CN117726422A - Bad account preparation amount calculation method and device, electronic equipment and storage medium - Google Patents

Abstract

The application provides a bad account preparation amount calculation method, a bad account preparation amount calculation device, electronic equipment and a storage medium, relates to the technical field of data processing, refines customer groups, determines bad account counting and lifting ratios corresponding to different types of customers, and further calculates bad account preparation amount according to the bad account counting and lifting ratios corresponding to the customers. The method comprises the following steps: obtaining credit information of a target credit customer, wherein the credit information comprises a customer scale, a customer type, a risk level and a loan amount; determining a target customer group corresponding to the target credit customer according to the customer scale, the customer type and the risk level of the target credit customer; determining a target repayment model corresponding to the target customer group from a mapping relation comprising a plurality of customer groups and a plurality of repayment models according to the target customer group; determining a metering and lifting proportion of a target credit customer based on the target repayment model; and determining bad account preparation amount corresponding to the second time node according to the proportion of the calculated and the loan amount.

Description

Bad account preparation amount calculation method and device, electronic equipment and storage medium

Technical Field

The present disclosure relates to the field of data processing technologies, and in particular, to a bad account preparation amount calculating method, a bad account preparation amount calculating device, an electronic device, and a storage medium.

Background

In order to objectively reflect the recovery condition of accounts receivable, the bad account preparation amount is calculated, the bad account calculation proportion corresponding to the customer needs to be determined, and when the bad account preparation amount is calculated in the prior art, after account age of the customer and corresponding account age amount data are determined, the preset bad account calculation proportion is used for all types of customers uniformly, so that the bad account preparation amount is calculated.

However, credit evaluation of different types of clients is different, the applicable bad account calculation proportion should be different, and the bad account preparation amount corresponding to different clients is calculated by using the same bad account calculation proportion in the prior art, so that the calculated bad account preparation amount is inaccurate.

Disclosure of Invention

The application provides a bad account preparation amount calculation method, a bad account preparation amount calculation device, electronic equipment and a storage medium, which are used for refining a customer group, determining bad account calculation proportion corresponding to different types of customers, and further calculating bad account preparation amount according to the bad account calculation proportion corresponding to the customers.

In order to achieve the above purpose, the present application adopts the following technical scheme:

in a first aspect, a method for calculating a bad account preparation amount is provided, the method including: obtaining credit information of a target credit customer; the credit information includes customer size, customer type, risk level, and loan amount; determining a target customer group corresponding to the target credit customer according to the customer scale, the customer type and the risk level of the target credit customer; the client scale, client type and risk level corresponding to different client groups are different; determining a target repayment model corresponding to the target customer group from a mapping relation comprising a plurality of customer groups and a plurality of repayment models according to the target customer group; a repayment model for reflecting repayment probabilities of a customer group at different time nodes; determining a metering and lifting proportion of a target credit customer based on the target repayment model; the calculating and extracting ratio is the ratio of the repayment probability of the first time node to the repayment probability of the second time node in the target repayment model; the second time node is any time node before the first time node; and determining bad account preparation amount corresponding to the second time node according to the proportion of the calculated and the loan amount.

Optionally, the method further comprises: obtaining repayment probability curves of a plurality of sample customer groups; clustering the multiple repayment probability curves according to a preset clustering algorithm to obtain a first repayment probability curve, a second repayment probability curve and a third repayment probability curve; the residual repayment amount of the sample customer group corresponding to the first repayment probability curve at the target time node is smaller than or equal to a first preset amount; the residual repayment amount of the sample customer group corresponding to the second repayment probability curve at the target time node is larger than the first preset amount and smaller than or equal to the second preset amount; the residual repayment amount of the sample customer group corresponding to the third repayment probability curve at the target time node is larger than the second preset amount; the second preset amount is larger than the first preset amount; and taking the first repayment probability curve, the second repayment probability curve and the third repayment probability curve as different repayment models.

Optionally, obtaining a payment probability curve for a plurality of sample customer groups includes: obtaining repayment record information of each sample customer group; the repayment record information is used for reflecting the residual repayment amounts of different months; for any sample customer group, determining the corresponding repayment probability of the sample customer group under different account ages based on repayment record information of the sample customer group, and obtaining a plurality of repayment probabilities; one account age corresponds to one time node; and fitting to obtain a repayment probability curve of the sample customer group based on the repayment probabilities.

Optionally, one account age is obtained by subtracting one billing month from one observation month, the billing month being used to reflect a repayment month of the customer group, the observation month being the month after the billing month; based on repayment record information of the sample customer group, determining repayment probabilities corresponding to the sample customer group at different account ages comprises the following steps: for any account age, comparing the remaining repayment amount of one bill month of the account age with the remaining repayment amount of the observation month corresponding to the upper bill month to obtain an amount ratio of the account age; determining the weight of each monetary ratio, and weighting a plurality of monetary ratios of the account age according to the weight of each monetary ratio to obtain the corresponding repayment probability of the sample customer group under the account age; the weight of one monetary ratio is positively correlated with the observed month of the monetary ratio.

In a second aspect, there is provided a bad account preparation amount calculating device, the calculating device including an acquisition unit, a determination unit; an acquisition unit configured to acquire credit information of a target credit customer; the credit information includes customer size, customer type, risk level, and loan amount; the determining unit is used for determining a target customer group corresponding to the target credit customer according to the customer scale, the customer type and the risk level of the target credit customer; the client scale, client type and risk level corresponding to different client groups are different; the determining unit is further used for determining a target repayment model corresponding to the target customer group from the mapping relation comprising the plurality of customer groups and the plurality of repayment models according to the target customer group; a repayment model for reflecting repayment probabilities of a customer group at different time nodes; the determining unit is further used for determining the metering and lifting proportion of the target credit client based on the target repayment model; the calculating and extracting ratio is the ratio of the repayment probability of the first time node to the repayment probability of the second time node in the target repayment model; the second time node is any time node before the first time node; and the determining unit is also used for determining bad account preparation amount corresponding to the second time node according to the proportion of the calculated and the loan amount.

Optionally, the acquiring unit is specifically configured to: obtaining repayment probability curves of a plurality of sample customer groups; clustering the multiple repayment probability curves according to a preset clustering algorithm to obtain a first repayment probability curve, a second repayment probability curve and a third repayment probability curve; the residual repayment amount of the sample customer group corresponding to the first repayment probability curve at the target time node is smaller than or equal to a first preset amount; the residual repayment amount of the sample customer group corresponding to the second repayment probability curve at the target time node is larger than the first preset amount and smaller than or equal to the second preset amount; the residual repayment amount of the sample customer group corresponding to the third repayment probability curve at the target time node is larger than the second preset amount; the second preset amount is larger than the first preset amount; and taking the first repayment probability curve, the second repayment probability curve and the third repayment probability curve as different repayment models.

Optionally, the acquiring unit is further specifically configured to: obtaining repayment record information of each sample customer group; the repayment record information is used for reflecting the residual repayment amounts of different months; for any sample customer group, determining the corresponding repayment probability of the sample customer group under different account ages based on repayment record information of the sample customer group, and obtaining a plurality of repayment probabilities; one account age corresponds to one time node; and fitting to obtain a repayment probability curve of the sample customer group based on the repayment probabilities.

Optionally, one account age is obtained by subtracting one billing month from one observation month, the billing month being used to reflect a repayment month of the customer group, the observation month being the month after the billing month; the determining unit is specifically configured to: for any account age, comparing the remaining repayment amount of one bill month of the account age with the remaining repayment amount of the observation month corresponding to the upper bill month to obtain an amount ratio of the account age; determining the weight of each monetary ratio, and weighting a plurality of monetary ratios of the account age according to the weight of each monetary ratio to obtain the corresponding repayment probability of the sample customer group under the account age; the weight of one monetary ratio is positively correlated with the observed month of the monetary ratio.

In a third aspect, there is provided an electronic device comprising: a processor, a memory for storing instructions executable by the processor; wherein the processor is configured to execute instructions to implement the bad account preparation amount calculation method of the first aspect described above.

In a fourth aspect, there is provided a computer-readable storage medium having instructions stored thereon that, when executed by a processor of an electronic device, enable the electronic device to perform the bad account preparation amount calculation method of the first aspect as described above.

The technical scheme provided by the embodiment of the application at least brings the following beneficial effects: the computing device acquires credit information of a target credit customer, wherein the credit information comprises a customer scale, a customer type, a risk level and a loan amount; and determining a target customer group corresponding to the target credit customer according to the customer scale, the customer type and the risk level of the target credit customer, and determining a target repayment model corresponding to the target customer group from a mapping relation comprising a plurality of customer groups and a plurality of repayment models. After the computing device determines the target repayment model corresponding to the target customer group, determining the calculation and withdrawal proportion of the target credit customer according to the target repayment model, wherein the calculation and withdrawal proportion is the ratio of the repayment probability of the first time node to the repayment probability of the second time node in the target repayment model. After the calculating device determines the proportion of the target credit customer, the bad account preparation amount corresponding to the second time point can be determined according to the proportion of the target credit customer and the loan amount. Compared with the prior art, after the account age of the client and the corresponding account age amount data are determined, the bad account preparation amount is directly calculated according to the unified preset bad account counting proportion of all the clients.

Drawings

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

Fig. 1 is a schematic structural diagram of a bad account preparation amount calculating system according to an embodiment of the present application;

fig. 2 is a schematic flow chart of a bad account preparation amount calculation method provided in an embodiment of the present application;

FIG. 3 is a schematic diagram of an average payment probability curve after clustering according to an embodiment of the present application;

fig. 4 is a second flow chart of a bad account preparation amount calculation method according to the embodiment of the present application;

fig. 5 is a schematic diagram of an accounting adjustment method according to an embodiment of the present application;

FIG. 6 is a schematic diagram of sample data of a group client according to an embodiment of the present application;

fig. 7 is a schematic diagram of mapping relationship between profile coefficients and the number of clusters according to an embodiment of the present application;

fig. 8 is a flowchart of a bad account preparation amount calculation method according to an embodiment of the present application;

FIG. 9 is a schematic diagram of a computing device according to an embodiment of the present disclosure;

fig. 10 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

Detailed Description

The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are only some, but not all, of the embodiments of the present application. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are within the scope of the present disclosure.

It should be noted that, in the embodiments of the present application, words such as "exemplary" or "such as" are used to mean serving as an example, instance, or illustration. Any embodiment or design described herein as "exemplary" or "for example" should not be construed as preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary" or "such as" is intended to present related concepts in a concrete fashion.

It should be noted that, in the embodiment of the present application, "english: of", "corresponding" and "corresponding" may sometimes be used in combination, and it should be noted that the meaning to be expressed is consistent when the distinction is not emphasized.

In order to clearly describe the technical solutions of the embodiments of the present application, in the embodiments of the present application, the terms "first", "second", and the like are used to distinguish the same item or similar items having substantially the same function and effect, and those skilled in the art will understand that the terms "first", "second", and the like are not limited in number and execution order.

Before explaining the embodiments of the present application in detail, some related technical terms and related technologies related to the embodiments of the present application are described.

The account age refers to the length of the account receivable delinquent time, the account age analysis refers to the judgment of the recoverable amount of the company and bad accounts, the bad accounts refer to the loss caused by the fact that the enterprise cannot recover or recovers the possible minimum account receivable, the bad account loss refers to the loss generated by the fact that bad accounts occur, the bad account preparation refers to the account receivable settlement of the enterprise, the account is prepared, the loss is estimated before the bad account loss occurs, the bad account preparation is formed, when the bad account loss occurs, the bad account preparation is directly washed out, and meanwhile, the corresponding account receivable amount is re-used.

In order to objectively reflect the recovery condition of accounts receivable, calculating bad account preparation amount, determining a bad account calculation proportion corresponding to a customer, when calculating the bad account preparation amount in the prior art, generally setting the bad account calculation proportion of the government enterprise customer according to an expected credit loss model after determining account age and corresponding account age amount data of the government enterprise customer, and using the set bad account calculation proportion to be uniformly used for all types of customers so as to calculate the bad account preparation amount.

However, credit risk loss characteristics of government enterprise customers are different, applicable bad account settlement proportion models are also different, the current receivables settlement models are too general for distinguishing the government enterprise customers, are not careful enough, the credit risk characteristics of the government enterprise customers cannot be distinguished finely, bad account settlement amounts are directly calculated without determining different bad account settlement proportions for different customer groups, and the situation that the bad account settlement amounts are calculated inaccurately exists is caused, so that financial information cannot reflect credit risks more comprehensively, and risk levels cannot be quantified accurately. And after implementing the new counting and extracting rule, the counting and extracting amount fluctuates greatly, and the counting and extracting amount rule is obviously not met for 3 years.

The core of credit loss accounting estimation using statistics is the parameter sequence { P ] _j Estimate of probability of repayment P of account age j-1 in next period _j In fact, the historical data has a plurality of statistics for selection, namely, the ratio of the account age j-1 at the previous period to the account amount of the account age j at the current period can be used, the ratio of the account age j-1 at the previous period to the account amount of the account age j at the previous period can also be used, and one ratio can be calculated in each period.

Let P be _j The numerator and denominator of each period can be summarized when the system is unchanged in different periods, and then the total ratio is calculated; but if P _j Itself a time-varying sequence, then all together and then computing the total ratio results in erroneous values being introduced. In fact, accounts receivable produced by customers in different periods are systematically different, and the repayment probability is affected by economic environment and other factors, so that P is caused _j There may be variations where computing the payment probability estimate credit loss with the data of a period of time is occasional, resulting in inaccurate accounting of bad account preparation amounts.

In view of the above problems, the present application provides a bad account preparation amount calculation method, in which a calculation device obtains credit information of a target credit customer, and determines a target customer group corresponding to the target credit customer according to a customer scale, a customer type, and a risk level of the target credit customer. After determining the target customer group corresponding to the target credit customer, the computing device determines a target repayment model corresponding to the target customer group from a mapping relation comprising a plurality of customer groups and a plurality of repayment models, and according to the target repayment model, the calculation and the collection proportion of the target credit customer can be determined, wherein the calculation and the collection proportion are the ratio of the repayment probability of a first time node to the repayment probability of a second time node in the target repayment model. The computing device may determine a bad account preparation amount corresponding to the second time node according to the proportion of the target credit customer and the loan amount.

The method for calculating the bad account preparation amount provided in the embodiment of the present application is described in detail below with reference to the accompanying drawings.

The bad account preparation amount calculating method provided by the embodiment of the application can be applied to a bad account preparation amount calculating system, and fig. 1 shows a schematic diagram of a structure of the bad account preparation amount calculating system. As shown in fig. 1, the bad account preparation amount calculation system 10 includes a bad account preparation amount calculation device (hereinafter referred to as calculation device) 101 and a server 102. The computing device 101 and the server 102 may be connected in a wired manner or may be connected in a wireless manner, which is not limited in the embodiment of the present application.

The server 102 may be configured to store credit information of the credit client and send the stored credit information of the credit client to the computing device 101 when the computing device 101 needs to acquire the credit information of the credit client.

The computing device 101 may obtain credit information of the target credit customer from the server 102, determine a target customer group corresponding to the target credit customer according to the customer scale, the customer type and the risk level of the target credit customer, determine a corresponding target repayment model from a mapping relationship including a plurality of customer groups and a plurality of repayment models, and further determine a metering proportion and calculate a bad account preparation amount according to the target repayment model corresponding to the target credit customer. The specific bad account preparation amount calculation method may refer to the following description of the embodiments, which is not repeated here.

The computing device 101 may be an electronic device with a data processing function, for example, the computing device 101 may be a computer, a tablet computer, a notebook computer, an ultra-mobile personal computer (UMPC), or the like, and the embodiment of the present application does not limit a specific type of the electronic device.

The server 102 may be a single server, a server cluster, a chip or a system on a chip in a server, etc., and the specific type of the server is not limited in the embodiments of the present application.

Fig. 2 is a flow diagram illustrating a bad account preparation amount calculation method according to some example embodiments. In some embodiments, the bad account preparation amount calculation method described above may be applied to a calculation apparatus as shown in fig. 1, and may also be applied to other similar devices.

As shown in fig. 2, the bad account preparation amount calculating method provided in the embodiment of the present application includes the following S201 to S205.

S201, obtaining credit information of a target credit customer.

Wherein the credit information includes customer size, customer type, risk level, loan amount.

As one possible implementation, the computing device obtains the customer scale information, the customer type information, the risk level information, and credit information such as the loan month, the loan amount, the account age, and the account age amount of the customer for the target credit customer.

The target credit client is a debt client of the corresponding proportion to be calculated; the customer scale information is used for reflecting the information such as the quantity range of products ordered by the customer, the amount of orders required to be spent by the customer for ordering the products, and the like, and different customer scale grades can be divided according to the ordered amount. The customer type includes a plurality of types such as a wisdom type customer (which is wished by own inherent thinking when selecting to purchase a product), an impulse type customer (which decides whether to purchase a product according to the mood of the user when purchasing the product), a benefit type customer (which decides the product to be purchased according to the benefit provided by a seller when purchasing the same product), a habitual customer (which is familiar and used by himself when purchasing the product), and a hesitant type customer (which is hesitant when purchasing the product), and the like, and is respectively provided with different identifications, and the clients of different types can be classified according to the attribute characteristics of the clients to adopt different service strategies. The risk level of a customer is generally classified into high risk, medium risk, low risk, medium and high risk, etc., and is usually assessed according to the type, background, purpose and nature of the customer, transparency of information, etc. The loan amount is used to reflect the amount of arrears that the customer has made after ordering the product.

S202, determining a target customer group corresponding to the target credit customer according to the customer scale, the customer type and the risk level of the target credit customer.

The client scale, the client type and the risk level corresponding to different client groups are different.

As one possible implementation manner, the computing device sets a classification level of the client scale, a classification condition of the client type and a classification condition of the client risk level, and classifies all clients into groups (into a plurality of sample client groups) according to three dimensions of the client scale, the client type and the risk level, namely, the clients meeting the same client scale, the same client type and the same risk level are classified into the same group. After the division of the group is completed, the computing device determines the group where the target credit client is located according to the client size, the client type and the risk level of the target credit client, and the group is the target client group corresponding to the target credit client.

Optionally, accounting estimations refer to predictions or inferences made regarding accounting events where some results are uncertain based on existing information, transactions that occur, and data internal and external to the company. The method for calculating the bad account preparation amount is a classical problem in accounting estimation, the problem can be solved by adopting a fixed proportion method, an account age analysis method and the like, and clients with similar repayment probability can be divided into a group by adopting a statistical grouping method in the account age analysis method according to a set grouping rule, namely, all clients are divided into specified groups according to the set grouping rule.

Illustratively, after sample data of each customer is obtained, the computing device sets a classification level of the customer scale, a classification condition of the customer type and a grading classification condition of the customer risk level, and classifies all customers into 180 subgroups according to three dimensions of the customer scale, the customer type and the risk level. For example, the computing device classifies the customer scale levels into five levels of a (order amount less than or equal to 30 ten thousand), B (order amount greater than 30 ten thousand, less than or equal to 50 ten thousand), C (order amount greater than 50 ten thousand, less than or equal to 70 ten thousand), D (order amount greater than 70 ten thousand, less than or equal to 100 ten thousand), E (order amount greater than 100 ten thousand); the client types are divided into a plurality of types such as a reasonable client (provided with a mark a), an impulse client (provided with a mark b), a profitability client (provided with a mark c), a habit client (provided with a mark d), a hesitant client (provided with a mark e) and the like; the risk classes are divided into five classes of very low risk, lower risk, medium and high risk. After the division is completed, the computing device sets an arrangement combination of three dimensions of a customer scale, a customer type and a risk level to obtain 180 preset subgroups, for example, the requirement that the customer scale level is A, the customer type is a reasonable customer, and the customer with extremely low risk level is divided into a first subgroup; the customer scale level B, the customer type being a wisdom type customer, the risk level being a very low risk customer being classified into a second group, etc. After the computing device determines the customer size, customer type, and risk level information for the target credit customer, a comparison is made with the 180 groups that were set up to determine the group in which the target credit customer is located.

S203, determining a target repayment model corresponding to the target customer group from the mapping relation comprising the plurality of customer groups and the plurality of repayment models according to the target customer group.

Wherein a repayment model is used to reflect repayment probabilities of a customer group at different time nodes.

As one possible implementation manner, after the computing device completes the division of the sample customer groups, payment record information of each sample customer group is obtained, and a payment probability curve of each sample customer group is determined according to the payment record information of each sample customer group. After obtaining a plurality of repayment probability curves, the computing device performs clustering processing on all repayment probability curves according to a K-Means (K-Means) clustering algorithm to obtain a first repayment probability curve, a second repayment probability curve and a third repayment probability curve, and the three repayment probability curves are used as different repayment models. After the computing device determines the target customer group corresponding to the target credit customer, determining a target repayment model corresponding to the target customer group according to the repayment model in which the repayment probability curve corresponding to the target customer group is located.

S204, determining the metering and lifting proportion of the target credit client based on the target repayment model.

The calculating and extracting ratio is the ratio of the repayment probability of the first time node to the repayment probability of the second time node in the target repayment model; the second time node is any time node before the first time node.

As one possible implementation manner, the computing device determines a target customer group corresponding to a target credit customer, further determines a corresponding target repayment model according to the target customer group, then uses repayment probability corresponding to a first time node (preset month) on the target repayment model as a numerator, and uses repayment probability corresponding to a second time node (month for which the calculation of the calculation proportion is required) as a denominator, so as to obtain a plurality of ratios, namely, obtain the calculation proportion corresponding to the month for which the calculation of the calculation proportion is required by the target credit customer.

For example, as shown in fig. 3, the computing device performs clustering processing on 180 repayment probability curves according to a K-Means (K-Means) clustering algorithm to obtain three types of repayment probability curves, and after determining a repayment probability curve corresponding to a target credit customer, the computing device may calculate a calculation and lifting proportion of the target credit customer according to the repayment probability curve. For example, assuming that the computing device determines that the target repayment model corresponding to the target credit customer is a second type repayment probability curve, when the calculation of the calculation proportion of the 6 th month of the customer is to be performed, the repayment probability corresponding to the 36 th month on the second type repayment probability curve is used as a numerator, the repayment probability corresponding to the 6 th month on the second type repayment probability curve is used as a denominator, and the obtained ratio is the calculation proportion of the 6 th month of the customer. Similarly, assuming that the computing device determines that the target repayment model corresponding to the target credit customer is a third-class repayment probability curve, when the 12 th month of the customer is to be calculated, using the repayment probability corresponding to the 36 th month on the third-class repayment probability curve as a molecule, and using the repayment probability corresponding to the 12 th month on the third-class repayment probability curve as a denominator, wherein the obtained ratio is the 12 th month of the customer.

S205, determining bad account preparation amount corresponding to the second time node according to the proportion of the calculated and the loan amount.

As one possible implementation manner, the computing device determines a bad account preparation amount of month corresponding to the billing proportion according to the billing proportion and the loan amount corresponding to the target credit customer.

For example, if the customer a pays 10 ten thousand yuan in the first year, pays 5 ten thousand yuan in the third year, and pays 10 ten thousand yuan in the fourth year, the account age information counted by the customer in the fourth year is: within 1 year, 10 ten thousand yuan, 1-2 years, 2-3 years, more than 3 years and 5 ten thousand yuan. Assuming that the calculated account proportion corresponding to 12 months of the customer a is 5% and the calculated account proportion corresponding to 36 months is 100%, bad account preparation corresponding to the account age amount within 1 year is 10 tens of thousands of 5% = 0.5 tens of thousands of yuan, bad account preparation corresponding to the account age amount above 3 years is 5 tens of thousands of times 100% = 5 tens of thousands of yuan, and bad account preparation amount of the customer a in the financial transaction information is 0.5+5=5.5 tens of thousands of yuan.

In one design, S203 may specifically include S2031-S2033 described below, as shown in fig. 4, in order to determine a target payment model corresponding to the target customer group.

S2031, obtaining repayment probability curves of a plurality of sample customer groups.

As one possible implementation manner, the computing device obtains repayment record information of each sample customer group; for any sample customer group, based on repayment record information of the sample customer group, determining repayment probabilities corresponding to the sample customer group at different account ages to obtain a plurality of repayment probabilities, fitting to obtain a repayment probability curve of the sample customer group based on the plurality of repayment probabilities, and sequentially calculating the repayment probabilities corresponding to each sample customer group by a calculating device to fit to obtain a plurality of repayment probability curves.

It should be noted that, the repayment record information of each sample customer group is sample data after being subjected to monotonous processing, the repayment record information is used for reflecting the residual repayment amounts of different months, one account age corresponds to one time node, one bill month is subtracted from one observation month to obtain, the bill month is used for reflecting the repayment month of the customer group, and the observation month is the month after the bill month is taken.

Optionally, the computing device obtains payment record information of a plurality of customers, collects an account age table according to the billing month and the observation month, collects accounts receivable amount, and because the integrity of the account age table cannot be completely ensured, the real data may also relate to the increase of logic of the account age table which is not matched with the account and is brought by the account adjustment, so that the accounts receivable amount data collected according to the account age table data cannot be used as customer sample data, and the conversion formula b needs to be used first _i ＝max _j≥i {a _j }，i∈N ^* A backward max-padded transformation (a) is applied to the sequence in which the increment exists _j Representing the sequence before transformation, b _i Representing the monotone sequence after transformation), namely changing the accounts receivable amount sequence with increasing into the monotone decreasing sequence, completing data processing to obtain sample data corresponding to each customer (the sample data is used for reflecting the repayment record information processed by each customer), and further obtaining the repayment record information of each sample customer group after the sample customer group is divided.

For example, a sequence constructed with sample data of a customer is shown in fig. 5, where a solid line represents an original sequence, a dotted line represents a monotonous sequence, and a horizontal x-axis represents a plurality of months recorded by the customer from a arrears month; the vertical y axis represents the amount of the arrears of the customers, the figure shows that the original sequence has the increase of the logic of the table which does not accord with the account number caused by the accounting, the original sequence needs to be processed to be changed into a monotonically decreasing sequence, the amount of the arrears of the customers on the monotonically decreasing sequence is reduced along with the increase of time, the data processing is completed to obtain the sample data of each customer, and the repayment record information of each sample customer group is obtained after the division of the sample customer group is completed.

Alternatively, the idea of contemporaneous group analysis is to divide users into different groups according to some common characteristic (e.g., time of first purchase, time of registration, age group, etc.)And (3) analyzing the variation of different populations along with the period. For a sequence { P } that varies continuously _j (available to reflect the probability of repayment of the account for ledger age j-1 in the next phase), estimate { P over using multi-phase data _j When in process, different periods are respectively given different weights, the closer to the current period, the more weight is given, the weight is continuously attenuated as the period to be calculated is continuously far away from the current period, and one common attenuation weight is that W in _k Represents the weight applicable in the lag phase k, M _k For the size of the observed amount in this period, q is an attenuation coefficient, which takes a value between 0 and 1 (the closer q is to 0, the more important the recent data is represented and the distant data is ignored; the faster the attenuation is to 1, the more tends to uniformly process the recent distant), the attenuation weight is attenuated in a certain proportion, and the sum is 1. The corresponding weighting statistic is +.>In the formula +.>Is the repayment proportion of the account age j-1 in the lag k period in the next month.

For example, after the computing device completes the division of 180 subgroups, the computing device obtains the repayment record information of each subgroup, as shown in fig. 6, takes sample data of one subgroup, takes a lag phase for the monotonous sequence of each billing month of the subgroup client, namely, moves each month backwards for one month on the current phase sequence, gathers the current phase sequence value of the subgroup client as a numerator, takes the lag phase sequence value as a denominator, and calculates the ratio. For example, a ratio is obtained by the corresponding amount of 1 month to the corresponding amount of 2 months, a ratio is obtained by the corresponding amount of 2 months to the corresponding amount of 3 months, and the ratio of the current-period sequence value to the lag-period sequence value is sequentially calculated to obtain a plurality of ratios.

Further, the computing device calculates the probability of repayment of the group of customers at different account ages, and exemplary, the payment is takenThe accounting age is 6 months, the calculating device determines a series of billing months and observation months corresponding to the accounting age of 6 months, and calculates the repayment probability according to the series of billing months and observation months corresponding to the accounting age. As shown in fig. 6, the bill month is taken 1 st year and 1 st month, then the observation month is taken 1 st year and 7 th months, the remaining repayment amount corresponding to 1 st year and 1 st month is compared with the remaining repayment amount corresponding to 1 st year and 7 th month, and an amount ratio with the account age of 6 months is obtained, according to the formula Determining the weight corresponding to the month of day 7 from the month of day 1 (the month of day is a preset month and can be the month of day 4 and the month of day 6); and similarly, comparing the residual repayment amount corresponding to the 2 nd month of the 1 st year with the residual repayment amount corresponding to the 8 th month of the 1 st year to obtain a further amount ratio of 6 months of account age, and determining the weight corresponding to the 8 th month of the observed month from the current month according to a weight formula. The calculation device sequentially calculates all the sum ratios with account ages of 6 months and the weights of the sum ratios, weights a plurality of sum ratios with account ages of 6 months according to the weights of the sum ratios after obtaining all the sum ratios and the weights of the sum ratios, obtains point estimation (calculated according to the independent event multiplication principle and obtained by accumulation) of the corresponding repayment probability of the next period when the account ages of the customers of the group are 6 months, and obtains the repayment probability corresponding to the 6 th month on the repayment probability curve of the group.

Similarly, the computing device sequentially computes the repayment probabilities of the group at different account ages to obtain a plurality of repayment probabilities, and fits the repayment probabilities to obtain a repayment probability curve of the group based on the repayment probabilities. After the computing device fits to obtain a repayment probability curve of a group, the repayment probabilities corresponding to the remaining groups are sequentially computed, and then 180 repayment probability curves are obtained through fitting.

S2032, clustering the plurality of repayment probability curves according to a preset clustering algorithm to obtain a first repayment probability curve, a second repayment probability curve and a third repayment probability curve, and taking the three repayment probability curves as different repayment models.

The method comprises the steps that a sample customer group corresponding to a first repayment probability curve has a residual repayment amount smaller than or equal to a first preset amount at a target time node; the residual repayment amount of the sample customer group corresponding to the second repayment probability curve at the target time node is larger than the first preset amount and smaller than or equal to the second preset amount; the residual repayment amount of the sample customer group corresponding to the third repayment probability curve at the target time node is larger than the second preset amount; the second preset amount is greater than the first preset amount.

As a possible implementation manner, the computing device invokes a K-Means clustering algorithm, performs clustering according to the similarity of the repayment probability curves, determines the number of clusters, and then clusters all the repayment probability curves according to the number of clusters to obtain three clusters of average repayment probability curves, thereby obtaining three different repayment models.

Alternatively, clustering is a standard practice in machine learning algorithms to solve the unsupervised grouping problem. Clustering is often used when it is desired to group a population into groups according to similarity and variability within the population, but without a correct classification result that can be referenced. It divides the data in the dataset into groups or clusters such that the data points within each cluster are highly similar, while the data points between different clusters are less similar. The calculation device can be mainly divided into a partition-based clustering mode, a density-based clustering mode, a hierarchical clustering mode and the like, and the calculation device can use a K-Means clustering algorithm in the partition-based clustering algorithm to cluster the repayment probability curve.

Illustratively, as shown in FIG. 7, the lateral direction represents the x-axis, representing the number of clusters; the vertical axis represents the y axis, and represents the contour coefficient, when the number of clusters is more than or equal to 5, the contour coefficient is lower and is not considered; when the number of clusters is 2, the contour coefficient is the largest, but the number of clusters is too small to consider; when the number of clusters is 3 or 4, the profile coefficients are similar, and for conciseness of business meaning, the computing device determines the number of clusters K=3 by combining business experience. After the number of clusters is determined, the computing device clusters 180 repayment probability curves according to a K-Means clustering algorithm to obtain three repayment probability curves, wherein the first repayment probability curve is positioned below the second repayment probability curve as shown in fig. 3, and the corresponding clients repayment rapidly and almost no residues exist after 12 months; the second type repayment probability curve is positioned between the first type repayment probability curve and the third type repayment probability curve, and a certain proportion of corresponding clients remain after 12 months; the third type of repayment probability curve is above the second type of repayment probability curve, and the corresponding customer repayment is slowest and the residual proportion is higher after 12 months.

S2033, determining a target repayment model corresponding to the target customer group from the mapping relation between the plurality of sample customer groups and the three repayment models.

As one possible implementation manner, after the computing device determines the target customer group corresponding to the target credit customer, determining a target repayment model corresponding to the target customer group according to the repayment model in which the repayment probability curve corresponding to the target customer group is located.

Illustratively, the computing device determines that the target credit customer is in a group 1 of the 180 groups, knowing that the repayment probability curve corresponding to group 1 is clustered into a first type repayment probability curve, and the target repayment model corresponding to the target credit customer is the first type repayment probability curve.

As shown in fig. 8, the computing device performs data preparation to obtain repayment record information of a plurality of clients, and obtains repayment record information processed by the clients through business interviews, demand investigation, exploratory data analysis (Exploratory Data Analysis, EDA), data processing (Extract-form-Load, ETL) and accounting reconciliation (backward maximum filling), thereby obtaining sample data of the clients. After sample data of a plurality of clients are obtained, a computing device builds an initial model through data analysis modeling, trains the initial model, namely processes the sample data of the clients in the initial model, divides all the clients into 180 subgroups according to three dimensions of client scale, client type and risk level by adopting a statistical grouping method in an account age analysis method, carries out contemporaneous group analysis on the sample data of the 180 subgroups to obtain repayment probabilities of the 180 subgroups under different account ages, and then fits to obtain 180 repayment probability curves.

Further, the computing device performs clustering processing on the 180 repayment probability curves according to a K-Means algorithm to obtain three types of repayment probability curves (three different repayment models), and a plurality of calculation and extraction ratios corresponding to each type of repayment probability curves can be calculated according to the three types of repayment probability curves and used as model output, training of the computing device on an initial model is completed, and a more accurate credit loss estimation model is generated. When new customer sample data is input into the trained model, the model firstly judges the group to which the customer belongs, further judges a repayment model corresponding to the customer, and takes a plurality of metering and lifting ratios corresponding to the repayment model as a plurality of metering and lifting ratios corresponding to the customer, further calculates bad account preparation amount according to the customer sample data and the metering and lifting ratios.

The technical scheme provided by the embodiment of the application at least brings the following beneficial effects: the computing device acquires credit information of a target credit customer, wherein the credit information comprises a customer scale, a customer type, a risk level and a loan amount; and determining a target customer group corresponding to the target credit customer according to the customer scale, the customer type and the risk level of the target credit customer, and determining a target repayment model corresponding to the target customer group from a mapping relation comprising a plurality of customer groups and a plurality of repayment models. After the computing device determines the target repayment model corresponding to the target customer group, determining the calculation and withdrawal proportion of the target credit customer according to the target repayment model, wherein the calculation and withdrawal proportion is the ratio of the repayment probability of the first time node to the repayment probability of the second time node in the target repayment model. After the calculating device determines the proportion of the target credit customer, the bad account preparation amount corresponding to the second time point can be determined according to the proportion of the target credit customer and the loan amount. Compared with the prior art, after the account age of the client and the corresponding account age amount data are determined, the bad account preparation amount is directly calculated according to the unified preset bad account counting proportion of all the clients.

The foregoing embodiments mainly describe the solutions provided in the embodiments of the present application from the perspective of the apparatus (device). It will be appreciated that, in order to implement the above-mentioned method, the apparatus or device includes hardware structures and/or software modules corresponding to each of the method flows, and these hardware structures and/or software modules corresponding to each of the method flows may constitute a material information determining apparatus. Those of skill in the art will readily appreciate that the algorithm steps of the examples described in connection with the embodiments disclosed herein may be implemented as hardware or a combination of hardware and computer software. Whether a function is implemented as hardware or computer software driven hardware depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.

The embodiment of the application may divide the functional modules of the apparatus or the device according to the above method example, for example, the apparatus or the device may divide each functional module corresponding to each function, or may integrate two or more functions into one processing module. The integrated modules may be implemented in hardware or in software functional modules. It should be noted that, in the embodiment of the present application, the division of the modules is schematic, which is merely a logic function division, and other division manners may be implemented in actual implementation.

FIG. 9 is a schematic diagram of a computing device shown according to an example embodiment. Referring to fig. 9, a computing device 30 provided in an embodiment of the present application includes an acquisition unit 301 and a determination unit 302.

An acquiring unit 301, configured to acquire credit information of a target credit client; the credit information includes customer size, customer type, risk level, and loan amount; a determining unit 302, configured to determine, according to a client size, a client type, and a risk level of the target credit client, a target client group corresponding to the target credit client; the client scale, client type and risk level corresponding to different client groups are different; the determining unit 302 is further configured to determine, according to the target customer group, a target payment model corresponding to the target customer group from a mapping relationship including a plurality of customer groups and a plurality of payment models; a repayment model for reflecting repayment probabilities of a customer group at different time nodes; a determining unit 302, configured to determine a proportion of the target credit customer based on the target repayment model; the calculating and extracting ratio is the ratio of the repayment probability of the first time node to the repayment probability of the second time node in the target repayment model; the second time node is any time node before the first time node; the determining unit 302 is further configured to determine a bad account preparation amount corresponding to the second time node according to the proportion of the calculated and the loan amount.

Optionally, the acquiring unit 301 is specifically configured to: obtaining repayment probability curves of a plurality of sample customer groups; clustering the multiple repayment probability curves according to a preset clustering algorithm to obtain a first repayment probability curve, a second repayment probability curve and a third repayment probability curve; the residual repayment amount of the sample customer group corresponding to the first repayment probability curve at the target time node is smaller than or equal to a first preset amount; the residual repayment amount of the sample customer group corresponding to the second repayment probability curve at the target time node is larger than the first preset amount and smaller than or equal to the second preset amount; the residual repayment amount of the sample customer group corresponding to the third repayment probability curve at the target time node is larger than the second preset amount; the second preset amount is larger than the first preset amount; and taking the first repayment probability curve, the second repayment probability curve and the third repayment probability curve as different repayment models.

Optionally, the obtaining unit 301 is further specifically configured to: obtaining repayment record information of each sample customer group; the repayment record information is used for reflecting the residual repayment amounts of different months; for any sample customer group, determining the corresponding repayment probability of the sample customer group under different account ages based on repayment record information of the sample customer group, and obtaining a plurality of repayment probabilities; one account age corresponds to one time node; and fitting to obtain a repayment probability curve of the sample customer group based on the repayment probabilities.

Optionally, one account age is obtained by subtracting one billing month from one observation month, the billing month being used to reflect a repayment month of the customer group, the observation month being the month after the billing month; the determining unit 302 is specifically configured to: for any account age, comparing the remaining repayment amount of one bill month of the account age with the remaining repayment amount of the observation month corresponding to the upper bill month to obtain an amount ratio of the account age; determining the weight of each monetary ratio, and weighting a plurality of monetary ratios of the account age according to the weight of each monetary ratio to obtain the corresponding repayment probability of the sample customer group under the account age; the weight of one monetary ratio is positively correlated with the observed month of the monetary ratio.

Fig. 10 is a schematic structural diagram of an electronic device provided in the present application. As shown in fig. 10, the electronic device 40 may include at least one processor 401 and a memory 402 for storing processor executable instructions, wherein the processor 401 is configured to execute the instructions in the memory 402 to implement the bad account preparation amount calculation method in the above embodiment.

In addition, the electronic device 40 may also include a communication bus 403 and at least one communication interface 404.

The processor 401 may be a processor (central processing units, CPU), a microprocessor unit, ASIC, or one or more integrated circuits for controlling the execution of the programs of the present application.

Communication bus 403 may include a pathway to transfer information between the aforementioned components.

The communication interface 404 uses any transceiver-like device for communicating with other devices or communication networks, such as ethernet, radio access network (radio access network, RAN), wireless local area network (wireless local area networks, WLAN), etc.

The memory 402 may be, but is not limited to, read-only memory (ROM) or other type of static storage device that can store static information and instructions, random access memory (random access memory, RAM) or other type of dynamic storage device that can store information and instructions, but may also be electrically erasable programmable read-only memory (EEPROM), compact disc-read only memory (compact disc read-only memory) or other optical disk storage, optical disk storage (including compact disc, laser disc, optical disc, digital versatile disc, blu-ray disc, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer. The memory may be implemented separately and coupled to the processor 401 via a bus. The memory may also be integrated with the processor 401.

The memory 402 is used for storing instructions for executing the embodiments of the present application, and the processor 401 controls the execution. The processor 401 is configured to execute instructions stored in the memory 402, thereby implementing the functions in the methods of the present application.

As an example, in connection with fig. 9, the acquisition unit 301, the determination unit 302 in the computing device 30 realize the same functions as the processor 401 in fig. 10.

In a particular implementation, as one embodiment, processor 401 may include one or more CPUs, such as CPU0 and CPU1 in FIG. 10.

In a particular implementation, electronic device 40 may include multiple processors, such as processor 401 and processor 407 in FIG. 10, as one embodiment. Each of these processors may be a single-core (single-CPU) processor or may be a multi-core (multi-CPU) processor. A processor herein may refer to one or more devices, circuits, and/or processing cores for processing data (e.g., computer program instructions).

In a particular implementation, electronic device 40 may also include an output device 405 and an input device 406, as one embodiment. The output device 405 communicates with the processor 401 and may display information in a variety of ways. For example, the output device 405 may be a liquid crystal display (liquid crystal display, LCD), a light emitting diode (light emitting diode, LED) display device, a Cathode Ray Tube (CRT) display device, or a projector (projector), or the like. The input device 406 is in communication with the processor 401 and may accept input of a user object in a variety of ways. For example, the input device 406 may be a mouse, keyboard, touch screen device, or sensing device, among others.

Those skilled in the art will appreciate that the structure shown in fig. 10 is not limiting of the electronic device 40 and may include more or fewer components than shown, or may combine certain components, or may employ a different arrangement of components.

In addition, the present application also provides a computer-readable storage medium, which when executed by a processor of an electronic device, enables the electronic device to perform the bad account preparation amount calculation method provided by the above embodiment.

In addition, the application also provides a computer program product comprising computer instructions which, when executed on an electronic device, cause the electronic device to perform the bad account preparation amount calculation method as provided in the above embodiments.

Other embodiments of the present application will be apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the application following, in general, the principles of the application and including such departures from the present disclosure as come within known or customary practice within the art to which the application pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the application being indicated by the following claims.

Claims (10)

1. A bad account preparation amount calculation method, the method comprising:

obtaining credit information of a target credit customer; the credit information comprises a customer scale, a customer type, a risk level and a loan amount;

determining a target customer group corresponding to the target credit customer according to the customer scale, the customer type and the risk level of the target credit customer; the client scale, client type and risk level corresponding to different client groups are different;

determining a target repayment model corresponding to the target customer group from a mapping relation comprising a plurality of customer groups and a plurality of repayment models according to the target customer group; a repayment model for reflecting repayment probabilities of a customer group at different time nodes;

determining a meter-to-offer ratio of the target credit customer based on the target repayment model; the calculating and lifting proportion is the ratio of the repayment probability of the first time node to the repayment probability of the second time node in the target repayment model; the second time node is any time node before the first time node;

and determining bad account preparation amount corresponding to the second time node according to the proportion and the loan amount.

2. The method according to claim 1, wherein the method further comprises:

obtaining repayment probability curves of a plurality of sample customer groups;

clustering the multiple repayment probability curves according to a preset clustering algorithm to obtain a first repayment probability curve, a second repayment probability curve and a third repayment probability curve; the residual repayment amount of the sample customer group corresponding to the first repayment probability curve at the target time node is smaller than or equal to a first preset amount; the residual repayment amount of the sample customer group corresponding to the second repayment probability curve at the target time node is larger than the first preset amount and smaller than or equal to the second preset amount; the residual repayment amount of the sample customer group corresponding to the third repayment probability curve at the target time node is larger than the second preset amount; the second preset amount is larger than the first preset amount;

and taking the first repayment probability curve, the second repayment probability curve and the third repayment probability curve as different repayment models.

3. The method of claim 2, wherein the obtaining a repayment probability curve for a plurality of sample customer populations comprises:

Obtaining repayment record information of each sample customer group; the repayment record information is used for reflecting the residual repayment amounts of different months;

for any sample customer group, determining the corresponding repayment probability of the sample customer group under different account ages based on repayment record information of the sample customer group, and obtaining a plurality of repayment probabilities; one account age corresponds to one time node;

and fitting to obtain a repayment probability curve of the sample customer group based on the repayment probabilities.

4. A method according to claim 3, wherein an account age is obtained by subtracting a billing month from an observed month, said billing month being used to reflect the repayment month of said customer base, said observed month taking the month following said billing month; the determining, based on the payment record information of the sample customer group, a payment probability corresponding to the sample customer group at different account ages includes:

for any account age, comparing the remaining repayment amount of one bill month of the account age with the remaining repayment amount of the observation month corresponding to the bill month to obtain an amount ratio of the account age;

Determining the weight of each monetary ratio, and weighting a plurality of monetary ratios of the account age according to the weight of each monetary ratio to obtain the corresponding repayment probability of the sample customer group under the account age; the weight of an monetary ratio is positively correlated with the observed month of said monetary ratio.

5. A bad account preparation amount calculating device, which is characterized in that the calculating device comprises an acquisition unit and a determination unit;

the acquisition unit is used for acquiring credit information of the target credit client; the credit information comprises a customer scale, a customer type, a risk level and a loan amount;

the determining unit is used for determining a target customer group corresponding to the target credit customer according to the customer scale, the customer type and the risk level of the target credit customer; the client scale, client type and risk level corresponding to different client groups are different;

the determining unit is further configured to determine, according to the target customer group, a target payment model corresponding to the target customer group from a mapping relationship including a plurality of customer groups and a plurality of payment models; a repayment model for reflecting repayment probabilities of a customer group at different time nodes;

The determining unit is further configured to determine a proportion of the target credit customer based on the target repayment model; the calculating and lifting proportion is the ratio of the repayment probability of the first time node to the repayment probability of the second time node in the target repayment model; the second time node is any time node before the first time node;

and the determining unit is further used for determining bad account preparation amount corresponding to the second time node according to the proportion and the loan amount.

6. The computing device according to claim 5, wherein the acquisition unit is specifically configured to: obtaining repayment probability curves of a plurality of sample customer groups; clustering the multiple repayment probability curves according to a preset clustering algorithm to obtain a first repayment probability curve, a second repayment probability curve and a third repayment probability curve; the residual repayment amount of the sample customer group corresponding to the first repayment probability curve at the target time node is smaller than or equal to a first preset amount; the residual repayment amount of the sample customer group corresponding to the second repayment probability curve at the target time node is larger than the first preset amount and smaller than or equal to the second preset amount; the residual repayment amount of the sample customer group corresponding to the third repayment probability curve at the target time node is larger than the second preset amount; the second preset amount is larger than the first preset amount; and taking the first repayment probability curve, the second repayment probability curve and the third repayment probability curve as different repayment models.

7. The computing device of claim 6, wherein the acquisition unit is further specifically configured to: obtaining repayment record information of each sample customer group; the repayment record information is used for reflecting the residual repayment amounts of different months; for any sample customer group, determining the corresponding repayment probability of the sample customer group under different account ages based on repayment record information of the sample customer group, and obtaining a plurality of repayment probabilities; one account age corresponds to one time node; and fitting to obtain a repayment probability curve of the sample customer group based on the repayment probabilities.

8. The computing device of claim 7, wherein an account age is derived from an observed month minus a billing month, the billing month being used to reflect a repayment month of the customer base, the observed month taking a month following the billing month; the determining unit is specifically configured to: for any account age, comparing the remaining repayment amount of one bill month of the account age with the remaining repayment amount of the observation month corresponding to the bill month to obtain an amount ratio of the account age; determining the weight of each monetary ratio, and weighting a plurality of monetary ratios of the account age according to the weight of each monetary ratio to obtain the corresponding repayment probability of the sample customer group under the account age; the weight of an monetary ratio is positively correlated with the observed month of said monetary ratio.

9. An electronic device, comprising: a processor, a memory for storing instructions executable by the processor; wherein the processor is configured to execute instructions to implement the bad account preparation amount calculation method of any of claims 1-4.

10. A computer readable storage medium having instructions stored thereon, which when executed by a processor of an electronic device, cause the electronic device to perform the bad account preparation amount calculation method of any of claims 1-4.