WO2021103572A1 - 资产投资建议信息的生成方法、装置和可读存储介质 - Google Patents

资产投资建议信息的生成方法、装置和可读存储介质 Download PDF

Info

Publication number
WO2021103572A1
WO2021103572A1 PCT/CN2020/101647 CN2020101647W WO2021103572A1 WO 2021103572 A1 WO2021103572 A1 WO 2021103572A1 CN 2020101647 W CN2020101647 W CN 2020101647W WO 2021103572 A1 WO2021103572 A1 WO 2021103572A1
Authority
WO
WIPO (PCT)
Prior art keywords
asset
investment
invested
sequence
generating
Prior art date
Application number
PCT/CN2020/101647
Other languages
English (en)
French (fr)
Inventor
林晓明
Original Assignee
华泰证券股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华泰证券股份有限公司 filed Critical 华泰证券股份有限公司
Publication of WO2021103572A1 publication Critical patent/WO2021103572A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
    • G06Q40/06—Asset management; Financial planning or analysis

Definitions

  • This application relates to the field of computer technology, and in particular to a method, device and readable storage medium for generating asset investment advice information.
  • the main purpose of this application is to provide a method, device and readable storage medium for generating asset investment advice information, aiming to solve the problem of poor reliability of capital investment methods in the prior art.
  • this application provides a method for generating asset investment advice information.
  • the method for generating asset investment advice information includes the following steps:
  • the step of generating the prompt information suggesting that the investment asset to be invested in the asset to be purchased in the predicted time period is divided into warehouses includes:
  • the number of capital splits and the number of investments of the asset to be invested in the forecast time period is equal to the ratio of the time length to the preset time period, and one investment amount in the forecast time period corresponds to one share of the investment. Sub-funds.
  • the step of determining the predicted value corresponding to the logarithmic difference data of the asset to be invested according to the merging sequence includes:
  • the predicted value corresponding to the asset to be invested is determined according to the extrapolated combination sequence and the linear regression coefficient.
  • the step of determining the predicted value corresponding to the asset to be invested according to the extrapolated combination sequence and the linear regression coefficient includes:
  • the step of determining the merging sequence corresponding to the logarithmic difference data in each preset observation period includes:
  • the step of determining the filter sequence corresponding to various types of assets in each preset observation period according to the year-on-year sequence data includes:
  • the present application also provides a device for generating asset investment advice information.
  • the device for generating asset investment advice information includes: a memory, a processor, and a memory, a processor, and a memory that is stored on the memory and can run on the processor.
  • a program for generating asset investment recommendation information which implements the steps of the method for generating asset investment recommendation information as described above when the program for generating asset investment recommendation information is executed by the processor.
  • the present application also provides a readable storage medium, the readable storage medium stores an asset investment advice information generation program, and the asset investment advice information generation program is executed by a processor to achieve the above The steps of the method for generating asset investment advice information described above.
  • the device determines the predicted time period of the asset to be invested and the time length of the predicted time period, so as to obtain the asset to be invested in the target according to the length of time.
  • the corresponding logarithmic difference data before the beginning of the investment period and determine the combined sequence of the logarithmic difference data in each preset observation period, and then determine the corresponding predicted value of the asset on the logarithmic difference data according to the combined sequence.
  • the predicted value is greater than the predicted value.
  • the device can predict the investment asset based on the logarithmic difference data of the asset to be invested to obtain the predicted value, if the predicted value is greater than the preset threshold, it indicates that the holding income of the asset to be invested after the predicted period of time is positive, thus recommending the user During the forecast period, the assets to be invested will be purchased in separate warehouses to provide users with investment advice on wealth management products with reliable and stable returns.
  • FIG. 1 is a schematic diagram of the hardware structure of the device for generating asset investment recommendation information involved in the scheme of the embodiment of the present application;
  • FIG. 2 is a schematic flowchart of a first embodiment of a method for generating investment advice information for assets in an application
  • FIG. 3 is a detailed flow diagram of determining the merge sequence corresponding to the logarithmic difference data in each preset observation period in step S10 in FIG. 2;
  • FIG. 3 is a detailed flow diagram of determining the merge sequence corresponding to the logarithmic difference data in each preset observation period in step S10 in FIG. 2;
  • FIG. 4 is a schematic diagram of the detailed flow of step S20 in FIG. 2;
  • FIG. 5 is a schematic flowchart of a second embodiment of a method for generating investment advice information of an application asset.
  • the main solution of the embodiment of the present application is to determine the predicted time period of the asset to be invested and the time length of the predicted time period, the asset to be invested is a major category of assets; and the time period of the asset to be invested is obtained according to the length of time.
  • the corresponding logarithmic difference data before the start of the target investment period and determine the merging sequence corresponding to the logarithmic difference data in each preset observation period; according to the merging sequence, determine that the asset to be invested is on the logarithmic difference data Corresponding predicted value; when the predicted value is greater than the preset threshold, a prompt message suggesting that the investment asset to be invested is purchased in the predicted time period is generated.
  • the device can predict the investment asset based on the logarithmic difference data of the asset to be invested to obtain the predicted value, if the predicted value is greater than the preset threshold, it indicates that the holding income of the asset to be invested after the predicted period of time is positive, thus recommending the user During the forecast period, the assets to be invested will be purchased in separate warehouses to provide users with investment advice on wealth management products with reliable and stable returns.
  • Fig. 1 is a schematic diagram of the hardware structure of the device for generating asset investment recommendation information involved in the solution of the embodiment of the present application.
  • the device for generating asset investment recommendation information may include: a processor 1001, such as a CPU, a communication bus 1002, and a memory 1003.
  • the communication bus 1002 is used to implement connection and communication between these components.
  • the memory 1003 may be a high-speed RAM memory, or a non-volatile memory (non-volatile memory), such as a magnetic disk memory.
  • the memory 1003 may also be a storage device independent of the aforementioned processor 1001.
  • the structure of the device for generating asset investment advice information shown in FIG. 1 does not constitute a limitation on the device for generating asset investment advice information, and may include more or less components than shown in the figure, or a combination Certain components, or different component arrangements.
  • the memory 1003 which is a computer storage medium, may include an operating system and a program for generating asset investment advice information.
  • the processor 1001 may be used to call the asset investment recommendation information generation program stored in the memory 1003, and perform the following operations:
  • the processor 1001 may call a processing program for generating asset investment recommendation information stored in the memory 1003, and also perform the following operations:
  • the processor 1001 may call a processing program for generating asset investment recommendation information stored in the memory 1003, and also perform the following operations:
  • the number of capital splits and the number of investments of the asset to be invested in the predicted time period is equal to the ratio of the time length to the preset time period, and one investment amount in the predicted time period corresponds to a piece of divided funds.
  • the processor 1001 may call a processing program for generating asset investment recommendation information stored in the memory 1003, and also perform the following operations:
  • the predicted value corresponding to the asset to be invested is determined according to the extrapolated combination sequence and the linear regression coefficient.
  • the processor 1001 may call a processing program for generating asset investment recommendation information stored in the memory 1003, and also perform the following operations:
  • the processor 1001 may call a processing program for generating asset investment recommendation information stored in the memory 1003, and also perform the following operations:
  • the processor 1001 may call a processing program for generating asset investment recommendation information stored in the memory 1003, and also perform the following operations:
  • the first embodiment of the present application provides a method for generating asset investment advice information, the method includes:
  • Step S10 determining the predicted time period of the asset to be invested and the time length of the predicted time period, the asset to be invested is a major category of assets
  • the execution subject is the device for generating asset investment recommendation information.
  • the device below refers to the device for generating asset investment recommendation information.
  • the device can be regarded as a server, and the device can be loaded into the client in the form of an APP, so that the client communicates with the device based on the APP.
  • the user can input the predicted time period to the device, and the predicted time period is the time period during which the user wants to invest in the asset.
  • the device can determine the length of time corresponding to the predicted time period.
  • Assets to be invested refer to broad categories of assets, such as commodities, stocks, etc.
  • the time length of the user’s investment plan for large-scale assets can be greater than or equal to the time length of the predicted time period.
  • the user plans to invest in commodities for 5 years, but the return of commodities in the next year can be predicted.
  • the time corresponding to 5 years is the investment time period of the investment plan, and the predicted next year is the forecast time period.
  • the preset time period can generally be 1 month to 12 months, preferably 12 months.
  • Step S20 Obtain the logarithmic difference data corresponding to the asset to be invested before the start of the target investment period according to the length of time, and determine the combined sequence of the logarithmic difference data corresponding to each preset observation period;
  • the assets to be invested may be large-scale assets such as funds, bonds, stocks, and commodities.
  • the user can open the APP loaded on the client.
  • the APP is the investment program of the wealth management product; the user can select the wealth management product that needs to be invested based on the APP. For example, if the user wants to invest in A stocks and B bonds, the user can select the interface based on the assets of the APP Select A stock and B bond on the APP, and then the APP sends the A stock and B bond selected by the user to the device, and the device uses the A stock and B bond as assets to be invested.
  • the device After determining the length of time, the device then determines the logarithmic difference data of the asset to be invested before the start of the target investment period according to the length of time.
  • the target investment period is generally the investment period corresponding to the future time, that is, the data generated before the start of the target investment period needs to be used as the logarithmic difference data.
  • the data included in the logarithmic difference data are all monthly difference data. Each monthly difference data is calculated in a set period.
  • the set period can be 1 month, 6 months, 12 months, etc.
  • the length of the set period is the same as the time length corresponding to the forecast period.
  • the ratio of the closing price of the stock at the end of September 2019 to the closing price of the stock at the end of March 2019 or the logarithm of the ratio is taken as September 2019
  • the logarithmic difference data of September 2019 is the observation time, that is, the target investment period is September 2019, the logarithmic difference data Calculated according to the following formula:
  • the device After determining the logarithmic difference data, the device determines the combined sequence of the logarithmic difference data in each preset observation period. For details, please refer to FIG. 3, that is, in step S20, it is determined that the logarithmic difference data is in each preset observation period.
  • the merge sequence corresponding to the cycle includes:
  • Step S21 Determine the filter sequence corresponding to each preset observation period of the asset to be invested according to the logarithmic difference data
  • the asset data sequence in a continuous period of time has a time dimension.
  • stock price data is a numerical sequence in continuous time
  • the resulting logarithmic difference data of the asset can be regarded as a time sequence.
  • the time series can be analogous to the time-domain signal generated by the movement of an economic and financial asset, and the time-domain signal has a Fourier series expression, so the logarithmic difference data can be Fourier transformed to obtain the corresponding Frequency domain data is analyzed and processed in the frequency domain.
  • Assets have an economic cycle in a unified financial economic system.
  • the economic cycle is the preset observation cycle.
  • the year-on-year data of assets shows regular changes within the preset observation cycle, and the preset observation cycle is 42 months and 100 months. And 200 months, the preset observation period is set in the device.
  • each preset observation period corresponds to a target frequency signal.
  • These target frequency signals are stable and continuous, while other unstable or unsustainable frequency signals can be regarded as noise. Therefore, the device performs frequency domain filtering on the logarithmic difference data of the asset to retain the target frequency signal and reduce noise interference, and the retention of the logarithmic difference data of the target frequency signal is the frequency domain data.
  • the logarithmic difference data is first filled with zeros, and the zero-filled logarithmic difference data is Fourier transformed to obtain the corresponding Frequency domain data.
  • the logarithmic difference data of the i-th asset As Represents the frequency domain data corresponding to the logarithmic difference data of the i-th asset.
  • a set of filter coefficients gauss win is determined according to each preset observation period period, and according to each set of filter coefficients gauss win and frequency domain data wave fft Get a set of intermediate sequences
  • nfft is the zero-padded length of the year-on-year data
  • the gauss index is a series of numbers from 1 to nfft
  • center frequency represents the center frequency, that is, the frequency corresponding to the period factor to be extracted
  • gauss alpha is a parameter that affects the Gaussian filter bandwidth.
  • the formula (5) is obtained through the transformation of the above formula.
  • the transformation process is: perform inverse Fourier transform on each set of intermediate sequences to obtain a set of second intermediate sequences; intercept data from the second intermediate sequences according to the preset sequence length LEN Click to get the filtering sequence of each type of asset under each preset observation period Among them, the preset sequence length LEN is equal to the sum of the logarithmic difference length L1 and the extrapolated length L2.
  • Real(Z) is the real part of Z.
  • Step S22 Combine the filter sequences corresponding to the various types of assets to obtain the combined sequence of each of the preset observation periods.
  • the device After determining the filtering sequence, the device will merge each filtering sequence to obtain a combined sequence.
  • assets are driven by the same economic cycle and exhibit extremely relevant behaviors. Therefore, other types of assets will affect the return of the assets to be invested. It is necessary to merge the filter sequences of the assets to be invested and other types of assets under the same preset observation period, that is, to merge their similar common economic cycle change characteristics.
  • the combined sequence can reflect the uniform system-level periodic movement in the market, so as to better fit the logarithmic difference data of various asset prices in subsequent processing.
  • the merging sequence of the preset observation period can be synthesized by referring to the following process:
  • Step A Obtain a first filter matrix according to the filter sequence of the various types of assets
  • various types of assets can be stocks, bonds, and commodities
  • the filter sequences of stocks, bonds, and commodities are Taking each filter sequence as a vector, the three vectors are combined to form a filter matrix M1, and M1 is the first filter matrix.
  • Step B Hilbert transform is performed on the first filter matrix to obtain a corresponding second filter matrix
  • Step C Perform iterative calculation of merging weights according to the second filter matrix.
  • the combined weight vector with the preset observation period is a vector of all 1s with length N, where N is the number of vectors in matrix M2, that is, the number of asset categories, and the value is 3 when the asset only includes stocks, bonds, and commodities.
  • weight m mean(weight) (7)
  • M*N represents the multiplication of the matrix M and N
  • (M)′ is the conjugate transpose of M
  • diag(W) contains W on the main diagonal Diagonal matrix
  • conj(M) is the complex conjugate of M
  • M.*N represents the dot product of matrix M and N
  • mean(M) is the mean value of each column of M.
  • the filter sequences of various assets can be regarded as time-domain signals, they are affected by the cycle of the entire economic and financial system and are similar to the principles of signal propagation. Most of them are interfered by strong noise, and the signal-to-noise ratio is usually not high.
  • the iterative calculation method of merging weights given in the step reduces the phase difference estimation error and obtains the optimal weights of the merging sequence of various assets, thereby effectively improving the signal-to-noise ratio and stability of the merging sequence of the filter sequence of various assets .
  • Step D Obtain the combined sequence according to the second filter matrix and the combined weight vector.
  • the synthesized sequence X period whose preset observation period takes the value of period is a vector, namely
  • abs(W) is the complex magnitude of each element of W (each element of W is a complex number), and sum(W) is the sum of the elements of W.
  • Step S30 Determine the predicted value corresponding to the logarithmic difference data of the asset to be invested according to the merging sequence
  • step S20 includes:
  • Step S31 Determine the output length of the combined sequence, and modify the output length with a preset length to obtain an extrapolated combined sequence
  • the device determines the output length of the combined sequence, and then modifies the output length by a preset length, thereby obtaining the extrapolated combined sequence.
  • the output length of a 42-month merged sequence is 120
  • the extrapolation filtering period is an extrapolated merged sequence with an output length of 121.
  • the extrapolation filtering period is the extrapolation length L2, and the extrapolation length L2 can be based on The actual demand is not limited to 1.
  • Step S32 determining a linear regression coefficient according to each of the logarithmic difference data and the combination sequence
  • the linear regression model is stored in the device, and the device can use the logarithmic difference data corresponding to various assets as the dependent variable, and the combined sequence as the independent variable, which is input into the linear regression model to obtain the linear regression coefficient.
  • the linear regression model has a corresponding formula, that is, the linear regression formula, and the linear regression formula is:
  • X 42 , X 100 , X 200 are the merged sequences corresponding to the preset observation period of 42 months, 100 months, and 200 months, respectively
  • b 1 is the intercept term
  • b 2 b 3 , and b 4 are linear regressions coefficient
  • It is logarithmic difference data.
  • the linear regression model uses the least squares estimation algorithm to obtain the estimated value of the linear regression coefficient according to the linear regression formula.
  • Step S33 Perform fitting prediction on the logarithmic difference data according to the extrapolated combined sequence and the linear regression coefficient to obtain a predicted value.
  • the device can perform logarithmic difference data fitting predictions for the investment assets to obtain predicted values based on the extrapolated combined sequence and the linear regression coefficients. Specifically, the device obtains the predicted value of the factor according to the extrapolated combined sequence as the independent variable, and then inputs the predicted value of the factor and the linear regression coefficient into the linear regression model to obtain the predicted value, the predicted value
  • step S40 when the predicted value is greater than the preset threshold, a prompt message is generated suggesting that the asset to be invested in the investment asset should be purchased in the predicted time period.
  • the device After determining the predicted value, the device needs to determine whether the predicted value is greater than the preset threshold, and the preset threshold may be zero. When the predicted value is greater than the preset threshold, it can indicate that the holding income of the asset to be invested after the investment period is positive.
  • the income of asset splitting purchase is relatively stable, and the risk is relatively small. Therefore, the device generates prompt information suggesting that the investment asset to be invested in the expected time period is divided into warehouses, so as to send the prompt information to the user's corresponding terminal, so that the user can invest in the investment asset according to the prompt information.
  • the sub-warehouse purchase of assets to be invested in the forecast time period refers to first dividing the funds into the number of shares corresponding to the time length corresponding to the forecast time period. For example, if the time length is 12 months, the funds are divided into 12 shares. Each fund corresponds to a corresponding purchase time period in the forecast time period, and each purchase time period constitutes the forecast time period.
  • the device determines the predicted time period of the asset to be invested and the time length of the predicted time period to obtain the logarithmic difference data corresponding to the asset to be invested in the target investment period according to the length of time, and determine the logarithm
  • the difference data corresponds to the combined sequence in each preset observation period, and then the predicted value of the asset on the logarithmic difference data is determined according to the combined sequence.
  • a suggestion is generated to split the investment assets in the predicted time period. Prompt information for purchase.
  • the device can predict the investment asset based on the logarithmic difference data of the asset to be invested to obtain the predicted value, if the predicted value is greater than the preset threshold, it indicates that the holding income of the asset to be invested after the predicted period of time is positive, so it is recommended that the user During the forecast period, the assets to be invested will be purchased in separate warehouses to provide users with investment advice on wealth management products with reliable and stable returns.
  • Fig. 5 is a second embodiment of the method for generating asset investment advice information of this application. Based on the first embodiment, the step S40 includes:
  • Step S41 Generate prompt information suggesting that funds should be split to invest in assets to be invested within the predicted time period every preset time interval.
  • the holding income of the asset to be invested after the predicted time period is positive. Therefore, if you need to purchase the assets to be invested, you need to start the purchase from the current period, that is, the user can start from the current period and purchase the assets to be invested within the investment period.
  • the investment of the assets to be invested can be made every preset time interval in the investment time period.
  • the preset duration can be one month.
  • the time length of the forecast time period is 6 months
  • the funds are first divided into 6 shares, and each month in the forecast time period corresponds to an investment of one fund.
  • the number of capital splits and the number of investments of the asset to be invested in the forecast time period is equal to the ratio of the time length to the preset time period, and the amount of one investment in the forecast time period corresponds to a split fund.
  • the shares of the assets after the split can be the same, that is, the investment amount of each split investment is the same. According to actual measurement, the risk of purchasing the same funds of the assets to be invested every one month is less risky and the returns are relatively stable.
  • the device when the device determines that the predicted value is greater than the preset threshold, it generates a prompt message suggesting that the funds should be split to invest in the asset to be invested within the predicted time period at intervals of a preset period of time. So as to provide users with cost-effective investment advice information.
  • each logarithmic difference data can be expressed as:
  • R K+1 is the normalization coefficient, which can prevent the weight range from becoming unstable due to continuous accumulation. It ensures that the sum of the squares of the weight coefficients of each logarithmic difference data is equal to the number of logarithmic difference data, that is
  • the weight coefficient of each logarithmic difference data is converged through a preset number of iterations, that is, the number of iterations for the weight coefficient convergence is determined in advance through experiments, and It is stored as a preset number of times.
  • the SUMPLE algorithm is used to synthesize each logarithmic difference data, when the number of iterations reaches the preset number of times, the iteration is stopped, and the synthesized combined sequence is output.
  • the system-level financial data movement law represented by the merged sequence is more stable and reliable, and more predictable.
  • the SUMPLE algorithm is suitable for the synthesis of low signal-to-noise ratio data
  • the optimal weight of each logarithmic difference data can be calculated, so that the combined sequence of the synthesis can be obtained.
  • the signal-to-noise ratio is higher.
  • n cor is the duration corresponding to the overall data
  • the relevant time interval n cor is the preset duration, which can be set according to the actual situation
  • the preset duration can be set to 4 months (or 120 days).
  • the number of iterations is limited by the total time length of the sampled data. Therefore, the number of iterations may be less due to insufficient data length, which affects the convergence of the weight coefficients.
  • the number of iterations of the iterative method is not limited by the length of the logarithmic difference data. Therefore, in this embodiment, preferably, when the SUMPLE algorithm is used to synthesize each logarithmic difference data, the overall iterative method is selected for synthesis, which can ensure that the weight coefficients of each logarithmic difference data converge.
  • the user may send a prediction instruction to the terminal, and the prediction instruction refers to an instruction to predict the return of the investment asset to be invested.
  • the user can input the forecast time period based on the APP interface, and the device loaded with the APP generates the forecast instruction according to the forecast time period.
  • the user does not need to enter the forecast time period into the APP interface, that is, the user defaults to the current time as the observation time to predict the set months in the future. It is understandable that the forecast instruction may or may not contain the forecast time period.
  • step S10 to step S40 are executed.
  • the logarithmic difference data of the asset to be invested is directly obtained.
  • the time length of the logarithmic difference data is preset, that is, the set period, and the device determines the logarithmic difference
  • the data is merged sequence corresponding to each preset observation period, so as to perform step S30 to obtain the predicted value, and when the predicted value is greater than the preset threshold, determine the preset time length corresponding to the logarithmic difference data; determine according to the preset time length For the forecast time period, the preset time length is equal to the time length of the default forecast time period, thereby generating prompt information suggesting that the investment asset to be invested in the investment asset is purchased in the forecast time period.
  • the preset time length corresponding to the logarithmic difference data is 12 months, and if the predicted value is greater than the preset threshold, it is determined that the holding income of the asset to be invested after 12 months is positive.
  • the device predicts the time period corresponding to the preset time length. For example, if the preset time length is 12 months, and the current time is March 2019, the forecast time period is March 2019-March 2020. It should be noted that the preset time length can be set by the user in the terminal. For example, if a user wants to know the return of stocks after 12 months, he can set the preset time length to 12 months. If he wants to know the return of stocks after 6 months, he can set the preset time length Set to 6 months.
  • the predicted value is small and the preset threshold can determine that the holding income of the asset to be invested after the length of time corresponding to the logarithmic difference data is negative.
  • the current time is March 2019, and the preset time length is 12 months , The holding income after March 2020 is negative, so it is possible to generate a reminder that it is not recommended to purchase assets to be invested from March 2019 to March 2020.
  • the device receives a prediction instruction to determine whether the user sets a prediction time period, and adopts different measures for different judgment results to predict the return of the asset to be invested, thereby providing the user with accurate investment advice .
  • the present application also provides a device for generating asset investment advice information.
  • the device for generating asset investment advice information includes: a memory, a processor, and asset investment advice information stored on the memory and running on the processor
  • the generating program of the asset investment recommendation information is executed by the processor to implement the steps of the method for generating asset investment recommendation information as described in the above embodiment.
  • the present application also provides a readable storage medium on which a program for generating asset investment advice information is stored.
  • a program for generating asset investment advice information is stored on a readable storage medium on which a program for generating asset investment advice information is stored.
  • the program for generating asset investment advice information is executed by a processor, the asset as described in the above embodiment is implemented. Steps of the method of generating investment advice information.

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Finance (AREA)
  • Accounting & Taxation (AREA)
  • Development Economics (AREA)
  • Operations Research (AREA)
  • Game Theory and Decision Science (AREA)
  • Human Resources & Organizations (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • Strategic Management (AREA)
  • Technology Law (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)

Abstract

一种资产投资建议信息的生成方法，所述资产投资建议信息的生成方法包括以下步骤：确定待投资资产的预测时间段以及所述预测时间段的时间长度，所述待投资资产为大类资产（S10）；根据所述时间长度获取所述待投资资产在目标投资期开始之前对应的对数差分数据，并确定所述对数差分数据在各个预设观测周期对应的合并序列（S20）；根据所述合并序列确定所述待投资资产在所述对数差分数据上对应的预测值（S30）；当所述预测值大于预设阈值时，生成建议在所述预测时间段对待投资资产进行分仓购买的提示信息（S40）。

Description

资产投资建议信息的生成方法、装置和可读存储介质
优先权信息
本申请要求于2019年11月25日申请的、申请号为201911169498.3、名称为“资产投资建议信息的生成方法、装置和可读存储介质”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及计算机技术领域，尤其涉及一种资产投资建议信息的生成方法、装置及可读存储介质。
背景技术
随着人们生活水平的提高，手上的流动资金越来越富余，相对于将资金存入银行，人们更愿意将资金投资于股票、基金、债券等收益较高的理财产品。
现有技术中，人们在进行理财产品的投资时，通过查看理财产品的风险以及回报说明，并基于自身的投资经验、理财产品的风险以及回报将资金进行投资。由此可知，用户的投资方式是通过用户的投资经验以及对理财产品的预期收益进行确定，这种投资方式可靠性较差，并不能为用户带来稳定的、可靠的投资收益。
上述内容仅用于辅助理解本申请的技术方案，并不代表承认上述内容是现有技术。
发明内容
本申请的主要目的在于提供一种资产投资建议信息的生成方法、装置及可读存储介质，旨在解决现有技术中资金的投资方式可靠性较差的问题。
为实现上述目的，本申请提供一种资产投资建议信息的生成方法，所述资产投资建议信息的生成方法包括以下步骤：
确定待投资资产的预测时间段以及所述预测时间段的时间长度，所述待 投资资产为大类资产；
根据所述时间长度获取所述待投资资产在目标投资期开始之前对应的对数差分数据，并确定所述对数差分数据在各个预设观测周期对应的合并序列；
根据所述合并序列确定所述待投资资产在所述对数差分数据上对应的预测值；
当所述预测值大于预设阈值时，生成建议在所述预测时间段对待投资资产进行分仓购买的提示信息。
在一实施例中，所述生成建议在所述预测时间段对待投资资产进行分仓购买的提示信息的步骤包括：
生成建议将资金拆分以每间隔预设时长在所述预测时间段内进行待投资资产的投资的提示信息。
在一实施例中，资金拆分的份数以及待投资资产在预测时间段内的投资次数等于所述时间长度与预设时长的比值，所述预测时间段内的一次投资金额对应一份拆分的资金。
在一实施例中，所述根据所述合并序列确定所述待投资资产在所述对数差分数据上对应的预测值步骤包括：
确定所述合并序列的输出长度，以对所述输出长度进行预设长度的修改以得到外推合并序列；
根据所述对数差分数据以及所述合并序列确定线性回归系数；
根据所述外推合并序列以及所述线性回归系数确定所述待投资资产对应的预测值。
在一实施例中，所述根据所述外推合并序列以及所述线性回归系数确定所述待投资资产对应的预测值的步骤包括：
根据所述外推合并序列确定因子预测值；
根据所述因子预测值以及所述线性回归系数输入线性回归模型，以得到所述待投资资产对应的预测值。
在一实施例中，所述确定所述对数差分数据在各个预设观测周期对应的合并序列的步骤包括：
根据所述同比序列数据确定各类资产在各个预设观测周期对应的滤波序列，其中，各类所述资产属于同一大类资产；
对各类所述资产对应的所述滤波序列进行合并得到各个所述预设观测周期的合并序列，
在一实施例中，所述根据所述同比序列数据确定各类资产在各个预设观测周期对应的滤波序列的步骤包括：
对所述对数差分数据进行傅里叶变换，得到对应的频域数据；
根据每一个所述预设观测周期确定一组滤波器系数，并根据所述滤波器系数与所述频域数据得到中间序列；
对所述中间序列进行逆傅里叶变换，得到资产在各个所述预设观测周期下的滤波序列。
为实现上述目的，本申请还提供一种资产投资建议信息的生成装置，所述资产投资建议信息的生成装置包括：存储器、处理器及存储在所述存储器上并可在所述处理器上运行的资产投资建议信息的生成程序，所述资产投资建议信息的生成程序被所述处理器执行时实现如上所述的资产投资建议信息的生成方法的步骤。
为实现上述目的，本申请还提供一种可读存储介质，所述可读存储介质上存储有资产投资建议信息的生成程序，所述资产投资建议信息的生成程序被处理器执行时实现如上所述的资产投资建议信息的生成方法的步骤。
本申请实施例提出的一种资产投资建议信息的生成方法、装置和可读存储介质，装置确定待投资资产的预测时间段以及预测时间段的时间长度，以根据时间长度获取待投资资产在目标投资期开始之前对应的对数差分数据，并确定对数差分数据在各个预设观测周期对应的合并序列，再根据合并序列确定资产在对数差分数据上对应的预测值，在预测值大于预设阈值，生成建议在预测时间段对待投资资产进行分仓购买的提示信息。由于装置可以根据待投资资产的对数差分数据对待投资资产进行预测以得到预测值，若是预测值大于预设阈值，则表明待投资资产在预测时间段后的持有收益为正，从而建议用户在预测时间段内对待投资资产进行分仓购买，为用户提供具有可靠稳定收益的理财产品的投资建议。
附图说明
图1是本申请实施例方案涉及的资产投资建议信息的生成装置的硬件结构示意图；
图2为本申请资产投资建议信息的生成方法第一实施例的流程示意图；
图3为图2中步骤S10确定所述对数差分数据在各个预设观测周期对应的合并序列的细化流程示意图；
图4为图2中步骤S20的细化流程示意图；
图5为本申请资产投资建议信息的生成方法第二实施例的流程示意图。
本申请目的的实现、功能特点及优点将结合实施例，参照附图做进一步说明。
具体实施方式
应当理解，此处所描述的具体实施例仅仅用以解释本申请，并不用于限定本申请。
本申请实施例的主要解决方案是：确定待投资资产的预测时间段以及所述预测时间段的时间长度，所述待投资资产为大类资产；根据所述时间长度获取所述待投资资产在目标投资期开始之前对应的对数差分数据，并确定所述对数差分数据在各个预设观测周期对应的合并序列；根据所述合并序列确定所述待投资资产在所述对数差分数据上对应的预测值；当所述预测值大于预设阈值时，生成建议在所述预测时间段对待投资资产进行分仓购买的提示信息。
由于装置可以根据待投资资产的对数差分数据对待投资资产进行预测以得到预测值，若是预测值大于预设阈值，则表明待投资资产在预测时间段后的持有收益为正，从而建议用户在预测时间段内对待投资资产进行分仓购买，为用户提供具有可靠稳定收益的理财产品的投资建议。
如图1所示，图1是本申请实施例方案涉及的资产投资建议信息的生成装置的硬件结构示意图。
如图1所示，资产投资建议信息的生成装置可以包括：处理器1001，例如CPU，通信总线1002，存储器1003。其中，通信总线1002用于实现这些组件之间的连接通信。存储器1003可以是高速RAM存储器，也可以是稳定的存储器(non-volatile memory)，例如磁盘存储器。存储器1003可选的还可以是独 立于前述处理器1001的存储装置。
本领域技术人员可以理解，图1中示出的资产投资建议信息的生成装置结构并不构成对资产投资建议信息的生成装置的限定，可以包括比图示更多或更少的部件，或者组合某些部件，或者不同的部件布置。
如图1所示，作为一种计算机存储介质的存储器1003中可以包括操作系统和资产投资建议信息的生成程序。
在图1所示的装置中，处理器1001可以用于调用存储器1003中存储的资产投资建议信息的生成程序，并执行以下操作：
确定待投资资产的预测时间段以及所述预测时间段的时间长度，所述待投资资产为大类资产；
根据所述时间长度获取所述待投资资产在目标投资期开始之前对应的对数差分数据，并确定所述对数差分数据在各个预设观测周期对应的合并序列；
根据所述合并序列确定所述待投资资产在所述对数差分数据上对应的预测值；
当所述预测值大于预设阈值时，生成建议在所述预测时间段对待投资资产进行分仓购买的提示信息。
在一实施例中，处理器1001可以调用存储器1003中存储的资产投资建议信息的生成处理程序，还执行以下操作：
生成建议将资金拆分以每间隔预设时长在所述预测时间段内进行待投资资产的投资的提示信息。
在一实施例中，处理器1001可以调用存储器1003中存储的资产投资建议信息的生成处理程序，还执行以下操作：
资金拆分的份数以及待投资资产在预测时间段内的投资次数等于所述时间长度与预设时长的比值，所述预测时间段内的一次投资金额对应一份拆分的资金。
在一实施例中，处理器1001可以调用存储器1003中存储的资产投资建议信息的生成处理程序，还执行以下操作：
确定所述合并序列的输出长度，以对所述输出长度进行预设长度的修改以得到外推合并序列；
根据所述对数差分数据以及所述合并序列确定线性回归系数；
根据所述外推合并序列以及所述线性回归系数确定所述待投资资产对应的预测值。
在一实施例中，处理器1001可以调用存储器1003中存储的资产投资建议信息的生成处理程序，还执行以下操作：
根据所述外推合并序列确定因子预测值；
根据所述因子预测值以及所述线性回归系数输入线性回归模型，以得到所述待投资资产对应的预测值。
在一实施例中，处理器1001可以调用存储器1003中存储的资产投资建议信息的生成处理程序，还执行以下操作：
根据所述同比序列数据确定各类资产在各个预设观测周期对应的滤波序列，其中，各类所述资产属于同一大类资产；
对各类所述资产对应的所述滤波序列进行合并得到各个所述预设观测周期的合并序列，
在一实施例中，处理器1001可以调用存储器1003中存储的资产投资建议信息的生成处理程序，还执行以下操作：
对所述对数差分数据进行傅里叶变换，得到对应的频域数据；
根据每一个所述预设观测周期确定一组滤波器系数，并根据所述滤波器系数与所述频域数据得到中间序列；
对所述中间序列进行逆傅里叶变换，得到资产在各个所述预设观测周期下的滤波序列。
基于上述硬件构建，提出本申请资产投资建议信息的生成方法的各个实施例。
参照图2，本申请第一实施例提供一种资产投资建议信息的生成方法，所述方法包括：
步骤S10，确定待投资资产的预测时间段以及所述预测时间段的时间长度，所述待投资资产为大类资产；
在本实施例中，执行主体为资产投资建议信息的生成装置，为了便于描述，以下以装置指代资产投资建议信息的生成装置。装置可视为服务端，装置通过APP的形式可装载于客户端中，使得客户端基于APP与装置通信连接。
用户可向装置输入预测时间段，预测时间段即为用户想要对资产进行投资的时间段。装置可确定预测时间段对应的时间长度。待投资资产指的是大类资产，例如，商品、股票等。需要说明的是，用户对于大类资产的投资计划的时间长度可以大于或等于预测时间段的时间长度，例如，用户计划对商品进行5年的投资，但可预测下一年的商品的收益，5年对应的时间为投资计划的投资时间段，而预测的下一年即为预测时间段，预设时间段一般可为1个月至12个月，优选12个月。
步骤S20，根据所述时间长度获取所述待投资资产在目标投资期开始之前对应的对数差分数据，并确定所述对数差分数据在各个预设观测周期对应的合并序列；
在本实施例中，待投资资产可为基金、债券、股票以及商品等大类资产。用户可打开客户端上装载的APP，APP即为理财产品的投资程序；用户可基于APP选择需要投资的理财产品，例如，用户想投资A股票和B债券，用户可基于APP的资产的选择界面上选择A股票以及B债券，APP再将用户选择的A股票以及B债券发送装置，装置将A股票以及B债券作为待投资资产。
在确定时间长度后，装置再根据时间长度确定待投资资产在目标投资期开始之前的对数差分数据。目标投资期一般为未来时间对应的投资期，也即需要将目标投资期开始之前已产生的数据作为对数差分数据。对数差分数据中所包含的数据均为月度差分数据。每一个月度差分数据以设定的周期进行计算，设定的周期可为1个月、6个月12个月等，设定的周期的时间长度与预测时间段对应的时间长度相同。例如，待投资资产为股票，设定的周期为6个月，则将2019年9月月末的股票收盘价格与2019年3月月末的股票收盘价格的比值或比值的对数值作为2019年9月的对数差分数据，2019年9月为观测时刻，也即目标投资期为2019年9月，该对数差分数据
Figure PCTCN2020101647-appb-000001
根据下述公式计算得到：
Figure PCTCN2020101647-appb-000002
t 0＝2019年3月,t 0+6＝2019年9月。
装置在确定对数差分数据后，在确定对数差分数据在各个预设观测周期对应的合并序列，具体的，请参照图3，即步骤S20中确定所述对数差分数据在各个预设观测周期对应的合并序列包括：
步骤S21，根据所述对数差分数据确定所述待投资资产在各个预设观测周 期对应的滤波序列；
在一段连续时间内的资产数据序列具有时间维度，例如股票价格数据为连续时间上的数值序列，因此由此而得的资产的对数差分数据可看作一个时间序列。该时间序列可以类比为一项经济金融资产运动产生的时域信号，而该时域信号具有傅里叶级数的表达方式，因此可对该对数差分数据进行傅里叶变换，得到对应的频域数据，在频域对其进行分析处理。
资产在统一的金融经济系统中有经济周期，经济周期即为预设观测周期，资产的同比数据在预设观测周期内呈现有规律的变化，而预设观测周期为42个月、100个月以及200个月，预设观测周期设置于装置内。在频域上，每一个预设观测周期对应着一个目标频率信号，这些目标频率信号是稳定且持续的，而其他不稳定或者不可持续的频率信号可视为噪声。故，装置对资产的对数差分数据进行频域滤波，以保留目标频率信号，降低噪声的干扰，保留目标频率信号的对数差分数据即为频域数据。
由于傅里叶变换中存在栅栏效应，在求取每一类资产的滤波序列时，先对对数差分数据进行补零，并对补零后的对数差分数据进行傅里叶变换，得到对应的频域数据。以
Figure PCTCN2020101647-appb-000003
代表第i类资产的对数差分数据，以
Figure PCTCN2020101647-appb-000004
代表第i类资产的对数差分数据对应的频域数据。
如下述公式(1)、(2)、(3)所示，根据每一个预设观测周期period确定一组滤波器系数gauss win，并根据每一组滤波器系数gauss win与频域数据wave fft得到一组中间序列
Figure PCTCN2020101647-appb-000005
Figure PCTCN2020101647-appb-000006
Figure PCTCN2020101647-appb-000007
Figure PCTCN2020101647-appb-000008
其中，nfft是同比数据补零后长度，gauss index是1至nfft的数列，center frequency代表中心频率即所要提取周期因子对应的频率，gauss alpha为影响高斯滤波带宽的参数。这些参数可优选设置为：nfft取4096，period取对应的42个月、100个月或200个月，gauss alpha取10。
需要说明的是，根据下面的公式(4)对
Figure PCTCN2020101647-appb-000009
进行共轭对称操作：
Figure PCTCN2020101647-appb-000010
接着对第i类资产中各个预设观测周期period的中间序列
Figure PCTCN2020101647-appb-000011
进行逆傅里叶变换，得到第i类资产在各个预设观测周期下的滤波序列
Figure PCTCN2020101647-appb-000012
通过上述公式的变换得到公式(5)，变换流程为：对每一组中间序列进行逆傅里叶变换，得到一组第二中间序列；根据预设序列长度LEN从第二中间序列中截取数据点得到每一类资产在各个预设观测周期下的滤波序列
Figure PCTCN2020101647-appb-000013
其中，预设序列长度LEN等于对数差分长度L1与外推长度L2之和。
Figure PCTCN2020101647-appb-000014
其中，Real(Z)是Z的实数部分。
步骤S22，对各类所述资产对应的所述滤波序列进行合并得到各个所述预设观测周期的合并序列。
装置在确定滤波序列后，会对各个滤波序列进行合并，从而得到合并序列。具体的，在面对同一个全球经济金融环境，资产受到同样的经济周期的驱使而呈现出相关性极强的行为。因此，其他类型的资产会影响待投资资产的收益，需要将同一预设观测周期下的待投资资产和其他类型的资产的滤波序列进行合并，即将它们相似的共同经济周期变化特征并进行合并，合并后的序列能够反映市场中统一的系统级别的周期运动，以供后续处理中较好地对各类资产的价格对数差分数据进行拟合。
在合并方式为同类型资产合并方式时，预设观测周期的合并序列的合成可参照如下流程：
步骤A、根据各类所述资产的滤波序列得到第一滤波矩阵；
在本实施例中，各类资产可为股票、债券以及商品，股票、债券以及商品的滤波序列分别为
Figure PCTCN2020101647-appb-000015
将每一个滤波序列作为一个向量，由三个向量组合形成一个滤波矩阵M1，M1即为 第一滤波矩阵。
步骤B，对所述第一滤波矩阵进行希尔伯特变换，得到对应的第二滤波矩阵；
在本步骤中，如下式所示，调用软件平台库函数hibert对M1进行希尔伯特变换得到第二滤波矩阵M2：
M2＝hilbert  (M1)
步骤C，根据第二滤波矩阵进行合并权重的迭代计算。
具体地，先初始化预设观测周期取值为period的合并权重向量
Figure PCTCN2020101647-appb-000016
为长度为N的全1的向量，其中，N为矩阵M2中向量数目，即资产的类别数，当资产只包含股票、债券以及商品时取值为3。
根据下述公式(6)、(7)、(8)进行合并权重向量
Figure PCTCN2020101647-appb-000017
的迭代计算：
Figure PCTCN2020101647-appb-000018
weight m＝mean(weight)   (7)
Figure PCTCN2020101647-appb-000019
其中,
Figure PCTCN2020101647-appb-000020
代表经过第k次迭代计算后的合并权重向量,M*N代表矩阵M和N相乘，(M)′是M的共轭转置，diag(W)是包含W在主对角线上的对角矩阵，conj(M)是M的复共轭，M.*N代表矩阵M和N点乘,mean(M)是M的每列均值。
当迭代计算次数到达预设迭代次数阈值dcnt，合并权重收敛，得到最终的合并权重向量
Figure PCTCN2020101647-appb-000021
在本实施中，可选取dcnt＝100。
由于各类资产的滤波序列可视为时域信号，其受到整个经济金融系统的周期的影响与信号传播的原理有相似之处，大多受到强烈的噪音干扰，信噪比通常不高，因此本步骤所给出的合并权重迭代计算方法通过减少相位差估计误差，得出各类资产的合并序列的最优权重，以此有效提高各类资产的滤波序列的合并序列的信噪比和稳定性。
步骤D，根据第二滤波矩阵和合并权重向量得到合并序列。
根据下述公式(9)得到的预设观测周期取值为period的合成序列X period是一个向量，即
Figure PCTCN2020101647-appb-000022
abs(W)是W的每个元素的复数幅值(W的每个元素是复数)，sum(W)是W的元素总和。
步骤S30，根据所述合并序列确定所述待投资资产在所述对数差分数据上对应的预测值；
装置在确定合并序列后，即可根据合并序列确定待投资资产在对数差分数据上对应的预测值，具体请参照图4，即步骤S20包括：
步骤S31，确定所述合并序列的输出长度，并对所述输出长度进行预设长度的修改以得到外推合并序列；
在本实施例中，装置在获取合并序列后，再确定合并序列的输出长度，进而对输出长度进行预设长度的修改，从而得到外推合并序列。例如，42个月的合并序列的输出长度为120，外推滤波一期，即得到输出长度为121的外推合并序列，外推滤波一期即为外推长度L2，外推长度L2可根据实际需求进行设置，并不限定于1。
步骤S32，根据各个所述对数差分数据以及所述合并序确定线性回归系数；
装置中存储有线性回归模型，装置可将各类资产对应的对数差分数据作为因变量，合并序列作为自变量，输入线性回归模型中，从而得到线性回归系数。具体的，线性回归模型具有对应的公式，也即线性回归公式，线性回归公式为：
Figure PCTCN2020101647-appb-000023
其中，X 42、X 100、X 200分别为预设观测周期42个月、100个月以及200个月对应的合并序列，b 1为截距项，b 2、b 3、b 4为线性回归系数，
Figure PCTCN2020101647-appb-000024
为对数差分数据。线性回归模型根据线性回归公式，采用最小二乘估计算法得到线性回归系数的估计值。
步骤S33，根据所述外推合并序列以及所述线性回归系数，对所述对数差分数据进行拟合预测，以得到预测值。
在确定线性回归系数后，装置即可根据外推合并序列以及线性回归系数对待投资资产进行对数差分数据的拟合预测得到预测值。具体的，装置根据外推合并序列作为自变量得到因子预测值，再将因子预测值以及线性回归系数输入线性回归模型中，从而得到预测值，预测值
Figure PCTCN2020101647-appb-000025
步骤S40，当所述预测值大于预设阈值时，生成建议在所述预测时间段对待投资资产进行分仓购买的提示信息。
装置在确定预测值后，需要判断预测值是否大于预设阈值，预设阈值可为零。在当预测值大于预设阈值，即可表明待投资资产在投资时间段后的持有收益为正。
资产的进行分仓购买的收益较为稳定，且风险较小。故装置生成建议在预测时间段对待投资资产进行分仓购买的提示信息，从而将该提示信息发送至用户对应的终端，使得用户根据提示信息对待投资资产进行投资。在预测时间段内待投资资产的分仓购买指的是，先将资金分成预测时间段对应的时间长度对应的份数，例如，时间长度为12个月，则将资金分成12份。每一份资金在预测时间段中对应相应的购买时间段，各个购买时间段构成预测时间段。
在本实施例提供的技术方案中，装置确定待投资资产的预测时间段以及预测时间段的时间长度，以根据时间长度获取待投资资产在目标投资期对应的对数差分数据，并确定对数差分数据在各个预设观测周期对应的合并序列，再根据合并序列确定资产在对数差分数据上对应的预测值，在预测值大于预设阈值，生成建议在预测时间段对待投资资产进行分仓购买的提示信息。由于装置可以根据待投资资产的对数差分数据对待投资资产进行预测以得到预测值，若是预测值大于预设阈值，则表明待投资资产在预测时间段后的持有收益为正，从而建议用户在预测时间段内对待投资资产进行分仓购买，为用户提供具有可靠稳定收益的理财产品的投资建议。
参照图5，图5为本申请资产投资建议信息的生成方法的第二实施例，基于第一实施例，所述步骤S40包括：
步骤S41，生成建议将资金拆分以每间隔预设时长在所述预测时间段内进 行待投资资产的投资的提示信息。
在本实施例中，预测值大于预设阈值时，待投资资产在预测时间段之后的持有收益为正。故，若需购买待投资资产，则需从当前期开始购买，也即用户可从当前期开始，在投资时间段内进行待投资资产的分仓购买。
具体的，可在投资时间段内每间隔预设时长进行待投资资产的投资。预设时长可为一个月。例如，预测时间段的时间长度为6个月，则先将资金拆分成6份，预测时间段的每个月对应投资一份资金。资金拆分的份数以及待投资资产在预测时间段内的投资次数等于时间长度与预设时长的比值，预测时间段内的一次投资的金额对应一份拆分的资金。
需要说明的是，拆分后的资产的份额可以相同，也即每次的分仓投资的投资金额相同。经实测，对每间隔1个月进行待投资资产的相同资金的购买，风险较小，且收益较为稳定。
在本实施例提供的技术方案中，装置在确定预测值大于预设阈值时，，生成建议将资金拆分以每间隔预设时长在所述预测时间段内进行待投资资产的投资的提示信息从而为用户提供性价比较高的投资建议信息。
在一实施例中，合并序列还可根据SUMPLE算法合成。具体的，每个对数差分数据可以表示为：
Figure PCTCN2020101647-appb-000026
式中k为时间变量，
Figure PCTCN2020101647-appb-000027
是第i个对数差分数据在k时刻的数据，
Figure PCTCN2020101647-appb-000028
为噪声。合成的权值系数表示为：
Figure PCTCN2020101647-appb-000029
其中K是以相关时间间隔n cor为单位的时间变量，即合成中的迭代次数，
Figure PCTCN2020101647-appb-000030
为理想权值，
Figure PCTCN2020101647-appb-000031
是噪声引起的权值估计误差，则合成的合成序列可以表示为：
Figure PCTCN2020101647-appb-000032
其中*为取复共轭，L为对数差分数据的总个数。如果将合成序列的输出表示成如下形式：
Figure PCTCN2020101647-appb-000033
那么，信号和噪声项分别为：
Figure PCTCN2020101647-appb-000034
Figure PCTCN2020101647-appb-000035
SUMPLE算法中的第K+1次的权值系数
Figure PCTCN2020101647-appb-000036
可由第K次的
Figure PCTCN2020101647-appb-000037
递推得到：
Figure PCTCN2020101647-appb-000038
式中R K+1为归一化系数，可以防止权值幅度因连续累加变得不稳定，它保证了各个对数差分数据的权值系数的平方和等于对数差分数据的数量，即
Figure PCTCN2020101647-appb-000039
上式的权值
Figure PCTCN2020101647-appb-000040
还可利用C k改写为：
Figure PCTCN2020101647-appb-000041
本实施例中，利用SUMPLE算法对各个对数差分数据进行合成时，通过预设次数的迭代使得各个对数差分数据的权值系数收敛，即预先通过实验确定权值系数收敛的迭代次数，并存储为预设次数，在利用SUMPLE算法对各个对数差分数据进行合成时，在迭代次数达到所述预设次数时，停止迭代，并输出合成的合并序列。合并序列代表的系统级别的金融数据运动规律，更加稳定可靠，可预测性更强。同时，由于SUMPLE算法适用于低信噪比的数据的合成，利用SUMPLE算法对各个对数差分数据合成时，可以计算出各个对数差分数据的最优权值，从而使得到的合成的合并序列信噪比更高。
本实施例中，在利用SUMPLE算法对各个对数差分数据进行合成时，可选用整体迭代或滚动迭代两种迭代方式进行迭代，使得各个对数差分数据的权值系数收敛。在通过整体迭代的方式进行合成时，进行一次采样，然后用整体数据对自身进行迭代更新，此时相关时间间隔n cor即为整体数据对应的时长；在通过滚动迭代的方式进行合成时，将采样窗口滚动向前，多次采样， 用下一时刻采样得到的序列更新上一时刻得到的权重系数，此时相关时间间隔n cor为预设时长，所述预设时长可根据实际情况进行设置，例如，所述预设时长可设置为4个月(或120天)。在实际合成中，由于数据长度有限，而在利用滚动迭代时，迭代的次数受到采样数据的总时长限制，因此可能由于数据长度不够导致迭代次数较少，从而影响权值系数的收敛，而整体迭代的方式的迭代次数不受对数差分数据长度的限制。因此，本实施例中，优选地，利用SUMPLE算法对各个对数差分数据进行合成时，选用整体迭代的方式进行合成，可以保证各个对数差分数据的权值系数收敛。
在一实施例中，用户可向终端发送预测指令，预测指令指的是对待投资资产进行收益预测的指令。用户可基于APP界面输入预测时间段，装载APP的装置即根据预测时间段生成预测指令。当然，用户也无需向APP界面输入预测时间段，也即用户默认以当前时间为观测时刻，对未来的设定的几个月进行预测。可以理解的是，预测指令中可含有预测时间段，也可不含有预测时间段。
在当预测指令中含有预测时间段时，则执行步骤S10-步骤S40。
而在当预测指令中不含有预测时间段，则直接获取待投资资产的对数差分数据，对数差分数据的时间长度为预设的，也即为设定的周期，装置再确定对数差分数据在各个预设观测周期对应的合并序列，从而执行步骤S30得到预测值，并在预测值大于预设阈值时，确定对数差分数据对应的预设的时间长度；根据预设的时间长度确定预测时间段，预设的时间长度等于默认的预测时间段的时间长度，从而生成建议在所述预测时间段对待投资资产进行分仓购买的提示信息。
例如，对数差分数据对应的预设的时间长度为12个月，若预测值大于预设阈值，则判定该待投资资产在12个月后的持有收益为正。
故，在预测值大于预设阈值后，装置根据预设的时间长度对应的预测时间段。例如，预设的时间长度为12个月，当前时间为2019年3月，则预测时间段为2019年3月-2020年3月。需要说明的是，预设的时间长度可由用户在终端进行设置。例如，用户想得知股票在12个月后的收益，则可将预设的时间长度设置为12个月，若想得知股票在6个月后的收益，则可将预设的 时间长度设置为6个月。
此外，预测值小预设阈值，可判定待投资资产在对数差分数据对应的时间长度后的持有收益为负，例如，当前时间为2019年3月，预设的时间长度为12个月，2020年3月后的持有收益为负，故可生成不建议在2019年3月-2020年3月购买待投资资产的提示信息。
在本实施例提供的技术方案中，装置接收预测指令，从而判断用户是否设置预测时间段，从而对不同的判断结果采用不同的措施进行待投资资产的收益预测，进而为用户提供准确的投资建议。
本申请还提供一种资产投资建议信息的生成装置，所述资产投资建议信息的生成装置包括：存储器、处理器及存储在所述存储器上并可在所述处理器上运行的资产投资建议信息的生成程序，所述资产投资建议信息的生成程序被所述处理器执行时实现如上实施例所述的资产投资建议信息的生成方法的步骤。
本申请还提供一种可读存储介质，所述可读存储介质上存储有资产投资建议信息的生成程序，所述资产投资建议信息的生成程序被处理器执行时实现如上实施例所述的资产投资建议信息的生成方法的步骤。
需要说明的是，在本文中，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、物品或者系统不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者系统所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括该要素的过程、方法、物品或者系统中还存在另外的相同要素。
上述本申请实施例序号仅仅为了描述，不代表实施例的优劣。
通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到上述实施例方法可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件，但很多情况下前者是更佳的实施方式。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体 现出来，该计算机软件产品存储在如上所述的一个存储介质(如ROM/RAM、磁碟、光盘)中，包括若干指令用以使得一台终端设备(可以是手机，计算机，服务器，空调器，或者网络设备等)执行本申请各个实施例所述的方法。
以上仅为本申请的优选实施例，并非因此限制本申请的专利范围，凡是利用本申请说明书及附图内容所作的等效结构或等效流程变换，或直接或间接运用在其他相关的技术领域，均同理包括在本申请的专利保护范围内。

Claims (9)

  1. 一种资产投资建议信息的生成方法，其中，所述资产投资建议信息的生成方法包括以下步骤：
    确定待投资资产的预测时间段以及所述预测时间段的时间长度，所述待投资资产为大类资产；
    根据所述时间长度获取所述待投资资产在目标投资期开始之前对应的对数差分数据，并确定所述对数差分数据在各个预设观测周期对应的合并序列；
    根据所述合并序列确定所述待投资资产在所述对数差分数据上对应的预测值；
    当所述预测值大于预设阈值时，生成建议在所述预测时间段对待投资资产进行分仓购买的提示信息。
  2. 如权利要求1所述的资产投资建议信息的生成方法，其中，所述生成建议在所述预测时间段对待投资资产进行分仓购买的提示信息的步骤包括：
    生成建议将资金拆分以每间隔预设时长在所述预测时间段内进行待投资资产的投资的提示信息。
  3. 如权利要求2所述的资产投资建议信息的生成方法，其中，资金拆分的份数以及待投资资产在预测时间段内的投资次数等于所述时间长度与预设时长的比值，所述预测时间段内的一次投资金额对应一份拆分的资金。
  4. 如权利要求1所述的资产投资建议信息的生成方法，其中，所述根据所述合并序列确定所述待投资资产在所述对数差分数据上对应的预测值步骤包括：
    确定所述合并序列的输出长度，以对所述输出长度进行预设长度的修改以得到外推合并序列；
    根据所述对数差分数据以及所述合并序列确定线性回归系数；
    根据所述外推合并序列以及所述线性回归系数确定所述待投资资产对应的预测值。
  5. 如权利要求4所述的资产投资建议信息的生成方法，其中，所述根据所述外推合并序列以及所述线性回归系数确定所述待投资资产对应的预测值的步骤包括：
    根据所述外推合并序列确定因子预测值；
    根据所述因子预测值以及所述线性回归系数输入线性回归模型，以得到所述待投资资产对应的预测值。
  6. 如权利要求1所述的资产投资建议信息的生成方法，其中，所述确定所述对数差分数据在各个预设观测周期对应的合并序列的步骤包括：
    根据所述同比序列数据确定各类资产在各个预设观测周期对应的滤波序列，其中，各类所述资产属于同一大类资产；
    对各类所述资产对应的所述滤波序列进行合并得到各个所述预设观测周期的合并序列。
  7. 如权利要求6所述的资产投资建议信息的生成方法，其中，所述根据所述同比序列数据确定各类资产在各个预设观测周期对应的滤波序列的步骤包括：
    对所述对数差分数据进行傅里叶变换，得到对应的频域数据；
    根据每一个所述预设观测周期确定一组滤波器系数，并根据所述滤波器系数与所述频域数据得到中间序列；
    对所述中间序列进行逆傅里叶变换，得到资产在各个所述预设观测周期下的滤波序列。
  8. 一种资产投资建议信息的生成装置，其中，所述资产投资建议信息的生成装置包括：存储器、处理器及存储在所述存储器上并可在所述处理器上运行的资产投资建议信息的生成程序，所述资产投资建议信息的生成程序被所述处理器执行时实现如权利要求1至7中任一项所述的资产投资建议信息的生成方法的步骤。
  9. 一种可读存储介质，其中，所述可读存储介质上存储有资产投资建议信息的生成程序，所述资产投资建议信息的生成程序被处理器执行时实现如权利要求1至7中任一项所述的资产投资建议信息的生成方法的步骤。
PCT/CN2020/101647 2019-11-25 2020-07-13 资产投资建议信息的生成方法、装置和可读存储介质 WO2021103572A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201911169498.3 2019-11-25
CN201911169498.3A CN111091466A (zh) 2019-11-25 2019-11-25 资产投资建议信息的生成方法、装置和可读存储介质

Publications (1)

Publication Number Publication Date
WO2021103572A1 true WO2021103572A1 (zh) 2021-06-03

Family

ID=70393723

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/101647 WO2021103572A1 (zh) 2019-11-25 2020-07-13 资产投资建议信息的生成方法、装置和可读存储介质

Country Status (2)

Country Link
CN (1) CN111091466A (zh)
WO (1) WO2021103572A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110458352A (zh) * 2019-08-06 2019-11-15 华泰证券股份有限公司 预测资产价格走势的方法、服务器及计算机可读存储介质
CN111091466A (zh) * 2019-11-25 2020-05-01 华泰证券股份有限公司 资产投资建议信息的生成方法、装置和可读存储介质
CN113656470A (zh) * 2020-05-12 2021-11-16 北京市天元网络技术股份有限公司 资产采购提示方法、装置、电子设备及存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120116996A1 (en) * 2010-11-04 2012-05-10 Investpic, Llc Method and system for analyzing investment information
CN108446984A (zh) * 2018-03-20 2018-08-24 张家林 一种投资数据管理方法及装置
CN109859052A (zh) * 2019-01-16 2019-06-07 平安科技（深圳）有限公司 一种投资策略的智能推荐方法、装置、存储介质和服务器
CN110458352A (zh) * 2019-08-06 2019-11-15 华泰证券股份有限公司 预测资产价格走势的方法、服务器及计算机可读存储介质
CN111091466A (zh) * 2019-11-25 2020-05-01 华泰证券股份有限公司 资产投资建议信息的生成方法、装置和可读存储介质

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120116996A1 (en) * 2010-11-04 2012-05-10 Investpic, Llc Method and system for analyzing investment information
CN108446984A (zh) * 2018-03-20 2018-08-24 张家林 一种投资数据管理方法及装置
CN109859052A (zh) * 2019-01-16 2019-06-07 平安科技（深圳）有限公司 一种投资策略的智能推荐方法、装置、存储介质和服务器
CN110458352A (zh) * 2019-08-06 2019-11-15 华泰证券股份有限公司 预测资产价格走势的方法、服务器及计算机可读存储介质
CN111091466A (zh) * 2019-11-25 2020-05-01 华泰证券股份有限公司 资产投资建议信息的生成方法、装置和可读存储介质

Also Published As

Publication number Publication date
CN111091466A (zh) 2020-05-01

Similar Documents

Publication Publication Date Title
WO2021103572A1 (zh) 资产投资建议信息的生成方法、装置和可读存储介质
Van Staden et al. Time-consistent mean–variance portfolio optimization: A numerical impulse control approach
US8036975B2 (en) Systems and methods for compound risk factor sampling with integrated market and credit risk
US20170213284A1 (en) Methods and systems for computing trading strategies for use in portfolio management and computing associated probability distributions for use in option pricing
Pang et al. Two-stage non-cooperative games with risk-averse players
He et al. Real time detection of structural breaks in GARCH models
Rompolis Retrieving risk neutral densities from European option prices based on the principle of maximum entropy
WO2021022977A1 (zh) 预测资产价格走势的方法、服务器及计算机可读存储介质
Pagliarani et al. The exact Taylor formula of the implied volatility
Cong et al. Accurate and robust numerical methods for the dynamic portfolio management problem
Qin et al. Nonparametric implied Lévy densities
WO2021103571A1 (zh) 资产投资建议信息的生成方法、装置和可读存储介质
WO2021051976A1 (zh) 金融时间序列的合成方法、装置和存储介质
Laumer et al. Does the government spending multiplier depend on the business cycle?
Albanese* et al. Discrete credit barrier models
Liu et al. Revisit of stochastic mesh method for pricing American options
Smith et al. Equity premium forecasts with an unknown number of structural breaks
Jang Structural estimation of the New-Keynesian model: A formal test of backward-and forward-looking behavior
Li et al. Do price trajectory data increase the efficiency of market impact estimation?
Ascione et al. Credit default swap spreads modeling and forecasting with a stochastic square-root three-factor model
Ayadi et al. Portfolio performance sensitivity for various asset-pricing kernels
Liao et al. Multivariate least squares forecasting averaging by vector autoregressive models
Atanassov et al. Sensitivity study of heston stochastic volatility model using GPGPU
Olschewsky et al. Finite difference methods for estimating marginal risk contributions in asset management
Bo et al. The pricing of basket options: A weak convergence approach

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20892679

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20892679

Country of ref document: EP

Kind code of ref document: A1