making full use of the real churn information statistics in the sequence data, which is more authentic and explanatory, making the calculation of the core deposit ratio more accurate and objective.
[0009] To achieve the above objectives, the technical solutions adopted by the present invention include:
[0010] A method for calculating the core deposit ratio of demand deposits, characterized by comprising:
[0011] S1. Determine the stable churn rate and stable deposits, calculate the churn rate at each window time point in the deposit sequence, sort the churn rates at the window time points, determine the stable churn rate and calculate the stable deposits;
[0012] S2. Determine the characteristic factors that affect the stable churn rate, and clean and normalize the data of the characteristic factors;
[0013] S3. Calculate the correlation coefficients between different characteristic factors, retain one of the characteristic factor pairs with higher correlation, and use the collinearity diagnostic tool to verify the collinearity of the remaining characteristic factors. According to the collinearity verification results, remove some of the remaining characteristic factors with high collinearity to obtain the remaining characteristic factors;
[0014] S4. fitting the relationship between the remaining characteristic factors and the stable deposits using a multiple linear regression model, and obtaining the regression coefficients of the remaining characteristic factors and the stable deposits by training the multiple linear regression model;
[0015] S5. Determine the core deposit ratio, remove the influence of the remaining characteristic factors at each window time point to obtain the core deposits at all the window time points, take the median of the core deposits at the window time points and divide the median by the original deposit value to obtain the core deposit ratio.
[0016] Furthermore, the churn rate at the window time point is an aggregated result of the churn rate at each time point.
[0017] Specifically, the churn rate values in each window are aggregated, such as calculating the average or median. To ensure that extreme cases are considered, a high quantile (such as 95% or 99%) is selected as the representative churn rate in the window. The churn rates of all windows are counted and sorted.
[0018] Furthermore, the data cleaning of S2 includes but is not limited to processing missing values, outliers and duplicate values.
[0019] Furthermore, if the correlation coefficient of the characteristic factor is within a set threshold range, one of the characteristic factor pairs is retained.
[0020] Furthermore, the multivariate linear regression model is trained using MSE as a loss function to obtain the regression coefficient between each of the characteristic factors and the stable deposit.
[0021] Furthermore, the collinearity diagnostic tool includes but is not limited to variance inflation factor (VIF).
[0022] The present invention also relates to a system for calculating the core deposit ratio of demand deposits, which is characterized by comprising:
[0023] A stability coefficient calculation module, used to calculate the churn rate at each window time point in the deposit sequence, sort the churn rates at the window time points, determine the stable churn rate and calculate the stable deposit;
[0024] A preprocessing characteristic factor module, used to determine the characteristic factors that affect the stable churn rate, and clean and normalize the data of the characteristic factors;
[0025] A characteristic factor screening module is used to calculate the correlation coefficients between different characteristic factors, retain one of the characteristic factor pairs with higher correlation, and use a collinearity diagnostic tool to verify the collinearity of the remaining characteristic factors. According to the collinearity verification results, some of the remaining characteristic factors with high collinearity are removed to obtain the remaining characteristic factors;
[0026] A regression coefficient acquisition module, used for fitting the relationship between the remaining characteristic factors and the stable deposits by using a multiple linear regression model, and acquiring the regression coefficients of the remaining characteristic factors and the stable deposits by training the multiple linear regression model;
[0027] The core deposit ratio determination module is used to remove the influence of the remaining characteristic factors at each window time point to obtain the core deposits at all the window time points, take the median of the core deposits at the window time points and divide the median by the original deposit value to obtain the core deposit ratio.
[0028] The present invention also provides a computer-readable storage medium, the characteristic factor of which is that a computer program is stored on the storage medium, and when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.
[0029] The present invention also provides an electronic device, the characteristic factor of which is that it includes a processor and a memory;
[0030] The memory is used to store computer programs;
[0031] The processor is configured to execute the method according to any one of claims 1 to 6 by calling a stored computer program.
[0032] The present invention also relates to a computer program product, comprising a computer program and / or instructions, wherein the computer program and / or instructions implement the steps of the above method when executed by a processor.
[0033] The beneficial effects of the present invention are as follows: by introducing a multivariate linear regression model, the present invention can more objectively quantify and analyze the influence of different factors on stable deposits, especially the influence of market interest rates on stable deposits. Compared with the traditional correlation coefficient calculation method, this solution provides a more accurate method for calculating the core deposit ratio, solving the problem of replacing the real influence with the closeness. In view of the assumption that the calculation of stable deposits relies on the residual as a stationary sequence, it is proposed to calculate the stable deposits at a point in time according to the real churn rate of the balance sequence, making full use of the real churn information statistics in the sequence data, which is more authentic and explanatory, making the calculation of the core deposit ratio more accurate and objective. BRIEF DESCRIPTION OF THE DRAWINGS
[0034] Figure 1 The present invention is a flowchart of a method for calculating the core deposit ratio of demand deposits.
[0035] Figure 2 The present invention is a schematic diagram of the system structure for calculating the core deposit ratio of demand deposits. DETAILED DESCRIPTION
[0036] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.
[0037] The following specific embodiments may be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.
[0038] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings. At the same time, in the description of the present invention, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.
[0039] The first aspect of the present invention relates to a process of steps as follows Figure 1 The method shown comprises:
[0040] S1. Determine the stable churn rate and stable deposits, calculate the churn rate at each window time point in the deposit sequence, sort the churn rates at the window time points, determine the stable churn rate and calculate the stable deposits;
[0041] Specifically, for each time point value in the sequence, calculate whether there is subsequent loss. If there is loss, calculate the loss value, if there is no loss, calculate it as zero. If the loss rate of the first point after the time point is calculated as the loss rate of the window of 1 under time point i, if the nth point after the time point, calculate it as the loss rate of the window of n under time point i, where the number of time points is L, the number of windows is L-1, and L represents the sequence length. Stable deposit calculation y_stable = time point balance * (1-stable loss rate).
[0042] loop_1: traverse the balance sequence V in chronological order;
[0043] Get the information Vi at each time point;
[0044] loop_2: traverse the window sequence [1,L] from small to large;
[0045] The current window length is k;
[0046] Calculate the minimum value of the subsequent sequence of length k at each time point Vi as M(i,k);
[0047] Calculate the difference DIFF(i,k) of VM(i,k), and set it to 0 if it is less than 0;
[0048] Calculate r(i,k)=DIFF(i,k) / Vi, which represents the churn rate of anchor point Vi in the kth month;
[0049] loop_2:end;
[0050] loop_1:end;
[0051] loop_3: traverse the window sequence [1,L] from small to large;
[0052] The current window length is k, which indicates the kth month of churn. r(k) indicates the churn rate under the k window. SUM DIFF is defined to indicate the cumulative sum of churn values, with an initial value of 0. SUM V is defined to indicate the sum of anchor points, with an initial value of 0.
[0053] If the mean churn rate is calculated:
[0054] loop_4: traverse DIFF(1,k), the value range of i is [1,L];
[0055] SUM DIFF+=(values greater than 0 in DIFF(i,k));
[0056] SUM V+=(the anchor point Vi corresponding to the value greater than 0 in DIFF(i,k));
[0057] Mean churn rate calculation: r(k) = SUM_DIFF / SUM_VELSE;
[0058] if quantile churn rate: quantile churn rate calculation: r(k) = quantile function ([r(1,k),…,ri,k),…,p), where p represents the quantile, such as 95%;
[0059] loop_3:end;
[0060] Calculate the final churn rate: Get a stable point in the churn rate sequence [r(1), …, rk), …]. After this point appears, the value in the subsequent sequence does not change as the window becomes larger.
[0061] S2. Determine characteristic factors that affect the stable churn rate, and clean and normalize the data of the characteristic factors.
[0062] Specifically, we identify various factors that affect stable deposits (such as product interest rates, market interest rates, market competition environment, GDP, housing price index, etc.), and clean the data of these factors, including processing missing values, outliers, and duplicate values. We normalize or standardize the data so that we can compare the impact of different factors on stable deposits in subsequent analysis.
[0063] S3. Calculate the correlation coefficients between different characteristic factors, retain one of the characteristic factor pairs with higher correlation, and use the collinearity diagnostic tool to verify the collinearity of the remaining characteristic factors. According to the collinearity verification results, remove some of the remaining characteristic factors with high collinearity to obtain the remaining characteristic factors.
[0064] Specifically, the Pearson correlation coefficient between different factors is calculated to measure the degree of linear correlation between them. For feature pairs with correlation coefficient values that are too high (such as greater than 0.9) or too low (such as less than -0.9), only one of the features is retained to avoid the impact of multicollinearity on the regression model. Use the variance inflation factor (VIF) or other collinearity diagnostic tools to check the collinearity between the remaining features and remove those with high collinearity to ensure the stability and accuracy of the regression model.
[0065] However, in most cases, in order to ensure the stability and accuracy of the regression model, it is preferred to remove some of the features with high collinearity to reduce the redundancy and correlation between the features. This can avoid problems such as overfitting, inaccurate parameter estimation, or difficulty in interpretation.
[0066] Let X represent the characteristic factor sequence, the number of characteristic factors is n, x(i) represents the i-th characteristic factor sequence, the characteristic factors include market interest rate (shibor overnight rate), M2, GDP, unemployment rate, etc., let Y_stable represent the stable deposit sequence, y_stable(k) represents the stable deposit value at time point k, and the items that must be retained when setting the characteristic factors.
[0067] loop1: correlation analysis - traverse the characteristic factor sequence, convenient range [1, n];
[0068] Let the selected characteristic factor be numbered i;
[0069] loop2: correlation analysis traverses the characteristic factor sequence, traversing the range (1, n];
[0070] Let's assume the selected characteristic factor number is j;
[0071] Calculate the Pearson correlation coefficient p(i,j) between the sequences x(1) and x(j);
[0072] Ifp(ij)>threshold, check whether xi and xj are items that must be retained, if not, only one of them is retained;
[0073] loop2:end;
[0074] loop1:end;
[0075] Obtain the feature factor set X_hter after correlation filtering.
[0076] S4. Use a multiple linear regression model to fit the relationship between the remaining characteristic factors and the stable deposits, and obtain the regression coefficients of the remaining characteristic factors and the stable deposits by training the multiple linear regression model.
[0077] Specifically, a multivariate linear regression model is used to fit the relationship between the preprocessed characteristic factors and stable deposits. The mean square error (MSE) is used as the loss function to optimize the fitting effect of the model. The regression coefficients between each characteristic factor and stable deposits are obtained through training, and these coefficients reflect the degree of influence of different factors on stable deposits. The obtained regression coefficients are normalized so that they range from 0 to 1. The normalized coefficients are easier to explain and compare the relative influence of different factors on stable deposits.
[0078] Suppose the feature set obtained after correlation filtering is X_hter, and the number of features is m;
[0079] loop1: collinearity analysis and traversal of feature sequence X_hter, traversal range [1,m];
[0080] Assume that the selected feature number is i; select features other than i, and use the multivariate linear regression equation to represent x_hlter(i), which is x_pred(i) to calculate the residual resid(i) = x_hlter(i) - x_pred(i), and perform normalization, and calculate the variance; calculate VIF (expansion factor) = 1 / (1-variance);
[0081] If the VIF (expansion factor) is greater than the threshold, it is considered that the characteristic factor i can be linearly represented by other characteristic factors, and the characteristic factor i is not retained, otherwise the characteristic factor i is retained. If the characteristic factor i is not retained, the Xhlter sequence is updated.
[0082] loop1:end;
[0083] The remaining X_hlter sequence is the final feature sequence X_hinal.
[0084] S5. Determine the core deposit ratio, remove the influence of the remaining characteristic factors at each window time point to obtain the core deposits at all the window time points, take the median of the core deposits at the window time points and divide the median by the original deposit value to obtain the core deposit ratio.
[0085] Specifically, for each time point, if we need to calculate the impact of the market interest rate on the stable deposits at that time point, we can calculate the product of the market interest rate and its regression coefficient. Subtract the impact of the market interest rate from the stable deposits at that time point to get the core deposits. The core deposits represent the part of the stable deposits that is not affected by the market interest rate.
[0086] Finally, the core deposits at each time point are divided by the original deposit value at the corresponding time point to obtain the core deposit ratio, and the median of these ratios is taken as the core deposit ratio of the entire deposit series.
[0087] Assume that the number of features of the remaining feature factor sequence X_hnal is m;
[0088] Through multivariate linear regression fitting sequence (X_hnal, Y_stable), the regression coefficient r(m) under different characteristics is obtained after fitting. Let the regression coefficient of various characteristic factors (Shibor overnight rate) be r(m1). For the stable value y_stable(i) at a time point, after removing the impact of characteristic factors such as market interest rate, it is the value at that time point that is not affected by various characteristic factors, that is, the core deposit at that time point. Core deposit (i) = y_stable(i) - absolute value (regression coefficient is r(m1)*x_final[m1](i)) Core deposit ratio (i) = core deposit (i) / sequence time point value V(i), and get the core deposit ratio sequence K. Take the median of the core deposit ratio sequence K to get the final core deposit ratio.
[0089] Another aspect of the present invention also relates to a system for calculating the core deposit ratio of demand deposits, the structure of which is as follows: Figure 2 As shown, including:
[0090] A stability coefficient calculation module, used to calculate the churn rate at each window time point in the deposit sequence, sort the churn rates at the window time points, determine the stable churn rate and calculate the stable deposit;
[0091] A preprocessing characteristic factor module, used to determine the characteristic factors that affect the stable churn rate, and clean and normalize the data of the characteristic factors;
[0092] A characteristic factor screening module is used to calculate the correlation coefficients between different characteristic factors, retain one of the characteristic factor pairs with higher correlation, and use a collinearity diagnostic tool to verify the collinearity of the remaining characteristic factors. According to the collinearity verification results, some of the remaining characteristic factors with high collinearity are removed to obtain the remaining characteristic factors;
[0093] A regression coefficient acquisition module, used for fitting the relationship between the remaining characteristic factors and the stable deposits by using a multiple linear regression model, and acquiring the regression coefficients of the remaining characteristic factors and the stable deposits by training the multiple linear regression model;
[0094] The core deposit ratio determination module is used to remove the influence of the remaining characteristic factors at each window time point to obtain the core deposits at all the window time points, take the median of the core deposits at the window time points and divide the median by the original deposit value to obtain the core deposit ratio.
[0095] By using this system, the above-mentioned operation processing method can be executed and the corresponding technical effects can be achieved.
[0096] An embodiment of the present invention also provides a computer-readable storage medium capable of implementing all the steps in the method in the above embodiment, wherein a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, all the steps in the method in the above embodiment are implemented.
[0097] An embodiment of the present invention also provides an electronic device for executing the above method. As an implementation device of the method, the electronic device has at least a processor and a memory, and in particular, the memory stores data and related computer programs required for executing the method, such as storing set calculation programs, etc., and the processor calls the data and programs in the memory to execute all steps of the implementation method and obtains the corresponding technical effect.
[0098] Preferably, the electronic device may include a bus architecture, the bus may include any number of interconnected buses and bridges, and the bus will include various circuits linked together by one or more processors and memories. The bus may also link together various other circuits such as peripherals, voltage regulators, and power management circuits, which are all well known in the art and are therefore not further described herein. The bus interface provides an interface between the bus and the receiver and transmitter. The receiver and transmitter may be the same component, i.e., a transceiver, providing a unit for communicating with various other systems over a transmission medium. The processor is responsible for managing the bus and general processing, while the memory may be used to store data used by the processor when performing operations.
[0099] In addition, the electronic device may further include components such as a communication module, an input unit, an audio processor, a display, and a power supply. The processor (or controller, operation control) used may include a microprocessor or other processor device and / or a logic device, which receives input and controls the operation of various components of the electronic device; the memory may be one or more of a cache, a flash memory, a hard drive, a removable medium, a volatile memory, a non-volatile memory, or other suitable devices, which may store the above-mentioned related data information, and may also store a program for executing related information, and the processor may execute the program stored in the memory to realize information storage or processing, etc.; the input unit is used to provide input to the processor, for example, it may be a key or a touch input device; the power supply is used to provide power to the electronic device; the display is used to display display objects such as images and text, for example, it may be an LCD display. The communication module is a transmitter / receiver that sends and receives signals via an antenna. The communication module (transmitter / receiver) is coupled to the processor to provide input signals and receive output signals, which may be the same as the case of a conventional mobile communication terminal. Based on different communication technologies, multiple communication modules may be provided in the same electronic device, such as a cellular network module, a Bluetooth module and / or a wireless local area network module, etc. The communication module (transmitter / receiver) is also coupled to a speaker and a microphone via an audio processor to provide an audio output via the speaker and receive an audio input from the microphone, thereby realizing a common telecommunication function. The audio processor may include any suitable buffer, decoder, amplifier, etc. In addition, the audio processor is also coupled to a central processor, so that recording can be performed on the machine through the microphone, and the sound stored on the machine can be played through the speaker.
[0100] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0101] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A system that specifies the functions of a box or multiple boxes.
[0102] These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a product including an instruction system, which is implemented in the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0103] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps of the functions specified in a box or multiple boxes. Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they know the basic creative concept. Therefore, the attached claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.
[0104] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.
Claims
1. A method for calculating the core deposit ratio of demand deposits, characterized in that: include: S1. Determine the stable churn rate and stable deposits, calculate the churn rate at each window time point in the deposit sequence, sort the churn rates at the window time points, determine the stable churn rate and calculate the stable deposits; S2. Determine the characteristic factors that affect the stable churn rate, and clean and normalize the data of the characteristic factors; S3. Calculate the correlation coefficients between different characteristic factors, retain one of the characteristic factor pairs with higher correlation, and use the collinearity diagnostic tool to verify the collinearity of the remaining characteristic factors. According to the collinearity verification results, remove some of the remaining characteristic factors with high collinearity to obtain the remaining characteristic factors; S4. fitting the relationship between the remaining characteristic factors and the stable deposits using a multiple linear regression model, and obtaining the regression coefficients of the remaining characteristic factors and the stable deposits by training the multiple linear regression model; S5. Determine the core deposit ratio, remove the remaining characteristic factors at each window time point to obtain the core deposits at all window time points, take the median of the core deposits at the window time points and divide the median by the original deposit value to obtain the core deposit ratio.
2. The method according to claim 1, characterized in that The churn rate at the window time point is the aggregated result of the churn rate at each time point.
3. The method according to claim 1, characterized in that The data cleaning of S2 includes but is not limited to processing missing values, outliers and duplicate values.
4. The method according to claim 1, characterized in that If the correlation coefficient of the characteristic factor is within a set threshold range, one of the characteristic factor pairs is retained.
5. The method according to claim 1, characterized in that The multivariate linear regression model is trained using MSE as a loss function to obtain the regression coefficient between each characteristic factor and the stable deposit.
6. The method according to claim 1, characterized in that The collinearity diagnostic tools include, but are not limited to, variance inflation factor (VIF).
7. A system for calculating the core deposit ratio of demand deposits, characterized in that: include: A stability coefficient calculation module, used to calculate the churn rate at each window time point in the deposit sequence, sort the churn rates at the window time points, determine the stable churn rate and calculate the stable deposit; A preprocessing characteristic factor module, used to determine the characteristic factors that affect the stable churn rate, and clean and normalize the data of the characteristic factors; A characteristic factor screening module is used to calculate the correlation coefficients between different characteristic factors, retain one of the characteristic factor pairs with higher correlation, and use a collinearity diagnostic tool to verify the collinearity of the remaining characteristic factors. According to the collinearity verification results, some of the remaining characteristic factors with high collinearity are removed to obtain the remaining characteristic factors; A regression coefficient acquisition module, used for fitting the relationship between the remaining characteristic factors and the stable deposits by using a multiple linear regression model, and acquiring the regression coefficients of the remaining characteristic factors and the stable deposits by training the multiple linear regression model; The core deposit ratio determination module is used to remove the influence of the remaining characteristic factors at each window time point to obtain the core deposits at all the window time points, take the median of the core deposits at the window time points and divide the median by the original deposit value to obtain the core deposit ratio.
8. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.
9. An electronic device, whose characteristic factor is that including a processor and a memory; The memory is used to store computer programs; The processor is configured to execute the method according to any one of claims 1 to 6 by calling a stored computer program.
10. A computer program product comprising a computer program and / or instructions, characterized in that: When the computer program and / or the instructions are executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.
