Risk assessment method and device, nonvolatile storage medium and computer equipment

Abstract

The invention discloses a risk assessment method and device, a nonvolatile storage medium and computer equipment. The method comprises the steps that initial parameters are obtained, the initial parameters comprise the current price of virtual resources, the historical price of the virtual resources, the fluctuation rate of total resources and the transaction volume of the total resources, and the total resources comprise the virtual resources; the initial parameters are input into a target risk assessment model to obtain risk parameters, a network structure of the target risk assessment model comprises a Fourier operator layer and a multi-layer perceptron layer, and the risk parameters represent the sensitive degree of the price of the virtual resources to the change of the total resources; and based on the risk parameter, determining a risk assessment result of the virtual resource, the size of the risk assessment result representing the risk of price reduction of the virtual resource caused by the change of the total resource. The technical problems that a traditional risk management method is not rapid enough in response and not accurate enough in evaluation are solved.

Description

Technical Field

[0001] The present invention relates to the field of big data and AI, and in particular to a risk assessment method, device, non-volatile storage medium and computer equipment. Background Art

[0002] In recent years, the complexity and dynamism of financial markets have increased significantly, requiring financial institutions and investors to assess and manage risk exposure more accurately and timely. Traditional financial models, such as the Black-Scholes model, can provide effective pricing and risk assessment in certain environments, but they have limitations in processing high-dimensional data and capturing complex market dynamics. The Black-Scholes model assumes that volatility is constant and ignores many market realities, such as volatility smile and stochastic volatility, which often leads to divergence between model predictions and actual market performance in practical applications.

[0003] With the rapid development of big data and artificial intelligence technologies, the financial industry has begun to explore the application of deep learning in risk management. However, the application of deep learning technology to the assessment of financial risk exposure faces many challenges. First, the high dimensionality and nonlinear relationship of financial market data require the model to have strong data processing capabilities. Second, the rapid changes in market conditions require the model to be able to adjust and update in real time to reflect the latest market dynamics. In addition, risk management is not just about predicting asset prices, but also about quantifying the sensitivity of assets to different risk factors, that is, risk exposure. Therefore, it is an urgent need for the current financial industry to develop a risk management method that can efficiently process high-dimensional data, capture complex market dynamics, and assess risk exposure in real time.

[0004] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0005] The embodiments of the present invention provide a risk assessment method, apparatus, non-volatile storage medium and computer equipment to at least solve the technical problems that traditional risk management methods do not respond quickly enough and do not assess accurately enough.

[0006] According to one aspect of an embodiment of the present invention, there is provided a risk assessment method, comprising: obtaining initial parameters, wherein the initial parameters include the current price of virtual resources, the historical price of virtual resources, the volatility of total resources and the transaction volume of total resources, and the total resources include virtual resources; inputting the initial parameters into a target risk assessment model to obtain risk parameters, wherein the network structure of the target risk assessment model includes a Fourier operator layer and a multi-layer perceptron layer, and the risk parameters represent the sensitivity of the price of virtual resources to changes in total resources; based on the risk parameters, determining the risk assessment result of the virtual resources, wherein the size of the risk assessment result represents the size of the risk of a price drop of the virtual resources due to changes in total resources.

[0007] Optionally, before inputting the initial parameters into the target risk assessment model, the process includes: constructing an initial risk assessment model, wherein the network structure of the initial risk assessment model includes a Fourier operator layer and a multi-layer perceptron layer; training the initial risk assessment model based on training data, and calculating the value of a loss function in the initial risk assessment model, wherein the training data includes historical prices of virtual resources, historical volatility of total resources, and historical transaction volume of total resources; when the value of the loss function is less than a preset threshold, determining that the model corresponding to the current training process is the target risk assessment model, wherein the loss function is a weighted sum of a first error and a second error based on preset weights, the first error being the output error of the Fourier operator layer, and the second error being the output error of the multi-layer perceptron layer.

[0008] Optionally, obtaining initial parameters includes: constructing an initial price model and a volatility model, wherein the initial price model is a Black-Scholes model, and the volatility model characterizes the relationship between the volatility of total resources and time; combining the initial price model with the volatility model to obtain a target price model; and determining the initial parameters based on the target price model.

[0009] Optionally, the initial parameters are input into the target risk assessment model to obtain the risk parameters, including: inputting the initial parameters into the Fourier operator layer to obtain the initial output results, wherein the initial output results include the option price of the virtual resources; inputting the initial output results into the multi-layer perceptron layer to obtain the risk parameters.

[0010] Optionally, a risk assessment result is determined based on risk parameters, wherein the risk parameters include multiple types of risk parameters, including: based on changes in total resources, setting corresponding weights for each of the multiple types of risk parameters; based on the weights, performing weighted summation of the multiple types of risk parameters to obtain a risk assessment result.

[0011] Optionally, after determining the risk assessment result of the virtual resource, the method further includes: comparing the risk assessment result with a preset level threshold to obtain a risk level corresponding to the risk assessment result; and determining a processing method for the virtual resource based on the risk level.

[0012] According to another aspect of an embodiment of the present invention, a risk assessment device is also provided, including: an acquisition module, used to acquire initial parameters, wherein the initial parameters include the current price of virtual resources, the historical price of virtual resources, the volatility of total resources and the trading volume of total resources, and the total resources include virtual resources; an assessment module, used to input the initial parameters into a target risk assessment model to obtain risk parameters, wherein the network structure of the target risk assessment model includes a Fourier operator layer and a multi-layer perceptron layer, and the risk parameters represent the sensitivity of the price of virtual resources to changes in total resources; a determination module, used to determine the risk assessment result of the virtual resources based on the risk parameters, wherein the size of the risk assessment result represents the risk size of the price drop of the virtual resources caused by the change of total resources.

[0013] According to another aspect of an embodiment of the present invention, a non-volatile storage medium is provided, the non-volatile storage medium including a stored program, wherein when the program is running, the device where the non-volatile storage medium is located is controlled to execute any one of the above risk assessment methods.

[0014] According to yet another aspect of an embodiment of the present invention, a computer device is provided. The computer device includes a processor, and the processor is used to run a program. When the program is run, any one of the above-mentioned risk assessment methods is executed.

[0015] According to yet another aspect of an embodiment of the present invention, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, any one of the above-mentioned risk assessment methods is implemented.

[0016] In an embodiment of the present invention, a risk assessment method is adopted, by obtaining initial parameters, wherein the initial parameters include the current price of virtual resources, the historical price of virtual resources, the volatility of total resources and the transaction volume of total resources, and the total resources include virtual resources; the initial parameters are input into a target risk assessment model to obtain risk parameters, wherein the network structure of the target risk assessment model includes a Fourier operator layer and a multi-layer perceptron layer, and the risk parameters characterize the sensitivity of the price of virtual resources to changes in total resources; based on the risk parameters, a risk assessment result of the virtual resources is determined, wherein the size of the risk assessment result characterizes the size of the risk of a price drop of virtual resources caused by changes in total resources, thereby achieving the purpose of accurately assessing the sensitivity of the price of virtual resources to changes in total resources, thereby effectively predicting the risk of a price drop of virtual resources, thereby achieving the technical effect of improving the accuracy and timeliness of risk assessment, and further solving the technical problems of traditional risk management methods that are not quick enough to respond and are not accurate enough in assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0018] Figure 1 A hardware structure block diagram of a computer terminal for implementing a risk assessment method is shown;

[0019] Figure 2 is a schematic diagram of a risk assessment method according to an embodiment of the present invention;

[0020] Figure 3 is a schematic diagram of a network structure of a Fourier operator provided according to an optional embodiment of the present invention;

[0021] Figure 4 is a schematic diagram of a neural network structure based on a Fourier operator according to an optional embodiment of the present invention;

[0022] Figure 5 is a schematic diagram of a neural network structure combining a Fourier operator and an MLP according to an optional embodiment of the present invention;

[0023] Figure 6 is a flow chart of a risk assessment method based on Fourier operator provided according to an optional embodiment of the present invention;

[0024] Figure 7 is a structural block diagram of a risk assessment device provided according to an embodiment of the present invention. DETAILED DESCRIPTION

[0025] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0026] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0027] According to an embodiment of the present invention, a risk assessment method embodiment is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0028] The method embodiment provided in the first embodiment of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Figure 1 FIG. 1 shows a hardware structure block diagram of a computer terminal for implementing a risk assessment method. Figure 1 As shown, the computer terminal 10 may include one or more (102a, 102b, ..., 102n are used to illustrate) processors (the processor may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 104 for storing data. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the BUS bus), a network interface, a power supply and / or a camera. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 1 More or fewer components as shown, or with Figure 1Different configurations are shown.

[0029] It should be noted that the one or more processors and / or other data processing circuits described above may generally be referred to herein as "data processing circuits". The data processing circuits may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuit may be a single independent processing module, or may be incorporated in whole or in part into any of the other components in the computer terminal 10. As described in the embodiments of the present application, the data processing circuit acts as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0030] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the risk assessment method in the embodiment of the present invention. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, the risk assessment method of the above-mentioned application is realized. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the computer terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0031] The display may be, for example, a touch screen liquid crystal display (LCD) that enables a user to interact with a user interface of the computer terminal 10 .

[0032] Figure 2 is a flow chart of a risk assessment method according to an embodiment of the present invention. Figure 2 As shown, the method comprises the following steps:

[0033] Step S201, obtaining initial parameters, wherein the initial parameters include the current price of virtual resources, the historical price of virtual resources, the volatility of total resources and the transaction volume of total resources, and the total resources include virtual resources.

[0034] In this step, the virtual resource can be an asset, and the total resource can be a market including the asset. For example, the current price of the virtual resource can be obtained by using real-time data streams or regularly updated data points to reflect the latest market value of the asset. The current price is usually the latest transaction price or settlement price at a specific time point. For non-real-time trading assets, such as certain bonds or funds, the most recent closing price may be used as an approximation of the current price; the historical price data of virtual resources provides the price change trajectory of the asset over a period of time, which is crucial for analyzing the volatility, trend and cyclicality of the asset; the volatility of the total resource can be the historical volatility calculated based on the historical price, or the market implied volatility (such as the VIX index), the latter of which more directly reflects the market's expectations of future volatility; the trading volume data reflects the market liquidity of the asset, that is, the activeness of the asset in the market. High trading volume usually means that the asset has good liquidity, while low trading volume may indicate a high liquidity risk.

[0035] Step S202, inputting the initial parameters into the target risk assessment model to obtain risk parameters, wherein the network structure of the target risk assessment model includes a Fourier operator layer and a multi-layer perceptron layer, and the risk parameters represent the sensitivity of the price of virtual resources to changes in total resources.

[0036] In this step, a deep neural network can be designed that contains an integral kernel parameterized in Fourier space to implement the Fourier neural operator. In this framework, the FNO (Fourier Neural Operator) does not focus on predicting prices alone, but works closely with the MLP (Multi-Layer Perceptron) layer to jointly process risk parameters, that is, risk exposure assessment problems. This structure ensures that the two are functionally inseparable and together constitute the core of the entire prediction model. The model input is the initial parameters, which can include historical data such as current price, strike price, risk-free rate, volatility, expiration time, and other factors that may affect the price. The Fourier neural operator part focuses on predicting prices from these inputs, while the MLP layer focuses on refining and optimizing key risk assessment parameters from the output of the Fourier neural operator. After the model is trained, an independent data set is used to verify the entire model (including the Fourier neural operator and the MLP layer) to ensure that it can provide appropriate risk parameters under different market conditions. The performance of the Fourier neural operator can also be tested in an actual trading environment to ensure its stability and accuracy in actual applications.

[0037] After the initial parameters such as current price, historical price, volatility and trading volume are input into the target risk assessment model, after deep processing by the Fourier operator layer and the multi-layer perceptron layer, the model will output a series of risk parameters. These parameters characterize the sensitivity of the price of virtual resources to changes in total resources, that is, the responsiveness of assets to risk factors such as market fluctuations, interest rate changes, and changes in macroeconomic indicators.

[0038] For example, the market sensitivity parameter (β), β represents the sensitivity of asset prices to key market factors (such as market index, interest rate), which is used to measure the impact of market changes on asset prices. Regression analysis can be performed through historical data to determine the relationship between asset prices and market factors, β = Cov(R_asset, R_market) / Var(R_market), where R_asset is the asset return rate and R_market is the market return rate; the asset correlation parameter (ρ), ρ measures the price correlation between different assets or asset portfolios, and is used for combined risk management and asset allocation strategies. Historical price data can be used to calculate the correlation coefficient between different assets, ρ = Cov(R_asset1, R_asset2) / (σ_asset1*σ_asset2), where σ is the standard deviation of the asset; the liquidity risk parameter (λ), λ reflects the liquidity of the asset, and factors such as the bid-ask spread and trading volume can be considered. Low liquidity assets may face greater price volatility and transaction difficulties in rapidly changing markets, λ = 1 / (average trading volume / average bid-ask spread), assets with large trading volumes and small bid-ask spreads have higher liquidity; credit risk parameter (δ), δ assesses the credit risk of an asset, such as the probability of default or changes in credit ratings, which is particularly important for bonds and other credit-sensitive assets. It can be estimated through historical default data and changes in market credit ratings. δ can also be the probability of default, which is estimated through credit ratings and market data; extreme event risk parameter (ξ), ξ is used to measure the performance of an asset under extreme market events (such as financial crises), reflecting the asset's performance during market collapses or extreme fluctuations. The risk exposure can be analyzed based on historical extreme event data. ξ can be the average loss ratio or maximum drawdown of an asset in past extreme events. The strategy adjustment parameter (γ) is defined as a market adaptability indicator, which quantifies the ability of an asset or portfolio to adjust its configuration under different market conditions. This parameter can help identify assets or strategies that require more attention during market fluctuations. By analyzing historical market data, γ can be calculated as the correlation between market conditions (such as volatility, interest rate levels) and asset performance (such as return rate, volatility). Time series analysis can be used to estimate the sensitivity of asset performance under different market conditions. γ = Corr (market conditions, asset performance).

[0039] Step S203, based on the risk parameters, determining the risk assessment result of the virtual resources, wherein the magnitude of the risk assessment result represents the magnitude of the risk of a price drop of the virtual resources due to the change of the total resources.

[0040] In this step, once the risk parameters are determined, the comprehensive risk assessment results of the virtual resources can be calculated. The size of the risk assessment results directly characterizes the risk of a decline in the price of virtual resources due to changes in total resources (including market fluctuations, interest rate changes, adjustments to macroeconomic indicators, etc.). Based on the calculated risk assessment results, investors and financial institutions can conduct more accurate risk management. For example, by monitoring numerical changes, investment strategies can be adjusted in a timely manner to increase or decrease exposure to high-risk assets. In addition, regularly updating risk parameters and recalculating risk assessment results helps maintain a continuously optimized risk management framework based on market dynamics and changes in the portfolio.

[0041] Through the above steps, the purpose of accurately assessing the sensitivity of virtual resource prices to changes in total resources is achieved, thereby effectively predicting the risk of a decline in virtual resource prices, thereby achieving the technical effect of improving the accuracy and timeliness of risk assessment, and further solving the technical problems of traditional risk management methods that are not responsive enough and are not accurate enough in assessment.

[0042] As an optional embodiment, before inputting the initial parameters into the target risk assessment model, it includes: constructing an initial risk assessment model, wherein the network structure of the initial risk assessment model includes a Fourier operator layer and a multi-layer perceptron layer; based on training data, training the initial risk assessment model, and calculating the value of the loss function in the initial risk assessment model, wherein the training data includes the historical price of virtual resources, the historical volatility of total resources, and the historical transaction volume of total resources; when the value of the loss function is less than a preset threshold, determining that the model corresponding to the current training process is the target risk assessment model, wherein the loss function is a weighted sum of a first error and a second error based on preset weights, the first error is the output error of the Fourier operator layer, and the second error is the output error of the multi-layer perceptron layer.

[0043] Optionally, the architecture of the model includes a Fourier operator layer and a multi-layer perceptron layer. The Fourier operator layer is used to capture and process the frequency domain features of the input data. By operating the data in the frequency domain, it can efficiently identify the complex dynamics of the financial market, especially for high-dimensional data and scenarios involving time series analysis. The multi-layer perceptron layer is responsible for nonlinear transformation and feature learning of the frequency domain features output by the Fourier operator layer, converting the frequency domain information into a high-dimensional feature representation that the model can understand, and then predicting the risk assessment parameters.

[0044] Specifically, Figure 3is a schematic diagram of a network structure of a Fourier operator provided according to an optional embodiment of the present invention, Figure 4 is a schematic diagram of a neural network structure based on a Fourier operator according to an optional embodiment of the present invention. As shown in the figure, the input initial parameters are first promoted to a higher-dimensional representation v0(x)=P(a(x)) through a local transformation P, where P is parameterized by a shallow fully connected neural network. Applying the Fourier transform F to v0(x) yields In Fourier space, a linear transformation R is applied to the low-frequency modes, namely Apply the inverse Fourier transform F to the transformed signal -1 , and we get v'(k). Among them, the Fourier transform and inverse transform can be defined as:

[0045]

[0046] In the integrator layer, v'(x) is processed through the integrator layer, each layer contains a global integrator and a local nonlinear activation function. For the i-th layer, the update of v'(k) can be expressed as:

[0047]

[0048] Among them, σ is a nonlinear activation function (such as ReLU), W i is a local linear transformation, K i is a parameterized kernel function that maps the input to the feature space to achieve nonlinear transformation, and Ω is the domain. The output v(x) of the last layer is projected back to the target dimension through the neural network Q to obtain the output u, that is, u = Q(v(x)).

[0049] Figure 5 is a schematic diagram of a neural network structure combining a Fourier operator and an MLP according to an optional embodiment of the present invention, such as Figure 5 As shown, the input layer of the MLP layer receives the outputs from the Fourier neural operators v(x) and u, and uses similar residual connections to extract more features. These outputs will be high-dimensional feature representations of asset value and risk assessment. The MLP layer includes one or more hidden layers for extracting and learning complex nonlinear relationships. Each hidden layer can be activated using ReLU or other activation functions. The output layer of the MLP will specifically predict the value of each risk parameter (e.g., β, ρ, λ, δ, ξ, γ). Linear or appropriate activation functions, such as Sigmoid or Tanh, can be used, adjusted according to the specific range and nature of the parameters.

[0050] In this embodiment, the loss function is composed of a weighted sum of a first error and a second error based on a preset weight. The first error corresponds to the output error of the Fourier operator layer, that is, the difference between the frequency domain representation predicted by the Fourier operator layer and the corresponding true frequency domain representation. Evaluating the above difference in the frequency domain helps to capture the frequency characteristics of the data and guide the model to learn more accurate spatial domain features. The second error corresponds to the output error of the multilayer perceptron layer, that is, the difference between the risk exposure parameters predicted by the multilayer perceptron layer and the true parameters. During the training process, the parameters of the model are adjusted according to the value of the loss function, and the goal is to minimize the loss function. The setting of the preset threshold provides a stopping condition for model training. When the value of the loss function is less than the preset threshold, it means that the prediction performance of the model has reached an acceptable level. At this time, the model is considered to be trained and can be used for risk exposure assessment. The Adam optimizer can be selected for network training because it performs well when processing large data sets. The data can be processed in batches, each batch containing a certain number of data points. In each iteration, the network updates its weights and biases by minimizing the loss function. During the training process, you can use the validation dataset to monitor model performance and adjust hyperparameters such as learning rate and batch size based on the validation results to optimize network performance.

[0051] After the model training is completed, the model can also be verified. For example, a traditional neural network can be used as a benchmark to compare its results with the FNO model. If the results of the FNO model are similar to or better than this benchmark, this will provide additional verification for it; different market scenarios can be simulated, such as stock surges, plunges, or extreme behaviors of other assets to test the robustness of the model under different conditions; the impact of different input variables (such as different macroeconomic indicators, different asset allocation parameters) on model predictions can be analyzed. Ideally, small input changes should only lead to small output changes unless they are close to the boundaries of the constraints.

[0052] As an optional embodiment, obtaining initial parameters includes: constructing an initial price model and a volatility model, wherein the initial price model is a Black-Scholes model, and the volatility model characterizes the relationship between the volatility of total resources and time; combining the initial price model with the volatility model to obtain a target price model; and determining the initial parameters based on the target price model.

[0053] Optionally, in this embodiment, since virtual resources can be assets and total resources can be markets, the Black-Scholes equation can be used as the initial price model. The Black-Scholes model is a mathematical model for pricing options and other derivative financial products. It is based on the geometric Brownian motion hypothesis and describes how option prices change with changes in underlying asset prices and time. The BS model is widely used in financial engineering and risk management. First, the basic parameters of the model can be defined, such as S current price, K strike price, r risk-free interest rate, σ volatility, T expiration time, etc., and historical data and market forecasts can be used to determine the specific values ​​of the above parameters. For example, data from the past year can be used to estimate volatility. In addition, parameters related to the volatility of total resources can also be introduced, such as trading volume, market sentiment indicators, macroeconomic indicators, etc.

[0054] According to the above expansion parameters, the standard Black-Scholes equation can be expanded into a more complex partial differential equation. A random volatility model can also be introduced so that volatility is no longer a fixed parameter, but a random process that changes over time. The Heston model can be used to describe this random volatility. The above two equations are shown as follows:

[0055] Current price dynamics:

[0056]

[0057] Dynamics of volatility:

[0058]

[0059] Among them, S t is the current price, v t is the variance of volatility, r is the risk-free rate, κ is the mean reversion rate of volatility, θ is the variance of long-term volatility, ξ is the volatility of volatility, and dW t 1 and dW t 2 are two Brownian motions with a correlation coefficient of ρ.

[0060] In the standard Black-Scholes model, volatility σ is a constant. But in the stochastic volatility Heston model, volatility is time-varying. Incorporating the volatility dynamics of the Heston model into the Black-Scholes model yields the following extended Black-Scholes PDE:

[0061]

[0062] Among them, the parameters of Black-Scholes PDE and the collection of other related factors (such as market volatility data) are the inputs of the target risk assessment model.

[0063] As an optional embodiment, the initial parameters are input into the target risk assessment model to obtain the risk parameters, including: inputting the initial parameters into the Fourier operator layer to obtain the initial output results, wherein the initial output results include the option price of the virtual resources; inputting the initial output results into the multi-layer perceptron layer to obtain the risk parameters.

[0064] Optionally, the output of the Fourier operator layer is a feature representation after frequency domain transformation, including the option price prediction of the virtual resource. This is to convert the frequency domain features back to the time domain through the inverse Fourier transform to obtain the predicted value of the option price. This initial output result is the basis for further processing by the multi-layer perceptron layer. It not only contains option pricing information, but also implies the market dynamic characteristics captured by the Fourier transform. The multi-layer perceptron (MLP) layer is a neural network structure that can process the high-dimensional features output by the Fourier operator layer and fit and predict complex data relationships through nonlinear transformations. In this embodiment, the MLP layer receives the initial output results obtained from the Fourier operator layer, including option price predictions and frequency domain features. Through the forward propagation of the multi-layer neural network, these features are further processed to extract high-order features that are closely related to risk exposure assessment.

[0065] As an optional embodiment, a risk assessment result is determined based on risk parameters, wherein the risk parameters include multiple types of risk parameters, including: based on the change of total resources, setting corresponding weights of the multiple types of risk parameters; based on the weights, performing weighted summation of the multiple types of risk parameters to obtain a risk assessment result.

[0066] Optionally, in this embodiment, six types of risk parameters can be considered, including: market sensitivity parameter (β), asset correlation parameter (ρ), liquidity risk parameter (λ), credit risk parameter (δ), extreme event risk parameter (ξ), and market adaptability index (γ). Each parameter reflects the specific type of risk faced by virtual resources under different market conditions. The "changes" of total resources refer to changes in background conditions in the financial market, including but not limited to market volatility, interest rate levels, changes in macroeconomic indicators, etc. For example, when market volatility is high, asset prices may be more sensitive to market changes, and credit risks may become more significant during economic recessions. In-depth analysis of these changes is the basis for determining the weights of risk parameters.

[0067] The setting of weights needs to consider the degree of impact of changes in total resources on each risk parameter. For example, when market volatility is high, the weights of market sensitivity β and extreme event risk ξ can be increased, because in a high-volatility market, asset prices are more susceptible to market fluctuations and the impact of extreme market events is more obvious. The weight of the market adaptability indicator γ can also be reduced, because in a volatile market, the adaptability of an asset may not be as important as its inherent characteristics; if interest rates rise, the weight of credit risk δ can be increased, especially for bonds and other fixed-income products; if interest rates fall, liquidity risk λ may need to be reassessed, especially for highly liquid assets; in times of economic slowdown or recession, the weight of extreme event risk ξ can be increased; in a period of economic growth, it is possible to consider reducing the weight of extreme event risk ξ and increasing the weight of the market adaptability indicator γ; for investment portfolios pursuing high growth, For portfolios that focus on value investing, increase the weight of market sensitivity β, because growth stocks are usually more sensitive to market changes. At the same time, you can also appropriately increase the weight of extreme event risk ξ, because growth strategies may be greatly affected when the market falls. For portfolios that focus on value investing, increase the weight of liquidity risk λ and credit risk δ, because value investing may involve some assets with low liquidity or low credit ratings. You can also reduce the weight of market sensitivity β, because value stocks are usually more stable. For strategies that pursue a balanced asset allocation, avoid all risk indicators from becoming extreme to reflect the characteristics of diversified investment. You can appropriately increase the weight of the market adaptability indicator γ to reflect the ability of a balanced strategy to adjust under different market conditions. You can also impose a penalty coefficient on indicators with excessively large scores.

[0068] Weighted summation is to multiply each type of risk parameter by its corresponding weight, and then add all the results to get a comprehensive risk assessment result. For example, R = w1·β+w2·ρ+w3·λ+w4·δ+w5·ξ+w6·γ, where w1, w2, w3, w4, w5 and w6 are the weights corresponding to each risk indicator, and these weights can be adjusted according to the asset type or investment strategy. These adjustment methods provide a flexible framework that allows risk assessment results to be dynamically adjusted according to changes in market scenarios and investment strategies. In this way, the risk assessment model not only takes into account the inherent risk characteristics of the asset, but also adapts to market changes and investment strategy needs, thereby providing more comprehensive and accurate risk management.

[0069] As an optional embodiment, after determining the risk assessment result of the virtual resource, it also includes: comparing the risk assessment result with a preset level threshold to obtain the risk level corresponding to the risk assessment result; and determining a processing method for the virtual resource based on the risk level.

[0070] Optionally, the risk assessment result is a quantitative indicator that characterizes the risk level faced by the virtual resource under current market conditions. The setting of risk level thresholds is usually based on risk analysis of historical data and the understanding and definition of different risk levels. For example, three level thresholds can be set, low risk, medium risk and high risk levels. The setting of these thresholds needs to take into account factors such as market conditions, investor risk preferences and target investment returns.

[0071] Based on the determined risk level, financial institutions or investors can formulate corresponding treatment methods. For example, for high-level risks, they can explore strategies to reduce risks, such as diversification of investments, use of hedging tools, etc., and adopt more conservative strategies; for low-level and medium-level risks, they can consider increasing investment to pursue higher returns, but at the same time, they need to pay attention to risk control and consider more active investment strategies.

[0072] As an optional embodiment, Figure 6 is a flow chart of a risk assessment method based on Fourier operator according to an optional embodiment of the present invention. Figure 6 As shown in the figure, the Fourier Neural Operator (FNO) is combined with the Multilayer Perceptron (MLP) to form a powerful hybrid network model, which not only improves the model's ability to process high-dimensional data and capture complex spatial dynamics, but also enhances the model's accuracy and efficiency in assessing the value and risk of financial assets. The model is able to adjust risk scores based on changing market scenarios (such as market volatility, interest rate changes, etc.) and different investment strategies (such as growth, value, or balanced investments). This makes risk management more flexible and adaptable to market changes, improving the quality of investment decisions.

[0073] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that the present invention is not limited by the described action sequence, because according to the present invention, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.

[0074] Through the description of the above implementation methods, those skilled in the art can clearly understand that the risk assessment method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present invention.

[0075] According to an embodiment of the present invention, a device for implementing the above risk assessment method is also provided. Figure 7 is a structural block diagram of a risk assessment device provided according to an embodiment of the present invention, such as Figure 7 As shown, the device includes: an acquisition module 71, an evaluation module 72 and a determination module 73. The device is described below.

[0076] The acquisition module 71 is used to acquire initial parameters, wherein the initial parameters include the current price of virtual resources, the historical price of virtual resources, the volatility of total resources and the transaction volume of total resources, and the total resources include virtual resources.

[0077] The evaluation module 72 is connected to the acquisition module 71, and is used to input the initial parameters into the target risk assessment model to obtain risk parameters, wherein the network structure of the target risk assessment model includes a Fourier operator layer and a multi-layer perceptron layer, and the risk parameters represent the sensitivity of the price of virtual resources to changes in total resources.

[0078] The determination module 73 is connected to the evaluation module 72 and is used to determine the risk evaluation result of the virtual resource based on the risk parameter, wherein the magnitude of the risk evaluation result represents the magnitude of the risk of price drop of the virtual resource due to the change of the total resources.

[0079] It should be noted that the acquisition module 71, the evaluation module 72 and the determination module 73 correspond to steps S201 to S203 in the embodiment, and the examples and application scenarios implemented by the multiple modules and the corresponding steps are the same, but are not limited to the contents disclosed in the above embodiment. It should be noted that the above modules, as part of the device, can be run in the computer terminal 10 provided in the embodiment.

[0080] An embodiment of the present invention may provide a computer device. Optionally, in this embodiment, the computer device may be located in at least one network device among multiple network devices of a computer network. The computer device includes a memory and a processor.

[0081] Among them, the memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the risk assessment method and device in the embodiment of the present invention. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, realizing the above-mentioned risk assessment method. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the computer terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0082] The processor can call the information and application programs stored in the memory through the transmission device to perform the following steps: obtain initial parameters, wherein the initial parameters include the current price of virtual resources, the historical price of virtual resources, the volatility of total resources and the trading volume of total resources, and the total resources include virtual resources; input the initial parameters into the target risk assessment model to obtain risk parameters, wherein the network structure of the target risk assessment model includes a Fourier operator layer and a multi-layer perceptron layer, and the risk parameters represent the sensitivity of the price of virtual resources to changes in total resources; based on the risk parameters, determine the risk assessment results of the virtual resources, wherein the size of the risk assessment results represents the risk of a price drop in the virtual resources due to changes in total resources.

[0083] The embodiment of the present invention provides a risk assessment method. By obtaining initial parameters, wherein the initial parameters include the current price of virtual resources, the historical price of virtual resources, the volatility of total resources and the transaction volume of total resources, and the total resources include virtual resources; the initial parameters are input into a target risk assessment model to obtain risk parameters, wherein the network structure of the target risk assessment model includes a Fourier operator layer and a multi-layer perceptron layer, and the risk parameters represent the sensitivity of the price of virtual resources to changes in total resources; based on the risk parameters, the risk assessment results of virtual resources are determined, wherein the size of the risk assessment results represents the size of the risk of price decline of virtual resources caused by changes in total resources, and the purpose of accurately assessing the sensitivity of virtual resource prices to changes in total resources is achieved, thereby effectively predicting the risk of price decline of virtual resources, thereby achieving the technical effect of improving the accuracy and timeliness of risk assessment, and thus solving the technical problems of insufficient response and inaccurate assessment of traditional risk management methods.

[0084] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a non-volatile storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0085] The embodiment of the present invention further provides a non-volatile storage medium. Optionally, in this embodiment, the non-volatile storage medium can be used to store the program code executed by the risk assessment method provided by the above embodiment.

[0086] Optionally, in this embodiment, the non-volatile storage medium may be located in any computer terminal in a computer terminal group in a computer network, or in any mobile terminal in a mobile terminal group.

[0087] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for executing the following steps: obtaining initial parameters, wherein the initial parameters include the current price of the virtual resources, the historical price of the virtual resources, the volatility of the total resources and the trading volume of the total resources, and the total resources include the virtual resources; inputting the initial parameters into the target risk assessment model to obtain risk parameters, wherein the network structure of the target risk assessment model includes a Fourier operator layer and a multi-layer perceptron layer, and the risk parameters characterize the sensitivity of the price of the virtual resources to changes in the total resources; based on the risk parameters, determining the risk assessment result of the virtual resources, wherein the size of the risk assessment result characterizes the risk of a price drop in the virtual resources due to changes in the total resources.

[0088] An embodiment of the present invention also provides a computer program product, including a computer program. Optionally, in this embodiment, when the computer program is executed by a processor, it can achieve: obtaining initial parameters, wherein the initial parameters include the current price of virtual resources, the historical price of virtual resources, the volatility of total resources and the trading volume of total resources, and the total resources include virtual resources; inputting the initial parameters into a target risk assessment model to obtain risk parameters, wherein the network structure of the target risk assessment model includes a Fourier operator layer and a multi-layer perceptron layer, and the risk parameters represent the sensitivity of the price of virtual resources to changes in total resources; based on the risk parameters, determining the risk assessment results of the virtual resources, wherein the size of the risk assessment results represents the risk size of the price drop of the virtual resources caused by the change of total resources.

[0089] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0090] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0091] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0092] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0093] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0094] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a non-volatile storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0095] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A risk assessment method, characterized in that: include: Acquiring initial parameters, wherein the initial parameters include a current price of a virtual resource, a historical price of the virtual resource, a volatility of total resources, and a transaction volume of the total resources, wherein the total resources include the virtual resource; Inputting the initial parameters into a target risk assessment model to obtain risk parameters, wherein the network structure of the target risk assessment model includes a Fourier operator layer and a multi-layer perceptron layer, and the risk parameters represent the sensitivity of the price of the virtual resource to the change of the total resource; Based on the risk parameter, a risk assessment result of the virtual resource is determined, wherein the magnitude of the risk assessment result represents the magnitude of the risk of a price drop of the virtual resource due to a change in the total resources.

2. The method according to claim 1, characterized in that Before inputting the initial parameters into the target risk assessment model, the method includes: Constructing an initial risk assessment model, wherein the network structure of the initial risk assessment model includes a Fourier operator layer and a multi-layer perceptron layer; Based on the training data, the initial risk assessment model is trained, and the value of the loss function in the initial risk assessment model is calculated, wherein the training data includes the historical price of the virtual resource, the historical volatility of the total resource, and the historical transaction volume of the total resource; When the value of the loss function is less than a preset threshold, the model corresponding to the current training process is determined to be the target risk assessment model, wherein the loss function is a weighted sum of a first error and a second error based on a preset weight, the first error is the output error of the Fourier operator layer, and the second error is the output error of the multilayer perceptron layer.

3. The method according to claim 1, characterized in that The obtaining of initial parameters includes: Constructing an initial price model and a volatility model, wherein the initial price model is a Black-Scholes model, and the volatility model characterizes the relationship between the volatility of the total resources and time; Combining the initial price model with the volatility model to obtain a target price model; The initial parameters are determined according to the target price model.

4. The method according to claim 1, characterized in that: The step of inputting the initial parameters into the target risk assessment model to obtain risk parameters includes: Inputting the initial parameters into the Fourier operator layer to obtain an initial output result, wherein the initial output result includes the option price of the virtual resource; The initial output result is input into the multi-layer perceptron layer to obtain the risk parameter.

5. The method according to claim 1, characterized in that The risk assessment result is determined based on the risk parameter, wherein the risk parameter includes multiple types of risk parameters, including: Based on the change of the total resources, setting the weights corresponding to the multiple types of risk parameters; Based on the weights, the multiple types of risk parameters are weighted and summed to obtain the risk assessment result.

6. The method according to claim 1, characterized in that After determining the risk assessment result of the virtual resource, the method further includes: Compare the risk assessment result with a preset level threshold to obtain a risk level corresponding to the risk assessment result; Based on the risk level, a method for handling the virtual resource is determined.

7. A risk assessment device, characterized in that: include: An acquisition module, used for acquiring initial parameters, wherein the initial parameters include the current price of the virtual resource, the historical price of the virtual resource, the volatility of the total resources and the transaction volume of the total resources, wherein the total resources include the virtual resource; An evaluation module, used for inputting the initial parameters into a target risk evaluation model to obtain risk parameters, wherein the network structure of the target risk evaluation model includes a Fourier operator layer and a multi-layer perceptron layer, and the risk parameters represent the sensitivity of the price of the virtual resources to the change of the total resources; A determination module is used to determine a risk assessment result of the virtual resource based on the risk parameter, wherein the magnitude of the risk assessment result represents the magnitude of the risk of a price drop of the virtual resource due to a change in the total resources.

8. A non-volatile storage medium, characterized in that: The non-volatile storage medium includes a stored program, wherein when the program is executed, the device where the non-volatile storage medium is located is controlled to execute the risk assessment method according to any one of claims 1 to 6.

9. A computer device, characterized in that: include: Memory and processor, The memory stores a computer program; The processor is used to execute the computer program stored in the memory, and when the computer program is run, the processor executes the risk assessment method according to any one of claims 1 to 6.

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

Images