Testing method and device for value-at-risk measurement model, medium and program product

By filtering and interpolation, a test data set that is closer to the current market fluctuations is constructed, which solves the problems of the differences between historical data and current market and high data storage in the existing technology, and improves the testing accuracy of the risk value measurement model and efficiency.

CN118967315APending Publication Date: 2024-11-15CHINA INT CAPITAL CORP LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410889532.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-03
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

The existing value-of-risk measurement model testing methods have problems such as historical deviations from the current market conditions, excessive data spans lead to differences, data breakpoints affect simulation, and high data collection and storage requirements.

Method used

By obtaining historical price data and dividing it into fitting intervals and benchmark intervals, similar fluctuations are selected based on the fluctuations of the benchmark intervals, interpolation processing is performed, and a test data set is constructed to test the risk value measurement model.

Benefits of technology

The built test data set can better reflect the current market price fluctuations, reduce the demand for historical data, reduce purchase and storage costs, and improve the accuracy and efficiency of testing.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118967315A_ABST
    Figure CN118967315A_ABST
Patent Text Reader

Abstract

The invention provides a test method for testing a value-at-risk measurement model. The test method comprises the following steps: acquiring historical price data of each resource in a to-be-analyzed resource combination in an evaluation time interval; for each resource, dividing the evaluation time interval into a fitting interval and a reference interval; for each resource, based on the fluctuation condition of the historical price data of the reference interval, screening out historical price data with similar fluctuation conditions from the historical price data of the fitting interval as screened price data; for each resource, performing interpolation on a discontinuous part in the screened price data to obtain continuous fitting price data; and testing the value-at-risk measurement model by using a test data set formed by fitting price data of each resource in the fitting interval and historical price data of each resource in the reference interval. The invention further relates to a computing device, a computer readable storage medium and a computer program product which implement the testing method.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to the field of data, and in particular, to a testing method for testing a risk value measurement model, as well as a computing device, a computer-readable storage medium, and a computer program product that implement the testing method. Background Art

[0002] Value at risk refers to the maximum potential loss of a portfolio of resources (including commodities, equities, stocks, bonds, options, etc. in the financial field) held by resource holders (including financial institutions in the financial field (such as securities companies, etc.) and individuals in the financial field) when market fluctuations (such as fluctuations in interest rates, exchange rates, and prices) occur under a specific holding period and confidence level. Since value at risk can comprehensively reflect the risk exposure of all assets of resource holders at a specific time in the future, it has become an important tool for regulatory agencies and financial institutions to conduct risk management.

[0003] When measuring risk value, a risk value measurement model is usually constructed, and then the risk value measurement model is tested using historical price data. The commonly used method is to compare the risk value obtained based on the risk value measurement model using historical price data over the past period with the profit and loss incurred over the past period, and check the deviation of different quantiles to verify the rationality of the model. However, this testing method has certain defects, including: first, there may be a large deviation between the historical situation and the current market situation (for example, there may have been a stock market crash in history, while the current market situation is relatively stable), so that when there is a large deviation between the historical situation and the current market situation, the historical price data of the past period is not suitable for testing the risk value measurement model; second, when the span of the test data set of the historical price data used for testing is too large, there may be a large difference between the early market conditions reflected by the historical price data and the current market conditions; third, there are usually breakpoints in the historical price data used for testing (such as new stocks or no transactions on trading days), which may make it impossible to simulate market fluctuations; fourth, a large amount of historical data is usually required to test the model, but the fees for financial data are relatively expensive, and the access of a large amount of data of different targets is not only very time-consuming, but also has high requirements for storage devices. Summary of the invention

[0004] According to a first aspect of the present disclosure, a testing method for testing a risk value measurement model is provided, the testing method comprising: obtaining historical price data of each resource in the resource combination to be analyzed in an evaluation time interval, wherein the evaluation time interval is from the evaluation start time point to the evaluation end time point; for each resource, obtaining an interval division time point between the evaluation start time point and the evaluation end time point, dividing the evaluation time interval into a fitting interval and a benchmark interval, wherein the fitting interval is from the evaluation start time point to the interval division time point, and the benchmark interval is from the interval division time point to the evaluation end time point; for each resource, based on the fluctuation of the historical price data in the benchmark interval, historical price data with similar fluctuation conditions are screened out from the historical price data in the fitting interval as screening price data; for each resource, interpolating the discontinuous parts in the screening price data to obtain continuous fitting price data; and testing the risk value measurement model using a test data set consisting of the fitting price data of each resource in the fitting interval and the historical price data in the benchmark interval.

[0005] According to some exemplary embodiments, the obtaining of historical price data of each resource in the resource combination to be analyzed in an evaluation time interval further includes: adding a label to each resource in the resource combination to be analyzed.

[0006] According to some exemplary embodiments, obtaining the interval division time point between the evaluation start time point and the evaluation end time point for each resource also includes: obtaining the interval division time point in the same batch for all resources with the same label in the resource combination to be analyzed.

[0007] According to some exemplary embodiments, for each resource, based on the fluctuation of the historical price data of the benchmark interval, selecting the historical price data with similar fluctuation as the screening price data from the historical price data of the fitting interval also includes: calculating the historical price data fluctuation vector of the benchmark interval based on the historical price data of the benchmark interval; dividing the fitting interval into a plurality of fitting sub-intervals in chronological order, the number of historical price data included in each fitting sub-interval being equal to the number of historical price data included in the benchmark interval, and there being a time shift of a preset time length between adjacent fitting sub-intervals in chronological order; calculating the historical price data fluctuation vector of each fitting sub-interval based on the historical price data of each fitting sub-interval; calculating the cosine similarity between the historical price data fluctuation vector of each fitting sub-interval and the historical price data fluctuation vector of the benchmark interval; comparing the cosine similarity corresponding to each fitting sub-interval with a preset similarity threshold, and determining the fitting sub-interval whose cosine similarity is greater than the similarity threshold; removing duplicate historical price data from the historical price data included in all the fitting sub-intervals whose determined cosine similarity is greater than the similarity threshold, thereby obtaining the screening price data.

[0008] According to some exemplary embodiments, the preset time shift is 1 day.

[0009] According to some exemplary embodiments, the similarity threshold is 0.8.

[0010] According to some exemplary embodiments, for each resource, interpolating the discontinuous parts of the filtered price data to obtain continuous fitting price data further comprises: based on equation Interpolate the discontinuous parts of the screened price data to obtain the fitted price data; wherein S t Represents the price at time point t; t represents the time point; Δt represents the time interval; u represents the expected value of the resource's rate of return; σ represents the volatility of the resource's rate of return; ∈ represents random disturbance.

[0011] According to some exemplary embodiments, interpolating the discontinuous parts in the filtered price data for each resource to obtain continuous fitting price data also includes: performing slope interpolation on the discontinuous parts based on the numerical values ​​at the breakpoints of the discontinuous parts in the filtered price data to obtain the fitting price data.

[0012] According to some exemplary embodiments, interpolating the discontinuous parts in the filtered price data for each resource to obtain continuous fitting price data also includes: using the numerical values ​​at the breakpoints of the discontinuous parts in the filtered price data as data in the discontinuous parts to obtain the fitting price data.

[0013] According to some exemplary embodiments, the testing of the risk value measurement model using the fitted price data of each resource in the fitting interval and the historical price data in the benchmark interval also includes: using the test data set to generate a profit and loss sequence of the resource combination to be analyzed; determining corresponding quantiles based on the profit and loss sequence of the resource combination to be analyzed, and using the values ​​of the quantiles as test benchmark values; inputting the test data set into the risk value measurement model to obtain the profit and loss sequence output by the model; and evaluating the risk value measurement model based on the profit and loss sequence output by the model and the quantiles.

[0014] According to a second aspect of the present disclosure, a computing device is provided, comprising a processor and a memory, wherein a computer program is stored in the memory, wherein the processor executes the computer program to implement the steps of the testing method according to the first aspect of the present disclosure and its various exemplary embodiments.

[0015] According to a third aspect of the present disclosure, a storage medium is provided, on which a computer program / instruction is stored, wherein the computer program / instruction, when executed by a processor, implements the steps of the testing method according to the first aspect of the present disclosure and its various exemplary embodiments.

[0016] According to a fourth aspect of the present disclosure, a computer program product is provided, comprising a computer program / instructions, wherein the computer program / instructions, when executed by a processor, implement the steps of the testing method according to the first aspect of the present disclosure and its various exemplary embodiments.

[0017] Therefore, in the test method for testing the risk value measurement model provided by the present disclosure, the historical price data in the fitting area is screened based on the historical price data of the benchmark interval that reflects the current market price fluctuations, the historical price data similar to the current market price fluctuations are retained, the historical price data that may have a large deviation between the historical situation and the current market situation are removed, and the interpolation method is used to supplement the breakpoints in the historical price data. Therefore, compared with the existing test methods, the test data constructed according to the test method provided by the present disclosure can better reflect the current market price fluctuations, thereby making the evaluation of the risk value measurement model more accurate. In addition, since the historical price data that may have a large deviation between the historical situation and the current market situation are removed from the newly constructed test data set, the final test data can better reflect the current market price fluctuations. Therefore, according to the test method provided by the present disclosure, much less historical price data can be used to construct the test data set for testing, which helps to save the cost of purchasing historical price data on the one hand, and can also reduce the time consumed by data access on the other hand, and can reduce the demand for storage devices for storing historical price data, reduce the demand for data transmission, and even reduce the demand for computing power. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Hereinafter, various exemplary embodiments of the present disclosure will be described in detail with reference to the accompanying drawings; in the accompanying drawings:

[0019] Figure 1 According to an exemplary embodiment of the present disclosure, an application scenario of the testing method according to the present disclosure is schematically shown;

[0020] Figure 2 According to an exemplary embodiment of the present disclosure, various steps of a testing method for testing a risk value measurement model are shown in the form of a flow chart;

[0021] Figure 3 Schematically showing a price fluctuation curve reflecting resource price fluctuations in a reference interval according to an exemplary embodiment of the present disclosure;

[0022] Figure 4 According to an exemplary embodiment of the present disclosure, it is further shown Figure 2 Details of the test methods shown;

[0023] Figure 5 Schematically showing a curve formed by selecting price data in a fitting interval according to an exemplary embodiment of the present disclosure;

[0024] Figure 6 According to an exemplary embodiment of the present disclosure, it is further shown Figure 2Details of the test methods shown;

[0025] Figure 7 According to an exemplary embodiment of the present disclosure, a hardware environment related to the testing of a risk value measurement model is schematically shown.

[0026] It should be understood that the accompanying drawings are only schematic illustrations of exemplary embodiments of the present disclosure, and are not intended to limit the present disclosure, nor are they necessarily drawn to scale. In addition, in the accompanying drawings of the present disclosure, the same or similar features are indicated by the same or similar reference numerals. DETAILED DESCRIPTION

[0027] Various exemplary embodiments of the present disclosure are described below in conjunction with the contents shown in the accompanying drawings, so that those skilled in the art can fully understand and implement the technical solutions according to the present disclosure.

[0028] It should be understood that the term "resource" used in this disclosure should be understood in a broad sense, including various commodities, equities, stocks, bonds, options, etc. in the general sense in the financial field. In addition, in the context of this disclosure, the risk value refers to the maximum potential loss of the resource portfolio held by the resource holder under a specific holding period and confidence level when market fluctuations (e.g., fluctuations in interest rates, exchange rates, prices) occur. The risk value can usually be expressed in quantiles, such as the 99% quantile refers to the maximum loss amount of the resource portfolio held by the resource holder in the next day under a 99% probability, and the 95% quantile refers to the maximum loss amount of the resource portfolio held by the resource holder in the next day under a 95% probability.

[0029] The term "database" used in this disclosure refers to a data system that can provide data for various objects. It should be understood that in this disclosure, the term should be understood in a broad sense, that is, all functional entities that can be used for data storage and data provision can be regarded as databases.

[0030] See also Figure 1 , which schematically illustrates an application scenario of the test method according to the present disclosure according to an exemplary embodiment of the present disclosure. Figure 1 As shown, the application scenario 100 may include: a management end 110 operated by an administrator 160, a risk value measurement model management platform 120, a database 130, and a user terminal 150 operated by a user 170, wherein the management end 110, the risk value measurement model management platform 120, the database 130 and the user terminal 150 can be connected to each other through a network 140.

[0031] The management terminal 110 may be a terminal device such as a smart phone, a tablet computer, a portable computer, a dedicated computer, etc. In some exemplary embodiments, the management terminal 110 may also be a server-side device, and it may be an independent physical system, or a system cluster or distributed system composed of multiple physical systems, or a cloud system that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, distribution networks (Content Delivery Network, CDN), and big data and artificial intelligence platforms.

[0032] The management terminal 110 may install and run an application for operating the risk value measurement model management platform 120, so that the required test data set can be constructed for the risk value measurement model running on the risk value measurement model management platform 120 based on the resource combination to be analyzed, and each risk value measurement model running can be tested based on the test data set. As a non-limiting example, the management terminal 110 may be a device used by the administrator 160, in which an administrator account may be logged in to the running application. Therefore, the administrator 160 can construct a test data set for the risk value measurement model running on the risk value measurement model management platform 120 through the management terminal 110, test the risk value measurement model based on the test data set to verify its rationality, and can also provide the auxiliary support required for the daily operation of the risk value measurement model on the risk value measurement model management platform 120. It should be understood that the term "administrator" mentioned in the present disclosure should be understood in a broad sense, and it can be a person or any functional entity that can operate an application, for example, a functional entity that triggers the operation of an application based on the analysis of a person's operating behavior. In addition, those skilled in the art may know that the number of the management terminals 110 in the application scenario 100 may be more or less. For example, there may be only one management terminal 110, or two or more management terminals 110. The present disclosure does not limit the specific number and type of the management terminals 110.

[0033] The risk value measurement model management platform 120 can generally be a server-side device, which can be an independent physical system, or a system cluster or distributed system composed of multiple physical systems. In addition, the risk value measurement model management platform 120 can also be a cloud system that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, distribution networks, and big data and artificial intelligence platforms. Corresponding risk value measurement models can be run on the risk value measurement model management platform 120, and these risk value measurement models can be tested to verify their rationality, and can also provide relevant services involving risk value measurement to corresponding users.

[0034] The database 130 is used to store the historical price data of various resources, and can provide the historical price data of the corresponding resources to the risk value measurement model management platform 120 as needed. The database 130 can be implemented as a separate storage device, or as a collection of multiple storage devices connected to a network. It should be understood that the database 130 can also be a collection of data that is stored together in a certain way, can be shared with multiple users, has as little redundancy as possible, and is independent of the application program. Therefore, in a broad sense, all functional entities that can be used for data storage and data provision can be regarded as databases.

[0035] The user terminal 150 can be operated by the user 170 so as to utilize the risk value measurement model running on the risk value measurement model management platform 120 to evaluate the maximum potential loss of the resource portfolio held by the resource holder when market fluctuations occur, thereby performing risk management. The user terminal 150 can also be a terminal device such as a smart phone, a tablet computer, a portable computer, a dedicated computer, etc. In addition, the user terminal 150 can also be a server-side device, and it can be an independent physical system, or a system cluster or distributed system composed of multiple physical systems, or a cloud system that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, distribution networks, and big data and artificial intelligence platforms.

[0036] As a non-limiting example, the user terminal 150 may be a device used by the user 170, in which a user account may be logged in to the running application. Therefore, the user 170 may use the risk value measurement model running on the server 120 to perform assessment and risk management through the user terminal 150. It should be understood that the term "user" mentioned in the present disclosure should be understood in a broad sense, and it may be a person, or any functional entity capable of operating an application, for example, a functional entity that triggers the operation of an application based on the analysis of a person's operating behavior. In addition, it is known to those skilled in the art that the number of the above-mentioned user terminals 150 in the application scenario 100 may be more or less. For example, the user terminal 150 may be only one, but it may also be two or more. The present disclosure does not limit the specific number and type of user terminals 150.

[0037] The network 140 communicatively connects the management terminal 110, the risk value measurement model management platform 120, the database 130, and the user terminal 150. As a non-limiting example, the network 140 may include any combination of a local area network (LAN), a wide area network (WAN), a personal area network (PAN), and / or a communication network such as the Internet. The present disclosure does not limit the specific form of the network 140.

[0038] The cloud computing mentioned above is a computing model that distributes computing tasks on a resource pool composed of a large number of computers, so that various application systems can obtain computing power, storage space and information services as needed. The network that provides resources is called the "cloud". The resources in the "cloud" are infinitely expandable in the eyes of users, and can be obtained at any time, used on demand, expanded at any time, and paid for by use. As a basic capability provider of cloud computing, a cloud computing resource pool (referred to as a cloud platform, generally referred to as an IaaS (Infrastructure as a Service) platform) will be established, and various types of virtual resources will be deployed in the resource pool for external customers to choose to use. The cloud computing resource pool mainly includes: computing devices (virtualized machines, including operating systems), storage devices, and network devices. According to the logical function division, the PaaS (Platform as a Service) layer can be deployed on the IaaS (Infrastructure as a Service) layer, and the SaaS (Software as a Service) layer can be deployed on the PaaS layer. SaaS can also be deployed directly on IaaS. PaaS is a platform for software operation, such as databases, web containers, etc. SaaS is a variety of business software, such as web portals, SMS mass senders, etc. Generally speaking, SaaS and PaaS are upper layers relative to IaaS. With the development of the Internet, real-time data streams, and the diversification of connected devices, as well as the promotion of search services, social networks, mobile commerce, and open collaboration, cloud computing has developed very rapidly, and it will promote revolutionary changes in the entire Internet model, application services (including financial services, etc.), business models, etc.

[0039] See also Figure 2 According to an exemplary embodiment of the present disclosure, the various steps of a testing method for testing a risk value measurement model are shown in the form of a flow chart. The testing method can be applied to Figure 1 The risk value measurement model management platform 120 in the application scenario shown. Figure 2 As shown, the test method 200 may include steps 210, 220, 230, 240 and 250. These steps of the test method 200 will be described in detail below.

[0040] The test method 200 can construct a test data set, and use the constructed test data set to test the risk value measurement model, and the risk value measurement model can be used to evaluate the risk value of the resource combination to be analyzed. In the present disclosure, as a non-limiting example, the resource combination to be analyzed can be represented as a resource list R, where R = [(u1, p1, q1), ..., (u n ,p n ,q n )], n is an integer, and n≥1. Therefore, the resource list R represents that there are n kinds of resources in the resource combination to be analyzed, wherein u represents the resource code, which can be a digital code for identifying the resource in the trading market (as a non-limiting example, such as a stock code, a bond code, etc.), p represents the price of the corresponding resource at the corresponding time point (as a non-limiting example, such as the closing price of the stock at the corresponding time point), and q represents the number of holdings of the resource in the resource combination. It should be noted that when the various steps of the test method 200 and the corresponding specific embodiments are described in detail below, the resource list R will continue to be used for illustration and explanation.

[0041] Continue to see Figure 2 In step 210, historical price data of each resource in the resource combination to be analyzed in an evaluation time interval is obtained, wherein the evaluation time interval is from the evaluation start time point to the evaluation end time point.

[0042] For example, for a resource list R, the evaluation time interval T of each resource can be determined first, and then the historical price data of each resource included in the evaluation time interval T can be obtained, where the evaluation time interval T starts from the evaluation start time point t start Until the end of the evaluation time t end Generally speaking, the historical price data used to test the risk value measurement model needs to be purchased, so the range of the evaluation time interval T can be selected according to actual needs and the cost that can be borne. For example, you can choose to start from the evaluation end time point t end The first three years are used as the evaluation time interval T. Then, the historical price data of various resources in the evaluation time interval T can be obtained from the corresponding database by, for example, purchasing.

[0043] In an exemplary embodiment, step 210 may also include adding tags to each resource in the resource combination to be analyzed. The added tags may be used to identify the characteristics of the corresponding resource price fluctuations with the market. As an example, when collecting historical transaction data of resources, the tags may be determined by the data source and the key fields of the data. The specific tag adding logic may be set according to different scenarios, as a non-limiting example, such as:

[0044] Resource type: major categories may include commodities, equity, interest rates, foreign exchange, etc., and minor categories may include stocks, bonds, equity options, commodity futures, etc.;

[0045] Resource trading market: on-site and off-site markets, main board / small and medium-sized board / Sci-Tech Innovation Board, etc.;

[0046] Industries covered by the resources: information, medicine, automobiles, etc.

[0047] After adding tags, each resource in the resource combination to be analyzed will have a corresponding tag. Taking resource list R as an example, after adding tags, each resource in resource list R can become [(u1, p1, q1, taglist1),…, (u n ,p n ,q n ,taglist n )], where n is an integer and n≥1, and taglist is the corresponding tag.

[0048] By adding corresponding tags to each resource in the resource combination to be analyzed, the test method 200 can perform batch operations on resources with the same tags, thereby improving the efficiency of building a test data set and testing the risk value measurement model. In addition, by adding tags, the market situation can be segmented so that the constructed test data set can be more in line with the market fluctuations.

[0049] Continue to see Figure 2 In step 220, for each resource, an interval division time point between the evaluation start time point and the evaluation end time point is obtained, and the evaluation time interval is divided into a fitting interval and a benchmark interval, wherein the fitting interval is from the evaluation start time point to the interval division time point, and the benchmark interval is from the interval division time point to the evaluation end time point.

[0050] The obtained interval division time point t d At the starting time t start The evaluation end time t end Therefore, the evaluation time interval T can be divided into a fitting interval T1 and a reference interval T2. Therefore, the fitting interval T1 = [t start ,t d ], and the reference interval T2 = [t d ,t end ] Interval division time point t d You can choose according to actual needs so that the historical price data in the benchmark interval T2 can best reflect the price fluctuations in the trading market at that time. Set the interval division time point t for each resourced , which is conducive to personalized screening of various resources, so as to screen out the historical price data that best fits the market fluctuations. However, if there are a large number of resources, then when each resource in the resource combination to be analyzed has been tagged separately, the interval division time point t can also be set in batches for resources with the same tag (that is, uniformly set in the same batch). d , which can greatly improve the processing efficiency of the test method.

[0051] See also Figure 3 , which schematically shows a price fluctuation curve reflecting resource price fluctuations in a reference interval according to an exemplary embodiment of the present disclosure. Figure 3 As shown, it shows that the reference interval T2 is the evaluation end time point t end 6 days before the end of the assessment end , that is, the reference interval T2 = [t end -6,t end ], a total of 7 days (that is, t d =t end -6), the curve in the figure reflects the corresponding resource u in these 7 days i As a non-limiting example, the i-th resource u i It can be stocks. In this case, Figure 3 The price fluctuation shown may be the price trend of the stock price in the reference interval T2, thereby reflecting the price fluctuation of the resource in the trading market during the period.

[0052] Continue to see Figure 2 In step 230, for each resource, based on the fluctuation of the historical price data in the benchmark interval, historical price data with similar fluctuation conditions are screened out from the historical price data in the fitting interval as screened price data.

[0053] Specifically, for each resource, the historical price data in the fitting interval T1 can be analyzed, and the screening price data can be determined based on the historical price data that constitutes a curve similar to the price fluctuation curve in the reference interval T2. Then, these screening price data can be used to determine the fitting price data, which will be described in detail below. It should be understood that any suitable method for calculating curve similarity can be used to calculate the curve similarity relative to the price fluctuation curve in step 230. As a non-limiting example, vector similarity calculation, machine learning method based on feature extraction, neural network method based on learning training, etc. can be used to calculate curve similarity.

[0054] See also Figure 4, which shows according to an exemplary embodiment of the present disclosure, Figure 2 The test method shown uses vector similarity calculation to determine the process of filtering price data. Figure 4 As shown, step 230 in the testing method 200 may include steps 231 , 232 , 233 , 234 , 235 and 236 .

[0055] Next, we will continue to use the resource list R and the evaluation time interval T as an example. Figure 4 The process shown in FIG. 1 is described, where the base interval T2 of the evaluation time interval T has a period length of Tbase days, and each resource in the resource list R is not tagged. Therefore, for the i-th resource u in the resource list R i (1≤i≤n), the historical price data in the benchmark interval T2 of the evaluation time interval T is [h i_T2_1 ,h i_T2_2 ,…,h i_T2_Tbase ].

[0056] In step 231, based on the historical price data of the reference interval, a fluctuation vector of the historical price data of the reference interval is calculated.

[0057] For example, for a resource list R, its i-th resource u i In the benchmark period T2, the historical price data [h i_T2_1 ,h i_T2_2 ,…,h i_T2_Tbase ], therefore, the resource u can be calculated i The fluctuation vector of historical price data in the benchmark interval T2 is V base =[h i_T2_2 -h i_T2_1 ,h i_T2_3 -h i_T2_2 ,…,h i_T2_Tbase -h i_T2_(Tbase-1) ].

[0058] In step 232, the fitting interval is divided into a plurality of fitting sub-intervals in chronological order, the number of historical price data included in each fitting sub-interval is equal to the number of historical price data included in the base interval, and there is a time shift of a preset length between adjacent fitting sub-intervals in chronological order.

[0059] For example, for the i-th resource u i , the fitting interval T1 of the evaluation time interval T can be divided into N fitting sub-intervals S1 to S N The number of historical price data included in each fitting sub-interval is equal to the number of historical price data included in the reference interval T2, and the fitting sub-intervals S1 to SN There is a time shift t of preset length between adjacent fitting subintervals shift = 1 day. In this case, N is an integer and N = t d –Tbase+1. Therefore, resource u i The historical price data included in each fitting sub-interval are as follows: Each fitting sub-interval includes Tbase historical price data, and the historical price data included in the fitting sub-interval S1 is [h i_T1_1 ,h i_T1_2 ,…,h i_T1_Tbase ], the historical price data included in the fitting sub-interval S2 is [h i_T1_2 ,h i_T1_3 ,…,h i_T1_(Tbase+1) ],, fitting subinterval S N The historical price data included is [h i_T1_N ,h i_T1_(N+1) ,…,h i_T1_(Tbase+N-1) ].

[0060] It should be understood that the time shift t shift It can be set according to actual needs. In some other exemplary embodiments of the present disclosure, the time shift t shift It can be greater than 1 day, such as 2 days or 3 days, etc. shift In the case of more than 1 day, the value of N can be (t d –Tbase) divided by the time shift t shift The obtained quotient is then added with 1. In the above example, the time shift t shift Setting the time length to 1 day is conducive to fully screening the historical price data in the fitting interval T1, so that the historical price data in the fitting interval T1 that are similar to the historical price data in the benchmark interval T2 can be screened out as much as possible, so that the fluctuation of the final fitting price data can be as similar as possible to the fluctuation of the historical price data in the benchmark interval T2. On the other hand, if the time shift t shift If the time length is set to be greater than 1 day, it is helpful to reduce the amount of calculation, thereby reducing the time required to execute the test method. Therefore, when the computing power is limited, or when you want to reduce the overall time required to execute the test method and the fitting accuracy of the price data is not high, you can also choose to set the time shift t to 1 day according to the actual situation. shift The duration is set to be greater than 1 day.

[0061] In step 233, based on the historical price data of each fitting sub-interval, the historical price data fluctuation vector of each fitting sub-interval is calculated.

[0062] Specifically, for the i-th resource u iFor each fitting subinterval of historical price data, the same calculation method as in step 231 can be used to calculate the historical price data fluctuation vector of each fitting subinterval. Taking the fitting subinterval S1 as an example, its historical price data fluctuation vector can be calculated as V S1 =[h i_T1_2 –h i_T1_1 ,h i_T1_3 –h i_T1_2 ,…,h i_T1_Tbase –h i_T1_(Tbase-1) ], and similarly, for the fitting subinterval S N For example, the historical price data fluctuation vector can be calculated as V SN =[h i_T1_(N+1) –h i_T1_N ,h i_T1_(N+2) –h i_T1_(N+1) ,…,h i_T1_(Tbase+N-1) –h i_T1_(Tbase+N-2) ]. The calculation method of the historical price data fluctuation vector of other fitting sub-intervals can be deduced in the same way, so it will not be repeated here.

[0063] In step 234, the cosine similarity between the historical price data fluctuation vector of each fitting sub-interval and the historical price data fluctuation vector of the reference interval is calculated.

[0064] To fit the historical price data fluctuation vector V of subinterval S1 S1 For example, its historical price data fluctuation vector V relative to the benchmark interval T2 base The cosine similarity can be calculated as follows: cos(θ1)=(V base ·V S1 ) / (V base ‖V S1 ). The cosine similarity between the historical price data fluctuation vectors of other fitting sub-intervals and the historical price data fluctuation vectors of the reference interval can be calculated in a similar manner, which will not be repeated here. i , for the fitting subinterval S1 to S N , the corresponding cosine similarities cos(θ1) to cos(θ N ).

[0065] In step 235, the cosine similarity corresponding to each fitting subinterval is compared with a preset similarity threshold, and the fitting subinterval whose cosine similarity is greater than the similarity threshold is determined.

[0066] As mentioned above, for the i-th resource u i , for its fitting subinterval S1 to S N, the corresponding cosine similarities cos(θ1) to cos(θ N ). The similarity threshold can be preset according to actual needs. As a non-limiting example, the similarity threshold can be set to 0.8. N ) are compared with the similarity threshold respectively, so as to fit the subintervals S1 to S N All fitting subintervals whose cosine similarity is greater than the similarity threshold are screened out.

[0067] In step 236, repeated historical price data are removed from the historical price data included in all fitting subintervals whose determined cosine similarity is greater than the similarity threshold, thereby obtaining the filtered price data.

[0068] From the above, the fitting sub-interval S1 to S N From the description of the division method, it can be seen that the historical price data of adjacent fitting subintervals have repeated parts. Therefore, after determining all fitting subintervals whose cosine similarity is greater than the similarity threshold, it is necessary to remove the repeated historical price data from the historical price data of these fitting subintervals. As an example, in the fitting interval T1, along the starting time point t start To the interval division time point t d The unique historical price data corresponding to the corresponding time point is determined from the historical price data of the selected fitting sub-intervals in the order of . After removing the duplicate historical price data, the obtained historical price data is the selected price data, which is the historical price data reflecting the historical price data with similar fluctuations in the fitting interval T1 as the historical price data of the benchmark interval T2.

[0069] See also Figure 5 , which schematically shows a curve formed by screening price data in a fitting interval according to an exemplary embodiment of the present disclosure. Figure 5 As shown, the i-th resource u in the resource list R i The screening price data in the fitting interval T1 has a fluctuation similar to the historical price data in the reference interval T2. In addition, because the fitting subintervals that are not similar to the fluctuation of the historical price data in the reference interval T2 are removed during the screening process, the screening price data determined in the fitting interval T1 are discontinuous.

[0070] Continue to see Figure 2 , in step 240, for each resource, the discontinuous part of the screened price data is interpolated to obtain continuous fitting price data.

[0071] As mentioned above and Figure 5As shown, the screening price data determined in the fitting interval T1 is discontinuous, so it is necessary to interpolate the discontinuous parts of the screening price data. One available interpolation method is to interpolate the discontinuous parts of the screening price data using the following equation to obtain the fitting price data:

[0072]

[0073] Among them, S t Represents the price at time point t; t represents the time point; Δt represents the time interval, which can be one-Nth of the duration between adjacent time points (N is an integer greater than 1). For example, the duration between adjacent time points can be one day, then Δt = 1 / N days, which means that one day can be divided into N random changes, and the result of accumulating N fluctuations is the price of the next day; u represents the expected value of the resource's yield; σ represents the volatility of the resource's yield; ∈ represents random disturbance, which is used to describe the randomness or uncertainty in the above resource price change model, and can be calculated using normal distribution or other specified functions.

[0074] Based on the above equation and the data at the breakpoints of the discontinuous part in the filtered price data, the filtered price data can be interpolated. After the interpolation process, continuous fitting price data can be obtained for each resource.

[0075] It should be understood that any suitable interpolation processing method can be used in the test method of the present disclosure to interpolate the discontinuous part of the screening price data. For example, it is also possible to perform slope interpolation based on the data at the breakpoints of the discontinuous part of the screening price data (i.e., interpolation is performed in an arithmetic progression based on the data at the breakpoints), or the data at the breakpoints can be directly used as the value of the discontinuous part. The present disclosure does not limit the specific type of interpolation method used.

[0076] Continue to see Figure 2 In step 250, the risk value measurement model is tested using a test data set consisting of the fitting price data of each resource in the fitting interval and the historical price data in the benchmark interval.

[0077] It can be seen that for each resource, the fitting price data in the fitting interval and the historical price data in the benchmark interval together constitute a test data set that can be used to test the risk value measurement model. Therefore, the newly constructed test data set can be used to test the risk value measurement model to verify its rationality.

[0078] See also Figure 6 , which shows according to an exemplary embodiment of the present disclosure Figure 2The process of testing the risk value measurement model using the obtained test data set in step 250 of the test method 200 shown in FIG. Figure 6 As shown, step 250 of the test method 200 may further include steps 251, 252, 253 and 254:

[0079] In step 251, the profit and loss sequence of the resource combination to be analyzed is generated using the test data set;

[0080] In the following, we will continue to use the resource list R and the evaluation time interval T as an example to Figure 6 The process shown in the figure is described, where the evaluation time interval T starts from the evaluation start time point t start Until the end of the evaluation time t end . Assume that for resource u in resource list R i , the new test data set constructed based on historical price data is H i =[h i_start ,.,h i_end ], where h i_start It is a resource i At the evaluation start time t start Price data at h i_end It is a resource i At the end of the evaluation time t end In step 251, the total resource value of the resource list R at each time point in the evaluation time interval T is calculated according to the newly constructed test data set. The sequence of the total resource value is shown in the following formula:

[0081]

[0082] Among them, q i is the i-th resource u in the resource list R i The number of holdings, therefore, Represents the resource list R at the evaluation start time t start The total value of resources at Represents the resource list R at the end of the evaluation time t end The total value of resources at.

[0083] Then, the profit and loss sequence of the resource list R can be obtained by subtracting the total asset value at the current time point from the total asset value at each time point in the evaluation time interval, as shown in the following formula:

[0084]

[0085] Among them, p i is the i-th resource u in the resource list R i The price at the current point in time, therefore, It represents the total resource value of the resource list R as the resource combination to be analyzed at the current time point. Represents the resource list R at the evaluation start time t start Profit and loss at It represents the resource list R at the end time t of the evaluation end Profit and loss at the place.

[0086] In step 252, based on the profit and loss sequence of the resource combination to be analyzed, a corresponding quantile is determined, and the value of the quantile is used as a test benchmark value;

[0087] For example, after obtaining the above profit and loss sequence of the resource list R, the items in the profit and loss sequence can be sorted from large to small, and the values ​​corresponding to the 99%, 95%, and 90% quantiles are taken out in turn as the test benchmark values. Assume that the evaluation time interval T starts from the evaluation start time point t start Until the end of the evaluation time t end With 1000 time points, the 99% quantile is the value corresponding to the 990th element in the sorted list, the 95% quantile is the value corresponding to the 950th element in the sorted list, and the 90% quantile is the value corresponding to the 900th element in the sorted list.

[0088] In step 253, the test data set is input into the risk value measurement model to obtain the profit and loss sequence output by the model;

[0089] In step 254, the risk value measurement model is evaluated based on the profit and loss sequence and the quantiles output by the model.

[0090] Specifically, the rationality of the risk value measurement model can be determined by comparing the numerical distribution of the profit and loss sequence output by the model with the benchmark values ​​of different quantiles. Therefore, the following Table 1 lists the schematic evaluation of the risk value measurement models A, B, and C, where the holding period is 1 day (the holding period refers to the time period during which the resource owner holds the asset portfolio. It is used to measure the duration of the resource owner's holding of the asset, thereby evaluating the risk exposure and the validity period of the investment strategy; for example: a holding period of 1 day means the loss of the current assets held by the resource owner in the next 1 day, and a holding period of 3 days means the loss of the current assets held by the resource owner in the next 3 days):

[0091]

[0092] Table 1 Evaluation results of risk value measurement models A, B, and C

[0093] It can be seen that according to the test method for testing the risk value measurement model provided by the present disclosure, the historical price data in the fitting area is screened based on the historical price data of the benchmark interval that reflects the current market price fluctuations, and the historical price data similar to the current market price fluctuations are retained, and the historical price data that may have a large deviation between the historical situation and the current market situation are removed, and the interpolation method is used to supplement the breakpoints in the historical price data. Therefore, compared with the existing test methods, the test data constructed according to the test method provided by the present disclosure can better reflect the current market price fluctuations, thereby making the evaluation of the risk value measurement model more accurate. In addition, since the historical price data that may have a large deviation between the historical situation and the current market situation are removed from the newly constructed test data set, the final test data can better reflect the current market price fluctuations. Therefore, according to the test method provided by the present disclosure, much less historical price data can be used to construct the test data set for testing, which helps to save the cost of purchasing historical price data on the one hand, and on the other hand, it can also reduce the time consumed by data access, and can reduce the demand for storage devices for storing historical price data, reduce the demand for data transmission, and even reduce the demand for computing power.

[0094] See also Figure 7 , which schematically illustrates a hardware environment related to the testing of the risk value measurement model according to an exemplary embodiment of the present disclosure. Figure 1 As described above, in the embodiment of the present disclosure, the testing method for the risk value measurement model can be executed on the risk value measurement model management platform 120. Figure 7 In the embodiment, the risk value measurement model management platform 120 is further shown as a device 300, which may include multiple components. Figure 7The risk value measurement model management platform 120 is only illustrated as a device 300, but in fact the risk value measurement model management platform 120 can also be a plurality of devices, each device 300 includes a processor 304, including a hardware element 310. The processor 304 includes, for example, one or more digital signal processors (DSPs), general-purpose microprocessors, application-specific integrated circuits (ASICs), field programmable logic arrays (FPGAs), or other equivalent integrated or discrete logic circuits. As used herein, the term "processor" may refer to any of the above structures or any other structures suitable for implementing the technology described herein. In addition, in some aspects, the functions described herein may be provided in a dedicated hardware and / or software module configured for testing the risk value measurement model, or incorporated in a combined hardware and / or software module. In addition, the functions described herein may also be fully implemented in one or more circuits or logic elements. The method described in the present disclosure may be implemented in various components, modules, or units, but does not necessarily have to be implemented by different hardware units. Rather, as described above, the various components, modules or units may be combined or provided by a collection of interoperating hardware units (including one or more processors as described above) in conjunction with appropriate software and / or firmware.

[0095] In one or more embodiments, the above combination Figures 2 to 6 The content described can be implemented based on hardware, software, firmware or any combination thereof. If implemented in software, its functions can be stored on or transmitted via a computer-readable medium 306 as one or more instructions or codes, and executed by a hardware-based processor. The computer-readable medium 306 may include a computer-readable storage medium corresponding to a tangible medium such as a data storage medium, or a communication medium containing any medium that facilitates the transmission of a computer program (including the one or more instructions or codes) from one place to another according to a communication protocol. In this way, the computer-readable medium 306 can generally correspond to (1) a non-temporary tangible computer-readable storage medium, or (2) a communication medium such as a signal or carrier wave. The data storage medium can be any available medium that can be read by one or more computers or one or more processors for the instructions, codes and / or data structures for implementing the technology described in the present disclosure. In addition, the computer-readable medium 306 and the one or more instructions or codes stored thereon can also be implemented as a computer program product.

[0096] As non-limiting examples, such computer-readable storage media may include RAM, ROM, EEPROM, CD-ROM or other optical disks, magnetic disk storage or other magnetic storage, flash memory, or any other memory 312 that can be used to store the desired program code in the form of instructions or data structures and can be read by the computer. In addition, any suitable connection may also be referred to as a computer-readable medium 306. For example, if a coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwaves are used to transmit instructions from a website, a settlement system, or other remote source, then the coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwaves are also included in the definition of the medium. However, it should be understood that computer-readable storage media and data storage media do not include connections, carrier waves, signals, or other transient media, but are directed to non-transient tangible storage media. As used herein, disks and optical disks include compact disks (CDs), laser disks, optical disks, digital versatile disks (DVDs), floppy disks, and Blu-ray disks, where disks usually reproduce data magnetically, while optical disks use lasers to reproduce data optically. Combinations of the above should also be included within the scope of computer-readable media 306 .

[0097] The device 300 may also include an I / O interface for transmitting data, and other functions 314. The device 300 may also be included in different terminals, such as a computer 316, a server 318, and other devices 320, etc. Each of these configurations includes devices that may have generally different structures and capabilities, and thus may be configured according to one or more of the different device categories for testing the risk value measurement model. In addition, the technical solution according to the present disclosure may also be implemented in whole or in part on the "cloud" 322 by using a distributed system, such as a platform 324 as described below.

[0098] Cloud 322 includes and / or represents a platform 324 for resources 326. Platform 324 abstracts the underlying functionality of the hardware (e.g., settlement system) and software resources of cloud 322. Resources 326 may include applications and / or data that may be used when executing computer processing on a data system remote from computing device 300. Resources 326 may also include services provided over the Internet and / or over a subscriber network such as a cellular or Wi-Fi network.

[0099] The platform 324 can abstract resources and functions to connect the computing device with other computing devices. The platform 324 can also be used to abstract the hierarchy of resources to provide a hierarchy of corresponding levels of demand encountered for resources 326 implemented via the platform 324. Therefore, in the interconnected device embodiment, the implementation of the functions described herein can be distributed throughout the device 300 or partially implemented on the computing device and in conjunction with the platform 324 that abstracts the functions of the cloud 322.

[0100] The terms used in the present disclosure are only used to describe the embodiments in the present disclosure and are not intended to limit the present disclosure. As used in the present disclosure, the singular forms "one", "one" and "the" are intended to also include plural forms, unless the context clearly indicates otherwise. It is also to be understood that the terms "include" and "comprise" when used in the present disclosure refer to the presence of the features described, but do not exclude the presence of one or more other features or add one or more other features. As used in the present disclosure, the term "and / or" includes any and all combinations of one or more of the associated listed items. It will be understood that although the terms "first", "second", "third", etc. can be used to describe various features in the present disclosure, these features should not be limited by these terms. These terms are only used to distinguish one feature from another feature.

[0101] Unless otherwise defined, all terms (including technical and scientific terms) used in this disclosure have the same meaning as commonly understood by those of ordinary skill in the art to which this disclosure belongs. It is also understood that terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the relevant art and / or the context of this specification, and will not be interpreted in an idealized or overly formal sense unless explicitly defined in this disclosure.

[0102] In the description of this specification, the description of the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present disclosure. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradicting each other and without violating technical principles, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, or may omit some technical features from the different embodiments or examples described in this specification, and the embodiments or examples obtained based on such combination, combination or omission are also considered to fall within the scope of the present disclosure.

[0103] The method described in the present disclosure includes one or more steps or actions. These method steps and / or actions do not have to be performed in the order described in the present disclosure, but can be performed in a different order, for example, they can be performed simultaneously or in a reverse order, as long as they do not conflict with the principle of the technical solution described in the present disclosure. In addition, according to actual needs, the steps or actions in the method described in the present disclosure can be replaced with different steps or actions, or can also include additional steps or actions.

[0104] Although the present disclosure has been described in detail in conjunction with some exemplary embodiments, it is not to be limited to the specific forms described in the present disclosure. Rather, the scope of the present disclosure is limited only by the appended claims.

Claims

1. A testing method for testing a risk value measurement model, characterized in that: The test method includes: Obtaining historical price data of each resource in the resource combination to be analyzed in an evaluation time interval, wherein the evaluation time interval is from the evaluation start time point to the evaluation end time point; For each resource, an interval division time point between the evaluation start time point and the evaluation end time point is obtained, and the evaluation time interval is divided into a fitting interval and a reference interval, wherein the fitting interval is from the evaluation start time point to the interval division time point, and the reference interval is from the interval division time point to the evaluation end time point; For each resource, based on the fluctuation of the historical price data in the benchmark interval, historical price data with similar fluctuations are screened out from the historical price data in the fitting interval as screening price data; For each resource, interpolating the discontinuous parts of the screened price data to obtain continuous fitting price data; The risk value measurement model is tested using a test data set consisting of the fitted price data of each resource in the fitting interval and the historical price data in the benchmark interval.

2. The testing method according to claim 1, characterized in that: The obtaining of the historical price data of each resource in the resource combination to be analyzed in the evaluation time interval further includes: adding a label to each resource in the resource combination to be analyzed.

3. The testing method according to claim 2, characterized in that: The step of obtaining, for each resource, an interval division time point between the evaluation start time point and the evaluation end time point further comprises: obtaining, for all resources with the same label in the resource combination to be analyzed, the interval division time point in the same batch.

4. The testing method according to claim 1, characterized in that: The step of selecting, for each resource, historical price data with similar fluctuation conditions from the historical price data in the fitting interval as the filtered price data based on the fluctuation conditions of the historical price data in the benchmark interval further includes: Based on the historical price data of the reference interval, calculating the historical price data fluctuation vector of the reference interval; Dividing the fitting interval into a plurality of fitting subintervals in chronological order, wherein the number of historical price data included in each fitting subinterval is equal to the number of historical price data included in the base interval, and there is a time shift of a preset time length between adjacent fitting subintervals in chronological order; Based on the historical price data of each fitting sub-interval, calculate the historical price data fluctuation vector of each fitting sub-interval; Calculate the cosine similarity between the historical price data fluctuation vector of each fitting sub-interval and the historical price data fluctuation vector of the benchmark interval; Compare the cosine similarity corresponding to each fitting subinterval with a preset similarity threshold, and determine the fitting subinterval whose cosine similarity is greater than the similarity threshold; The repeated historical price data are removed from the historical price data included in all fitting subintervals whose determined cosine similarity is greater than the similarity threshold, thereby obtaining the filtered price data.

5. The testing method according to claim 4, characterized in that: The preset time shift is 1 day.

6. The testing method according to claim 4, characterized in that: The similarity threshold is 0.

8.

7. The testing method according to claim 1, characterized in that: For each resource, interpolating the discontinuous part of the filtered price data to obtain continuous fitting price data also includes: based on the equation Interpolate the discontinuous parts of the screened price data to obtain the fitted price data; wherein S t Represents the price at time point t; t represents the time point; Δt represents the time interval; u represents the expected value of the resource's rate of return; σ represents the volatility of the resource's rate of return; ∈ represents random disturbance.

8. The testing method according to claim 1, characterized in that: For each resource, interpolating the discontinuous part in the filtered price data to obtain continuous fitting price data also includes: performing slope interpolation on the discontinuous part based on the numerical value at the breakpoint of the discontinuous part in the filtered price data to obtain the fitting price data.

9. The testing method according to claim 1, characterized in that: For each resource, interpolating the discontinuous part in the filtered price data to obtain continuous fitting price data also includes: using the numerical value at the breakpoint of the discontinuous part in the filtered price data as the data in the discontinuous part to obtain the fitting price data.

10. The testing method according to claim 1, characterized in that: The testing of the risk value measurement model using the fitted price data of each resource in the fitting interval and the historical price data in the benchmark interval further includes: Using the test data set, generating a profit and loss sequence of the resource combination to be analyzed; Based on the profit and loss sequence of the resource combination to be analyzed, a corresponding quantile is determined, and the value of the quantile is used as a test benchmark value; Inputting the test data set into the risk value measurement model to obtain the profit and loss sequence output by the model; The risk value measurement model is evaluated based on the profit and loss sequence and the quantiles output by the model.

11. A computing device comprising a processor and a memory, wherein a computer program is stored in the memory, wherein: The processor executes the computer program to implement the steps of the testing method according to any one of claims 1 to 10.

12. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: The computer program / instructions implement the steps of the testing method according to any one of claims 1 to 10 when executed by a processor.

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