Dynamic asset allocation intelligent investment advisor system based on artificial intelligence

Abstract

The invention discloses a dynamic asset allocation intelligent investment advisor system based on artificial intelligence, particularly relates to the field of intelligent investment, and comprises a data acquisition module, a data processing module, a data analysis module and a judgment module. The data acquisition module acquires financial market data, macroeconomic data, enterprise financial data and investor personal data and transmits the data to the data processing module. The data processing module carries out cleaning and standardization operation on the data and then transmits the data to the data analysis module. And the data analysis module constructs a mean variance model based on the processed data to obtain an initial optimal portfolio weight, and then calculates and optimizes the risk contribution of each asset. And the judgment module generates asset allocation suggestions based on the analysis result and dynamically adjusts the suggestions. According to the system, personalized suggestions are generated according to investor risk preferences and investment targets through multi-module cooperation in combination with a mean variance model and a risk evaluation model, the situation changes of markets and investors can be responded in time, and the accuracy and adaptability of asset allocation are improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent investment technology, and more specifically, to an artificial intelligence-based dynamic asset allocation intelligent investment advisory system. Background Art

[0002] With the continuous development and innovation of the financial market, investment products are becoming increasingly diverse, and the market environment is becoming increasingly complex and volatile. Investors are increasingly demanding professionalism and timeliness in asset allocation. The traditional manual investment advisory model, with its high service costs, limited coverage, and slow response times, is no longer able to meet the needs of the majority of investors.

[0003] Against this backdrop, robo-advisory systems have emerged. Leveraging information technology and financial theory, robo-advisory systems provide investors with automated, personalized asset allocation advice, gradually becoming a key development in the financial technology sector. Currently, a variety of robo-advisory systems are available on the market. These systems collect market data and investor information, apply relevant models for analysis, and generate asset allocation plans for investors.

[0004] However, in actual use, it still has some shortcomings. For example, existing smart investment advisory systems often have problems with single data sources and incomplete data types in terms of data collection. Some systems only collect a small amount of financial market data, and the collection of macroeconomic data and corporate financial data is not comprehensive. The data is updated infrequently, and cannot reflect market dynamics in a timely manner. As a result, the asset allocation recommendations generated based on this data are inaccurate and ineffective. In the data processing stage, existing systems do not pay enough attention to data cleaning and standardization, and the processing methods are relatively simple. There is a lack of scientific and reasonable means to deal with missing values ​​and outliers, which may lead to data distortion. The lack of standardization of data makes the comparability between different types of data poor, affecting the effectiveness of subsequent model analysis. Many smart investment advisory systems only use a single mean-variance model for asset allocation analysis. Although this model considers the balance between risk and return, it fails to fully consider the balance of risk contributions between assets. When the market fluctuates greatly, it is easy for the risk of a single asset to be over-concentrated, making it difficult to effectively control the overall risk of the investment portfolio. The asset allocation recommendations generated by existing systems often lack a deep match with investors' risk preferences and investment goals, and cannot provide personalized solutions based on the specific needs of different types of investors, resulting in low practicality and operability of the recommendations. The dynamic adjustment mechanism of some smart investment advisory systems is not flexible and timely enough, and the adjustment trigger conditions are set unreasonably or single, which cannot quickly respond to major changes in the market environment and changes in investors' personal circumstances, making it difficult for asset allocation plans to adapt to changes in actual conditions. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a dynamic asset allocation smart investment advisory system based on artificial intelligence, which solves the problems raised in the above-mentioned background technology through the following scheme.

[0006] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: an artificial intelligence-based dynamic asset allocation smart investment advisory system, comprising a data acquisition module, a data processing module, a data analysis module, and a judgment module; The data acquisition module is used to collect financial market data, macroeconomic data, corporate financial data and investor personal data, and transmit the data to the data processing module; The data processing module is used to perform data cleaning and data standardization operations on the collected data, and transmit the processed data to the data analysis module; The data analysis module is used to construct a mean-variance model based on the processed data to calculate the initial optimal investment portfolio weights, and calculate the risk contribution of each asset based on the initial weights; The judgment module is used to generate asset allocation suggestions based on the analysis results and make dynamic adjustments.

[0007] Preferably, the method for collecting financial market data is as follows: by signing a data purchase agreement with a professional financial data provider and using the API interface provided by it, real-time stock price data of major global stock markets is obtained, including opening price, closing price, highest price, lowest price, trading volume, and trading amount. At the same time, yield data of various types of bonds in the bond market and net asset value data of different types of funds are obtained. The data update frequency is set to once per minute to ensure the timeliness of the data; The macroeconomic data collection method is as follows: GDP growth rate, inflation rate, and interest rate data are regularly obtained from official government websites, websites of international organizations, and official central bank websites. The GDP growth rate and CPI data for the previous month are updated at the beginning of each month. Interest rate data is updated in real time when central banks make adjustments or when important market nodes change; The method for collecting the company's financial data is as follows: Relying on the Wind database, obtain the financial data from the annual, semi-annual, and quarterly reports regularly released by listed companies, covering revenue, profit, balance sheet, and cash flow statement, and complete the data update within one business day after the listed company's financial report is released; The method for collecting the investor's personal data is as follows: when investors register to use the intelligent investment advisory system, they need to fill out a detailed questionnaire, which includes asset size, risk preference, and investment goals. When investors actively modify their personal information or the regular return visit time set by the system arrives, the investor's personal data will be updated.

[0008] Preferably, the data cleaning operation is as follows: use Python's pandas library to write a data cleaning program. For the collected data, first use the drop_duplicates() method to remove duplicate data. For missing values, if it is numerical data, use the mean filling method, use the fillna() method, and fill it with the mean of the numerical data. If it is categorical data, use mode filling, that is, count the categories with the highest frequency of all categorical data to fill it. For outliers, calculate the Z-score to identify them. For stock price and yield data, if the absolute value of the Z-score of a data point is greater than 3, it is determined to be an outlier and replaced with the median of the stock price and yield data.

[0009] Preferably, the data standardization operation is as follows: for numerical data, the pandas library is combined with the NumPy library, and the Z-score standardization method is adopted. First, the mean and standard deviation of each column of data are calculated, and then each data point is transformed as follows: , where x is the original data point, μ is the mean of the column data, and σ is the standard deviation. The data is converted to a standard normal distribution with a mean of 0 and a standard deviation of 1. For non-numeric data, the OneHotEncoder in the scikit-learn library is used for one-hot encoding conversion to convert each category into a binary vector.

[0010] Preferably, the mean-variance model calculation process is as follows: Data input: The stock prices, bond yields, and fund net assets obtained after data processing are used as the basic data for calculating asset returns, and historical data from the past five years are selected; Calculation of expected rate of return: For each asset i, use the arithmetic mean method to calculate the expected rate of return E(R i ), the formula is , where R it is the rate of return of asset i in period t, and n is the number of data periods. Through this calculation, the expected level of return of each asset is clarified; Covariance matrix construction: Use Python's NumPy library to calculate the covariance Cov between assets (R i ,R j ), the elements of the covariance matrix Σ are: , where i,j=1,2, ,n, n is the number of asset types; Optimal weight solution: In the mean-variance model, the expected return E(Rp) and risk (variance) of the portfolio They are: ; , where wi is the weight of asset i in the portfolio, and , using the minimize function in Python's SciPy library, taking the maximum volatility that investors can bear as the risk constraint, maximize the expected return of the portfolio and solve for the initial optimal portfolio weights.

[0011] Preferably, the risk contribution is calculated as follows: Based on the initial weights obtained from the mean-variance model, the risk contribution of each asset is calculated. The risk contribution RCi of asset i is calculated as follows: ,in , the calculation of the above formula is realized through Python programming, the contribution of each asset to the total risk of the investment portfolio is clarified, and a basis for risk optimization is provided. The iterative program is written in Python for risk assessment adjustment, and the average risk contribution is set to 1 / n of the total risk, where n is the number of asset types. Starting from the initial weight, in each iteration, the difference between the risk contribution of each asset and the average risk contribution is calculated. If the risk contribution of an asset is higher than the average risk contribution, its weight is reduced according to the difference ratio. Otherwise, its weight is increased according to the difference ratio. The risk contribution is recalculated after each adjustment, and the iterative process is repeated until the absolute value of the difference between the risk contribution of each asset and the average risk contribution is less than the set accuracy threshold, and the optimized asset allocation weight is obtained.

[0012] Preferably, the method for generating the asset allocation recommendation is as follows: the analysis results of the mean-variance model and the risk parity model are deeply matched with the investor's risk preference and investment objectives. For conservative investors, on the premise of ensuring that the annualized rate of return is not lower than the rate of return of treasury bonds in the same period, bonds and money market funds are given priority, and their weight in the investment portfolio is not less than 70%. For aggressive investors, within the range of an acceptable annualized volatility of no more than 30%, an annualized rate of return of more than 15% is pursued, and the weight of stock funds and stocks in the investment portfolio can reach more than 70%. Through the configuration algorithm of the system background, personalized asset allocation recommendations are generated that include specific investment proportions and targets of various assets.

[0013] Preferably, the dynamic adjustment is as follows: the system sets up a market data monitoring module and an investor personal data monitoring module. The market data monitoring module tracks the rise and fall of the stock market index in real time. If the rise and fall in a single day exceeds 5%, or the cumulative rise and fall in 10 trading days exceeds 10%, as well as interest rate data, if the benchmark interest rate is adjusted by more than 0.5 percentage points in a single time, it triggers a mechanism for judging major changes in the market environment. The investor personal data monitoring module triggers a mechanism for judging changes in the investor's personal situation when the investor modifies the investment target or risk preference. Once the above mechanism is triggered, the system automatically re-runs the mean-variance model and the risk parity model, and dynamically adjusts the asset allocation recommendations. The adjusted recommendations will be pushed to investors within 24 hours, and the reasons for the adjustment and a market analysis report will be attached to help investors understand the relationship between market changes and investment adjustments.

[0014] Technical effects and advantages of the present invention: 1. The data acquisition module of the present invention can collect multiple types of data, including financial market data, macroeconomic data, corporate financial data, and investor personal data. It also clarifies the specific collection sources and update frequencies of each type of data. For example, financial market data is updated every minute, macroeconomic data is updated monthly or in real time, and corporate financial data is updated within one business day after the financial report is released. This ensures the comprehensiveness and timeliness of the data, providing a reliable data foundation for subsequent analysis. 2. The judgment module of the present invention deeply matches the analysis results with the investor's risk preferences and investment goals. For conservative investors, low-risk assets are prioritized with a weight of no less than 70%. For aggressive investors, the weight of equity assets is increased to more than 70% within the acceptable risk range. Personalized recommendations are generated, including specific investment ratios and targets, to meet the needs of different investors. 3. The data analysis module of the present invention integrates the mean-variance model and the risk parity model. The mean-variance model is used to obtain the initial optimal portfolio weights. The risk contribution of each asset is then calculated based on the initial weights. Risk parity is then achieved through iterative adjustments, ensuring a balanced distribution of risk across different assets. This multi-model integration overcomes the limitations of a single model, optimizing the risk-return characteristics of the portfolio while effectively controlling overall risk. 4. The present invention sets up a market data monitoring module and an investor personal data monitoring module, and clarifies the specific conditions for triggering dynamic adjustments. When the stock market index fluctuation or interest rate adjustment reaches the set threshold, or the investor's personal situation changes, the system automatically re-runs the model and pushes the adjusted recommendations within 24 hours, ensuring that the asset allocation plan can adapt to market changes and changes in investor needs in a timely manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1It is a schematic diagram of the overall structure of the present invention. DETAILED DESCRIPTION

[0016] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 making creative efforts are within the scope of protection of the present invention.

[0017] As attached Figure 1 The illustrated system is a dynamic asset allocation smart investment advisory system based on artificial intelligence, comprising a data acquisition module, a data processing module, a data analysis module, and a judgment module; The data acquisition module is used to collect financial market data, macroeconomic data, corporate financial data and investor personal data, and transmit the data to the data processing module; The financial market data collection method is as follows: by signing data purchase agreements with professional financial data providers such as Bloomberg and Reuters, and utilizing the API interfaces provided by them, real-time stock price data from major global stock markets (such as the New York Stock Exchange, Nasdaq, Shanghai Stock Exchange, Shenzhen Stock Exchange, etc.) is obtained, including opening price, closing price, highest price, lowest price, trading volume, and transaction amount. The yield data of various types of bonds (government bonds, corporate bonds, financial bonds, etc.) in the bond market, as well as the net asset value data of different types of funds (stock funds, bond funds, hybrid funds, and money market funds), are also obtained. The data update frequency is set to once per minute to ensure the timeliness of the data. The macroeconomic data collection method is as follows: regularly obtain data such as GDP growth rate, inflation rate (Consumer Price Index (CPI), and interest rates (benchmark interest rate, interest rates of various maturities on the Treasury yield curve) from official government websites (such as the U.S. Bureau of Economic Analysis and the National Bureau of Statistics of China), websites of international organizations (such as the World Bank and the International Monetary Fund), and official websites of central banks (such as the Federal Reserve and the People's Bank of China). The GDP growth rate and CPI data for the previous month are updated at the beginning of each month, and interest rate data are updated in real time when central bank adjustments are made or when important market nodes change. The method for collecting the company's financial data is as follows: Relying on the Wind database, obtain financial data from the annual, semi-annual, and quarterly reports regularly released by listed companies, covering key indicators such as revenue, profit (net profit, operating profit), balance sheet (total assets, total liabilities, shareholders' equity), and cash flow statement (cash flow from operating activities, cash flow from investing activities, and cash flow from financing activities). Data updates are completed within one business day after the listed company's financial report is released; The method for collecting the personal data of investors is as follows: When investors register to use the intelligent investment advisory system, they need to fill out a detailed questionnaire, which includes the asset size (accurate to 10,000 yuan), risk preference (quantified as conservative, stable, balanced, growth, and aggressive through a risk assessment questionnaire), and investment goals (short-term speculation, long-term investment appreciation, asset preservation, children's education reserves, retirement reserves, etc.). When the investor actively modifies his or her personal information or the regular return visit time set by the system (once every six months) arrives, the investor's personal data will be updated.

[0018] The data processing module is used to perform data cleaning and data standardization operations on the collected data, and transmit the processed data to the data analysis module; The data cleaning operation is as follows: Use Python's pandas library to write a data cleaning program. For the collected data, first use the drop_duplicates() method to remove duplicate data. For missing values, if it is numerical data (such as stock prices, financial indicators, etc.), use the mean filling method and fillna() method to fill with the mean of the numerical data; if it is categorical data (such as risk preference type, investment target type), use mode filling, that is, count all categorical data with the highest frequency of filling. For outliers, calculate the Z-score to identify them. For data such as stock prices and yields, if the absolute value of the Z-score of a data point is greater than 3, it is determined to be an outlier and replaced with the median of the stock price and yield data to ensure the stability and reliability of the data. The data standardization operation is as follows: For numerical data, the pandas library is combined with the NumPy library, and the Z-score standardization method is adopted. First, the mean and standard deviation of each column of data are calculated, and then each data point is transformed as follows: , where x is the original data point, μ is the mean of the column data, and σ is the standard deviation. The data is converted to a standard normal distribution with a mean of 0 and a standard deviation of 1. For non-numeric data, such as industry classification and enterprise nature, the OneHotEncoder in the scikit-learn library is used for one-hot encoding conversion, converting each category into a binary vector for subsequent model processing.

[0019] The data analysis module is used to construct a mean-variance model based on the processed data to calculate the initial optimal investment portfolio weights, and calculate the risk contribution of each asset based on the initial weights; The mean-variance model calculation process is as follows: Data input: Financial market data such as stock prices, bond yields, and fund net assets obtained after data processing are used as the basic data for calculating asset returns. Historical data from the past five years is selected (if the asset has been listed for less than five years, its entire historical data is used); Calculation of expected rate of return: For each asset i, use the arithmetic mean method to calculate the expected rate of return E(R i ), the formula is , where R it is the rate of return of asset i in period t, and n is the number of data periods. Through this calculation, the expected level of return of each asset is clarified; Covariance matrix construction: Use Python's NumPy library to calculate the covariance Cov between assets (R i ,R j ), the elements of the covariance matrix Σ are: , where i,j=1,2, ,n, n is the number of asset types. The covariance matrix reflects the linkage relationship between the returns of different assets and is an important basis for measuring the risk of asset portfolios; Optimal weight solution: In the mean-variance model, the expected return E(Rp) and risk (variance) of the portfolio They are: ; , where wi is the weight of asset i in the portfolio, and Using the minimize function in Python's SciPy library, we use the investor's maximum tolerable volatility as a risk constraint to maximize the expected return of the portfolio and obtain the initial optimal portfolio weights. This process transforms the investor's risk preference into a specific asset allocation ratio constraint, achieving a preliminary balance between risk and return. The risk contribution is calculated as follows: Based on the initial weights obtained from the mean-variance model, the risk contribution of each asset is calculated. The risk contribution RCi of asset i is calculated as follows: ,in , Python programming is used to calculate the above formula, clarify the contribution of each asset to the total risk of the investment portfolio, provide a basis for risk optimization, and use Python to write an iterative program to perform risk assessment adjustments. The average risk contribution is set to 1 / n of the total risk (n is the number of asset types). Starting from the initial weight, in each iteration, the difference between the risk contribution of each asset and the average risk contribution is calculated. If the risk contribution of an asset is higher than the average risk contribution, its weight is reduced according to the difference ratio. Otherwise, its weight is increased according to the difference ratio. The risk contribution is recalculated after each adjustment, and the iterative process is repeated until the absolute value of the difference between the risk contribution of each asset and the average risk contribution is less than 0.01 (the set accuracy threshold). The optimized asset allocation weights are obtained. This process achieves a balanced distribution of risk among different assets through refined iterative adjustments, and further optimizes the risk structure of the investment portfolio.

[0020] The judgment module is used to generate asset allocation suggestions based on the analysis results and make dynamic adjustments; The asset allocation recommendation is generated as follows: the analysis results of the mean-variance model and the risk parity model are deeply matched with the investor's risk preference and investment objectives. For conservative investors, under the premise of ensuring that the annualized rate of return is not lower than the rate of return of government bonds during the same period, priority is given to low-risk assets such as bonds and money market funds, and their weight in the investment portfolio is not less than 70%. For aggressive investors, within the range of an annualized volatility that can be tolerated not exceeding 30%, an annualized rate of return of more than 15% is sought, and the weight of equity assets such as stock funds and stocks in the investment portfolio can reach more than 70%. Through the configuration algorithm in the background of the system, a personalized asset allocation recommendation is generated, which includes the specific investment proportions and targets of each asset type. The dynamic adjustment is as follows: the system sets up a market data monitoring module and an investor personal data monitoring module. The market data monitoring module tracks the rise and fall of the stock market index in real time. If the single-day rise or fall exceeds 5%, or the cumulative rise or fall within 10 trading days exceeds 10%, as well as interest rate data, if the single adjustment of the benchmark interest rate exceeds 0.5 percentage points, it triggers a mechanism for judging major changes in the market environment. The investor personal data monitoring module triggers a mechanism for judging changes in investor personal circumstances when investors modify their investment goals (from long-term investment appreciation to short-term speculation) or risk preferences (from conservative to growth). Once the above mechanism is triggered, the system automatically re-runs the mean-variance model and risk parity model, dynamically adjusts the asset allocation recommendations, and pushes the adjusted recommendations to investors within 24 hours, with the reasons for the adjustment and a market analysis report attached, to help investors understand the relationship between market changes and investment adjustments.

[0021] Secondly: The drawings of the embodiments disclosed in the present invention only involve structures related to the embodiments disclosed in the present invention. Other structures may refer to conventional designs. The same embodiment and different embodiments of the present invention may be combined with each other without conflict. Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An artificial intelligence-based dynamic asset allocation smart investment advisory system, characterized by: include: Data acquisition module, data processing module, data analysis module and judgment module; The data acquisition module is used to collect financial market data, macroeconomic data, corporate financial data and investor personal data, and transmit the data to the data processing module; The data processing module is used to perform data cleaning and data standardization operations on the collected data, and transmit the processed data to the data analysis module; The data analysis module is used to construct a mean-variance model based on the processed data to calculate the initial optimal investment portfolio weights, and calculate the risk contribution of each asset based on the initial weights; The judgment module is used to generate asset allocation suggestions based on the analysis results and make dynamic adjustments.

2. The AI-based dynamic asset allocation smart investment advisory system according to claim 1, characterized in that: The financial market data collection method is as follows: by signing a data purchase agreement with a professional financial data provider and using the API interface provided by the provider, real-time stock price data of major global stock markets is obtained, including opening price, closing price, highest price, lowest price, trading volume, and transaction amount. At the same time, the yield data of various types of bonds in the bond market and the net asset value data of different types of funds are obtained. The data update frequency is set to once per minute to ensure the timeliness of the data; The macroeconomic data collection method is as follows: GDP growth rate, inflation rate, and interest rate data are regularly obtained from official government websites, websites of international organizations, and official central bank websites. The GDP growth rate and CPI data for the previous month are updated at the beginning of each month. Interest rate data is updated in real time when central banks make adjustments or when important market nodes change; The method for collecting the company's financial data is as follows: Relying on the Wind database, obtain the financial data from the annual, semi-annual, and quarterly reports regularly released by listed companies, covering revenue, profit, balance sheet, and cash flow statement, and complete the data update within one business day after the listed company's financial report is released; The method for collecting the investor's personal data is as follows: when registering to use the intelligent investment advisory system, investors need to fill out a detailed questionnaire, which includes asset size, risk preference, and investment goals. When the investor actively modifies personal information or the regular return visit time set by the system arrives, the investor's personal data will be updated.

3. The AI-based dynamic asset allocation smart investment advisory system according to claim 1, characterized in that: The data cleaning operation is as follows: use Python's pandas library to write a data cleaning program. For the collected data, first use the drop_duplicates() method to remove duplicate data. For missing values, if it is numerical data, use the mean filling method, and use the fillna() method to fill it with the mean of the numerical data. If it is categorical data, use the mode filling method, that is, count the categories with the highest frequency in all categorical data to fill it. For outliers, they are identified by calculating the Z-score. For stock price and yield data, if the absolute value of the Z-score of a data point is greater than 3, it is determined to be an outlier and replaced with the median of all stock price and yield data.

4. The AI-based dynamic asset allocation smart investment advisory system according to claim 1, characterized in that: The data standardization operation is as follows: For numerical data, the pandas library is combined with the NumPy library, and the Z-score standardization method is adopted. First, the mean and standard deviation of each column of data are calculated, and then each data point is transformed as follows: , where x is the original data point, μ is the mean of the column data, and σ is the standard deviation. The data is converted to a standard normal distribution with a mean of 0 and a standard deviation of 1. For non-numeric data, the OneHotEncoder in the scikit-learn library is used for one-hot encoding conversion to convert each category into a binary vector.

5. The AI-based dynamic asset allocation smart investment advisory system according to claim 1, characterized in that: The mean-variance model calculation process is as follows: Data input: The stock prices, bond yields, and fund net assets obtained after data processing are used as the basic data for calculating asset returns, and historical data from the past five years are selected; Calculation of expected rate of return: For each asset i, use the arithmetic mean method to calculate the expected rate of return E(R i ), the formula is , where R it is the rate of return of asset i in period t, and n is the number of data periods. Through this calculation, the expected level of return of each asset is clarified; Covariance matrix construction: Use Python's NumPy library to calculate the covariance Cov between assets (R i ,R j ), the elements of the covariance matrix Σ are: , where i,j=1,2, ,n, n is the number of asset types; Optimal weight solution: In the mean-variance model, the expected return E(Rp) and risk (variance) of the portfolio They are: ； , where wi is the weight of asset i in the portfolio, and , using the minimize function in Python's SciPy library, taking the maximum volatility that investors can bear as the risk constraint, maximize the expected return of the portfolio and solve for the initial optimal portfolio weights.

6. The AI-based dynamic asset allocation smart investment advisory system according to claim 1, characterized in that: The risk contribution is calculated as follows: Based on the initial weights obtained from the mean-variance model, the risk contribution of each asset is calculated. The risk contribution RCi of asset i is calculated as follows: ,in , the calculation of the above formula is realized through Python programming, the contribution of each asset to the total risk of the investment portfolio is clarified, and a basis for risk optimization is provided. The iterative program is written in Python for risk assessment adjustment, and the average risk contribution is set to 1 / n of the total risk, where n is the number of asset types. Starting from the initial weight, in each iteration, the difference between the risk contribution of each asset and the average risk contribution is calculated. If the risk contribution of an asset is higher than the average risk contribution, its weight is reduced according to the difference ratio. Otherwise, its weight is increased according to the difference ratio. The risk contribution is recalculated after each adjustment, and the iterative process is repeated until the absolute value of the difference between the risk contribution of each asset and the average risk contribution is less than the set accuracy threshold, and the optimized asset allocation weight is obtained.

7. The AI-based dynamic asset allocation smart investment advisory system according to claim 1, characterized in that: The method for generating the asset allocation recommendation is as follows: the analysis results of the mean-variance model and the risk parity model are deeply matched with the investor's risk preference and investment objectives. For conservative investors, on the premise of ensuring that the annualized rate of return is not lower than the rate of return of treasury bonds in the same period, bonds and money market funds are given priority, and their weight in the investment portfolio is not less than 70%. For aggressive investors, within the range of an acceptable annualized volatility of no more than 30%, an annualized rate of return of more than 15% is pursued, and the weight of stock funds and stocks in the investment portfolio can reach more than 70%. Through the configuration algorithm in the background of the system, personalized asset allocation recommendations are generated that include the specific investment proportions and targets of various assets.

8. The artificial intelligence-based dynamic asset allocation smart investment advisory system according to claim 1, characterized in that: The dynamic adjustment is as follows: the system sets up a market data monitoring module and an investor personal data monitoring module. The market data monitoring module tracks the rise and fall of the stock market index in real time. If the rise and fall in a single day exceeds 5%, or the cumulative rise and fall in 10 trading days exceeds 10%, as well as interest rate data, if the benchmark interest rate is adjusted by more than 0.5 percentage points in a single time, it triggers a mechanism for judging major changes in the market environment. The investor personal data monitoring module triggers a mechanism for judging changes in the investor's personal situation when the investor modifies the investment target or risk preference. Once the above mechanism is triggered, the system automatically re-runs the mean-variance model and the risk parity model, and dynamically adjusts the asset allocation recommendations. The adjusted recommendations will be pushed to investors within 24 hours, and the reasons for the adjustment and a market analysis report will be attached to help investors understand the relationship between market changes and investment adjustments.

Images