Currency resource processing method and device, storage medium and electronic equipment

Abstract

The invention discloses a currency resource processing method and device, a storage medium and electronic equipment. The method comprises the steps that a transaction data set is determined according to an enterprise number, and the transaction data set is used for indicating transaction information corresponding to the enterprise number; the transaction data set is extracted, a reference transaction feature set is obtained, a target transaction feature set is determined from the reference transaction feature set based on the enterprise number, and transaction features are used for indicating features of the transaction data set; and inputting the target transaction feature set into a quota prediction model to obtain a paid quota corresponding to the enterprise number, and determining a processing proportion of the currency resources according to the paid quota and the transaction data set, the quota prediction model being matched with the enterprise number. Through the method and the device, the problem of relatively low accuracy of a currency resource processing proportion confirmation method in related technologies is solved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence, and in particular, to a method and device for processing currency resources, a storage medium, and an electronic device. Background Art

[0002] Currently, when a bank is dealing with currency resources (overdue currency resources) that cannot be returned in the predetermined amount within the predetermined period, it can dispose of the overdue currency resources by reducing or canceling the overdue currency resources, so as to quickly resolve the overdue currency resources and recover the reserve funds.

[0003] If the processing ratio of the overdue currency resources is too low, it will increase the burden on the debtor, making it unable to repay and the overdue currency resources cannot be recovered. In addition, reducing or canceling the overdue currency resources will reduce the bank's interest income. If the processing ratio of the currency resources is too high, it will cause losses to the bank's income. Therefore, it is important to determine a reasonable processing ratio. However, the current method for determining the processing ratio of currency resources is confirmed by manual experience, with a large degree of subjectivity and unable to give a reasonable processing ratio.

[0004] Aiming at the problem of low accuracy of the method for confirming the processing ratio of currency resources in the related art, no effective solution has been proposed yet. Summary of the Invention

[0005] The main purpose of the present application is to provide a method and device for processing currency resources, a storage medium, and an electronic device, so as to solve the problem of low accuracy of the method for confirming the processing ratio of currency resources in the related art.

[0006] To achieve the above purpose, according to one aspect of the present application, a method for processing currency resources is provided. The method includes: determining a transaction data set according to the enterprise number, where the transaction data set is used to indicate the information of the transaction corresponding to the enterprise number; extracting the transaction data set to obtain a reference transaction feature set, and determining a target transaction feature set from the reference transaction feature set based on the enterprise number, where the transaction feature is used to indicate the feature of the transaction data set; inputting the target transaction feature set into an amount prediction model to obtain the recoverable amount corresponding to the enterprise number, and determining the processing ratio of the currency resources according to the recoverable amount and the transaction data set, where the amount prediction model is matched with the enterprise number.

[0007] To achieve the above object, according to another aspect of the present application, there is provided a processing device for monetary resources. The device includes: a data set determination unit configured to determine a transaction data set according to an enterprise number, where the transaction data set is used to indicate information on transactions corresponding to the enterprise number; a feature extraction unit configured to extract the transaction data set to obtain a reference transaction feature set, and determine a target transaction feature set from the reference transaction feature set based on the enterprise number, where the transaction feature is used to indicate the feature of the transaction data set; a ratio determination unit configured to input the target transaction feature set into a quota prediction model to obtain the receivable quota corresponding to the enterprise number, and determine the processing ratio of the monetary resources according to the receivable quota and the transaction data set, where the quota prediction model is matched with the enterprise number.

[0008] Optionally, the above feature extraction unit includes: an input feature determination module configured to determine a quota prediction model matched with the enterprise number and determine the input features of the quota prediction model; a target feature determination module configured to determine the transaction features corresponding to the input features in the reference transaction feature set as target transaction features to obtain a target transaction feature set.

[0009] Optionally, the above feature extraction unit further includes: a feature extraction module configured to extract a historical transaction data set to obtain a first transaction feature set, where the historical transaction data set is matched with the enterprise number; a feature screening module configured to screen out a second transaction feature set from the first transaction feature set, where the second transaction feature set is related to the receivable quota corresponding to the enterprise number; a feature confirmation module configured to confirm a third transaction feature set from the second transaction feature set; a model determination module configured to determine a quota prediction model according to the third transaction feature set.

[0010] Optionally, the above-mentioned feature screening module is further configured to: cluster the first transaction feature set to obtain M categories of transaction feature sets; sequentially determine the M categories of transaction feature sets as the Pth transaction feature set and perform the following operations: determine the difference between the ith transaction feature in the Pth transaction feature set and the mean of the Pth transaction feature set as the ith first difference, and determine the difference between the ith compensation amount corresponding to the ith transaction feature and the mean of the compensation amounts as the ith second difference, where each category of transaction feature set includes n transaction features; determine the product of the n first differences and the second differences as the first product, and determine the square root of the product of the sum of the squares of the n first differences and the sum of the squares of the n second differences as the second product; determine the ratio of the first product to the second product as the correlation coefficient of the Pth transaction feature set, and in the case where the correlation coefficient is greater than a preset coefficient threshold, incorporate the Pth transaction feature set into the second transaction feature set, where M is an integer greater than 0, P is an integer greater than 0 and less than or equal to M, i is an integer greater than 0 and less than or equal to n, and n is an integer greater than 0.

[0011] Optionally, the above-mentioned feature confirmation module is further configured to: sequentially determine the M categories of transaction feature sets as the Pth transaction feature set and perform the following operations: determine the statistical coefficient corresponding to the Pth transaction feature set according to the correlation coefficient of the Pth transaction feature set and the number of transaction features included in each category of transaction feature set; determine the statistical probability corresponding to the Pth transaction feature set based on the statistical coefficient; and in the case where the statistical probability corresponding to the Pth transaction feature set is less than the probability threshold, incorporate the Pth transaction feature set into the third transaction feature set.

[0012] Optionally, the above-mentioned model determination module is further configured to: determine the number of calculation coefficients included in the calculation coefficient set based on the number of categories of transaction feature sets included in the third transaction feature set; and use the third transaction feature set to determine the calculation coefficient set to obtain the amount prediction model.

[0013] In an embodiment of the present application, a transaction data set is determined according to an enterprise number, where the transaction data set is used to indicate information about transactions corresponding to the enterprise number; the transaction data set is extracted to obtain a reference transaction feature set, and a target transaction feature set is determined from the reference transaction feature set based on the enterprise number, where transaction features are used to indicate features of the transaction data set; the target transaction feature set is input into a quota prediction model to obtain the receivable quota corresponding to the enterprise number, and a processing ratio of monetary resources is determined according to the receivable quota and the transaction data set, where the quota prediction model is matched with the enterprise number. For different enterprises, their transaction data can be determined according to the enterprise number, the transaction features therein can be determined, and the corresponding target transaction features can be selected from the transaction features based on the enterprise number to accurately calculate the receivable quota and processing ratio corresponding to the enterprise. Furthermore, the technical problem of low accuracy of the processing ratio caused by manually determining the processing ratio of monetary resources is solved. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The drawings constituting a part of this application are used to provide a further understanding of this application. The schematic embodiments and descriptions thereof of this application are used to explain this application and do not constitute an improper limitation of this application. In the drawings:

[0015] Figure 1 The hardware structure block diagram of a computer terminal for implementing a method for processing monetary resources is shown;

[0016] Figure 2 is a flowchart of a method for processing monetary resources according to an embodiment of this application;

[0017] Figure 3 is a schematic diagram of a device for processing monetary resources according to an embodiment of this application;

[0018] Figure 4 is a structure block diagram of an electronic device according to an embodiment of this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] In order to enable those skilled in the art to better understand the solution of this application, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of this application.

[0020] It should be noted that the terms "first", "second", etc. in the description, claims and the above-mentioned drawings of this application are used to distinguish similar objects, and do not necessarily describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of this application described here can be implemented in an order other than those illustrated or described here. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or are inherent to these processes, methods, products or devices.

[0021] It should be noted that the information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for display, data for analysis, etc.) involved in this application are information and data that have been authorized by the user or fully authorized by all parties. Moreover, for the processing of relevant data such as collection, storage, use, processing, transmission, provision, disclosure and application, all comply with relevant laws, regulations and standards, adopt necessary confidentiality measures, do not violate public order and good customs, and provide corresponding operation entrances for users to choose to authorize or refuse. For example, there is an interface between this system and relevant users or institutions, providing corresponding operation entrances for users to choose to agree or refuse the results of automated decision-making; if the user chooses to refuse, the expert decision-making process will be entered.

[0022] Embodiment 1

[0023] According to the embodiments of this application, an embodiment of a method for processing currency resources is also provided. It should be noted that the steps shown in the flowchart of the drawings can be executed in a computer system such as a set of computer-executable instructions. And although the logical order is shown in the flowchart, in some cases, the steps shown or described here can be executed in a different order than here.

[0024] The method embodiment provided by the first embodiment of this application can be executed on a mobile terminal, a computer terminal or a similar computing device. Figure 1 The hardware structure block diagram of a computer terminal (or mobile device) for implementing the method for processing currency resources is shown. As Figure 1As shown, the computer terminal 10 (or mobile device) may include one or more processors 102 (shown as 102a, 102b, ……, 102n in the figure) (the processor 102 may include, but is not limited to, a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may further include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the BUS bus), a network interface, a power supply, and / or a camera. Those of ordinary skill in the art can understand that Figure 1 the structure shown is only schematic and does not limit the structure of the above-mentioned electronic device. For example, the computer terminal 10 may further include more or fewer components than those Figure 1 shown in, or have a different configuration from that Figure 1 shown.

[0025] It should be noted that the above one or more processors 102 and / or other data processing circuits can generally be referred to as "data processing circuits" herein. The data processing circuit may be embodied in software, hardware, firmware, or any combination thereof, in whole or in part. In addition, the data processing circuit may be a single independent processing module, or be incorporated in whole or in part into any one of the other elements in the computer terminal 10 (or mobile device). As involved in the embodiments of the present application, the data processing circuit is a processor control (such as the selection of a variable resistance terminal path connected to an interface).

[0026] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the processing method of currency resources in the embodiments of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, implements the above-mentioned processing method of currency resources. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memories, or other non-volatile solid-state memories. In some instances, the memory 104 may further include a memory remotely set relative to the processor 102, and these remote memories can be connected to the computer terminal 10 through a network. Examples of the above network include, but are not limited to, the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0027] The transmission device 106 is used to receive or send data via a network. Specific examples of the above-mentioned network may include a wireless network provided by the communication provider of the computer terminal 10. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a Radio Frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0028] The display can be, for example, a touch-screen liquid crystal display (LCD), which enables users to interact with the user interface of the computer terminal 10 (or mobile device).

[0029] Under the above operating environment, this application provides a method for processing currency resources as Figure 2 shown. Figure 2 It is a flowchart of the method for processing currency resources according to Embodiment 1 of this application.

[0030] Step S101, determine a transaction data set according to the enterprise number, where the transaction data set is used to indicate information about transactions corresponding to the enterprise number;

[0031] It should be noted that the enterprise number can be a digital or alphabetic code used to uniquely identify an enterprise. In the system of a bank or financial institution, each enterprise has a specific enterprise number for tracking and managing all transactions and data related to that enterprise. The transaction data set can refer to the set of all transaction information associated with a specific enterprise number. This information may include details such as transaction date, transaction amount, transaction type (such as loan disbursement, receipt, interest payment), transaction account, etc. In data management and analysis, the transaction data set is an important basis for evaluating an enterprise's creditworthiness, financial health, and repayment ability. The transaction data set can include not only loan-type transactions but also other types of transactions to facilitate a complete evaluation of the enterprise later.

[0032] In an alternative embodiment, all relevant transaction information of the enterprise is retrieved and determined through the enterprise's unique identifier (enterprise number). This process is usually carried out in the database query or data screening link to ensure the accuracy and pertinence of the data for subsequent analysis.

[0033] When processing financial data related to a company's overdue monetary resources, the first step is to accurately extract all transaction data related to a specific company from the database. Because subsequent analysis depends on the integrity, accuracy, and timeliness of this transaction data. Through the company number, the system can quickly and precisely locate the company's transaction records, thus forming a transaction data set. This set contains detailed information on all the company's transactions, providing a solid data foundation for subsequent financial analysis, credit assessment, and the construction of prediction models.

[0034] In an alternative implementation, assume that a bank has a large-scale database storing transaction data related to all its customers. The bank is developing a system to predict the processing ratio of overdue monetary resources in order to improve the recovery rate of overdue monetary resources. Step S101 is implemented as follows: Determination of company number: The bank first determines a company number, such as "123456", which corresponds to a specific company with overdue monetary resources (such as having long-term loans, etc.). Extraction of data set: The system queries all relevant transactions in the database according to the company number "123456". Here, the transaction data set includes all transaction records of the company over a certain period in the past, such as loan disbursement, principal repayment, interest payment, changes in guarantee information, etc. Information integration: The system integrates the extracted transaction data to form a detailed transaction data set. This set contains information such as the date, amount, type, and account information of each transaction, which will be used for subsequent analysis and model construction, etc.

[0035] By implementing step S101, it is ensured that the starting point of the analysis is accurate and comprehensive, avoiding analysis biases caused by incomplete or incorrect data. Next, using this transaction data set, through factor analysis and machine learning algorithms, the reasonable processing ratio of monetary resources in the disposal of the company's overdue monetary resources can be predicted, thereby reducing the bank's financial losses while increasing the asset recovery rate.

[0036] In step S102, the transaction data set is extracted to obtain a reference transaction feature set, and a target transaction feature set is determined from the reference transaction feature set based on the company number, where the transaction feature is used to indicate the feature of the transaction data set;

[0037] It should be noted that the transaction feature can refer to specific attributes or indicators in the transaction data set that can describe the nature of the transaction and the company's financial status. The target transaction feature set can indicate a set of transaction features selected from the reference transaction feature set that are most relevant to the transaction information corresponding to the company number or can best reflect the company's credit status.

[0038] Optionally, the following reference transaction characteristics can be extracted for customer information: the proportion of the balance of normal-class loans in the total loan balance, the proportion of the balance of watch-class loans in the total loan balance, the proportion of the balance of substandard-class loans in the total loan balance, the proportion of the balance of doubtful-class loans in the total loan balance, the proportion of the balance of loss-class loans in the total loan balance, the proportion of overdue interest on on-balance-sheet loans in the total loan balance, the proportion of overdue interest on off-balance-sheet loans in the total loan balance, the proportion of the loan balance in the peak loan balance, the proportion of the principal recovered before transferring to non-performing loans in the total loan balance, the proportion of the principal recovered after transferring to non-performing loans in the total loan balance, the proportion of the total profit of the previous year in the total loan balance, the proportion of the cash inflow from operating activities of the previous year in the total loan balance, the proportion of the cash inflow from investing activities of the previous year in the total loan balance, the proportion of the cash inflow from financing activities of the previous year in the total loan balance, the proportion of the total effective liabilities in the total loan balance, etc.

[0039] Among them, they can be classified into those related to loan quality classification, overdue interest, other financial indicators, profitability, cash flow, and liabilities.

[0040] Those related to loan quality classification can include: the proportion of the balance of normal-class loans in the total loan balance. Normal-class loans refer to loans where the borrower can fulfill the contract and there is no sufficient reason to suspect that the principal and interest of the loan cannot be repaid on time and in full. This proportion reflects the proportion of the higher-quality part of the bank's loans; the proportion of the balance of watch-class loans in the total loan balance. Watch-class loans refer to loans where, although the borrower currently has the ability to repay the principal and interest of the loan, there are some factors that may have an adverse impact on repayment. This proportion is used to monitor the level of potential risks of the loan; the proportion of the balance of substandard-class loans in the total loan balance. Substandard-class loans refer to loans where the borrower's repayment ability has obvious problems and it is impossible to fully repay the principal and interest of the loan solely relying on its normal operating income. Even if the guarantee is enforced, certain losses may still be incurred; the proportion of the balance of doubtful-class loans in the total loan balance. Doubtful-class loans refer to loans where the borrower is unable to fully repay the principal and interest of the loan. Even if the guarantee is enforced, significant losses will definitely be incurred; the proportion of the balance of loss-class loans in the total loan balance. Loss-class loans refer to loans where, after taking all possible measures or all necessary legal procedures, the principal and interest still cannot be recovered, or only a very small part can be recovered.

[0041] Those related to overdue interest can include: the proportion of overdue interest on on-balance-sheet loans in the total loan balance. On-balance-sheet overdue interest refers to the accrued but unpaid interest accounted for in the bank's balance sheet, usually referring to overdue interest with a relatively short overdue period (generally not exceeding 90 days); the proportion of overdue interest on off-balance-sheet loans in the total loan balance. Off-balance-sheet overdue interest refers to overdue interest with a relatively long overdue period (exceeding 90 days) or overdue interest not included in the balance sheet for other reasons, usually reflecting the high-risk status of the loan.

[0042] Other financial indicators may include: the ratio of the loan balance to the peak loan balance, where the peak loan balance refers to the highest point of the bank's loan balance in history. This ratio is used to measure the proportion of the current loan balance relative to the historical highest level and reflects the changing trend of the bank's credit scale; the ratio of the principal recovered before becoming non-performing to the loan balance, where the principal recovered before becoming non-performing refers to the part of the principal that has been recovered before the loan is identified as a non-performing loan. This ratio is used to evaluate the bank's ability to recover the principal before the loan quality deteriorates; the ratio of the principal recovered after becoming non-performing to the loan balance, where the principal recovered after becoming non-performing refers to the part of the principal that has been recovered through collection and other means after the loan is identified as non-performing. This ratio reflects the bank's disposal effect on non-performing loans.

[0043] Those related to profitability may include: the ratio of the total profit of the previous year to the loan balance. This indicator is used to measure the relationship between the total profit of the bank in the previous year and the loan balance and reflects the matching degree between the bank's profitability and the credit scale.

[0044] Those related to cash flow may include: the ratio of the cash inflow from operating activities of the previous year to the loan balance. This indicator reflects the ratio of the cash inflow generated by the bank's operating activities in the previous year to the loan balance and is used to evaluate the cash flow status of the bank's operating activities; the ratio of the cash inflow from investing activities of the previous year to the loan balance. This indicator reflects the ratio of the cash inflow generated by the bank's investing activities in the previous year to the loan balance and is used to evaluate the contribution of the investing activities to the bank's funds; the ratio of the cash inflow from financing activities of the previous year to the loan balance. This indicator reflects the ratio of the cash inflow generated by the bank's financing activities in the previous year to the loan balance and is used to evaluate the strength of the bank's financing activities in providing fund support.

[0045] Those related to liabilities may include: the ratio of the total effective liabilities to the loan balance, where effective liabilities refer to the part of the bank's liabilities that can actually be used to support the loan business. This ratio is used to measure the balance relationship between the bank's liabilities and loans and reflects the stability of the bank's funding sources.

[0046] Regarding the guarantee information, the following reference transaction characteristics can be extracted: the ratio of the guarantee amount to the loan balance, the ratio of the total guarantee amount of the guarantor to the loan balance, the ratio of the total external guarantee amount of the guarantor to its net assets, etc.

[0047] Among them, the ratio of the amount guaranteed to the loan balance refers to the ratio of the amount covered by the guarantee method in the loan to the total loan balance. This indicator reflects the degree of protection provided by the guarantee method in the loan. The higher the ratio, the greater the proportion of the loan that relies on the guarantee, and it may be necessary to further evaluate the guarantee ability of the guarantor. The ratio of the total guarantee amount of the guarantor to the loan balance refers to the ratio of the total guarantee amount provided by a certain guarantor to the total loan balance. This indicator is used to evaluate the contribution degree of a certain guarantor to the loan. If the ratio is too high, it may mean that the loan overly relies on a single guarantor, resulting in concentration risk. The ratio of the total external guarantee amount of the guarantor to its net assets refers to the ratio of the total external guarantee amount provided by the guarantor to its net assets. This indicator is used to evaluate whether the guarantee ability of the guarantor exceeds its tolerance range. If the ratio is too high, it indicates that the guarantor may face high guarantee risks, and its net assets may not be sufficient to cover its external guarantee obligations. Generally speaking, the regulatory department or financial institution will set a reasonable upper limit (such as 50% or 60%) to ensure that the guarantee behavior of the guarantor is within its financial tolerance range.

[0048] Optionally, after determining the transaction data set, a series of reference transaction characteristics are extracted from these transaction data. This is a data preprocessing stage, aiming to identify which transaction characteristics are the most important for predicting the processing ratio of overdue monetary resources. Then, according to the enterprise number, a set of target transaction characteristics specific to the enterprise is further selected from these reference characteristics, and these characteristics will be directly used for subsequent analysis.

[0049] In an alternative embodiment, based on step S101, it is necessary to analyze the transaction data set more deeply to determine which transaction characteristics play a key role in predicting the processing ratio in the disposal of overdue monetary resources. This step S102 involves extracting a set of reference transaction characteristics from the original data, which is a large set containing all possible transaction characteristics. Then, through further analysis of the reference characteristic set and combined with the enterprise number, a more refined set of target transaction characteristics can be determined. This set contains those transaction characteristics that can best describe the credit status and financial capabilities of the enterprise and is an important input for subsequent predictive analysis.

[0050] In an alternative embodiment, as described above, in the overdue monetary resource disposal prediction system of a bank, after step S101 where the system has obtained the transaction data set of enterprise number "123456", step S102 is implemented: Reference transaction feature extraction: The system extracts a series of transaction features from the transaction data set, including but not limited to the proportion of normal loan balance, the proportion of watched loan balance, the proportion of interest in and out of the statement, the proportion of loan peak, the proportion of principal recovered, the ratio of profit cash inflow, etc. These features together constitute the reference transaction feature set. Target transaction feature determination: Next, based on the enterprise number "123456", the system further analyzes and filters the reference transaction feature set. Suppose the analysis result shows that the proportion of normal loan balance, the proportion of interest in and out of the statement, and the proportion of principal recovered are the most important features, then the system determines these features as the target transaction feature set.

[0051] Step S102 is the key to accurately predicting the disposal ratio in the overdue monetary resource disposal of the bank. By extracting and determining the target transaction feature set from the transaction data set, the bank ensures that the modeling process focuses on the financial indicators that have the greatest impact on the prediction result. This not only improves the accuracy of the prediction but also enhances the interpretability of the model, enabling the bank to make more informed decisions when disposing of overdue monetary resources.

[0052] In step S103, the target transaction feature set is input into the amount prediction model to obtain the recoverable amount corresponding to the enterprise number, and the disposal ratio of the monetary resource is determined based on the recoverable amount and the transaction data set, where the amount prediction model is matched with the enterprise number.

[0053] It should be noted that the amount prediction model is a model constructed based on machine learning or statistical analysis, used to predict the recoverable amount that can be brought by the transaction feature set related to the enterprise number. This model can predict the specific amount of loans or debts that an enterprise can repay under a given transaction feature set by learning the patterns in historical data. It can also be used as a model to predict the disposal ratio of applicable monetary resources in the disposal of overdue monetary resources. The model is based on the target transaction feature set and historical recoverable data, and through training, it learns the relationship between the disposal ratio of monetary resources and the financial condition of the enterprise, so as to predict the disposal ratio of monetary resources for the enterprise or transaction.

[0054] The amount prediction model can be corresponding to each enterprise individually, or corresponding to enterprises with similar businesses jointly, or all enterprises can share one model.

[0055] In an alternative embodiment, the set of target transaction characteristics determined in step S102 is input into a quota prediction model, which predicts the specific amount of the loan or debt that the enterprise can repay based on these characteristics, i.e., the receivable quota. Subsequently, based on the predicted receivable quota and the original transaction data set of the enterprise, the system or analyst calculates a reasonable processing ratio. The quota prediction model mentioned here is matched with the enterprise number, which means that the prediction of the model is carried out for the actual situation of a specific enterprise, taking into account the transaction characteristics and financial status of the enterprise.

[0056] In the prediction process of overdue monetary resource disposal, step S103 is a key decision support link. The system first uses the set of target transaction characteristics as input and predicts the loan or debt quota receivable by the enterprise through the quota prediction model. This prediction is based on the patterns of historical data and machine learning algorithms, and can provide a quantitative assessment of the enterprise's debt repayment ability. Then, the system calculates an appropriate processing ratio based on the predicted receivable quota and the enterprise's transaction data set. The quota prediction model is matched with the enterprise number because it takes into account the specific transaction characteristics and financial status of the enterprise, providing a personalized prediction result rather than a general average. This enables the bank to make accurate waiver decisions based on the actual situation of the enterprise when dealing with overdue monetary resources, ensuring both the recovery of overdue monetary resources and avoiding financial losses caused by excessive waivers.

[0057] In an alternative embodiment, assume that a bank is using a machine learning-based overdue monetary resource disposal prediction system that has completed steps S101 and S102, i.e., it has determined the transaction data set for enterprise number "123456" and extracted the target transaction feature set from it, including features such as the proportion of normal-class loan balance, the proportion of overdue interest inside and outside the statement, the proportion of principal recovered from collection, the ratio of profit cash inflow, etc. Step S103 is implemented as follows: Quota prediction: The system inputs the target transaction feature set into the quota prediction model. This model is trained based on historical data and can predict the loan or debt quota that enterprise "123456" can receive in the disposal of overdue monetary resources according to the input transaction features. Assume that the predicted receivable quota is 80% of the loan principal. Determination of the processing ratio: Based on the predicted receivable quota (80% principal) and the enterprise's transaction data set, the system calculates the processing ratio. For example, if the total loan principal of enterprise "123456" is 1 million yuan and the predicted receivable quota is 800,000 yuan, then the bank can, based on this prediction, set an appropriate processing ratio, such as reducing the interest by 20%, to increase the likelihood of receivables. Decision application: The bank adjusts the disposal strategy of overdue monetary resources according to the processing ratio obtained from the quota prediction model. For example, to encourage enterprise "123456" to actively repay its debts, the bank decides to grant a 20% interest reduction. In this way, the enterprise only needs to pay 800,000 yuan in principal plus 80% of the interest, reducing the enterprise's financial burden and increasing the likelihood of the bank recovering overdue monetary resources.

[0058] Through the above embodiments, not only can a prediction of the enterprise's receivables ability be obtained, but also a reasonable disposal strategy for monetary resources can be formulated based on this prediction. This enables the bank to more accurately balance risks and returns in the disposal of overdue monetary resources and improve the recovery efficiency of overdue monetary resources. By analyzing customer information, transaction features related to the receivable quota are extracted, and irrelevant transaction features are removed through analysis to improve the model accuracy.

[0059] Optionally, in the method for processing monetary resources provided in the embodiments of the present application, determining the target transaction feature set from the reference transaction feature set based on the enterprise number includes: determining the quota prediction model that matches the enterprise number and determining the input features of the quota prediction model; determining the transaction features corresponding to the input features in the reference transaction feature set as the target transaction features to obtain the target transaction feature set.

[0060] It should be noted that the quota prediction model is an analytical tool that predicts the expected quota of an enterprise's future repayment based on the enterprise's transaction characteristics through machine learning or statistical methods. Such a model usually needs to be trained to learn the relationship between transaction characteristics and repayment quotas, so as to make predictions for new enterprises or transactions. Input features refer to specific transaction characteristics required by the quota prediction model during operation, and these characteristics are considered to have an important impact on predicting the repayment quota. For example, loan balance, interest payment situation, enterprise cash inflow, etc. are all possible input features.

[0061] In the overdue monetary resource disposal prediction process, a quota prediction model matching the enterprise number can be determined, and this model is trained on historical data. Then, the input features required by the model for predicting the repayment quota are identified. Subsequently, the transaction characteristics matching these input features are screened out from the reference transaction feature set, and these characteristics form the target transaction feature set, which will be used for subsequent prediction analysis.

[0062] In an alternative implementation, assume that a bank is using a machine learning-based overdue monetary resource disposal system, and the goal of this system is to predict the loan or debt quota that the bank expects to be repaid during the disposal of the overdue monetary resources of enterprise number "7890". First, the quota prediction model can be determined: The system first determines a quota prediction model matching the enterprise number "7890". This model may be trained on a large number of historical overdue monetary resource disposal cases and can predict the future repayment quota based on the enterprise's transaction characteristics. Determine the input features: The system identifies the input features required by the quota prediction model for predicting the repayment quota, such as the ratio of loan balance to peak loan balance, the ratio of interest owed (both on and off the balance sheet) to loan balance, the ratio of historical principal recovered, etc. These features are considered key indicators affecting the enterprise's repayment ability. Construct the target transaction feature set: Based on the above-determined input features, the system screens out the transaction characteristics corresponding to the input features from the reference transaction feature set of enterprise number "7890". For example, considering the features of the ratio of loan balance to peak loan balance, the ratio of interest owed (both on and off the balance sheet) to loan balance, and the ratio of historical principal recovered, the system extracts these specific transaction characteristics from the reference set to construct the target transaction feature set. Make a prediction: Finally, the target transaction feature set is input into the quota prediction model, and the model predicts the possible repayment quota of enterprise "7890" during the disposal of overdue monetary resources based on these features. The bank can formulate a reasonable strategy for handling monetary resources based on this prediction, combined with the enterprise's actual transaction data, to improve the recovery rate of overdue monetary resources and control the bank's losses at the same time.

[0063] Through the above embodiments of the present application, it is ensured that only those transaction features that are most critical for predicting the repayment amount are retained, thereby improving the efficiency of the prediction model and the accuracy of the prediction results. The amount prediction model, as the core of this process, can learn and understand the complex relationship between transaction features and the repayment amount. By determining the input features of the model, the system can accurately locate which transaction features are the key information required by the prediction model, thereby constructing a target transaction feature set. Through this series of steps, the bank can make more accurate predictions on the process of disposing of overdue monetary resources of a specific enterprise based on machine learning and data analysis technologies, and thus formulate more effective disposal strategies.

[0064] Optionally, in the method for processing monetary resources provided in the embodiments of the present application, before determining the target transaction feature set from the reference transaction feature set based on the enterprise number, it includes: extracting from the historical transaction data set to obtain a first transaction feature set, where the historical transaction data set matches the enterprise number; screening out a second transaction feature set from the first transaction feature set, where the second transaction feature set is related to the repayment amount corresponding to the enterprise number; identifying a third transaction feature set from the second transaction feature set; and determining the amount prediction model according to the third transaction feature set.

[0065] It should be noted that the historical transaction data set refers to the summary of all transaction records in the past period of time that match the enterprise number. These data contain all the financial activity information between the enterprise and the bank, and are an important basis for analyzing the financial health of the enterprise and predicting the repayment amount. The historical transaction data can be the individual historical data of the enterprise, the historical data of other enterprises with similar businesses to the enterprise, or the historical data of some or all enterprises within a certain period of time. The first transaction feature set can be extracted from the historical transaction data set and initially includes a series of features reflecting the enterprise's transaction behavior and financial status, which may be the initial data source for model training. The second transaction feature set can be a set formed by screening out the features related to the enterprise's repayment amount on the basis of the first transaction feature set. The third transaction feature set can be further identified from the second transaction feature set and is the key feature set finally used for constructing the amount prediction model. These features have undergone further analysis and verification to ensure that they have a direct relationship with the enterprise's repayment amount and can provide sufficient information to support the accuracy of the prediction model.

[0066] In an alternative embodiment, before determining the target transaction feature set based on the enterprise number, the system undergoes a series of preprocessing and feature selection steps to ensure that the feature set used for prediction is both comprehensive and accurate. First, the system extracts the first transaction feature set from the historical transaction data set, which is a feature set containing an overview of the enterprise's past transaction behaviors. Then, by analyzing the first transaction feature set, the system filters out the features highly correlated with the enterprise's repayment amount to form the second transaction feature set. This step is to remove irrelevant or redundant features and reduce the complexity of model training. Finally, based on the second transaction feature set, the system identifies the third transaction feature set with the most predictive value, which will be directly used for constructing the amount prediction model. Through this series of feature selection processes, the system ensures that the final feature set used for prediction is refined and can provide the most direct and effective information for predicting the enterprise's repayment amount.

[0067] In an alternative embodiment, assume that a bank is developing a system for predicting the repayment amount of overdue monetary resources based on enterprise financial data, and the goal of this system is to predict the repayment amount of enterprise number "XYZ". Steps implementation: Historical data extraction: First, the system extracts all historical transaction data from the bank database based on enterprise number "XYZ" to form the historical transaction data set. First transaction feature set construction: The system extracts a series of transaction features from the historical transaction data set, such as the type, amount, repayment record, overdue days of each loan, etc., to construct the first transaction feature set. Second transaction feature set screening: Through correlation analysis and statistical tests, the system filters out the features with relatively high correlation with the repayment amount in the first transaction feature set, such as the change rate of loan balance, cash inflow, the proportion of historical recovered principal, etc., to form the second transaction feature set. Third transaction feature set confirmation: Based on the second transaction feature set, the system further identifies the features that have the most influence on the amount prediction. For example, through principal component analysis or feature importance evaluation, it is confirmed that the change rate of loan balance and cash inflow are used as the final third transaction feature set. Amount prediction model construction: According to the third transaction feature set, the bank constructs an amount prediction model, which may use the random forest algorithm or support vector machine algorithm. Through learning from historical data, the model can predict the expected repayment amount of enterprise "XYZ" when disposing of overdue monetary resources.

[0068] Through the above embodiments of the present application, it is possible to construct an accurate amount prediction model based on the historical transaction data of enterprise number "XYZ" to predict the recovery situation of overdue monetary resources. This not only improves the bank's decision-making support ability in the disposal of overdue monetary resources but also enhances the accuracy of prediction and the practicality of the model.

[0069] Optionally, in the method for processing currency resources provided in the embodiments of the present application, screening out the second transaction feature set from the first transaction feature set includes: clustering the first transaction feature set to obtain M categories of transaction feature sets; sequentially determining the P-th transaction feature set from the M categories of transaction feature sets for the following operations: determining the difference between the i-th transaction feature in the P-th transaction feature set and the mean value of the P-th transaction feature set as the i-th first difference, and determining the difference between the i-th compensation amount corresponding to the i-th transaction feature and the mean value of the compensation amounts as the i-th second difference, where each category of transaction feature set includes n transaction features; determining the product of the n first differences and second differences as the first product, and determining the square root of the product of the sum of the squares of the n first differences and the sum of the squares of the n second differences as the second product; determining the ratio of the first product to the second product as the correlation coefficient of the P-th transaction feature set, and in the case where the correlation coefficient is greater than the preset coefficient threshold, incorporating the P-th transaction feature set into the second transaction feature set, where M is an integer greater than 0, P is an integer greater than 0 and less than or equal to M, i is an integer greater than 0 and less than or equal to n, and n is an integer greater than 0.

[0070] It should be noted that clustering is an unsupervised machine learning technique used to divide the samples in a dataset into multiple categories, such that the samples within the same category have a relatively high similarity, while the similarity between different categories is relatively low. In this context, it is used to group the features in the first transaction feature set according to their similarity or relevance to form M categories of transaction feature sets. It can also be that when extracting transaction features, they are directly classified to obtain M categories of transaction feature sets after extracting the transaction features. The P-th transaction feature set: After clustering, each category forms a transaction feature set, and the P-th transaction feature set refers to any one of these M categories, where P is an integer identifying a specific category. The correlation coefficient is used to measure the strength of the linear relationship between two variables, and its value range is from -1 to 1. In this embodiment, the correlation coefficient is used to evaluate the relevance between the transaction feature set and the compensation amount. When the correlation coefficient is greater than the preset coefficient threshold, it means that the transaction feature set has a significant impact on predicting the compensation amount. The preset coefficient threshold is set in advance and is a benchmark value used to determine whether the relevance between the transaction feature set and the compensation amount is strong enough. Only when the correlation coefficient exceeds this threshold is a specific transaction feature set considered to have sufficient importance and thus incorporated into the second transaction feature set.

[0071] Specifically, first, the first transaction feature set is divided into M categories through cluster analysis, and then the transaction features in each category (i.e., the Pth transaction feature set) are evaluated for correlation. The specific evaluation method is to calculate the difference between the transaction feature value and the mean (i.e., the first difference) and the difference between the repayment amount value and the mean (i.e., the second difference). Then, the first product is calculated by multiplying the first difference by the second difference, and the second product is calculated by taking the square root of the product of the sum of the squares of the first difference and the second difference. The correlation coefficient is the calculation result of the ratio of the first product to the second product. If the correlation coefficient exceeds the preset threshold, then the Pth transaction feature set is considered to be significantly correlated with the repayment amount and is incorporated into the second transaction feature set. The whole process ensures that the final second transaction feature set contains transaction features valuable for predicting the repayment amount.

[0072] In an alternative embodiment, in the prediction process of overdue monetary resource disposal, screening out the second transaction feature set from the first transaction feature set is a crucial feature selection step. First, cluster analysis is used to classify all transaction features in the first transaction feature set, forming M categories, and each category represents a group of transaction features with similar properties. Then, the system will conduct in-depth analysis for each category, that is, the Pth transaction feature set, to calculate the correlation between each transaction feature and the repayment amount. The process of calculating the correlation includes calculating the first difference and the second difference, and then calculating the first product and the second product based on these two differences, and finally determining the correlation coefficient. Only the transaction feature sets with a correlation coefficient higher than the preset coefficient threshold will be added to the second transaction feature set. This screening process ensures that each feature set in the second transaction feature set is closely related to the repayment amount, thereby improving the accuracy and reliability of the amount prediction model.

[0073] Specifically, let the borrower's repayment amount (repayment amount) be y and the transaction feature be x. The specific formula for calculating the correlation coefficient is as follows:

[0074]

[0075] For factors with |r| > 0.5, it shows a strong correlation, and its significance is further calculated.

[0076] In an alternative embodiment, assume that a bank is developing a system for predicting the amount of overdue monetary resources recoverable from enterprise "XYZ". First, the bank extracts a series of transaction features from the historical transaction data set of enterprise "XYZ", such as loan type, loan balance, repayment frequency, days overdue, etc. These features constitute the first transaction feature set. The system uses a clustering algorithm to classify these features into M categories, for example, M = 5, and each category represents transaction features with similar properties. Correlation evaluation: The system sequentially performs correlation evaluation on the transaction feature sets of the M categories to determine which categories have a significant association with the recoverable amount. For the transaction feature set of the P-th category (for example, P = 2), the system calculates the difference between the transaction feature and the mean of the set, and the difference between the recoverable amount and the mean of the recoverable amounts, further calculates the first product and the second product, and finally determines the correlation coefficient. Screening and merging: Assume that the correlation coefficient of the transaction feature set of the P-th category is 0.8, which exceeds the preset coefficient threshold of 0.7, indicating that this category of transaction features has a significant correlation with the recoverable amount. Therefore, the transaction feature set of the P-th category will be incorporated into the second transaction feature set. Construction of the amount prediction model: Through the above steps, the bank can construct a second transaction feature set containing transaction features significantly related to the recoverable amount, and this set will be used as the input of the amount prediction model. By learning the relationship between these features and the historical recoverable amounts, the model can predict new transaction data, helping the bank more accurately estimate the possibility of recovering overdue monetary resources.

[0077] Through the above embodiments of the present application, the most predictive transaction features can be screened out from complex historical transaction data, and a more accurate and reliable amount prediction model can be constructed, providing strong support for the disposal decision of overdue monetary resources. This process not only improves the accuracy of prediction but also reduces the complexity of the model, making the model easier to understand and contributing to the risk management and strategy formulation of the bank.

[0078] Optionally, in the method for processing monetary resources provided in the embodiments of the present application, identifying a third transaction feature set in the second transaction feature set includes: sequentially determining the transaction feature set of the P-th category as the P-th transaction feature set and performing the following operations: determining the statistical coefficient corresponding to the P-th transaction feature set according to the correlation coefficient of the P-th transaction feature set and the number of transaction features included in the transaction feature set of each category; determining the statistical probability corresponding to the P-th transaction feature set based on the statistical coefficient; and incorporating the P-th transaction feature set into the third transaction feature set when the statistical probability corresponding to the P-th transaction feature set is less than the probability threshold.

[0079] It should be noted that the statistical coefficient is an indicator calculated by combining the correlation coefficient and the number of transaction characteristics, and is used to evaluate the comprehensive influence of each set of transaction characteristics in predicting the repayment amount. It takes into account the correlation of the characteristics and the impact of the number of characteristics on the model's prediction ability. The statistical probability is a probability value calculated based on the statistical coefficient, which reflects the credibility or importance of the P-th set of transaction characteristics in predicting the repayment amount. The magnitude of the statistical probability can help determine which sets of transaction characteristics should be further retained and used. The probability threshold is the threshold used when screening the third set of transaction characteristics, and is used to judge whether the statistical probability is high enough to determine whether the set of transaction characteristics should be incorporated into the third set of transaction characteristics. Only when the statistical probability exceeds this threshold will the set of transaction characteristics be retained, which ensures that the final third set of transaction characteristics contains characteristics that have a significant impact on predicting the repayment amount.

[0080] Specifically, the system will traverse each set of transaction characteristics of M categories (i.e., the P-th set of transaction characteristics), calculate the statistical coefficient based on the correlation coefficient and the number of transaction characteristics, and then calculate the statistical probability based on the statistical coefficient. If the statistical probability of a certain set of transaction characteristics is lower than the preset probability threshold, then this set of transaction characteristics will be incorporated into the third set of transaction characteristics. The core purpose of this process is to further refine the set of transaction characteristics and ensure that only those characteristics with the highest credibility in predicting the repayment amount are retained, thereby improving the accuracy and reliability of the amount prediction model.

[0081] In an alternative implementation, in the process of predicting the repayment amount of overdue monetary resources based on machine learning, screening out the third set of transaction characteristics from the second set of transaction characteristics is an important step to ensure the prediction performance of the model. The second set of transaction characteristics contains multiple sets of transaction characteristics that have been preliminarily screened and are significantly related to the repayment amount. To further determine which characteristic sets are the most critical for predicting the repayment amount, the system will calculate the statistical coefficient of each set of transaction characteristics. This coefficient combines the influence of the correlation coefficient and the number of transaction characteristics and reflects the comprehensive importance of the characteristic set. Subsequently, based on the statistical coefficient, the system calculates the statistical probability, and this probability value evaluates the credibility of the set of transaction characteristics in predicting the repayment amount. Only the set of transaction characteristics with a statistical probability lower than the probability threshold will be incorporated into the third set of transaction characteristics to form the characteristic set finally used in the amount prediction model. The setting of this probability threshold ensures that the third set of transaction characteristics contains only those transaction characteristics that have a decisive impact on predicting the repayment amount, removing possible noise or irrelevant characteristics and improving the prediction accuracy of the model.

[0082] Specifically, for the aforementioned obtained correlation coefficient r, a statistical coefficient t can be constructed:

[0083]

[0084] The t-statistic follows a t-distribution, which is a distribution that describes the standard error of parameter estimation in the case of small samples. The t-distribution table provides a set of t-values based on degrees of freedom and significance levels. These t-values can help determine whether the t-statistic is significantly different from 0, and thus whether the correlation is significant. Therefore, the t-distribution table can be consulted to obtain statistical probabilities.

[0085] In an alternative embodiment, assume that a bank is using a machine learning algorithm to predict the amount of overdue monetary resources recovered for enterprise number "12345" and has completed the screening of the first transaction feature set and the second transaction feature set. The banking system calculates the statistical coefficient for each set based on the correlation coefficient and the number of transaction features in each P-th transaction feature set among M categories. For example, if the correlation coefficient of the second transaction feature set is 0.8 and it contains 10 transaction features, the system combines the correlation coefficient and the number of features to calculate the statistical coefficient. Statistical probability evaluation: Based on the calculated statistical coefficient, the system further calculates the statistical probability for each P-th transaction feature set. This probability reflects the credibility of the transaction feature set in predicting the recovered amount. Probability threshold screening: Assume the probability threshold set by the system is 0.2. After evaluating the statistical probabilities of all transaction feature sets, the system finds that the statistical probability of the second transaction feature set is 0.15, which is lower than the probability threshold. Therefore, the second transaction feature set is incorporated into the third transaction feature set. Quota prediction model construction: Through the above screening process, the bank can construct a third transaction feature set that contains the most predictive transaction features. Based on this set, the bank can train and optimize the quota prediction model to improve the accuracy and stability of predicting the recovered amount.

[0086] Through the above embodiments of the present application, not only can the most relevant transaction features be extracted from historical transaction data, but also the feature set can be further refined to ensure that the quota prediction model is based on the most accurate and important information. This helps the bank make more accurate predictions and formulate more effective strategies in the disposal of overdue monetary resources, thereby improving the asset recovery rate and reducing the negative impact of non-performing loans.

[0087] Optionally, in the method for processing monetary resources provided in the embodiments of the present application, determining a quota prediction model according to the third transaction feature set includes: determining the number of calculation coefficients included in the calculation coefficient set based on the number of categories of transaction feature sets included in the third transaction feature set; and using the third transaction feature set to determine the calculation coefficient set to obtain the quota prediction model.

[0088] It should be noted that the calculation coefficient set can be the parameter set in the amount prediction model, which is used to quantify the influence degree of each transaction feature in the third transaction feature set on the predicted amount of compensation. These calculation coefficients are optimized during the model training process to ensure that the model can accurately predict the future compensation amount based on historical data. The number of calculation coefficients can refer to the number of parameters included in the calculation coefficient set, and this number corresponds to the number of categories in the third transaction feature set. Usually, each category of transaction feature set will have one or more calculation coefficients to represent its contribution to the predicted amount of compensation.

[0089] The system determines the size of the calculation coefficient set, that is, the number of calculation coefficients, according to the number of categories in the third transaction feature set. This means that each transaction feature category will have one or more calculation coefficients to represent its weight or influence in the prediction. Subsequently, the system uses the transaction feature data in the third transaction feature set to train the model, thereby determining the specific parameters of the calculation coefficient set, and finally constructs the amount prediction model. This process ensures that the model can make full use of the transaction features most relevant to the predicted amount of compensation, improving the prediction accuracy and reliability of the model.

[0090] In an alternative implementation, in the prediction process of overdue monetary resource disposal, determining the amount prediction model based on the third transaction feature set is the final key step. The third transaction feature set has undergone multiple rounds of screening, retaining the transaction feature set that has the most influential on the predicted amount of compensation. First, the system determines the size of the calculation coefficient set based on the structure of the third transaction feature set, which means that each transaction feature category will be assigned one or more calculation coefficients to represent its prediction contribution to the amount of compensation. Then, the system uses the specific data in the third transaction feature set to train a machine learning model, such as a regression model or a neural network, to determine the specific calculation coefficient values in the calculation coefficient set. These calculation coefficients are optimized during the model training process to ensure that the model can accurately predict the future compensation amount based on historical data. Finally, through the third transaction feature set and the calculation coefficient set, the system constructs a complete and accurate amount prediction model, which can provide strong support for the bank's decision-making in the process of overdue monetary resource disposal.

[0091] Specifically, after obtaining the third transaction feature set a1, a2, a3,... a q it is possible to set the same number of coefficients θ1, θ2, θ3,... θ q as the number of categories, and construct the model:

[0092] y = θ1x1 + θ2x2 + θ3x3 +...

[0093] Then, the result of the calculation coefficient set is obtained through the third transaction feature set to obtain the quota prediction model.

[0094] In an alternative embodiment, the process of calculating the calculation coefficient can be as follows:

[0095] As mentioned above, assume that the amount received by the borrower is y, and the various types of transaction features in the third transaction feature set are a1, a2, a3,... a q , and the coefficients are θ1, θ2, θ3,... θ q , and the model is y = θ1x1 + θ2x2 + θ3x3 +...

[0096] The L1 constraint can be defined as The constraint condition ||θ||1 ≤ R, where is the L1 norm. This constraint is used to achieve the sparsity of the model, that is, to make some coefficients zero, so as to achieve the purpose of dimensionality reduction. R is a given constant, and J LS (θ) is the least squares loss function. To find the optimal solution under the constraint condition, the Lagrange multiplier method can be used to construct the Lagrange multiplier J(θ) = J LS (θ) + λ||θ||1.

[0097] In the case of solving the optimization problem, the inequality can be considered, where c j > 0. Through machine learning operations, the current solution can be used to replace c j , that is That is is the generalized inverse. Then the optimization problem can be expressed as

[0098]

[0099] where is the element diagonal matrix, is the constant for this round of calculation.

[0100] It can be iteratively solved according to the following formula

[0101]

[0102] where y is the sample value and Φ is the sample value matrix of x.

[0103] When the value is very small, it can be set to 0 to achieve factor dimensionality reduction (such as ). Select the condition R and give the initial value. Through machine learning calculation, the regression equation y = θ1x1 + θ2x2 + θ3x3 +... is obtained.

[0104] In an alternative embodiment, it is assumed that the bank has completed the screening of transaction characteristics based on the enterprise number "12345" and formed a third transaction characteristic set containing 5 categories (i.e., M = 5). The size of the calculation coefficient set can be determined: The banking system first determines the size of the calculation coefficient set based on the information that the third transaction characteristic set contains 5 categories. Training the quota prediction model: The system uses the transaction characteristic data in the third transaction characteristic set and the historical recovered quota data to train the quota prediction model. During the training process, the model will learn and optimize the calculation coefficients in the calculation coefficient set to ensure that each calculation coefficient can accurately reflect the impact of its corresponding transaction characteristic category on the recovered quota. Model verification and application: After training, the system will verify the accuracy of the quota prediction model and apply it to the prediction of new cases. For example, when the enterprise number "12345" has another overdue monetary resource disposal, the bank can use this model to predict the recovered quota and provide data support for the disposal strategy.

[0105] Through the above embodiments of the present application, an accurate and efficient quota prediction model based on the third transaction characteristic set can be constructed. This model can not only provide personalized recovered quota predictions for enterprise numbers but also serve as a reference to help banks make more accurate predictions when dealing with other enterprises' overdue monetary resources, improving the efficiency of overdue monetary resource disposal and the bank's financial risk management capabilities.

[0106] In the embodiment of the present application, a transaction data set is determined according to the enterprise number, where the transaction data set is used to indicate information on transactions corresponding to the enterprise number; the transaction data set is extracted to obtain a reference transaction characteristic set, and a target transaction characteristic set is determined from the reference transaction characteristic set based on the enterprise number, where the transaction characteristics are used to indicate the characteristics of the transaction data set; the target transaction characteristic set is input into the quota prediction model to obtain the recovered quota corresponding to the enterprise number, and the processing ratio of the monetary resource is determined according to the recovered quota and the transaction data set, where the quota prediction model matches the enterprise number. For different enterprises, their transaction data can be determined according to the enterprise number, and the transaction characteristics therein can be determined, and the corresponding target transaction characteristics can be selected from the transaction characteristics based on the enterprise number to accurately calculate the recovered quota and processing ratio corresponding to the enterprise. Through the regression operation of historical data, the prediction of the transaction price becomes objective, avoiding the arbitrariness of subjective prediction. Furthermore, the technical problem of low accuracy of the processing ratio caused by manually determining the processing ratio is solved.

[0107] It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0108] Example 2

[0109] The embodiment of the present application further provides a processing device for monetary resources. It should be noted that the processing device for monetary resources in the embodiment of the present application can be used to execute the processing method for monetary resources provided in the embodiment of the present application. The following introduces the processing device for monetary resources provided in the embodiment of the present application.

[0110] According to the embodiment of the present application, there is also provided a device for implementing the above-mentioned processing method for monetary resources, as Figure 3 shown, the device includes:

[0111] A data set determination unit 302, configured to determine a transaction data set according to an enterprise number, where the transaction data set is used to indicate information on transactions corresponding to the enterprise number;

[0112] A feature extraction unit 304, configured to extract the transaction data set to obtain a reference transaction feature set, and determine a target transaction feature set from the reference transaction feature set based on the enterprise number, where the transaction feature is used to indicate the feature of the transaction data set;

[0113] A ratio determination unit 306, configured to input the target transaction feature set into a quota prediction model to obtain the receivable quota corresponding to the enterprise number, and determine the processing ratio of monetary resources according to the receivable quota and the transaction data set, where the quota prediction model matches the enterprise number.

[0114] The processing device for monetary resources provided in the embodiment of the present application determines a transaction data set according to an enterprise number, where the transaction data set is used to indicate information on transactions corresponding to the enterprise number; extracts the transaction data set to obtain a reference transaction feature set, and determines a target transaction feature set from the reference transaction feature set based on the enterprise number, where the transaction feature is used to indicate the feature of the transaction data set; inputs the target transaction feature set into a quota prediction model to obtain the receivable quota corresponding to the enterprise number, and determines the processing ratio of monetary resources according to the receivable quota and the transaction data set, where the quota prediction model matches the enterprise number. For different enterprises, their transaction data can be determined according to the enterprise number, and the transaction features therein can be determined, and the corresponding target transaction features can be selected from the transaction features based on the enterprise number to accurately calculate the receivable quota and the processing ratio of monetary resources corresponding to the enterprise. Furthermore, the technical problem of low accuracy of the processing ratio caused by manually determining the processing ratio is solved.

[0115] Optionally, the above-mentioned feature extraction unit 304 includes: an input feature determination module, configured to determine a quota prediction model that matches the enterprise number, and determine the input features of the quota prediction model; a target feature determination module, configured to determine the transaction features corresponding to the input features in the reference transaction feature set as target transaction features, so as to obtain a target transaction feature set.

[0116] Optionally, the above-mentioned feature extraction unit 304 further includes: a feature extraction module, configured to extract from the historical transaction data set to obtain a first transaction feature set, where the historical transaction data set matches the enterprise number; a feature screening module, configured to screen out a second transaction feature set from the first transaction feature set, where the second transaction feature set is related to the compensation quota corresponding to the enterprise number; a feature confirmation module, configured to confirm a third transaction feature set from the second transaction feature set; a model determination module, configured to determine a quota prediction model according to the third transaction feature set.

[0117] Optionally, the above-mentioned feature screening module is further configured to: cluster the first transaction feature set to obtain M categories of transaction feature sets; sequentially determine the P-th transaction feature set from the M categories of transaction feature sets for the following operations: determine the difference between the i-th transaction feature in the P-th transaction feature set and the mean value of the P-th transaction feature set as the i-th first difference, and determine the difference between the i-th compensation quota corresponding to the i-th transaction feature and the mean value of the compensation quotas as the i-th second difference, where each category of transaction feature set includes n transaction features; determine the product of the n first differences and the second differences as the first product, and determine the square root of the product of the sum of the squares of the n first differences and the sum of the squares of the n second differences as the second product; determine the ratio of the first product and the second product as the correlation coefficient of the P-th transaction feature set, and in the case where the correlation coefficient is greater than a preset coefficient threshold, incorporate the P-th transaction feature set into the second transaction feature set, where M is an integer greater than 0, P is an integer greater than 0 and less than or equal to M, i is an integer greater than 0 and less than or equal to n, and n is an integer greater than 0.

[0118] Optionally, the above-mentioned feature confirmation module is further configured to: sequentially determine the P-th transaction feature set from the M categories of transaction feature sets for the following operations: determine the statistical coefficient corresponding to the P-th transaction feature set according to the correlation coefficient of the P-th transaction feature set and the number of transaction features included in each category of transaction feature set; determine the statistical probability corresponding to the P-th transaction feature set based on the statistical coefficient; and in the case where the statistical probability corresponding to the P-th transaction feature set is less than the probability threshold, incorporate the P-th transaction feature set into the third transaction feature set.

[0119] Optionally, the above model determination module is further configured to: determine the number of calculation coefficients included in the calculation coefficient set based on the number of categories of the transaction feature sets included in the third transaction feature set; use the third transaction feature set to determine the calculation coefficient set to obtain a quota prediction model.

[0120] It should be noted here that the above data set determination unit 302 to the ratio determination unit 306 correspond to steps S101 to S103 in Embodiment 1. The two modules have the same implemented examples and application scenarios as the corresponding steps, but are not limited to the content disclosed in the above Embodiment 1. It should be noted that the above modules or units may be hardware components or software components stored in a memory (for example, memory 104) and processed by one or more processors (for example, processors 102a, 102b,..., 102n). The above modules may also be part of a device and can run in the computer terminal 10 provided in Embodiment 1.

[0121] Embodiment 3

[0122] Embodiments of the present application may provide an electronic device. Figure 4 It is a structural block diagram of an electronic device according to an embodiment of the present application. As Figure 4 shown, the electronic device may include: one or more ( Figure 4 only one is shown in the figure) processors 1002, a memory 1004, a storage controller, and a peripheral interface. Among them, the peripheral interface is connected to a radio frequency module, an audio module, and a display.

[0123] Among them, the memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the methods and devices in the embodiments of the present application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, implements the above methods. The memory may include a high-speed random access memory, and may further include a non-volatile memory, such as one or more magnetic storage devices, a flash memory, or other non-volatile solid-state memories. In some instances, the memory may further include a memory remotely disposed relative to the processor, and these remote memories may be connected to the terminal through a network. Examples of the above network include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0124] The processor may call the information and application programs stored in the memory through a transmission device to perform the following steps:

[0125] S1, determine a transaction data set according to the enterprise number, where the transaction data set is used to indicate information about transactions corresponding to the enterprise number;

[0126] S2. Extract the transaction data set to obtain a reference transaction feature set, and determine a target transaction feature set from the reference transaction feature set based on the enterprise number, where the transaction feature is used to indicate the feature of the transaction data set.

[0127] S3. Input the target transaction feature set into the quota prediction model to obtain the compensation quota corresponding to the enterprise number, and determine the processing ratio of the monetary resources according to the compensation quota and the transaction data set, where the quota prediction model is matched with the enterprise number.

[0128] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: determine the quota prediction model matched with the enterprise number, and determine the input features of the quota prediction model; determine the transaction features corresponding to the input features in the reference transaction feature set as the target transaction features to obtain the target transaction feature set.

[0129] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: extract the historical transaction data set to obtain a first transaction feature set, where the historical transaction data set is matched with the enterprise number; screen out a second transaction feature set from the first transaction feature set, where the second transaction feature set is related to the compensation quota corresponding to the enterprise number; confirm a third transaction feature set from the second transaction feature set; determine the quota prediction model according to the third transaction feature set.

[0130] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: cluster the first transaction feature set to obtain M sets of transaction features of different categories; sequentially determine the Pth set of transaction features from the M sets of transaction features of different categories to perform the following operations: determine the difference between the ith transaction feature in the Pth set of transaction features and the mean value of the Pth set of transaction features as the ith first difference, and determine the difference between the ith compensation quota corresponding to the ith transaction feature and the mean value of the compensation quotas as the ith second difference, where each set of transaction features of different categories includes n transaction features; determine the product of the n first differences and the second differences as the first product, and determine the square root of the product of the sum of the squares of the n first differences and the sum of the squares of the n second differences as the second product; determine the ratio of the first product to the second product as the correlation coefficient of the Pth set of transaction features, and in the case where the correlation coefficient is greater than the preset coefficient threshold, incorporate the Pth set of transaction features into the second transaction feature set, where M is an integer greater than 0, P is an integer greater than 0 and less than or equal to M, i is an integer greater than 0 and less than or equal to n, and n is an integer greater than 0.

[0131] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: sequentially determine the transaction feature sets of M categories as the Pth transaction feature set and perform the following operations: determine the statistical coefficient corresponding to the Pth transaction feature set according to the correlation coefficient of the Pth transaction feature set and the number of transaction features included in each category of transaction feature sets; determine the statistical probability corresponding to the Pth transaction feature set based on the statistical coefficient; in the case where the statistical probability corresponding to the Pth transaction feature set is less than the probability threshold, incorporate the Pth transaction feature set into the third transaction feature set.

[0132] The processor can also call the information and application programs stored in the memory through the transmission device to perform the following steps: determine the number of calculation coefficients included in the calculation coefficient set based on the number of categories of transaction feature sets included in the third transaction feature set; use the third transaction feature set to determine the calculation coefficient set to obtain a quota prediction model.

[0133] In the embodiment of the present application, determine a transaction data set according to the enterprise number, where the transaction data set is used to indicate the information of the transaction corresponding to the enterprise number; extract the transaction data set to obtain a reference transaction feature set, and determine a target transaction feature set from the reference transaction feature set based on the enterprise number, where the transaction feature is used to indicate the feature of the transaction data set; input the target transaction feature set into the quota prediction model to obtain the compensation quota corresponding to the enterprise number, and determine the processing ratio of the monetary resources according to the compensation quota and the transaction data set, where the quota prediction model is matched with the enterprise number. For different enterprises, their transaction data can be determined according to the enterprise number, and the transaction features therein can be determined, and the corresponding target transaction features can be selected from the transaction features based on the enterprise number to accurately calculate the compensation quota and the processing ratio of the monetary resources corresponding to the enterprise. Furthermore, the technical problem of low accuracy of the processing ratio caused by manually determining the processing ratio is solved.

[0134] Those of ordinary skill in the art can understand that Figure 4 The structure shown is only schematic, and the electronic device can also be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a handheld computer, and terminal devices such as Mobile Internet Devices (MID), PAD, etc. Figure 4 It does not limit the structure of the above electronic device. For example, the electronic device may further include more or fewer components (such as a network interface, a display device, etc.) than those shown Figure 4 in, or have a different configuration from that shown Figure 4 in.

[0135] Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the relevant hardware of the terminal device through a program, and the program can be stored in a computer-readable storage medium. The storage medium can include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, an optical disc, etc.

[0136] Embodiment 4

[0137] An embodiment of the present application also provides a storage medium. Optionally, in this embodiment, the above storage medium can be used to store the program code executed by the method for processing currency resources provided in the first embodiment above.

[0138] Optionally, in this embodiment, the above storage medium can be located in any one of the computer terminals in the computer terminal group in the computer network, or in any one of the mobile terminals in the mobile terminal group.

[0139] The present application also provides a computer program product, which is adapted to execute the program of the method for processing currency resources when executed on a data processing device.

[0140] The serial numbers of the above embodiments of the present application are only for description and do not represent the advantages and disadvantages of the embodiments.

[0141] In the above embodiments of the present application, the descriptions of the respective embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0142] In the several embodiments provided by the present application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be through some interfaces. The indirect coupling or communication connection of the units or modules can be in an electrical or other form.

[0143] The units described as separate components may or may not be physically separated. The components shown as units may or may not be physical units, that is, they can be located in one place, or they can be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0144] In addition, in each embodiment of the present application, the functional units can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0145] If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present application. The foregoing storage medium includes: various media such as USB flash drives, read-only memories (ROMs), random access memories (RAMs), mobile hard disks, magnetic disks, or optical discs that can store program codes.

[0146] The above are only the preferred embodiments of the present application. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present application, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present application.

Claims

1. A method for processing currency resources, characterized in that, Including: Determine a transaction data set according to the enterprise number, where the transaction data set is used to indicate information about the transaction corresponding to the enterprise number; Extract the transaction data set to obtain a reference transaction feature set, and determine a target transaction feature set from the reference transaction feature set based on the enterprise number, where the transaction feature is used to indicate the feature of the transaction data set; Input the target transaction feature set into a quota prediction model to obtain the receivable quota corresponding to the enterprise number, and determine the processing ratio of monetary resources according to the receivable quota and the transaction data set, where the quota prediction model matches the enterprise number.

2. The method according to claim 1, wherein The determining the target transaction feature set from the reference transaction feature set based on the enterprise number includes: Determine the quota prediction model that matches the enterprise number, and determine the input features of the quota prediction model; Determine the transaction features corresponding to the input features in the reference transaction feature set as the target transaction features to obtain the target transaction feature set.

3. The method according to claim 2, wherein Before determining the target transaction feature set from the reference transaction feature set based on the enterprise number, it includes: Extract a historical transaction data set to obtain a first transaction feature set, where the historical transaction data set matches the enterprise number; Filter out a second transaction feature set from the first transaction feature set, where the second transaction feature set is related to the receivable quota corresponding to the enterprise number; Identify a third transaction feature set from the second transaction feature set; Determine the quota prediction model according to the third transaction feature set.

4. The method according to claim 3, characterized in that, The filtering out the second transaction feature set from the first transaction feature set includes: Cluster the first transaction feature set to obtain M categories of transaction feature sets; Successively determine the P-th transaction feature set among the M categories of transaction feature sets for the following operations: Determine the difference between the i-th transaction feature in the P-th transaction feature set and the mean of the P-th transaction feature set as the i-th first difference, and determine the difference between the i-th receivable quota corresponding to the i-th transaction feature and the mean of the receivable quotas as the i-th second difference, where each category of transaction feature set includes n transaction features; Determine the product of n first differences and the second differences as the first product, and determine the square root of the product of the sum of the squares of n first differences and the sum of the squares of n second differences as the second product; Determine the ratio of the first product to the second product as the correlation coefficient of the P-th transaction feature set. When the correlation coefficient is greater than the preset coefficient threshold, incorporate the P-th transaction feature set into the second transaction feature set, where M is an integer greater than 0, P is an integer greater than 0 and less than or equal to M, i is an integer greater than 0 and less than or equal to n, and n is an integer greater than 0.

5. The method according to claim 4, characterized in that, The identifying the third transaction feature set from the second transaction feature set includes: Successively determine the transaction feature sets of the M categories as the P-th transaction feature set and perform the following operations: Determine the statistical coefficient corresponding to the P-th transaction feature set according to the correlation coefficient of the P-th transaction feature set and the number of transaction features included in the transaction feature set of each category; Determine the statistical probability corresponding to the P-th transaction feature set based on the statistical coefficient; In the case where the statistical probability corresponding to the P-th transaction feature set is less than the probability threshold, incorporate the P-th transaction feature set into the third transaction feature set.

6. The method according to claim 3, characterized in that, The determining of the amount prediction model according to the third transaction feature set includes: Determine the number of calculation coefficients included in the calculation coefficient set based on the number of categories of the transaction feature sets included in the third transaction feature set; Use the third transaction feature set to determine the calculation coefficient set to obtain the amount prediction model.

7. A processing device for currency resources, characterized in that, Including: A data set determination unit, configured to determine a transaction data set according to an enterprise number, where the transaction data set is used to indicate information about transactions corresponding to the enterprise number; A feature extraction unit, configured to extract the transaction data set to obtain a reference transaction feature set, and determine a target transaction feature set from the reference transaction feature set based on the enterprise number, where the transaction feature is used to indicate a feature of the transaction data set; A ratio determination unit, configured to input the target transaction feature set into an amount prediction model to obtain the receivable amount corresponding to the enterprise number, and determine the processing ratio of the monetary resources according to the receivable amount and the transaction data set, where the amount prediction model matches the enterprise number.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored executable program, where, when the executable program runs, it controls the device where the computer-readable storage medium is located to execute the method for processing monetary resources according to any one of claims 1 to 6.

9. An electronic device, characterized in that, Including: A memory, storing an executable program; A processor, configured to run the program, where, when the program runs, it executes the method according to any one of claims 1 to 6.

10. A computer program product, comprising computer instructions, characterized in that, The computer instructions, when executed by the processor, implement the steps of the method according to any one of claims 1 to 6.

Images