Asset analysis method and device based on artificial intelligence, computer equipment and medium

Abstract

The invention belongs to the technical field of artificial intelligence, and relates to an asset analysis method and device based on artificial intelligence, computer equipment and a storage medium, and the method comprises the steps: receiving an income analysis request for a target asset input by a user; wherein the income analysis request carries an analysis time period; obtaining asset related data corresponding to the target asset, and obtaining factor data associated with the analysis time period; integrating the asset related data and the factor data to obtain integrated data, and preprocessing the integrated data to obtain target data; performing income prediction processing on the target data based on a called preset risk income prediction model to obtain a risk income prediction result corresponding to the analysis time period; and carrying out display processing on the risk income prediction result. In addition, the risk income prediction result can be stored in the block chain. The method can be applied to an asset analysis scene in the field of financial science and technology, and the processing efficiency and accuracy of risk income analysis are effectively improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence, and can be applied to the field of financial technology, in particular to an asset analysis method and device based on artificial intelligence, a computer device and a storage medium. BACKGROUND

[0002] In the field of traditional financial asset risk and return analysis, the related tools and platforms on the market currently mostly use traditional statistical methods, such as the mean-variance model. These methods mainly estimate the yield and volatility of different assets based on historical data, and then analyze the risk and return characteristics thereof. However, this analysis method has obvious defects, resulting in low accuracy and processing efficiency of asset risk and return analysis.

[0003] Specifically, the traditional statistical method often assumes that the market environment and asset characteristics will remain similar to the past in the future, ignoring the influence of factors such as market dynamic changes, unexpected events, and complex correlation between assets. For example, in the stock investment analysis scenario, the traditional method only uses the yield and volatility data of a certain stock in the past period of time, uses the mean-variance model to calculate the risk and return characteristics thereof, and gives investment suggestions based thereon. However, when major policy adjustments, industry crises and other events occur in the market, the risk and return characteristics of the stock will change dramatically. The traditional method cannot capture these dynamic changes in time, and the investment suggestions given are often greatly deviated from the actual situation, which cannot provide accurate and effective decision support for investors. At the same time, the traditional method has high computational complexity and slow processing speed when dealing with large-scale, multi-dimensional financial data, and cannot meet the real-time analysis requirements.

[0004] Therefore, it is urgent to provide a new asset risk and return analysis method to improve the accuracy and processing efficiency of asset analysis and provide more reliable decision basis for investors. SUMMARY

[0005] The purpose of the embodiments of the present application is to provide an asset analysis method, device, computer device and storage medium based on artificial intelligence, to solve the technical problem of low accuracy and processing efficiency of existing asset risk and return analysis.

[0006] In a first aspect, an asset analysis method based on artificial intelligence is provided, comprising:

[0007] receiving a yield analysis request for a target asset input by a user; wherein the yield analysis request carries an analysis time period;

[0008] obtaining asset-related data corresponding to the target asset based on a preset data type, and obtaining factor data associated with the analysis time period;

[0009] Integrate the asset-related data and the factor data to obtain integrated data, and pre-process the integrated data to obtain corresponding target data;

[0010] Call a preset risk and return prediction model; wherein the risk and return prediction model is a model generated by training a preset deep learning model based on asset sample data subjected to feature enhancement processing;

[0011] Perform return prediction processing on the target data based on the risk and return prediction model to obtain a risk and return prediction result corresponding to the analysis time period;

[0012] Perform display processing on the risk and return prediction result.

[0013] In a second aspect, an asset analysis device based on artificial intelligence is provided, comprising:

[0014] A receiving module configured to receive a return analysis request for a target asset input by a user; wherein the return analysis request carries an analysis time period;

[0015] A first obtaining module configured to obtain asset-related data corresponding to the target asset based on a preset data type, and obtain factor data associated with the analysis time period;

[0016] A first processing module configured to integrate the asset-related data and the factor data to obtain integrated data, and pre-process the integrated data to obtain corresponding target data;

[0017] A calling module configured to call a preset risk and return prediction model; wherein the risk and return prediction model is a model generated by training a preset deep learning model based on asset sample data subjected to feature enhancement processing;

[0018] A prediction module configured to perform return prediction processing on the target data based on the risk and return prediction model to obtain a risk and return prediction result corresponding to the analysis time period;

[0019] A first display module configured to perform display processing on the risk and return prediction result.

[0020] In a fourth aspect, a computer readable storage medium is provided, which stores a computer program, and the computer program is executed by a processor to implement the steps of the above-mentioned asset analysis method based on artificial intelligence.

[0021] In the aforementioned scheme implemented by the AI-based asset analysis method, device, computer equipment, and storage medium, the following steps are taken: First, a user-inputted request for return analysis of a target asset is received; wherein the return analysis request carries an analysis time period; then, asset-related data corresponding to the target asset and factor data associated with the analysis time period are obtained based on a preset data type; subsequently, the asset-related data and the factor data are integrated to obtain integrated data, and the integrated data is preprocessed to obtain the corresponding target data; next, a preset risk-return prediction model is invoked; wherein, the risk-return prediction model is a model generated by training a preset deep learning model on asset sample data obtained through feature enhancement processing; and the target data is processed for return prediction based on the risk-return prediction model to obtain a risk-return prediction result corresponding to the analysis time period; finally, the risk-return prediction result is displayed. Based on the above processing flow, this application obtains asset-related data corresponding to the target asset and factor data associated with the analysis period, integrates and preprocesses the asset-related data and factor data to obtain target data, and then uses a risk-return prediction model to perform return prediction processing on the target data and generate risk-return prediction results. This enables efficient and accurate risk-return analysis of the target asset, improves the processing efficiency of asset risk-return analysis, and ensures the accuracy and effectiveness of the obtained risk-return prediction results. Attached Figure Description

[0022] To more clearly illustrate the solutions in this application, the accompanying drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is an exemplary system architecture diagram to which this application can be applied;

[0024] Figure 2 This is a flowchart of an embodiment of the AI-based asset analysis method according to this application;

[0025] Figure 3 This is a schematic diagram of the structure of an embodiment of the AI-based asset analysis device according to this application;

[0026] Figure 4 This is a schematic diagram of the structure of one embodiment of the computer device according to this application. Detailed Implementation

[0027] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs; the terms used in the specification are intended to describe the particular embodiments and are not intended to limit the application; the terms "include" and "have" and their any variations used in the specification and the claims and the above description of drawings are intended to cover the non-exclusive inclusion; the terms "first", "second" and the like used in the specification and the claims and the above description of drawings are intended to distinguish different objects, not to describe a particular order.

[0028] Reference herein to "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment can be included in at least one embodiment of the application. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are they necessarily mutually exclusive of one another. It is expressly understood that the embodiments described herein are merely examples and are not a limitation as to the scope of the application.

[0029] In order to make the person skilled in the art better understand the scheme of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely in conjunction with the drawings below.

[0030] As shown in Figure 1 The system architecture 100 can include a terminal device 101, a network 102 and a server 103, and the terminal device 101 can be a notebook computer 1011, a tablet computer 1012 or a mobile phone 1013. The network 102 is a medium for providing a communication link between the terminal device 101 and the server 103. The network 102 can include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0031] The user can use the terminal device 101 to interact with the server 103 through the network 102 to receive or send messages, etc. Various communication client applications can be installed on the terminal device 101, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0032] The terminal device 101 can be various electronic devices with a display screen and supporting web browsing, in addition to the notebook computer 1011, the tablet computer 1012 or the mobile phone 1013, the terminal device 101 can also be an electronic book reader, an MP3 player (Moving Picture Experts Group Audio Layer III), an MP4 player (Moving Picture Experts Group Audio Layer IV), a laptop computer and a desktop computer, etc.

[0033] The server 103 can be a server providing various services, for example, a background server providing support for a page displayed on the terminal device 101.

[0034] It should be noted that the asset analysis method based on artificial intelligence provided in the embodiments of the present application is generally executed by a server / terminal device, and accordingly, the asset analysis device based on artificial intelligence is generally arranged in a server / terminal device.

[0035] It should be understood that Figure 1 The number of terminal devices, networks and servers in

[0036] With reference to Figure 2 , a flowchart of one embodiment of the asset analysis method based on artificial intelligence according to the present application is shown. The order of the steps in the flowchart can be changed according to different needs, and some steps can be omitted. The asset analysis method based on artificial intelligence provided in the embodiments of the present application can be applied to any scenario requiring asset analysis, and then the asset analysis method based on artificial intelligence can be applied to products in these scenarios, for example, asset analysis scenarios in the field of financial technology. The asset analysis method based on artificial intelligence includes the following steps:

[0037] Step S201, receiving a user inputted income analysis request for a target asset; wherein the income analysis request carries an analysis time period.

[0038] In the present embodiment, the electronic device (for example Figure 1The server / terminal device shown) can obtain a user inputted yield analysis request for a target asset through a wired connection or a wireless connection. It should be noted that the wireless connection can include, but is not limited to, 3G / 4G / 5G connection, WiFi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB (ultra wide band) connection, and other now known or future developed wireless connection. The execution subject of the present application is an asset analysis system, which can be referred to as a system. The yield analysis request can be a processing request inputted by a user in the system according to actual query requirements, which needs to query the yield analysis result of the target asset in a specified analysis time period. Exemplarily, the target asset can be stocks, bonds, funds, etc. The analysis time period can be one week, one month, one year, etc. in the future.

[0039] Among them, the present application can be applied to the asset analysis scene in the field of financial technology. For example, in the financial field, the content of the yield analysis request can include that an investor pays attention to the stock of a certain technology company and hopes to understand the yield potential of the stock in the future year and the possible risks. Or, the content of the yield analysis request can also include that an investment institution manages a diversified investment portfolio containing stocks, bonds and funds, and the market fluctuates greatly recently, and the institution hopes to optimize the investment portfolio. The specific request includes: analyzing whether the allocation ratio of various assets in the current investment portfolio is reasonable, whether it needs to be adjusted according to the market trend; predicting the expected yield and risk level of the optimized investment portfolio in the future half year; evaluating the contribution of different asset categories to the overall yield and risk of the investment portfolio, so as to rebalance the assets if necessary.

[0040] For another example, the content of the yield analysis request can also include that a certain individual investor holds an ETF fund centered on the new energy industry and a combination of part of blue-chip stocks, and hopes to perform yield analysis recently due to the adjustment of new energy policy and the frequent rotation of stock sectors. The specific request includes: evaluating whether the yield performance of the new energy ETF and the blue-chip stock in the current holding meets the expectation, judging whether the proportion of the two needs to be adjusted; predicting the yield fluctuation range and potential downward risk of the adjusted combination in the future three months; analyzing the marginal influence of the new energy ETF and the blue-chip stock on the combination yield and risk under different market environments (such as policy benefits or economic fluctuations), providing basis for subsequent dynamic rebalancing. Explanation: The request focuses on the individual investment portfolio, and clearly states the current holding structure (new energy ETF + blue-chip stock) and market background (policy adjustment, sector rotation). The core demands include: 1) evaluating the yield performance and proportion rationality of the existing holding; 2) predicting the short-term yield and risk after adjustment; 3) quantifying the marginal influence of different asset categories on the combination, providing decision support for dynamic rebalancing, covering the three dimensions of yield analysis, risk assessment and strategy optimization.

[0041] In step S202, asset-related data corresponding to the target asset is obtained based on a preset data type, and factor data associated with the analysis time period is obtained.

[0042] In the embodiment, the data type at least refers to a type of key indicator that needs to be collected corresponding to an analysis target of the asset type of the target asset, for example, can include a yield, a volatility, a VaR (Value at Risk), and a maximum drawdown. Correspondingly, the asset-related data can refer to yield, volatility, VaR (Value at Risk), and maximum drawdown data of the target asset. In addition, the factor data refers to external factors related to the target asset and the analysis time period, for example, can include macroeconomic indicators (GDP growth rate, inflation rate, interest rate, etc.), market sentiment indexes (investor confidence index, market sentiment analysis on social media, etc.).

[0043] In step S203, the asset-related data and the factor data are integrated to obtain integrated data, and the integrated data is preprocessed to obtain corresponding target data.

[0044] In the embodiment, the specific implementation process of preprocessing the integrated data to obtain the corresponding target data will be further described in detail in subsequent specific embodiments, and will not be described here.

[0045] In step S204, a preset risk and return prediction model is called; wherein the risk and return prediction model is a model generated by training a preset deep learning model based on asset sample data obtained through feature enhancement processing.

[0046] In the embodiment, the specific construction process of the risk and return prediction model will be further described in detail in subsequent specific embodiments, and will not be described here.

[0047] In step S205, the target data is processed for yield prediction based on the risk and return prediction model, to obtain a risk and return prediction result corresponding to the analysis time period.

[0048] In the embodiment, the target data can be input into a pre-constructed risk and return prediction model, which will predict the risk and return of the target data to obtain a prediction result of the risk and return characteristics of the asset corresponding to the analysis time period, that is, the risk and return prediction result.

[0049] In step S206, the risk and return prediction result is processed for display.

[0050] In the present embodiment, the display of the risk and return prediction result includes designing a simple and intuitive user interface to help users efficiently analyze risk and return and make investment decisions. The display of the risk and return prediction result can be completed by displaying the risk and return prediction result in the user interface. The display template can be designed according to the type of the risk and return prediction result. For example, key indicators (yield, volatility, VaR, maximum drawdown) can be displayed using a table, asset price trends can be displayed using a line chart, and risk and return distribution can be displayed using a scatter plot. Important information can be highlighted through color, font size, or icons. For example, red can represent negative returns and green can represent positive returns; key indicators can be displayed in bold font. Users can view more detailed information, such as detailed transaction data for a certain day or the calculation method of a certain indicator, by clicking on the entries in the chart or table.

[0051] In addition, interactive controls (such as sliders and input boxes) can be provided to allow users to adjust analysis parameters. For example, users can adjust the length of the time window using a slider or modify the weight of macroeconomic indicators using an input box. After the user adjusts the parameters, the system should update the analysis results in real time to reflect the impact of the parameter changes. For example, after adjusting the time window, the interface should immediately display the analysis results for the new time window. Users can save analysis results for different parameter combinations for easy comparison and analysis later. For example, users can save the current parameter settings and give them a name to quickly call them up.

[0052] The application first receives a user inputted return analysis request for a target asset; wherein the return analysis request carries an analysis time period; then acquires asset related data corresponding to the target asset based on a preset data type, and acquires factor data associated with the analysis time period; thereafter integrates the asset related data and the factor data to obtain integrated data, and pre-processes the integrated data to obtain corresponding target data; subsequently calls a preset risk return prediction model; wherein the risk return prediction model is a model generated by training a preset deep learning model based on asset sample data obtained through feature enhancement processing; and performs return prediction processing on the target data based on the risk return prediction model to obtain a risk return prediction result corresponding to the analysis time period; finally performs display processing on the risk return prediction result. Based on the above processing flow, the application acquires asset related data corresponding to the target asset and factor data associated with the analysis time period, integrates and pre-processes the asset related data and the factor data to obtain target data, and then performs return prediction processing on the target data based on the use of the risk return prediction model to generate a risk return prediction result, so that efficient and accurate risk return analysis corresponding to the target asset can be realized, the processing efficiency of risk return analysis is improved, and the accuracy and effectiveness of the obtained risk return prediction result are ensured.

[0053] In some optional implementations, before step S204, the above-mentioned electronic device can further perform the following steps:

[0054] Collecting historical related data of multiple assets from a preset data source.

[0055] In the embodiment, the data source can be a professional financial data provider or an official financial regulatory agency. The asset type to be analyzed, such as stocks, bonds, funds, etc., can be determined. According to the analysis target, the key indicators required to be collected, such as yield, volatility, VaR (Value at Risk) and maximum drawdown, are determined. Then, the historical related data of the assets are acquired from the selected data source through data interface, data download tool or negotiation with the data supplier, and it is ensured that the acquired data covers the required time range and the data format meets the requirements of subsequent processing. In addition, the historical related data can be pre-processed (handling of missing values, handling of outliers and standardization) to improve the quality and standardization of the data.

[0056] Performing feature extraction on the historical related data based on a preset feature extraction algorithm to obtain first feature data related to risk return.

[0057] In this embodiment, the feature extraction algorithm described above can specifically adopt a deep learning method, such as a deep neural network (DNN), an autoencoder, etc. These algorithms can automatically learn complex patterns and features in the data. The pre-processed data can be input into the selected machine learning model for training. During the training process, the model will automatically learn features related to risk and return, such as the non-linear relationship between yield and volatility, the impact of macroeconomic indicators on asset performance, etc. Further, these features can be obtained through the output of the intermediate layers of the model or specific feature extraction techniques.

[0058] Obtain external information related to the asset.

[0059] In this embodiment, external information related to the asset can be collected, such as macroeconomic indicators (GDP growth rate, inflation rate, interest rate, etc.), market sentiment indices (investor confidence index, market sentiment analysis on social media, etc.).

[0060] Based on the external information, the first feature data is enhanced using a preset feature enhancement strategy to obtain processed second feature data.

[0061] In this embodiment, the specific implementation process of the above based on the external information, the first feature data is enhanced using a preset feature enhancement strategy to obtain processed second feature data will be further described in detail in subsequent specific embodiments, and will not be described in detail here.

[0062] The second feature data is used as the asset sample data.

[0063] The present application collects historical related data of multiple assets from a preset data source; then based on a preset feature extraction algorithm, the historical related data is feature extracted to obtain first feature data related to risk and return; then external information related to the asset is obtained; and based on the external information, the first feature data is enhanced using a preset feature enhancement strategy to obtain processed second feature data; subsequently, the second feature data is used as the asset sample data. Based on the above processing procedure, the present application can efficiently and accurately construct the required asset sample data by feature extraction and feature enhancement of the collected historical related data of multiple assets. Through the use of feature extraction algorithms, complex non-linear relationships and potential features in the historical related data can be discovered. Feature enhancement enriches the dimensions of the data by introducing external information, enabling the model to more comprehensively consider factors affecting asset risk and return. Further, through feature extraction and enhancement, more abundant feature inputs can be provided for constructing more accurate risk and return prediction models, which helps to improve the prediction performance of the risk and return prediction model.

[0064] In some optional implementations of the embodiment, the first feature data is enhanced using a preset feature enhancement strategy based on the external information to obtain processed second feature data, including the following steps:

[0065] A plurality of fusion methods are obtained.

[0066] In the embodiment, the fusion methods can include at least simple concatenation, weighted average, and multi-modal feature fusion. Specifically, the simple concatenation method includes directly concatenating the external information and the original data to form an extended feature vector. This method is simple and easy to implement, but may lead to an increase in feature dimension and redundancy. The weighted average method includes assigning different weights to the external information and the original data and calculating the weighted average. The determination of the weights can be based on expert experience or determined through an optimization algorithm (such as grid search). The multi-modal feature fusion method includes using more complex fusion algorithms, such as principal component analysis (PCA), canonical correlation analysis (CCA), or multi-modal fusion networks in deep learning. These methods can capture complex relationships between different data sources and improve the fusion effect.

[0067] A target fusion method is selected from all the fusion methods.

[0068] In the embodiment, one of the fusion methods can be randomly selected as the target fusion method according to actual business needs.

[0069] The external information and the first feature data are fused based on the target fusion method to obtain corresponding fusion data.

[0070] In the embodiment, the fusion process includes data alignment: ensuring that the external information and the original data are aligned in time and frequency. For example, if the original data is daily data, the external information also needs to be converted to daily frequency. Feature fusion: according to the selected fusion method, the external information and the original data are fused. For example, PCA is used to reduce the dimension of high-dimensional features and fuse them, or a multi-modal fusion network is used to learn the joint representation of different data sources.

[0071] The fusion effect of the fusion data is verified.

[0072] In the embodiment, the fusion effect of the fusion data can be verified by a statistical analysis method, and a corresponding verification result is generated. The content of the verification result includes that the fusion effect of the fusion data passes the verification, or that the fusion effect of the fusion data does not pass the verification.

[0073] If the fusion effect of the fusion data passes the verification, the fusion data is used as the second feature data.

[0074] In the embodiment, the parameters of the fusion method can also be adjusted according to the verification result, such as the number of principal components of PCA, the weight of weighted average, and the like. In addition, if the current fusion method is not ideal, other fusion methods or a combination of multiple methods are tried. For example, weighted average is used for preliminary fusion, and then PCA is used for further dimension reduction.

[0075] The application obtains a plurality of preset fusion methods, selects a target fusion method from all the fusion methods, performs fusion processing on the external information and the first feature data based on the target fusion method, obtains corresponding fusion data, verifies the fusion effect of the fusion data, and takes the fusion data as the second feature data if the fusion effect of the fusion data passes the verification. Based on the above processing flow, the use of the target fusion method can effectively fuse the external information and the first feature data, effectively enhance the performance of the first feature data, and further improve the prediction accuracy of the risk and return prediction model for asset risk and return.

[0076] In some optional implementations, before step S204, the above electronic device can further perform the following steps:

[0077] A corresponding deep learning model is determined based on a preset asset analysis requirement.

[0078] In the embodiment, the data characteristics of the asset risk and return characteristics are analyzed, such as the time sequence, volatility, periodicity, and the like of the data, and then a suitable model is selected as the deep learning model according to the data characteristics. For example, if the data has obvious local volatility mode, CNN can be more suitable; if the data has long-term dependence relationship, LSTM can be more suitable.

[0079] The asset sample data is divided into training data and verification data.

[0080] In the embodiment, the division ratio of the training data and the verification data can be determined according to the data volume and the model requirement, for example, 70% training data and 30% verification data. The asset sample data can be divided into training data and verification data by using random sampling or stratified sampling, so as to ensure that the divided data sets are consistent in distribution and avoid introducing bias due to data division.

[0081] The deep learning model is trained based on the training data to obtain a corresponding first processing model.

[0082] In this embodiment, the selected deep learning model can be initialized by parameter initialization, such as random initialization or pre-trained weights. The hyperparameters of the training are set, such as learning rate, batch size, number of training rounds, etc. The learning rate affects the update speed of the model parameters, the batch size affects the stability and efficiency of the training, and the number of training rounds affects the convergence of the model. Then the above deep learning model is trained using the above training data, the gradient is calculated by the back propagation algorithm and the model parameters are updated. During the training process, the changes of training loss and validation loss are monitored to prevent overfitting or underfitting of the model, until a trained model is obtained and used as the corresponding first processing model.

[0083] The first processing model is evaluated and optimized based on the validation data to obtain a corresponding second processing model.

[0084] In this embodiment, appropriate evaluation indicators can be selected according to the analysis target, such as mean square error (MSE), mean absolute error (MAE), R 2 These indicators can measure the difference between the model prediction results and the true values. Then the trained first processing model is evaluated using the validation data, and the value of the evaluation indicator is calculated. According to the evaluation result, it is judged whether the model has overfitting or underfitting problem. And according to the evaluation result, the hyperparameters of the first processing model are adjusted, such as learning rate, network layer number, neuron number, etc. The optimal combination of hyperparameters can be found by grid search, random search, etc. If the model performance is not ideal, the model structure is adjusted, such as increasing or decreasing the number of network layers, changing the type of activation function, etc., until a model that meets the requirements is obtained and used as the second processing model.

[0085] The second processing model is optimized based on a preset multi-factor analysis strategy to obtain a corresponding third processing model.

[0086] In this embodiment, the strategy content of the above multi-factor analysis strategy includes determining the analysis factors and the comprehensive evaluation model. Specifically, determining the analysis factors includes: identifying key factors: combining economic indicators (such as GDP growth rate, inflation rate, interest rate), market sentiment (such as investor confidence index, social media sentiment), and policy factors (such as monetary policy, fiscal policy) to determine the key factors that affect the risk and return characteristics of assets. Collect factor data: collect data related to these key factors to ensure that the time range and frequency of the data are consistent with the asset data.

[0087] The comprehensive evaluation model includes: selecting an evaluation method: selecting a suitable comprehensive evaluation method according to the analysis requirements, such as weighted comprehensive evaluation, analytic hierarchy process (AHP), principal component analysis (PCA), etc. These methods can quantify the influence of different factors and integrate them into a whole evaluation. Perform comprehensive evaluation: include different factors in the model analysis, and use the selected evaluation method to comprehensively evaluate the risk and return characteristics of different categories of assets in the model. For example, through weighted comprehensive evaluation, each factor can be assigned a weight, and the weighted score can be calculated to evaluate the risk and return characteristics of the asset. Analyze the evaluation results: analyze the comprehensive evaluation results to understand the influence of different factors on the risk and return characteristics of the asset. Through multi-factor analysis, a more comprehensive understanding of the risk and return characteristics of the asset can be obtained, providing more information for investment decision-making.

[0088] Among them, the multi-factor analysis is carried out in the training stage of the risk and return prediction model, mainly to optimize the input features of the model, improve the model's ability to capture the risk and return characteristics of the asset, so that the model learns more comprehensive and accurate data patterns. Further, in the model application stage of the risk and return prediction model, multi-factor analysis can automatically and intelligently adjust and explain the prediction results of the model according to the current market environment and actual situation, thereby providing more accurate basis for investment decision-making.

[0089] Generating the risk and return prediction model based on the third processing model.

[0090] In this embodiment, the specific implementation process of generating the risk and return prediction model based on the third processing model will be further described in detail in the subsequent specific embodiments, and will not be described here.

[0091] The application determines a corresponding deep learning model based on a preset asset analysis requirement; then divides the asset sample data into training data and verification data; then trains the deep learning model based on the training data to obtain a corresponding first processing model; and evaluates and optimizes the first processing model based on the verification data to obtain a corresponding second processing model; subsequently optimizes the second processing model based on a preset multi-factor analysis strategy to obtain a corresponding third processing model; and finally generates the risk and return prediction model based on the third processing model. Based on the above processing procedure, since the deep learning model can process complex data patterns and nonlinear relationships, the application can efficiently and accurately construct a risk and return prediction model capable of generating accurate risk and return prediction results by training and optimizing the deep learning model using asset sample data, thereby improving the model construction efficiency of the risk and return prediction model and ensuring the model prediction effect of the risk and return prediction model. Moreover, the multi-factor analysis considers multiple factors affecting asset risk and return, so that the risk and return analysis result is more comprehensive and reliable.

[0092] In some optional implementations, the generating the risk and return prediction model based on the third processing model comprises the following steps:

[0093] optimizing the model parameters and analysis method of the third processing model based on a preset reinforcement learning algorithm to obtain a corresponding fourth processing model.

[0094] In this embodiment, the model parameters and analysis method of the third processing model can be adjusted in real time by applying a reinforcement learning algorithm, such as automatically adjusting the time window or analysis interval for analyzing risk and return characteristics according to market conditions.

[0095] Specifically, the process of optimizing the model parameters and analysis method of the third processing model comprises: 1. Determine the reinforcement learning framework. Environment definition: regard the financial market as the environment of reinforcement learning, wherein the state (State) includes the current market data (such as price, volatility, trading volume, etc.), macroeconomic indicators, market sentiment index, etc. Agent (Agent): responsible for selecting the action (Action) of adjusting the model parameters or analysis method according to the state of the current environment, so as to optimize the prediction performance of the model. Reward (Reward): define the reward function to evaluate whether the action taken by the agent is effective. For example, the reward can be based on the accuracy of the model prediction (such as the deviation of the predicted yield from the actual yield), the risk-adjusted yield (such as the Sharpe ratio), etc.

[0096] 2. State representation and feature extraction. Feature engineering: Extract key features from real-time data, such as the current asset's price and its rate of change. Volatility indicators (e.g., historical volatility, implied volatility). Macroeconomic indicators (e.g., GDP growth rate, inflation rate, interest rate). Market sentiment indicators (e.g., investor confidence index, social media sentiment analysis). State encoding: Encode the extracted features into a state representation that the reinforcement learning agent can process, such as using a vector or matrix form.

[0097] 3. Action space design. Parameter adjustment actions: Define model parameters that can be adjusted, such as learning rate, regularization coefficient, number of network layers, etc. Analysis method adjustment actions: Define analysis methods that can be adjusted, such as time window size (e.g., switching from daily data to hourly data), feature selection strategy (e.g., increasing or decreasing the influence weight of certain external information).

[0098] 4. Reward function design. Accuracy reward: Give positive or negative rewards based on the deviation of the model's prediction results from the actual market results. Risk-adjusted reward: Adjust the reward by incorporating risk indicators (e.g., VaR, maximum drawdown) to encourage the model to improve returns while keeping risk under control. Stability reward: Reward the stability of model parameters to avoid frequent and large adjustments that can destabilize the model.

[0099] 5. Reinforcement learning algorithm selection and training. Algorithm selection: Choose appropriate reinforcement learning algorithms, such as Deep Q Network (DQN), Policy Gradient, or Proximal Policy Optimization (PPO), which can handle high-dimensional state spaces and continuous action spaces. Training process: Offline training: Perform initial training on historical data to allow the agent to learn basic strategies. Online learning: In real-time environments, the agent continuously updates its strategy based on new market data, gradually optimizing model parameters and analysis methods.

[0100] 6. Real-time adjustment and optimization. Real-time data stream processing: Establish a real-time data pipeline to continuously obtain the latest market data, macroeconomic indicators, and market sentiment information. State updating: Update the current state representation based on real-time data. Action selection and execution: The agent selects the optimal action (adjusts model parameters or analysis methods) based on the current state and executes it. Feedback loop: Calculate the reward based on the model's performance under the new parameters or methods and update the agent's strategy.

[0101] Obtain the preset real-time update strategy.

[0102] In this embodiment, the policy content of the real-time updating strategy includes: 1. Real-time data acquisition and processing: establish a connection with real-time data sources (such as exchange API, news API, social media API), ensure that the latest market data and related information can be obtained in time. And the real-time data is preprocessed in the same way as the training data, including cleaning, standardization and feature extraction. 2. Model state monitoring: real-time monitoring of model prediction performance, such as prediction error, risk indicators, etc. And detect changes in market environment, such as sudden increase in volatility, sharp changes in market sentiment, etc. 3. Dynamic adjustment strategy: when significant changes in market environment or model performance are detected, trigger reinforcement learning agent to adjust the strategy. Then, the agent selects the action of adjusting the model parameters or analysis method according to the current state, and executes the adjustment. 4. Model updating and verification: use incremental learning method, only update part of the model parameters, to reduce the computational burden and maintain the real-time performance of the model. And in a short period of time after adjustment, verify the performance of the model under the new parameters or method, to ensure that the adjustment is effective. 5. Feedback and optimization cycle: based on the real-time verification results, continuously optimize the reinforcement learning strategy to improve the model's adaptability to market changes. And record the history and effect of model adjustment for subsequent analysis and strategy optimization.

[0103] Based on the real-time updating strategy, the fourth processing model is updated to obtain a corresponding fifth processing model.

[0104] In this embodiment, the fourth processing model can be updated based on the policy content of the real-time updating strategy, and the updated fifth processing model can be used as the final risk and return prediction model.

[0105] The fifth processing model is used as the risk and return prediction model.

[0106] The application optimizes the model parameters and analysis methods of the third processing model based on the preset reinforcement learning algorithm to obtain a corresponding fourth processing model. Then a preset real-time updating strategy is obtained. Based on the real-time updating strategy, the fourth processing model is updated to obtain a corresponding fifth processing model. The fifth processing model is used as the risk and return prediction model. Based on the implemented processing flow, the application uses reinforcement learning algorithm to adjust the model parameters and analysis methods of the third processing model in real time, and the real-time updating mechanism of the model, so that the finally generated risk and return prediction model can better adapt to the dynamically changing market environment, and provide more accurate asset risk and return analysis, improving the asset risk analysis effect of the risk and return prediction model.

[0107] In some optional implementations of the embodiment, the pre-processing of the integrated data in step S203 obtains corresponding target data, including the following steps:

[0108] The integrated data is processed for missing values to obtain corresponding first processing data.

[0109] In the embodiment, the missing value processing includes checking whether there is a missing value in the data. For a small amount of missing values, interpolation methods such as linear interpolation, spline interpolation, etc. are used to fill in the missing values according to the values of adjacent data points. For a large amount of missing values, the data point or the data of the asset is deleted, and the influence on the analysis result is carefully evaluated.

[0110] The first processing data is processed for outliers to obtain corresponding second processing data.

[0111] In the embodiment, the outlier processing includes identifying outliers in the data, which can be determined by statistical methods (such as standard deviation method, box plot method) or based on business knowledge. For outliers, they can be corrected (such as replaced with reasonable values) or deleted, but the processing process needs to be recorded for subsequent analysis and verification.

[0112] The second processing data is standardized based on a preset standardization strategy to obtain corresponding third processing data.

[0113] In the embodiment, the strategy content of the standardization strategy includes selecting a standardization method, including Z-score standardization, Min-Max standardization, etc. Z-score standardization converts data to a distribution with a mean of 0 and a standard deviation of 1; Min-Max standardization linearly transforms data to a specified range (such as [0, 1]). Apply the standardization method: according to the selected standardization method, calculate and convert the original data to make the data meet the input requirements of the machine learning model, and avoid affecting the training effect of the model due to different data dimensions.

[0114] The third processing data is used as the target data.

[0115] In the embodiment, data cleaning (missing value processing and outlier processing) ensures the quality of the data, removes noise and error data, and makes the analysis result more accurate. Standardization processing solves the problem of inconsistent dimensions of different indicators, so that the model can better learn and fit the patterns in the data.

[0116] The application obtains corresponding first processing data by performing missing value processing on the integrated data. Then, the application obtains corresponding second processing data by performing outlier processing on the first processing data. Then, the application obtains corresponding third processing data by performing standardization processing on the second processing data based on a preset standardization strategy. Subsequently, the third processing data is taken as the target data. Based on the above processing procedure, the application can efficiently and accurately complete the preprocessing of the integrated data by performing missing value processing, outlier processing, and standardization processing, so as to obtain comprehensive and reliable target data. The target data obtained based on the data preprocessing is beneficial to providing a high-quality data basis for subsequent feature extraction and model prediction processing.

[0117] In some optional implementations of the embodiment, after step S205, the electronic device can further perform the following steps:

[0118] Obtain investment preference information of the user.

[0119] In the embodiment, the investment preference information (such as risk tolerance, investment goal, and investment period) of the user is collected through a questionnaire or an interactive interface.

[0120] Generate an investment suggestion result corresponding to the user based on the investment preference information and the risk and return prediction result.

[0121] In the embodiment, the investment suggestion result can be generated according to the obtained investment preference information of the user and the generated risk and return prediction result. Specifically, an investment portfolio matched with the user is selected according to the risk tolerance, investment goal, and investment period of the user. For example, for a conservative user, a low-risk and low-yield asset or a conservative investment portfolio is recommended, and capital preservation and stability are emphasized. For a moderate user, a moderate-risk and moderate-yield asset or a moderate investment portfolio is recommended, and the balance between yield and risk is emphasized. For an aggressive user, a high-risk and high-yield asset or an aggressive investment portfolio is recommended, and higher yield is pursued while the corresponding risk is borne. In addition, the content can be dynamically adjusted and suggested according to the changes in the market environment and the real-time performance of the user's investment portfolio. For example, when the market fluctuates greatly, it is suggested that the conservative user increase the proportion of cash or treasury bonds.

[0122] Obtain a preset suggestion display mode.

[0123] In the embodiment, the selection of the above-mentioned suggestion display mode is not specifically limited, and can be set according to actual display requirements. For example, the display mode of displaying through a card layout can be used.

[0124] Display process the investment suggestion result based on the suggestion display mode.

[0125] In the present embodiment, the investment suggestion result can be displayed according to the selected suggestion display mode. In addition, an interactive function is provided to allow the user to click on the card to view more detailed information, such as historical performance, fee structure, etc. At the same time, the user is allowed to adjust the investment suggestion, for example, modify the asset allocation ratio or replace the portfolio.

[0126] In addition, feedback of the user on the personalized suggestion can be collected, and the suggestion generation logic is continuously optimized according to the user feedback to improve the accuracy of the suggestion and the user satisfaction. For example, it is found that conservative users tend to choose assets with stable dividends, and the allocation ratio of such assets in the conservative portfolio can be increased.

[0127] The present application obtains the investment preference information of the user; then generates the investment suggestion result corresponding to the user based on the investment preference information and the risk and return prediction result; then obtains a preset suggestion display mode; and subsequently displays the investment suggestion result based on the suggestion display mode. Based on the above processing flow, the present application can automatically and accurately provide personalized investment suggestion results for the user according to the investment preference information and the risk and return prediction result, ensuring the intelligence and accuracy of the obtained investment suggestion results. Moreover, it will be displayed to the user in a clear and intuitive manner, helping the user to make a more intelligent investment decision, thereby improving the user's experience.

[0128] In some optional implementations, the obtained user information seeks the consent of the user and complies with relevant laws and relevant policies.

[0129] In addition, the non-company software tools or components appearing in the present application embodiment are only examples for introduction and do not represent actual use.

[0130] It should be understood that the size of the serial number of each step in the above embodiments does not mean the order of execution, and the execution order of each process should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0131] It should be emphasized that, in order to further ensure the privacy and security of the above risk and return prediction result, the above risk and return prediction result can also be stored in a node of a block chain.

[0132] The blockchain referred to in the present application is a new application mode of distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm and other computer technologies. The blockchain is essentially a decentralized database, which is a series of data blocks associated using cryptographic methods, each data block containing information of a batch of network transactions, used to verify the validity (anti-fake) of the information and generate the next block. The blockchain can include a blockchain underlying platform, a platform product service layer, and an application service layer.

[0133] The embodiments of the present application can acquire and process related data based on artificial intelligence technology. Artificial intelligence (AI) is the use of digital computers or computer-controlled machines to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0134] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by computer readable instructions instructing related hardware, and the computer readable instructions can be stored in a computer readable storage medium. The program can include the processes of the above-mentioned embodiments when executed.

[0135] It should be understood that although each step in the flowchart of the accompanying drawings is displayed in sequence according to the direction of the arrow, these steps are not necessarily executed in sequence according to the direction of the arrow. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and they can be executed in other orders. Moreover, at least part of the steps in the flowchart of the accompanying drawings can include multiple sub-steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence is not necessarily sequential, but can be executed alternately or alternately with at least part of other steps or sub-steps or stages of other steps.

[0136] Further referring to Figure 3 , as an implementation of the method shown in Figure 2 , the present application provides an embodiment of an asset analysis device based on artificial intelligence. The device embodiment corresponds to the method embodiment shown in Figure 2 , and the device can be specifically applied to various electronic devices.

[0137] As Figure 3As shown, the asset analysis device based on artificial intelligence 300 described in this embodiment comprises a receiving module 301, a first obtaining module 302, a first processing module 303, a calling module 304, a prediction module 305 and a first display module 306. Wherein:

[0138] The receiving module 301 is configured to receive a user inputted income analysis request for a target asset; wherein the income analysis request carries an analysis time period;

[0139] The first obtaining module 302 is configured to obtain asset related data corresponding to the target asset based on a preset data type, and obtain factor data associated with the analysis time period;

[0140] The first processing module 303 is configured to integrate the asset related data and the factor data to obtain integrated data, and pre-process the integrated data to obtain corresponding target data;

[0141] The calling module 304 is configured to call a preset risk income prediction model; wherein the risk income prediction model is a model generated by training a preset deep learning model based on asset sample data obtained through feature enhancement processing;

[0142] The prediction module 305 is configured to perform income prediction processing on the target data based on the risk income prediction model, to obtain a risk income prediction result corresponding to the analysis time period;

[0143] The first display module 306 is configured to perform display processing on the risk income prediction result.

[0144] In this embodiment, the above-mentioned modules or units are respectively used for performing operations corresponding to the steps of the asset analysis method based on artificial intelligence of the foregoing embodiments, which will not be repeated here.

[0145] In some optional implementation manners of this embodiment, the asset analysis device based on artificial intelligence further comprises:

[0146] The collection module is configured to collect historical related data of a plurality of assets from a preset data source;

[0147] The extraction module is configured to perform feature extraction on the historical related data based on a preset feature extraction algorithm, to obtain first feature data related to risk income;

[0148] The second obtaining module is configured to obtain external information related to assets;

[0149] The enhancement module is configured to perform enhancement processing on the first feature data based on a preset feature enhancement strategy using the external information, to obtain processed second feature data;

[0150] The first determining module is configured to take the second feature data as the asset sample data.

[0151] In some optional implementations of the embodiment, the enhancing module comprises:

[0152] The first obtaining sub-module is configured to obtain a plurality of preset fusion methods.

[0153] The screening sub-module is configured to screen a target fusion method from all the fusion methods.

[0154] The fusion sub-module is configured to perform fusion processing on the external information and the first feature data based on the target fusion method, to obtain corresponding fusion data.

[0155] The verification sub-module is configured to verify the fusion effect of the fusion data.

[0156] The first determining sub-module is configured to take the fusion data as the second feature data if the fusion effect of the fusion data passes the verification.

[0157] In some optional implementations of the embodiment, the asset analysis device based on artificial intelligence further comprises:

[0158] The second determining module is configured to determine a corresponding deep learning model based on a preset asset analysis requirement.

[0159] The dividing module is configured to divide the asset sample data into training data and verification data.

[0160] The training module is configured to train the deep learning model based on the training data, to obtain a corresponding first processing model.

[0161] The second processing module is configured to evaluate and optimize the first processing model based on the verification data, to obtain a corresponding second processing model.

[0162] The optimization module is configured to optimize the second processing model based on a preset multi-factor analysis strategy, to obtain a corresponding third processing model.

[0163] The first generating module is configured to generate the risk and return prediction model based on the third processing model.

[0164] In some optional implementations of the embodiment, the first generating module comprises:

[0165] The optimization sub-module is configured to optimize the model parameters and analysis methods of the third processing model based on a preset reinforcement learning algorithm, to obtain a corresponding fourth processing model.

[0166] The second obtaining sub-module is configured to obtain a preset real-time updating strategy.

[0167] The updating sub-module is configured to update the fourth processing model based on the real-time updating strategy to obtain a corresponding fifth processing model.

[0168] The second determining sub-module is configured to take the fifth processing model as the risk and return prediction model.

[0169] In some optional implementation manners of the embodiment, the first processing module 303 comprises:

[0170] The first processing sub-module is configured to perform missing value processing on the integrated data to obtain corresponding first processing data.

[0171] The second processing sub-module is configured to perform abnormal value processing on the first processing data to obtain corresponding second processing data.

[0172] The third processing sub-module is configured to perform standardization processing on the second processing data based on a preset standardization strategy to obtain corresponding third processing data.

[0173] The third determining sub-module is configured to take the third processing data as the target data.

[0174] In some optional implementation manners of the embodiment, the asset analysis apparatus based on artificial intelligence further comprises:

[0175] The third obtaining module is configured to obtain investment preference information of the user.

[0176] The second generating module is configured to generate an investment suggestion result corresponding to the user based on the investment preference information and the risk and return prediction result.

[0177] The fourth obtaining module is configured to obtain a preset suggestion display mode.

[0178] The second display module is configured to perform display processing on the investment suggestion result based on the suggestion display mode.

[0179] To solve the above technical problems, the embodiment of the present application further provides a computer device. For details, please refer to Figure 4 , Figure 4 The basic structure block diagram of the computer device of the embodiment is shown in FIG. 1.

[0180] The computer device 4 includes a memory 41, a processor 42, and a network interface 43, which are communicatively connected by a system bus. It should be noted that only the computer device 4 with components 41-43 is shown in the figure, but it should be understood that not all of the shown components are required to be implemented, and more or fewer components can be alternatively implemented. Among them, those skilled in the art can understand that the computer device herein is a device capable of automatically performing numerical calculation and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc.

[0181] The computer device can be a desktop computer, a notebook computer, a palm computer, a cloud server, and the like. The computer device can interact with the user through a keyboard, a mouse, a remote controller, a touchpad, a voice control device, and the like.

[0182] The memory 41 includes at least one type of readable storage medium, including a flash memory, a hard disk, a multimedia card, a card-type memory (e.g., an SD or DX memory, etc.), a random access memory (RAM), a static random access memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, an optical disk, and the like. In some embodiments, the memory 41 can be an internal storage unit of the computer device 4, such as a hard disk or a memory of the computer device 4. In other embodiments, the memory 41 can also be an external storage device of the computer device 4, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, and the like. Of course, the memory 41 can also include both the internal storage unit and the external storage device of the computer device 4. In the present embodiment, the memory 41 is generally used to store an operating system and various application software installed in the computer device 4, such as computer readable instructions of the asset analysis method based on artificial intelligence, and the like. In addition, the memory 41 can also be used to temporarily store various data that have been output or will be output.

[0183] The processor 42 may, in some embodiments, be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip. The processor 42 is generally used to control the overall operation of the computer device 4. In the present embodiment, the processor 42 is configured to execute computer-readable instructions stored in the memory 41 or to process data, such as computer-readable instructions of the artificial intelligence-based asset analysis method.

[0184] The network interface 43 may include a wireless network interface or a wired network interface, and is generally used to establish a communication connection between the computer device 4 and other electronic devices.

[0185] The present application also provides another embodiment, i.e., a computer-readable storage medium storing computer-readable instructions executable by at least one processor to cause the at least one processor to perform the steps of the artificial intelligence-based asset analysis method as described above.

[0186] From the above description of the embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software and a general hardware platform as required, and of course, can also be implemented by hardware, but in many cases, the former is a better embodiment. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, which is stored in a storage medium (such as a ROM / RAM, a magnetic disk, or an optical disk) and includes a number of instructions for causing a terminal device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device) to perform the methods described in the various embodiments of the present application.

[0187] Obviously, the above-described embodiments are only some of the embodiments of the present application, rather than all the embodiments, and the preferred embodiments of the present application are given in the drawings, but do not limit the patent scope of the present application. The present application can be implemented in many different forms, and conversely, the purpose of providing these embodiments is to make the disclosure of the present application more thorough and comprehensive. Although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing specific embodiments, or make equivalent replacements to some technical features. Any equivalent structure made by referring to the contents of the specification and drawings, directly or indirectly applied to other related technical fields, is also within the scope of the patent protection of the present application.

Claims

1. An artificial intelligence-based asset analysis method, characterized by, The method comprises the following steps: receiving a user inputted return analysis request for a target asset; wherein the return analysis request carries an analysis time period; acquiring asset-related data corresponding to the target asset based on a preset data type, and acquiring factor data associated with the analysis time period; integrating the asset-related data and the factor data to obtain integrated data, and preprocessing the integrated data to obtain corresponding target data; calling a preset risk-return prediction model; wherein the risk-return prediction model is a model generated by training a preset deep learning model based on asset sample data obtained through feature enhancement processing; performing return prediction processing on the target data based on the risk-return prediction model to obtain a risk-return prediction result corresponding to the analysis time period; performing display processing on the risk-return prediction result. 2.The artificial intelligence-based asset analysis method of claim 1, wherein, Before the step of calling the preset risk-return prediction model, the method further comprises: collecting historical related data of multiple assets from a preset data source; performing feature extraction on the historical related data based on a preset feature extraction algorithm to obtain first feature data related to risk-return; acquiring external information related to assets; based on the external information, performing enhancement processing on the first feature data using a preset feature enhancement strategy to obtain processed second feature data; using the second feature data as the asset sample data. 3.The artificial intelligence-based asset analysis method of claim 2, wherein, The step of performing enhancement processing on the first feature data using a preset feature enhancement strategy based on the external information to obtain processed second feature data specifically comprises: acquiring a plurality of preset fusion methods; selecting a target fusion method from all the fusion methods; performing fusion processing on the external information and the first feature data based on the target fusion method to obtain corresponding fusion data; verifying the fusion effect of the fusion data; if the fusion effect of the fusion data passes the verification, using the fusion data as the second feature data. 4.The artificial intelligence-based asset analysis method of claim 1, wherein, Before the step of calling the preset risk-return prediction model, the method further comprises: determining a corresponding deep learning model based on a preset asset analysis requirement; dividing the asset sample data into training data and verification data; training the deep learning model based on the training data to obtain a corresponding first processing model; evaluating and optimizing the first processing model based on the verification data to obtain a corresponding second processing model; performing model optimization on the second processing model based on a preset multi-factor analysis strategy to obtain a corresponding third processing model; generating the risk-return prediction model based on the third processing model. 5.The artificial intelligence-based asset analysis method of claim 4, wherein, The step of generating the risk-return prediction model based on the third processing model specifically comprises: optimizing the model parameters and analysis method of the third processing model based on a preset reinforcement learning algorithm to obtain a corresponding fourth processing model; acquiring a preset real-time update strategy; updating the fourth processing model based on the real-time update strategy to obtain a corresponding fifth processing model; using the fifth processing model as the risk-return prediction model. 6.The artificial intelligence-based asset analysis method of claim 1, wherein, The step of preprocessing the integrated data to obtain corresponding target data specifically includes: Performing missing value processing on the integrated data to obtain corresponding first processing data; Performing abnormal value processing on the first processing data to obtain corresponding second processing data; Performing standardized processing on the second processing data based on a preset standardized strategy to obtain corresponding third processing data; Taking the third processing data as the target data. 7.The artificial intelligence-based asset analysis method of claim 1, wherein, After the step of performing yield prediction processing on the target data based on the risk yield prediction model to obtain a risk yield prediction result corresponding to the analysis time period, the method further includes: Obtaining investment preference information of the user; Based on the investment preference information and the risk yield prediction result, generating an investment suggestion result corresponding to the user; Obtaining a preset suggestion display mode; Based on the suggestion display mode, performing display processing on the investment suggestion result.

8. An artificial intelligence-based asset analysis device, characterized by comprising: The method includes: A receiving module configured to receive a yield analysis request for a target asset input by a user, wherein the yield analysis request carries an analysis time period; A first obtaining module configured to obtain asset-related data corresponding to the target asset based on a preset data type, and obtain factor data associated with the analysis time period; A first processing module configured to integrate the asset-related data and the factor data to obtain integrated data, and preprocess the integrated data to obtain corresponding target data; A calling module configured to call a preset risk yield prediction model, wherein the risk yield prediction model is a model generated by training a preset deep learning model based on asset sample data obtained through feature enhancement processing; A prediction module configured to perform yield prediction processing on the target data based on the risk yield prediction model to obtain a risk yield prediction result corresponding to the analysis time period; A first display module configured to perform display processing on the risk yield prediction result.

9. A computer device, comprising: The method includes a memory and a processor, the memory stores computer readable instructions, and the processor executes the computer readable instructions to realize the steps of the asset analysis method based on artificial intelligence according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer readable instructions, and the computer readable instructions are executed by the processor to realize the steps of the asset analysis method based on artificial intelligence according to any one of claims 1 to 7.

Images