WO2021051976A1 - 金融时间序列的合成方法、装置和存储介质 - Google Patents

金融时间序列的合成方法、装置和存储介质 Download PDF

Info

Publication number
WO2021051976A1
WO2021051976A1 PCT/CN2020/101798 CN2020101798W WO2021051976A1 WO 2021051976 A1 WO2021051976 A1 WO 2021051976A1 CN 2020101798 W CN2020101798 W CN 2020101798W WO 2021051976 A1 WO2021051976 A1 WO 2021051976A1
Authority
WO
WIPO (PCT)
Prior art keywords
financial
indicator data
time series
synthetic
synthesizing
Prior art date
Application number
PCT/CN2020/101798
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 WO2021051976A1 publication Critical patent/WO2021051976A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/24—Querying
    • G06F16/245—Query processing
    • G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
    • G06F16/2474—Sequence data queries, e.g. querying versioned data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/21—Design, administration or maintenance of databases
    • G06F16/211—Schema design and management
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/21—Design, administration or maintenance of databases
    • G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
    • 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

Definitions

  • This application relates to the field of computer technology, in particular to a method, device and storage medium for synthesizing financial time series.
  • the main purpose of this application is to provide a method, device and storage medium for synthesizing financial time series, aiming to solve the technical problem that the current synthesizing method cannot identify and process financial signals with a low signal-to-noise ratio.
  • the present application provides a method for synthesizing financial time series.
  • the method for synthesizing financial time series includes the following steps:
  • the SUMPLE algorithm is used to synthesize each of the financial indicator data to obtain a synthetic financial time series.
  • the present application also provides a financial time series synthesis device
  • the financial time series synthesis device includes: a memory, a processor, and stored on the memory and capable of running on the processor
  • the financial time series synthesis program is executed by the processor to realize the steps of the financial time series synthesis method described in any one of the preceding items.
  • the present application also provides a storage medium on which a financial time series synthesis program is stored, and the financial time series synthesis program is executed by a processor to achieve the same as described in any one of the preceding items.
  • the steps of the synthetic method of the financial time series are described in any one of the preceding items.
  • a method, device, and storage medium for synthesizing financial time series proposed in the embodiments of the present application obtain observation time points, and obtain multiple financial indicator data up to the observation time point; and use the SUMPLE algorithm to compare each of the financial indicator data Perform synthesis to obtain a synthetic financial time series. Since the SUMPLE algorithm used for synthesis is suitable for the synthesis of low signal-to-noise ratio data, when the SUMPLE algorithm is used to synthesize each year-on-year sequence, the optimal weight of the year-on-year sequence of each financial indicator data can be calculated, so that the resulting synthetic financial The time series signal-to-noise ratio is high.
  • FIG. 1 is a schematic diagram of a terminal structure of a hardware operating environment involved in a solution of an embodiment of the present application
  • FIG. 2 is a schematic flowchart of a first embodiment of a method for synthesizing financial time series of the application
  • FIG. 3 is a schematic flowchart of a second embodiment of a method for synthesizing financial time series of the application
  • FIG. 4 is a schematic flowchart of a third embodiment of a method for synthesizing financial time series of this application.
  • FIG. 5 is a schematic flowchart of a fourth embodiment of a method for synthesizing financial time series in this application:
  • Fig. 6 is a schematic flowchart of a fifth embodiment of a method for synthesizing financial time series in this application.
  • FIG. 7 is a schematic flowchart of a sixth embodiment of a method for synthesizing financial time series of this application.
  • the SUMPLE algorithm is used to synthesize each of the financial indicator data to obtain a synthetic financial time series.
  • This application provides a solution. Since the SUMPLE algorithm used for synthesis is suitable for the synthesis of low signal-to-noise ratio data, when the SUMPLE algorithm is used to synthesize each year-on-year sequence, the optimal weight of the year-on-year sequence of each financial indicator data can be calculated Value, so that the resulting synthetic financial time series has a high signal-to-noise ratio.
  • FIG. 1 is a schematic diagram of a terminal structure of a hardware operating environment involved in a solution of an embodiment of the present application.
  • the terminal may include: a processor 1001, such as a CPU, a communication bus 1002, a user interface 1003, and a memory 1004.
  • the communication bus 1002 is used to implement connection and communication between these components.
  • the user interface 1003 may include a display screen (Display) and an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface.
  • the memory 1004 may be a high-speed RAM memory, or a non-volatile memory (non-volatile memory), such as a magnetic disk memory.
  • the memory 1004 may also be a storage device independent of the foregoing processor 1001.
  • terminal structure shown in FIG. 1 does not constitute a limitation on the terminal, and may include more or fewer components than shown in the figure, or combine some components, or arrange different components.
  • the memory 1004 which is a computer storage medium, may include an operating system, a network communication module, a user interface module, and a synthetic program of financial time series.
  • the user interface 1003 is mainly used to connect to the client (user terminal), and to communicate data with the client; and the processor 1001 can be used to call the synthetic program of the financial time series stored in the memory 1004, And do the following:
  • the SUMPLE algorithm is used to synthesize each of the financial indicator data to obtain a synthetic financial time series.
  • processor 1001 may call the financial time series synthesis program stored in the memory 1004, and also perform the following operations:
  • the SUMPLE algorithm is used to synthesize each of the year-on-year sequences to obtain a synthetic financial time sequence.
  • processor 1001 may call the financial time series synthesis program stored in the memory 1004, and also perform the following operations:
  • the logarithm of each of the ratios is used as the year-on-year sequence of the financial indicator data.
  • processor 1001 may call the financial time series synthesis program stored in the memory 1004, and also perform the following operations:
  • the SUMPLE algorithm is used to synthesize each of the periodic filtering sequences to obtain a synthetic financial time sequence corresponding to the observation period.
  • processor 1001 may call the financial time series synthesis program stored in the memory 1004, and also perform the following operations:
  • the phase difference between each of the financial indicator data and the synthetic financial time series is determined according to the weight coefficient.
  • processor 1001 may call the financial time series synthesis program stored in the memory 1004, and also perform the following operations:
  • the weight coefficients corresponding to each of the financial indicator data are converged through overall iteration or rolling iteration.
  • processor 1001 may call the financial time series synthesis program stored in the memory 1004, and also perform the following operations:
  • the SUMPLE algorithm is used to synthesize the first synthetic sequence of various financial indicator data to obtain a synthetic financial time sequence.
  • processor 1001 may call the financial time series synthesis program stored in the memory 1004, and also perform the following operations:
  • the SUMPLE algorithm is used to synthesize the second synthetic sequence of financial indicator data in each region to obtain a synthetic financial time series.
  • the observation time point is obtained, and multiple financial indicator data up to the observation time point is obtained; the SUMPLE algorithm is used to synthesize each of the financial indicator data to obtain a synthetic financial time series. Since the SUMPLE algorithm used for synthesis is suitable for the synthesis of low signal-to-noise ratio data, when the SUMPLE algorithm is used to synthesize each year-on-year sequence, the optimal weight of the year-on-year sequence of each financial indicator data can be calculated, so that the resulting synthetic financial The time series signal-to-noise ratio is high.
  • Fig. 2 is a schematic flow chart of the first embodiment of a method for synthesizing financial time series according to this application.
  • the method for synthesizing financial time series includes:
  • Step S10 Obtain the observation time point, and acquire multiple financial indicator data up to the observation time point;
  • the financial indicator data in this embodiment includes various assets and various economic indicators, such as stock market indexes, bond market indexes, and commodity indexes of various countries or regions (for example, economic cooperation regions), as well as various macroeconomic variables.
  • M1 narrow money supply
  • M2 broad money supply
  • PMI purchasing manager index
  • CPI consumer price index
  • PPI producer price index
  • economic prosperity index of each major country.
  • the financial indicator data can be obtained by performing statistics on a monthly basis, and the monthly average value or month-end value of the financial data can be used as the financial indicator data.
  • the stock market index for each month can be expressed by the average price of the stock for that month or the month-end price.
  • step S20 the SUMPLE algorithm is used to synthesize each of the financial indicator data to obtain a synthetic financial time series.
  • each financial indicator data can be expressed as:
  • K is the time variable in the unit of the relevant time interval n cor , that is, the number of iterations in the synthesis, Is the ideal weight, Is the weight estimation error caused by noise, the synthesized synthetic financial time series can be expressed as:
  • R K+1 is the normalized coefficient, which can prevent the weight range from becoming unstable due to continuous accumulation. It ensures that the square sum of the weight coefficient of each financial indicator data is equal to the number of financial indicator data, namely
  • the weight coefficients of each year-on-year sequence are 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 stored as a preset Assuming the number of times, when the SUMPLE algorithm is used to synthesize various financial indicator data, when the number of iterations reaches the preset number of times, the iteration is stopped, and the synthesized financial time series is output.
  • the system-level financial data movement law represented by the synthetic financial time series 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 financial indicator data can be calculated, so that the resulting synthetic financial time series signal noise Than higher.
  • the SUMPLE algorithm when used to synthesize various financial indicator data, two iterative methods can be selected, either the overall iteration or the rolling iteration, so that the weight coefficients of each financial indicator data converge.
  • the relevant time interval n cor is the duration corresponding to the overall data; when the synthesis is performed by rolling iteration, The sampling window is scrolled forward, samples are taken multiple times, and the sequence obtained at the next time is used to update the weight coefficient obtained at the previous time.
  • the relevant time interval n cor is the preset duration, and the preset duration 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 financial indicator data. Therefore, in this embodiment, preferably, when the SUMPLE algorithm is used to synthesize various financial indicator data, the overall iterative method is selected for synthesis, which can ensure the convergence of the weight coefficients of each financial indicator data.
  • the observation time point is obtained, and multiple financial indicator data up to the observation time point is obtained; the SUMPLE algorithm is used to synthesize each of the financial indicator data to obtain a synthetic financial time series. Since the SUMPLE algorithm used for synthesis is suitable for the synthesis of low signal-to-noise ratio data, when the SUMPLE algorithm is used to synthesize each year-on-year sequence, the optimal weight of the year-on-year sequence of each financial indicator data can be calculated, so that the resulting synthetic financial The time series signal-to-noise ratio is high.
  • step S20 includes:
  • Step S21 Calculate the year-on-year sequence of each of the financial indicator data according to a preset time interval
  • step S22 the SUMPLE algorithm is used to synthesize each of the year-on-year sequences to obtain a synthetic financial time sequence.
  • the year-on-year sequence is calculated for each financial indicator data first, and then the SUMPLE algorithm is used to synthesize the year-on-year sequence.
  • the ratios of two financial data separated by a preset time interval in the same financial indicator data can be calculated separately, and the calculated ratios can be used as the year-on-year sequence of the financial indicator data.
  • the preset time interval can be set by itself according to actual conditions, and there is no specific limitation here.
  • the preset time interval can be set to 12 months. For example, there are 304 months of financial indicator data from January 1992 to April 2017. The preset interval is set to 12 months.
  • the financial indicator data in April of 2016, we will push forward 12 months to obtain the financial indicator data of April 2016, and calculate the ratio of the financial indicator data of April 2017 and April 2016; after obtaining the financial indicator data of January 2017
  • the financial indicator data of January 2016 is obtained by pushing forward 12 months, and the ratio of the financial indicator data of January 2017 and January 2016 is calculated; and so on, the number of the same financial indicator data is calculated.
  • Each ratio is used as the year-on-year sequence of the financial indicator data.
  • the logarithmic value of each of the ratios can be further calculated, and then the logarithm value of each of the ratios is used as a year-on-year sequence of the financial data.
  • the logarithm with the base 10 can be obtained, or the natural logarithm can be obtained.
  • the financial indicator data may be obtained by performing statistics on a one-month period, such as the monthly average price of stocks, the monthly average price of bonds, or monthly M1 data.
  • the year-on-year data reflects the change rate of the variable compared to the same period value of the preset interval. For example, when the preset interval is 12 months, it reflects the comparison of the variable compared to the previous year. The rate of change of the value over the same period.
  • the obtained year-on-year data can eliminate the seasonal and monthly effects in the financial indicator data, and remove the possible interference in the financial indicator data Invalid information of the change trend judgment result.
  • Fig. 4 is a schematic flowchart of a third embodiment of a method for synthesizing financial time series of this application. Based on the first or second embodiment, the step S20 includes:
  • Step S23 Obtain an observation period, and perform Gaussian filtering on each of the financial indicator data according to the observation period to obtain a periodic filtering sequence of each of the financial indicator data;
  • step S24 the SUMPLE algorithm is used to synthesize each of the periodic filter sequences to obtain a synthetic financial time sequence corresponding to the observation period.
  • the observation period is obtained, and then Gaussian filtering is performed on each of the financial indicator data according to the observation period to obtain the periodic filtering sequence of each financial indicator data, so as to retain these
  • the target frequency signal containing important information about the economic cycle of the asset reduces noise interference.
  • the SUMPLE algorithm is used to synthesize each of the periodic filtering sequences to obtain the synthetic financial time sequence corresponding to the observation period.
  • the synthesized financial time series obtained after synthesis can reflect the uniform system-level periodic movement in the financial system, so as to better fit the price year-on-year series of various financial indicator data in subsequent processing.
  • the observation period is one of 42 months, 100 months, and 200 months, and the synthetic financial time series corresponding to the respective synthesis periods of 42 months, 100 months, and 200 months are obtained. , 100 months and 200 months corresponding cycle factor.
  • the year-on-year sequence of each financial indicator data may be calculated first, and then Gaussian filtering is performed on each of the year-on-year sequences according to the observation period to obtain the periodic filtering sequence of each year-on-year sequence. Then, the SUMPLE algorithm is further used to synthesize each of the periodic filtering sequences to obtain the synthetic financial time sequence corresponding to the observation period.
  • the observation period is obtained, and Gaussian filtering is performed on each financial indicator data according to the observation period to obtain the periodic filtering sequence of each financial indicator data, and then use SUMPLE
  • the algorithm synthesizes each of the periodic filter sequences to obtain the synthetic financial time sequence corresponding to the observation period, which can better fit the price year-on-year sequence of various financial indicator data in subsequent processing.
  • FIG. 5 is a schematic flowchart of a fourth embodiment of a method for synthesizing financial time series of this application. Based on any one of the first to third embodiments, after the step S20, the method further includes :
  • Step S30 Obtain the weight coefficient of each of the financial indicator data
  • Step S40 Determine the phase difference between each of the financial indicator data and the synthetic financial time series according to the weight coefficient.
  • each financial indicator data In reality, it is often necessary to predict the change trend of each financial indicator data.
  • the time series of each financial indicator data and the synthesized synthetic financial time series have the same changing law and common economic cycle, but each financial indicator data
  • the synthetic financial time series with the synthetic financial indicator data will not be completely synchronized, and there is a phase difference between the two, that is, the specific financial indicator data is ahead or lagging relative to the synthetic financial time series.
  • the weight coefficient of each financial indicator data is obtained, and then the sum of each financial indicator data is determined according to the weight coefficient.
  • the phase difference of the synthetic financial time series is a complex number
  • the argument angle of the weight coefficient (complex number) is the phase difference angle between the financial indicator data corresponding to the year-on-year sequence and the synthetic financial time sequence.
  • the SUMPLE algorithm is used to synthesize each of the financial indicator data, and after the synthetic financial time series is obtained, the weight coefficient of each financial indicator data is obtained, and the weight coefficient of each financial indicator data is determined according to the weight coefficient.
  • the phase difference of the synthetic financial time series in turn, it is possible to predict various financial indicators based on the synthetic financial time series obtained by synthesis and the phase difference.
  • FIG. 6 is a schematic flowchart of a fifth embodiment of a method for synthesizing financial time series of this application. Based on any one of the first to fourth embodiments, the step S20 includes:
  • Step S25 using the SUMPLE algorithm to synthesize various financial indicator data of the same type to obtain a first composite sequence of each type of financial indicator data;
  • Step S26 Use the SUMPLE algorithm to synthesize the first synthetic sequence of various financial indicator data to obtain a synthetic financial time sequence.
  • each type of financial index includes multiple economic and financial targets.
  • a stock index includes multiple stock data.
  • the year-on-year sequence data of multiple financial targets included in it showed a strong correlation.
  • the stock markets of various countries around the world show a strong correlation, and there are also strong correlations between various financial and economic indicators of different countries. This means that in the face of the same global economic and financial environment, various countries and major markets have shown extremely relevant behaviors.
  • the SUMPLE algorithm is used to synthesize each financial indicator data of the same type to obtain the first composite sequence of each type of financial indicator data.
  • Each synthesized first synthesized sequence can reflect the system-level operation of the same type of financial indicator data.
  • the SUMPLE algorithm is further used to synthesize the first synthetic sequence of various financial indicator data to obtain a synthetic financial time series. For example, after obtaining multiple financial indicator data up to the observation time point, obtain each stock market index (which belongs to the stock market index type) in the financial indicator data, and use the SUMPLE algorithm to synthesize the year-on-year sequence of each stock market index. Obtain the first synthetic sequence of the stock market index.
  • the bond market index, commodity index, M1 (narrow money supply), M2, PMI, CPI, PPI, and economic prosperity index are used to obtain the corresponding first synthesis sequence using the same synthesis method.
  • the SUMPLE algorithm uses the SUMPLE algorithm to analyze financial indicators such as stock market index, bond market index, commodity index, M1 (narrow money supply), M2, PMI, CPI, PPI, and economic prosperity index (it needs to be understood that in this example
  • the financial indicators of is only an example. In actual use, it can be set according to actual needs) to synthesize the first synthetic sequence to obtain a synthetic financial time sequence.
  • the year-on-year sequence of each financial indicator data can be calculated first, and then the SUMPLE algorithm is used to perform the year-on-year sequence of financial indicator data of the same type. Synthesize to obtain the first synthetic sequence of each type of financial indicator data. Then the SUMPLE algorithm is further used to synthesize the first synthetic sequence of various financial indicator data to obtain a synthetic financial time series.
  • the SUMPLE algorithm is used to synthesize each financial indicator data of the same type to obtain the first composite sequence of each type of financial indicator data, and the SUMPLE algorithm is used
  • the first synthetic sequence of various financial indicator data is synthesized to obtain a synthetic financial time series, and a more accurate synthetic financial time series can be obtained.
  • FIG. 7 is a schematic flowchart of a sixth embodiment of a method for synthesizing financial time series of this application. Based on any one of the first to fourth embodiments, the step S20 includes:
  • Step S27 using the SUMPLE algorithm to synthesize various financial indicator data in the same area to obtain a second composite sequence of financial indicator data in each area;
  • step S28 the SUMPLE algorithm is used to synthesize the second synthetic sequence of the financial indicator data of each region to obtain a synthetic financial time sequence.
  • the global financial indicator data includes financial indicator data of multiple regions.
  • the financial indicator data of the same region exhibits extremely strong correlation behavior. It should be understood that the regions are regions classified according to financial and economic relevance, such as different countries or different large market economies.
  • the SUMPLE algorithm is used to synthesize each financial indicator data in the same area to obtain a second composite sequence of financial indicator data in each area.
  • the second synthetic sequence obtained by synthesis can reflect the system-level operation of financial indicator data in the same area.
  • the SUMPLE algorithm is further used to synthesize the second synthetic sequence of financial indicator data in each region to obtain a synthetic financial time sequence. For example, after obtaining multiple financial indicator data up to the observation time point, first use the SUMPLE algorithm to analyze the U.S.
  • the year-on-year sequence of each financial indicator data can be calculated first, and then the year-on-year sequence of each financial target in the same region in the financial indicator data can be obtained.
  • the SUMPLE algorithm is used to synthesize the year-on-year sequence of each of the financial targets to obtain the second synthetic sequence of the financial indicator data of each region. Then the SUMPLE algorithm is further used to synthesize the second synthetic sequence of financial indicator data in each region to obtain a synthetic financial time sequence.
  • the SUMPLE algorithm is used to synthesize each financial indicator data in the same area to obtain a second composite sequence of financial indicator data in each area, using SUMPLE
  • the algorithm synthesizes the second synthetic sequence of financial indicator data in various regions to obtain a synthetic financial time series, which can obtain a more accurate synthetic financial time series.
  • the technical solution of this application essentially or the part that contributes to the existing technology can be embodied in the form of a software product, and the computer software product is stored in a storage medium (such as ROM/RAM) as described above. , Magnetic disks, optical disks), including several instructions to make a terminal device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) execute the method described in each embodiment of the present application.
  • a terminal device which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Probability & Statistics with Applications (AREA)
  • Development Economics (AREA)
  • Fuzzy Systems (AREA)
  • Software Systems (AREA)
  • Computational Linguistics (AREA)
  • Quality & Reliability (AREA)
  • Accounting & Taxation (AREA)
  • Mathematical Physics (AREA)
  • Economics (AREA)
  • Finance (AREA)
  • Marketing (AREA)
  • Strategic Management (AREA)
  • Technology Law (AREA)
  • General Business, Economics & Management (AREA)
  • Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)

Abstract

一种金融时间序列的合成方法，所述金融时间序列的合成方法包括以下步骤：获取观测时间点，并获取截至所述观测时间点的多个金融指标数据（S10）；利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列（S20）。

Description

金融时间序列的合成方法、装置和存储介质
本申请要求2019年9月18日申请的，申请号为201910885881.2，名称为“金融时间序列的合成方法、装置和存储介质”的中国专利申请的优先权，在此将其全文引入作为参考。
技术领域
本申请涉及计算机技术领域，尤其涉及金融时间序列的合成方法、装置和存储介质。
背景技术
研究发现，金融系统中存在一个统一的运行周期，各个金融指标的运行均受到统一周期的影响，因此，对各个金融指标合成合成金融时间序列有利于金融经济的预测。目前，合成金融时间序列的合成一般通过等权合成或按照成交量加权合成；等权合成是赋予每个经济指标相同的权重，然后将它们合成得到合成金融时间序列；按照成交量加权合成则赋予成交量高的金融指标大的权重，进而合成合成金融时间序列。
然而，由于金融分析所采用的金融信号大多收到强烈的噪音干扰，信噪比较低，无论是等权合成还是加权合成，均无法识别并处理低信噪比的金融信号。
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金融时间序列的合成方法、装置和存储介质，旨在解决目前的合成方法无法识别并处理低信噪比的金融信号的技术问题。
为实现上述目的，本申请提供一种金融时间序列的合成方法，所述金融时间序列的合成方法包括以下步骤：
获取观测时间点，并获取截至所述观测时间点的多个金融指标数据；以 及
利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列。
此外，为实现上述目的，本申请还提供一种金融时间序列的合成装置，所述金融时间序列的合成装置包括：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金融时间序列的合成程序，所述金融时间序列的合成程序被所述处理器执行时实现如上任一项所述的金融时间序列的合成方法的步骤。
此外，为实现上述目的，本申请还提供一种存储介质，所述存储介质上存储有金融时间序列的合成程序，所述金融时间序列的合成程序被处理器执行时实现如上任一项所述的金融时间序列的合成方法的步骤。
本申请实施例提出的一种金融时间序列的合成方法、装置和存储介质，获取观测时间点，并获取截至所述观测时间点的多个金融指标数据；利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列。由于用于合成的SUMPLE算法适用于低信噪比的数据的合成，利用SUMPLE算法对各个同比序列合成时，可以计算出各个金融指标数据的同比序列的最优权值，从而使得到的合成金融时间序列信噪比高。
附图说明
图1是本申请实施例方案涉及的硬件运行环境的终端结构示意图；
图2为本申请金融时间序列的合成方法第一实施例的流程示意图；
图3为本申请金融时间序列的合成方法第二实施例的流程示意图；
图4为本申请金融时间序列的合成方法第三实施例的流程示意图；
图5为本申请金融时间序列的合成方法第四实施例的流程示意图：
图6为本申请金融时间序列的合成方法第五实施例的流程示意图；以及
图7为本申请金融时间序列的合成方法第六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：
获取观测时间点，并获取截至所述观测时间点的多个金融指标数据；
利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列。
由于现有技术中的合成方法无法识别并处理低信噪比的金融信号。
本申请提供一种解决方案，由于用于合成的SUMPLE算法适用于低信噪比的数据的合成，利用SUMPLE算法对各个同比序列合成时，可以计算出各个金融指标数据的同比序列的最优权值，从而使得到的合成金融时间序列信噪比高。
如图1所示，图1是本申请实施例方案涉及的硬件运行环境的终端结构示意图。
如图1所示，该终端可以包括：处理器1001，例如CPU，通信总线1002，用户接口1003，存储器1004。其中，通信总线1002用于实现这些组件之间的连接通信。用户接口1003可以包括显示屏(Display)、输入单元比如键盘(Keyboard)，可选用户接口1003还可以包括标准的有线接口、无线接口。存储器1004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1004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中示出的终端结构并不构成对终端的限定，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1004中可以包括操作系统、网络通信模块、用户接口模块以及金融时间序列的合成程序。
在图1所示的终端中，用户接口1003主要用于连接客户端(用户端)，与客户端进行数据通信；而处理器1001可以用于调用存储器1004中存储的金融时间序列的合成程序，并执行以下操作：
获取观测时间点，并获取截至所述观测时间点的多个金融指标数据；
利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列。
进一步地，处理器1001可以调用存储器1004中存储的金融时间序列的合成程序，还执行以下操作：
根据预设时间间隔计算各个所述金融指标数据的同比序列；
利用SUMPLE算法对各个所述同比序列进行合成，得到合成金融时间序列。
进一步地，处理器1001可以调用存储器1004中存储的金融时间序列的合成程序，还执行以下操作：
根据预设时间间隔计算各个所述金融指标数据的比值；
将各个所述比值的对数值作为所述金融指标数据的同比序列。
进一步地，处理器1001可以调用存储器1004中存储的金融时间序列的合成程序，还执行以下操作：
获取观测周期，根据所述观测周期分别对各个所述金融指标数据进行高斯滤波，得到各个所述金融指标数据的周期滤波序列；
利用SUMPLE算法对各个所述周期滤波序列进行合成，得到所述观测周期对应的合成金融时间序列。
进一步地，处理器1001可以调用存储器1004中存储的金融时间序列的合成程序，还执行以下操作：
获取各个所述金融指标数据的权重系数；
根据所述权重系数确定每个所述金融指标数据与所述合成金融时间序列的相位差。
进一步地，处理器1001可以调用存储器1004中存储的金融时间序列的合成程序，还执行以下操作：
利用SUMPLE算法对各个所述金融指标数据进行合成时，通过整体迭代或滚动迭代使各个所述金融指标数据对应的权重系数收敛。
进一步地，处理器1001可以调用存储器1004中存储的金融时间序列的合成程序，还执行以下操作：
利用SUMPLE算法对同一类型的各个金融指标数据进行合成，得到每一类金融指标数据的第一合成序列；
利用SUMPLE算法对各类金融指标数据的第一合成序列进行合成，得到合成金融时间序列。
进一步地，处理器1001可以调用存储器1004中存储的金融时间序列的合成程序，还执行以下操作：
利用SUMPLE算法对同一地区的各个金融指标数据进行合成，得到每一地区的金融指标数据的第二合成序列；
利用SUMPLE算法对各个地区的金融指标数据的第二合成序列进行合成，得到合成金融时间序列。
根据上述方案，获取观测时间点，并获取截至所述观测时间点的多个金融指标数据；利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列。由于用于合成的SUMPLE算法适用于低信噪比的数据的合成，利用SUMPLE算法对各个同比序列合成时，可以计算出各个金融指标数据的同比序列的最优权值，从而使得到的合成金融时间序列信噪比高。
参照图2，图2为本申请金融时间序列的合成方法第一实施例的流程示意图，所述金融时间序列的合成方法包括：
步骤S10，获取观测时间点，并获取截至所述观测时间点的多个金融指标数据；
本实施例的金融指标数据包括各类资产和各种经济指标，比如各个国家或地区(例如经济合作区域)的股票市场指数、债券市场指数以及大宗商品指数等，还包括各种宏观经济变量，比如各个主要国家的M1(狭义货币供应量)、M2(广义货币供应量)、PMI(采购经理指数)、CPI(消费者物价指数)、PPI(生产者物价指数)和经济景气指数等。
本实施例中，金融指标数据可以月为周期进行统计得到，可以使用金融数据的月度平均值或月末值作为金融指标数据。例如，每个月的股票市场指数可通过股票该月的平均价格或月末价格表示。
步骤S20，利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列。
通过研究发现，以股票市场来说，全球各个国家的股票市场表现出极强的相关性；另一方面，不同国家各金融经济指标之间也存在较强的相关性。这也就是说，在面对同一个全球经济金融环境，各个国家及主要市场都表现出了相关性极强的行为。
因此，本实施例中，通过SUMPLE算法对各个金融指标数据进行合成， 得到合成金融时间序列。具体地，每个金融指标数据可以表示为：
Figure PCTCN2020101798-appb-000001
式中k为时间变量，
Figure PCTCN2020101798-appb-000002
是第i个金融指标在k时刻的数据，
Figure PCTCN2020101798-appb-000003
为噪声。合成的权值系数表示为：
Figure PCTCN2020101798-appb-000004
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 PCTCN2020101798-appb-000005
为理想权值，
Figure PCTCN2020101798-appb-000006
是噪声引起的权值估计误差，则合成的合成金融时间序列可以表示为：
Figure PCTCN2020101798-appb-000007
其中*为取复共轭，L为金融指标数据的总个数。如果将合成输出表示成如下形式：
Figure PCTCN2020101798-appb-000008
那么，信号和噪声项分别为：
Figure PCTCN2020101798-appb-000009
Figure PCTCN2020101798-appb-000010
SUMPLE算法中的第K+1次的权值系数
Figure PCTCN2020101798-appb-000011
可由第K次的
Figure PCTCN2020101798-appb-000012
递推得到：
Figure PCTCN2020101798-appb-000013
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，它保证了各个金融指标数据的权值系数的平方和等于金融指标数据的数量，即
Figure PCTCN2020101798-appb-000014
上式的权值
Figure PCTCN2020101798-appb-000015
还可利用C k改写为：
Figure PCTCN2020101798-appb-000016
本实施例中，利用SUMPLE算法对各个金融指标数据进行合成时，通过预设次数的迭代使得各个同比序列的权值系数收敛，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算法对各个金融指标数据进行合成时，在迭代次数达到所述预设次数时，停止迭代，并输出合成金融时间序列。所述合成金融时间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算法对各个金融指标数据合成时，可以计算出各个金融指标数据的最优权值，从而使得到的合成金融时间序列信噪比更高。
本实施例中，在利用SUMPLE算法对各个金融指标数据进行合成时，可选用整体迭代或滚动迭代两种迭代方式进行迭代，使得各个金融指标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金融指标数据长度的限制。因此，本实施例中，优选地，利用SUMPLE算法对各个金融指标数据进行合成时，选用整体迭代的方式进行合成，可以保证各个金融指标数据的权值系数收敛。
本实施例中，获取观测时间点，并获取截至所述观测时间点的多个金融指标数据；利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列。由于用于合成的SUMPLE算法适用于低信噪比的数据的合成，利用SUMPLE算法对各个同比序列合成时，可以计算出各个金融指标数据的同比序列的最优权值，从而使得到的合成金融时间序列信噪比高。
进一步地，请参照图3，图3为本申请金融时间序列的合成方法第二实施例的流程示意图，基于第一实施例，所述步骤S20包括：
步骤S21，根据预设时间间隔计算各个所述金融指标数据的同比序列；
步骤S22，利用SUMPLE算法对各个所述同比序列进行合成，得到合成金融时间序列。
为了通过监测金融指标数据的变化来分析其周期变化规律，本实施例中，先对各个金融指标数据计算同比序列，然后再利用SUMPLE算法对同比序列进行合成。可选地，可分别计算同一金融指标数据中相隔预设时间间隔的两个金融数据的比值，将计算得到的各个比值作为该金融指标数据的同比序列。其中，所述预设时间间隔可根据实际情况自行设置，在此不做具体限制。可选地，所述预设时间间隔可设置为12个月，例如有1992年1月到2017年4月共304个月的金融指标数据，预设间隔设置为12个月，在获取到2017年4月的金融指标数据时，往前推12个月获取2016年4月的金融指标数据，计算2017年4月和2016年4月的金融指标数据的比值；在获取到2017年1月的金融指标数据时，往前推12个月获取2016年1月的金融指标数据，计算2017年1月和2016年1月的金融指标数据的比值；以此类推，计算得到同一金融指标数据的多个比值，将各个比值作为该金融指标数据的同比序列。可选地，在按照预设时间间隔计算得到各个金融数据的比值后，还可进一步计算各个所述比值的对数值，然后将各个所述比值的对数值作为所述金融数据的同比序列。可以理解的是，计算所述比值的对数值时，可以求取以10为底数的对数，也可以求取自然对数。可选地，所述金融指标数据可以以一个月为周期进行统计而得，例如股票的月度平均价格，债券的月度平均价格或者月度M1数据等。在将金融指标数据视为一个变量时，同比数据反映了该变量相比预设间隔的同期值的变化率，例如当预设间隔为12个月时，反映了该变量相比上一年的同期值的变化率。当金融指标数据以一个月为周期进行统计而得，并且预设间隔周期为12个月时，所获得的同比数据可以消除金融指标数据中的季节效应和月度效应，去除金融指标数据中可能干扰变化趋势判断结果的无效信息。
本实施例中，在获取到截至观察时间点的金融指标数据后，先根据预设时间间隔计算各个所述金融指标数据的同比序列，然后再利用SUMPLE算法对各个所述同比序列进行合成，得到合成金融时间序列，从而能够进一步减小噪音信号的影响。
进一步地，请参照图4，图4为本申请金融时间序列的合成方法第三实施 例的流程示意图，基于第一或第二实施例，所述步骤S20包括：
步骤S23，获取观测周期，根据所述观测周期分别对各个所述金融指标数据进行高斯滤波，得到各个所述金融指标数据的周期滤波序列；
步骤S24，利用SUMPLE算法对各个所述周期滤波序列进行合成，得到所述观测周期对应的合成金融时间序列。
在对金融指标数据研究中发现，不同类型的金融指标数据在统一的经济金融系统中拥有共同的经济周期，即本实施例中的观测周期，其对应的金融指标数据的同比序列在该观测周期内呈现有规律的变化。
本实施例中，在得到各个金融指标数后，获取观测周期，然后根据所述观测周期分别对各个所述金融指标数据进行高斯滤波，得到各个所述金融指标数据的周期滤波序列，以保留这些包含着资产的经济周期重要信息的目标频率信号，降低噪声的干扰。进一步地，利用SUMPLE算法对各个所述周期滤波序列进行合成，得到所述观测周期对应的合成金融时间序列。合成后得到的合成金融时间序列能够体现金融系统中统一的系统级别的周期运动，以供后续处理中更好地对各类金融指标数据的价格同比序列进行拟合。可选地，所述观测周期为42个月、100个月以及200个月中的一个，分别合成周期为42个月、100个月和200个月对应的合成金融时间序列，获取42个月、100个月和200个月对应的周期因子。
可选地，本实施例中，还可先计算各个金融指标数据的同比序列，然后根据观测周期分别对各个所述同比序列进行高斯滤波，得到各个同比序列的周期滤波序列。然后进一步利用SUMPLE算法对各个所述周期滤波序列进行合成，得到所述观测周期对应的合成金融时间序列。
本实施例中，在计算得到各个金融指标数据后，获取观测周期，根据所述观测周期分别对各个所述金融指标数据进行高斯滤波，得到各个所述金融指标数据的周期滤波序列，然后利用SUMPLE算法对各个所述周期滤波序列进行合成，得到所述观测周期对应的合成金融时间序列，能够使得后续处理中更好地对各类金融指标数据的价格同比序列进行拟合。
进一步地，请参照图5，图5为本申请金融时间序列的合成方法第四实施例的流程示意图，基于第一至第三实施例中的任一实施例，所述步骤S20之后，还包括：
步骤S30，获取各个所述金融指标数据的权重系数；
步骤S40，根据所述权重系数确定每个所述金融指标数据与所述合成金融时间序列的相位差。
在现实情况中，常需要对各个金融指标数据的变化趋势进行预测，各个金融指标数据的时间序列与合成的所述合成金融时间序列具有相同的变化规律和共同的经济周期，但各个金融指标数据与合成的金融指标数据的合成金融时间序列之间不会完全同步，二者存在相位差，即具体的金融指标数据相对于合成金融时间序列存在超前或滞后。
本实施例中，在利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列后，获取各个金融指标数据的权重系数，然后根据所述权重系数确定每个所述金融指标数据和所述合成金融时间序列的相位差。在利用SUMPLE算法进行合成时，所述权重系数为一个复数，所述权重系数(复数)的幅角角度即为所述同比序列对应的金融指标数据和所述合成金融时间序列的相位差角度。得到金融指标数据和所述合成金融时间序列的相位差后，即可根据合成金融时间序列和所述相位差确定各个金融指标数据的拟合时间序列，从而对各个金融指标进行预测。
本实施例中，利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列后，获取各个所述金融指标数据的权重系数，根据所述权重系数确定每个所述金融指标数据与所述合成金融时间序列的相位差；进而使得能够根据合成得到的合成金融时间序列和所述相位差对各个金融指标进行预测。
进一步地，请参照图6，图6为本申请金融时间序列的合成方法第五实施例的流程示意图，基于第一至第四实施例中的任一实施例，所述步骤S20包括：
步骤S25，利用SUMPLE算法对同一类型的各个金融指标数据进行合成，得到每一类金融指标数据的第一合成序列；
步骤S26，利用SUMPLE算法对各类金融指标数据的第一合成序列进行合成，得到合成金融时间序列。
在本实施例中，每一类金融指标中包括多个经济金融标的，例如，在股票指数包括多支股票数据。在研究中发现，在面对同一个全球经济金融环境， 对于每一类金融指标，其所包括的多个金融标的的同比序列数据表现出极强的相关性。例如，全球各个国家的股票市场表现出极强的相关性，不同国家各金融经济指标之间也存在较强的相关性。这也就是说，在面对同一个全球经济金融环境，各个国家及主要市场都表现出了相关性极强的行为。
因此，本实施例中，在获取到截至观测时间点的多个金融指标数据后，先利用SUMPLE算法对同一类型的各个金融指标数据进行合成，得到每一类金融指标数据的第一合成序列。合成得到的各个第一合成序列能够反映同一类金融指标数据的系统级别的运行。然后再进一步利用SUMPLE算法对各类金融指标数据的第一合成序列进行合成，得到合成金融时间序列。例如，在获取到截至观测时间点的多个金融指标数据后，获取金融指标数据中的各个股票市场指数(同属于股票市场指数类型)，利用SUMPLE算法对各个股票市场指数的同比序列进行合成，得到股票市场指数的第一合成序列。同样地，利用相同的合成方式得到债券市场指数、大宗商品指数、M1(狭义货币供应量)、M2、PMI、CPI、PPI和经济景气指数等对应的第一合成序列。然后进一步利用SUMPLE算法对股票市场指数、债券市场指数、大宗商品指数、M1(狭义货币供应量)、M2、PMI、CPI、PPI和经济景气指数等金融指标(需要理解的是，本实施例中的金融指标仅为举例，实际使用中可根据实际需要进行设置)的第一合成序列进行合成，得到合成金融时间序列。
可选地，本实施例中，获取到截至观测时间点的多个金融指标数据后，可先计算各个金融指标数据的同比序列，然后利用SUMPLE算法对属于同一类型的金融指标数据的同比序列进行合成，得到每一类金融指标数据的第一合成序列。然后再进一步利用SUMPLE算法对各类金融指标数据的第一合成序列进行合成，得到合成金融时间序列。
本实施例中，在获取到截至观测时间点的多个金融指标数据后，利用SUMPLE算法对同一类型的各个金融指标数据进行合成，得到每一类金融指标数据的第一合成序列，利用SUMPLE算法对各类金融指标数据的第一合成序列进行合成，得到合成金融时间序列，可以得到更加精确的合成金融时间序列。
进一步地，请参照图7，图7为本申请金融时间序列的合成方法第六实施例的流程示意图，基于第一至第四实施例中的任一实施例，所述步骤S20包 括：
步骤S27，利用SUMPLE算法对同一地区的各个金融指标数据进行合成，得到每一地区的金融指标数据的第二合成序列；
步骤S28，利用SUMPLE算法对各个地区的金融指标数据的第二合成序列进行合成，得到合成金融时间序列。
本实施例中，全球的金融指标数据包括多个地区的金融指标数据，在面对同一个全球经济金融环境，同一个地区的金融指标数据表现出极强的相关性行为。需要理解的是，所述地区为根据金融经济相关性划分的地区，例如不同国家或不同的大市场经济。
因此，本实施例中，在获取到截至观测时间点的多个金融指标数据后，先利用SUMPLE算法对同一地区的各个金融指标数据进行合成，得到每一地区的金融指标数据的第二合成序列，合成得到的第二合成序列能够反映同一地区的金融指标数据的系统级别的运行。然后再进一步利用SUMPLE算法对各个地区的金融指标数据的第二合成序列进行合成，得到合成金融时间序列。例如，在获取到截至观测时间点的多个金融指标数据后，先利用SUMPLE算法对美国的股票市场指数、债券市场指数、大宗商品指数、M1(狭义货币供应量)、M2、PMI、CPI、PPI和经济景气指数等金融指标数据进行合成，得到美国的金融指标数据的第二合成序列；同样地，利用相同的方式合成得到中国、日本等地区(需要理解的是，本实施例中的地区仅为举例，实际使用中可根据实际需要进行设置)的金融指标数据的第二合成序列。然后再进一步利用SUMPLE算法对美国、中国和日本等地区的金融指标数据的第二合成序列进行合成，得到合成金融时间序列。
可选地，本实施例中，获取到截至观测时间点的多个金融指标数据后，可先计算各个金融指标数据的同比序列，然后获取金融指标数据中同一地区的各个金融标的的同比序列，利用SUMPLE算法对各个所述金融标的的同比序列进行合成，得到每一地区的金融指标数据的第二合成序列。然后再进一步利用SUMPLE算法对各个地区的金融指标数据的第二合成序列进行合成，得到合成金融时间序列。
本实施例中，在获取到截至观测时间点的多个金融指标数据后，利用SUMPLE算法对同一地区的各个金融指标数据进行合成，得到每一地区的金 融指标数据的第二合成序列，利用SUMPLE算法对各个地区的金融指标数据的第二合成序列进行合成，得到合成金融时间序列，可以得到更加精确的合成金融时间序列。
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。
以上仅为本申请的可选实施例，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 (20)

  1. 一种金融时间序列的合成方法，其中，所述金融时间序列的合成方法包括以下步骤：
    获取观测时间点，并获取截至所述观测时间点的多个金融指标数据；以及
    利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列。
  2. 如权利要求1所述的金融时间序列的合成方法，其中，所述利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列的步骤包括：
    根据预设时间间隔计算各个所述金融指标数据的同比序列；以及
    利用SUMPLE算法对各个所述同比序列进行合成，得到合成金融时间序列。
  3. 如权利要求2所述的金融时间序列的合成方法，其中，所述根据预设时间间隔计算各个所述金融指标数据的同比序列的步骤包括：
    根据预设时间间隔计算各个所述金融指标数据的比值；以及
    将各个所述比值的对数值作为所述金融指标数据的同比序列。
  4. 如权利要求1所述的金融时间序列的合成方法，其中，所述利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列的步骤包括：
    获取观测周期，根据所述观测周期分别对各个所述金融指标数据进行高斯滤波，得到各个所述金融指标数据的周期滤波序列；以及
    利用SUMPLE算法对各个所述周期滤波序列进行合成，得到所述观测周期对应的合成金融时间序列。
  5. 如权利要求1所述的金融时间序列的合成方法，其中，所述利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列的步骤之后，还包括：
    获取各个所述金融指标数据的权重系数；以及
    根据所述权重系数确定每个所述金融指标数据与所述合成金融时间序列 的相位差。
  6. 如权利要求1所述的金融时间序列的合成方法，其中，利用SUMPLE算法对各个所述金融指标数据进行合成时，通过整体迭代或滚动迭代使各个所述金融指标数据对应的权重系数收敛。
  7. 如权利要求1-6中任一项所述的金融时间序列的合成方法，其中，所述利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列的步骤包括：
    利用SUMPLE算法对同一类型的各个金融指标数据进行合成，得到每一类金融指标数据的第一合成序列；以及
    利用SUMPLE算法对各类金融指标数据的第一合成序列进行合成，得到合成金融时间序列。
  8. 如权利要求1-6中任一项所述的金融时间序列的合成方法，其中，利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列的步骤包括：
    利用SUMPLE算法对同一地区的各个金融指标数据进行合成，得到每一地区的金融指标数据的第二合成序列；以及
    利用SUMPLE算法对各个地区的金融指标数据的第二合成序列进行合成，得到合成金融时间序列。
  9. 一种金融时间序列的合成装置，其中，所述金融时间序列的合成装置包括：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金融时间序列的合成程序，所述金融时间序列的合成程序被所述处理器执行时实现以下步骤：
    获取观测时间点，并获取截至所述观测时间点的多个金融指标数据；以及
    利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列。
  10. 如权利要求9所述的金融时间序列的合成装置，其中，所述金融时间序列的合成程序被所述处理器执行时实现的所述利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列的步骤包括：
    根据预设时间间隔计算各个所述金融指标数据的同比序列；以及
    利用SUMPLE算法对各个所述同比序列进行合成，得到合成金融时间序列。
  11. 如权利要求10所述的金融时间序列的合成装置，其中，所述金融时间序列的合成程序被所述处理器执行时实现的所述根据预设时间间隔计算各个所述金融指标数据的同比序列的步骤包括：
    根据预设时间间隔计算各个所述金融指标数据的比值；以及
    将各个所述比值的对数值作为所述金融指标数据的同比序列。
  12. 如权利要求9所述的金融时间序列的合成装置，其中，所述金融时间序列的合成程序被所述处理器执行时实现的所述利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列的步骤包括：
    获取观测周期，根据所述观测周期分别对各个所述金融指标数据进行高斯滤波，得到各个所述金融指标数据的周期滤波序列；以及
    利用SUMPLE算法对各个所述周期滤波序列进行合成，得到所述观测周期对应的合成金融时间序列。
  13. 如权利要求9所述的金融时间序列的合成装置，其中，所述利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列的步骤之后，所述金融时间序列的合成程序被所述处理器执行时还实现以下步骤：
    获取各个所述金融指标数据的权重系数；以及
    根据所述权重系数确定每个所述金融指标数据与所述合成金融时间序列的相位差。
  14. 如权利要求9所述的金融时间序列的合成装置，其中，利用SUMPLE算法对各个所述金融指标数据进行合成时，通过整体迭代或滚动迭代使各个所述金融指标数据对应的权重系数收敛。
  15. 一种存储介质，其中，所述存储介质上存储有金融时间序列的合成程序，所述金融时间序列的合成程序被处理器执行时实现以下步骤：
    获取观测时间点，并获取截至所述观测时间点的多个金融指标数据；以及
    利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列。
  16. 如权利要求15所述的存储介质，其中，所述金融时间序列的合成程序被处理器执行时实现的所述利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列的步骤包括：
    根据预设时间间隔计算各个所述金融指标数据的同比序列；以及
    利用SUMPLE算法对各个所述同比序列进行合成，得到合成金融时间序列。
  17. 如权利要求16所述的存储介质，其中，所述金融时间序列的合成程序被处理器执行时实现的所述根据预设时间间隔计算各个所述金融指标数据的同比序列的步骤包括：
    根据预设时间间隔计算各个所述金融指标数据的比值；以及
    将各个所述比值的对数值作为所述金融指标数据的同比序列。
  18. 如权利要求15所述的存储介质，其中，所述金融时间序列的合成程序被处理器执行时实现的所述利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列的步骤包括：
    获取观测周期，根据所述观测周期分别对各个所述金融指标数据进行高斯滤波，得到各个所述金融指标数据的周期滤波序列；以及
    利用SUMPLE算法对各个所述周期滤波序列进行合成，得到所述观测周期对应的合成金融时间序列。
  19. 如权利要求15所述的存储介质，其中，所述利用SUMPLE算法对各个所述金融指标数据进行合成，得到合成金融时间序列的步骤之后，所述金融时间序列的合成程序被处理器执行时还实现以下步骤：
    获取各个所述金融指标数据的权重系数；以及
    根据所述权重系数确定每个所述金融指标数据与所述合成金融时间序列的相位差。
  20. 如权利要求15所述的存储介质，其中，利用SUMPLE算法对各个所述金融指标数据进行合成时，通过整体迭代或滚动迭代使各个所述金融指标数据对应的权重系数收敛。
PCT/CN2020/101798 2019-09-18 2020-07-14 金融时间序列的合成方法、装置和存储介质 WO2021051976A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910885881.2A CN110688365A (zh) 2019-09-18 2019-09-18 金融时间序列的合成方法、装置和存储介质
CN201910885881.2 2019-09-18

Publications (1)

Publication Number Publication Date
WO2021051976A1 true WO2021051976A1 (zh) 2021-03-25

Family

ID=69109613

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/101798 WO2021051976A1 (zh) 2019-09-18 2020-07-14 金融时间序列的合成方法、装置和存储介质

Country Status (2)

Country Link
CN (1) CN110688365A (zh)
WO (1) WO2021051976A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110688365A (zh) * 2019-09-18 2020-01-14 华泰证券股份有限公司 金融时间序列的合成方法、装置和存储介质

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120290501A1 (en) * 2011-05-13 2012-11-15 Chao-Hung Chang Generating method for transaction models with indicators for option
CN105980941A (zh) * 2014-02-10 2016-09-28 欧姆龙株式会社 监视装置及监视方法
CN106973023A (zh) * 2017-02-14 2017-07-21 中国科学院声学研究所 一种信号合成方法及系统
CN107425928A (zh) * 2017-05-05 2017-12-01 中国科学院声学研究所 一种信号合成方法及系统
CN108615096A (zh) * 2018-05-10 2018-10-02 平安科技（深圳）有限公司 服务器、金融时序数据的处理方法及存储介质
CN109492838A (zh) * 2019-01-16 2019-03-19 中国地质大学（武汉） 一种基于深度循环神经网络的股指价格预测方法
CN109933834A (zh) * 2018-12-26 2019-06-25 阿里巴巴集团控股有限公司 一种时序数据预测的模型创建方法及装置
CN110688365A (zh) * 2019-09-18 2020-01-14 华泰证券股份有限公司 金融时间序列的合成方法、装置和存储介质

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7702548B2 (en) * 2000-05-01 2010-04-20 Zumbach Gilles O Methods for analysis of financial markets
CN106658638B (zh) * 2016-12-16 2020-02-07 天津大学 一种基于hh-vbf的时间序列预测方法
CN107067096A (zh) * 2016-12-27 2017-08-18 河南理工大学 基于分形与混沌理论相结合的金融时间序列短期预测

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120290501A1 (en) * 2011-05-13 2012-11-15 Chao-Hung Chang Generating method for transaction models with indicators for option
CN105980941A (zh) * 2014-02-10 2016-09-28 欧姆龙株式会社 监视装置及监视方法
CN106973023A (zh) * 2017-02-14 2017-07-21 中国科学院声学研究所 一种信号合成方法及系统
CN107425928A (zh) * 2017-05-05 2017-12-01 中国科学院声学研究所 一种信号合成方法及系统
CN108615096A (zh) * 2018-05-10 2018-10-02 平安科技（深圳）有限公司 服务器、金融时序数据的处理方法及存储介质
CN109933834A (zh) * 2018-12-26 2019-06-25 阿里巴巴集团控股有限公司 一种时序数据预测的模型创建方法及装置
CN109492838A (zh) * 2019-01-16 2019-03-19 中国地质大学（武汉） 一种基于深度循环神经网络的股指价格预测方法
CN110688365A (zh) * 2019-09-18 2020-01-14 华泰证券股份有限公司 金融时间序列的合成方法、装置和存储介质

Also Published As

Publication number Publication date
CN110688365A (zh) 2020-01-14

Similar Documents

Publication Publication Date Title
Li et al. The co-movement and causality between the US housing and stock markets in the time and frequency domains
Andersen et al. Primal-dual simulation algorithm for pricing multidimensional American options
US11928620B2 (en) Method for estimating amount of task objects required to reach target completed tasks
WO2021103572A1 (zh) 资产投资建议信息的生成方法、装置和可读存储介质
US20140279304A1 (en) Method, System and Program Product for Matching of Transaction Records
US10521437B2 (en) Resource portfolio processing method, device, apparatus and computer storage medium
US20210342867A1 (en) Methods for predicting economic state and establishing economic state prediction model and corresponding apparatuses
Chen et al. Performance of information criteria for selection of Hawkes process models of financial data
Barhoumi et al. Testing the number of factors: An empirical assessment for a forecasting purpose
WO2023050649A1 (zh) 基于数据补全的esg指数确定方法及相关产品
Nason et al. Should we sample a time series more frequently?: decision support via multirate spectrum estimation
WO2021051976A1 (zh) 金融时间序列的合成方法、装置和存储介质
WO2020000715A1 (zh) 基于指数特征提取的股指预测方法、服务器及存储介质
WO2021103571A1 (zh) 资产投资建议信息的生成方法、装置和可读存储介质
CN116362823A (zh) 用于行为稀疏场景的推荐模型训练方法、推荐方法及装置
JP2009032237A (ja) ポートフォリオの信用リスクの計算方法および装置
KR20180059468A (ko) 글로벌적으로 검색된 데이터를 기반으로 한 글로벌 비즈니스 랭킹의 실시간 생성을 위한 글로벌 네트워킹 시스템
Jiang et al. Intertemporal pricing via nonparametric estimation: Integrating reference effects and consumer heterogeneity
CN116630051A (zh) 收益率曲线和债券估值实时计算系统、方法及设备
Ren et al. A multi-market comparison of the intraday lead–lag relations among stock index-based spot, futures and options
Ye et al. Two-step estimation of the volatility functions in diffusion models with empirical applications
Morris et al. Further evidence of long memory in the South African stock market
Lu et al. Testing unit roots and long range dependence of foreign exchange
US20130262340A1 (en) Equipment for data processing and a method for determining the weightings of constituents of a target portfolio
Jang Structural estimation of the New-Keynesian model: A formal test of backward-and forward-looking behavior

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: 20865569

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: 20865569

Country of ref document: EP

Kind code of ref document: A1