Time series data processing method and device, storage medium and electronic equipment

Abstract

The invention discloses a time series data processing method and device, a storage medium and electronic equipment, and relates to the field of financial science and technology. The method comprises the steps of setting a sliding window for a target data sequence; based on the target data sequence, sliding the sliding window M times from front to back to obtain M sub-data sequences; taking data in the middle position in each sub-data sequence as target data of the sub-data sequence; determining correction data of each sub-data sequence according to the sliding window; and replacing the target data of each sub-data sequence with the correction data of the sub-data sequence, and performing synchronous data updating on the target data sequence according to the target data replacement result of each sub-time sequence. According to the method and the device, the technical problem that in the prior art, financial data in a time sequence form is mainly analyzed in a manual mode, so that fund loss is easily caused to a transaction party due to the fact that an analysis result is not determined in time in a scene with very strong transaction timeliness is solved.

Description

Technical Field

[0001] The present application relates to the field of financial technology, and specifically, to a method, device, storage medium and electronic device for processing time series data. Background Art

[0002] In the financial field, time series data such as stock prices, futures contracts, foreign exchange rates, gold quotes, etc., form the basis of market analysis and forecasting. However, traditionally, the analysis of this data often relies on the intuition and experience of human experts to make decisions by manually drawing charts, identifying trends, and calculating complex statistical indicators. Although this method can provide deep insights in some cases, its inherent inefficiency and high latency limit its application in high-speed trading environments.

[0003] As the speed of financial market transactions continues to accelerate, especially with the rise of high-frequency trading, the dynamic changes in financial data require analytical algorithms to have the ability to respond in real time. Manual analysis is difficult to keep up with the pace of data updates, resulting in a significant lag in the determination of analytical results. This lag is particularly fatal in time-sensitive trading scenarios. Traders may make decisions based on outdated or insufficient information, which not only reduces trading efficiency, but also increases systemic risks, which may lead to significant losses of funds.

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

[0005] The present application provides a method, device, storage medium and electronic device for processing time series data, so as to at least solve the technical problem that in the prior art, financial data in time series form is mainly analyzed manually, resulting in low analysis efficiency. As a result, in some transaction scenarios with strong timeliness, problems such as untimely determination of analysis results are prone to occur, which may cause systemic risks in financial transactions and cause financial losses to the transaction parties.

[0006] According to one aspect of the present application, a method for processing time series data is provided, comprising: setting a sliding window for a target data sequence, wherein the target data sequence includes N data collected at N moments, N is an integer greater than 1, and the sliding window is used to determine the number of data selected continuously from the target data sequence each time; based on the target data sequence, sliding the sliding window M times from front to back to obtain M sub-data sequences; taking the data in the middle position of each sub-data sequence as the target data of the sub-data sequence; determining the correction data of each sub-data sequence according to the sliding window, wherein the correction data is used to reduce the volatility between multiple data in each sub-data sequence; replacing the target data of each sub-data sequence with the correction data of the sub-data sequence, and synchronously updating the target data sequence according to the target data replacement result of each sub-time series.

[0007] Optionally, setting a sliding window for the target data sequence includes: calculating a standard deviation of N data in the target data sequence; and determining the sliding window according to the standard deviation.

[0008] Optionally, determining the sliding window according to the standard deviation includes: obtaining the fluctuation range and fluctuation frequency of N data in the target data sequence; adjusting the standard deviation according to the fluctuation range and fluctuation frequency of the N data, and using the adjustment result as the sliding window.

[0009] Optionally, based on the target data sequence, sliding the sliding window M times from front to back to obtain M sub-data sequences, including: determining the number of sliding times of the sliding window according to the total number of data in the sliding window and the target data sequence; based on the target data sequence, sliding the sliding window M times from front to back according to the target sliding strategy and the number of sliding times, to obtain M sub-data sequences, wherein M represents the number of sliding times, and the target sliding strategy includes: taking the first element in the target data sequence as the first element of the first sub-data sequence, and determining the target data corresponding to the first sub-data sequence according to the sliding window; calculating the difference between the total number of data in the target data sequence and the sliding window, and determining the target data corresponding to the last sub-data sequence according to the difference.

[0010] Optionally, the corrected data of each sub-data sequence is determined according to the sliding window, including: taking each data except the target data in the i-th sub-data sequence as the first data, where i is a positive integer less than or equal to M; determining the weight value corresponding to each first data according to the difference between each first data and the target data and the sliding window, where the weight value corresponding to each first data is less than the weight value corresponding to the target data; determining the corrected data of each sub-data sequence according to the weight values ​​corresponding to all the data in the i-th sub-data sequence.

[0011] Optionally, the corrected data of each sub-data sequence is determined according to the weight values ​​corresponding to all the data in the i-th sub-data sequence, including: summing the weight values ​​corresponding to all the data in the i-th sub-data sequence to obtain a first numerical value; calculating the ratio of the weight value corresponding to each data in the i-th sub-data sequence to the first numerical value to obtain a target weight value corresponding to each data in the i-th sub-data sequence; and determining the corrected data of each sub-data sequence according to the target weight value corresponding to each data in the i-th sub-data sequence.

[0012] Optionally, the corrected data of each sub-data sequence is determined according to the target weight value corresponding to each data in the i-th sub-data sequence, including: performing weighted average calculation on all data in the i-th sub-data sequence according to the target weight value corresponding to each data in the i-th sub-data sequence, and determining the corrected data of the i-th sub-data sequence with the calculation result.

[0013] Optionally, the method for processing time series data also includes: when it is detected that the data source is switched or the data collection time exceeds a preset time, an update operation is performed on the target data sequence, wherein the update operation is used to delete at least one historical data with the earliest collection time in the target data sequence, and add the newly collected data to the target data sequence.

[0014] According to another aspect of the present application, a time series data processing device is also provided, including: a first setting unit, setting a sliding window for a target data sequence, wherein the target data sequence includes N data collected at N time moments, N is an integer greater than 1, and the sliding window is used to determine the number of data selected continuously from the target data sequence each time; a first processing unit, based on the target data sequence, sliding the sliding window M times from front to back to obtain M sub-data sequences; a first determination unit, taking the data in the middle position of each sub-data sequence as the target data of the sub-data sequence; a second determination unit, determining the correction data of each sub-data sequence according to the sliding window, wherein the correction data is used to reduce the volatility between multiple data in each sub-data sequence; a second processing unit, replacing the target data of each sub-data sequence with the correction data of the sub-data sequence, and synchronously updating the target data sequence according to the target data replacement result of each sub-time series.

[0015] Optionally, the first setting unit includes: a first calculating subunit and a first determining subunit, wherein the first calculating subunit is used to calculate the standard deviation of N data in the target data sequence; and the first determining subunit is used to determine the sliding window according to the standard deviation.

[0016] Optionally, the first determination subunit includes: a first acquisition module and a first determination module. The first acquisition module is used to acquire the fluctuation range and fluctuation frequency of N data in the target data sequence; the first determination module is used to adjust the standard deviation according to the fluctuation range and fluctuation frequency of the N data, and use the adjustment result as the sliding window.

[0017] Optionally, the first processing unit includes: a second determining subunit and a first processing subunit. The second determining subunit is used to determine the number of sliding times of the sliding window according to the total number of data in the sliding window and the target data sequence; the first processing subunit is used to slide the sliding window M times from front to back based on the target data sequence according to the target sliding strategy and the number of sliding times to obtain M sub-data sequences, wherein M represents the number of sliding times, and the target sliding strategy includes: taking the first element in the target data sequence as the first element of the first sub-data sequence, and determining the target data corresponding to the first sub-data sequence according to the sliding window; calculating the difference between the total number of data in the target data sequence and the sliding window, and determining the target data corresponding to the last sub-data sequence according to the difference.

[0018] Optionally, the second determination unit includes: a third determination subunit, a fourth determination subunit, and a fifth determination subunit. The third determination subunit is used to take each data except the target data in the i-th sub-data sequence as the first data, wherein i is a positive integer less than or equal to M; the fourth determination subunit is used to determine the weight value corresponding to each first data according to the difference between each first data and the target data and the sliding window, wherein the weight value corresponding to each first data is less than the weight value corresponding to the target data; the fifth determination subunit is used to determine the corrected data of each sub-data sequence according to the weight values ​​corresponding to all the data in the i-th sub-data sequence.

[0019] Optionally, the fifth determination subunit includes: a first calculation subunit, a second calculation subunit, and a sixth determination subunit. The first calculation subunit is used to sum the weight values ​​corresponding to all the data in the i-th sub-data sequence to obtain a first value; the second calculation subunit is used to calculate the ratio of the weight value corresponding to each data in the i-th sub-data sequence to the first value to obtain a target weight value corresponding to each data in the i-th sub-data sequence; the sixth determination subunit is used to determine the correction data of each sub-data sequence according to the target weight value corresponding to each data in the i-th sub-data sequence.

[0020] Optionally, the sixth determination subunit includes: a second determination module, used to perform weighted average calculation on all data in the i-th sub-data sequence according to the target weight value corresponding to each data in the i-th sub-data sequence, and determine the corrected data of the i-th sub-data sequence based on the calculation result.

[0021] Optionally, the time series data processing device also includes: an updating unit, which is used to update the target data sequence when it is detected that the data source is switched or the data collection time exceeds a preset time, wherein the updating operation is used to delete at least one historical data with the earliest collection time in the target data sequence, and add the newly collected data to the target data sequence.

[0022] According to another aspect of the present application, a computer-readable storage medium is also provided, wherein a computer program is stored in the computer-readable storage medium, wherein when the computer program is executed, the device where the computer-readable storage medium is located executes the above-mentioned method for processing time series data.

[0023] According to another aspect of the present application, an electronic device is also provided, which includes one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors execute the above-mentioned method for processing time series data.

[0024] According to another aspect of the present application, a computer program product is also provided, including computer instructions, which implement the steps of the above-mentioned time series data processing method when executed by a processor.

[0025] In the present application, a sliding window is first set for a target data sequence, wherein the target data sequence includes N data collected at N moments, N being an integer greater than 1, and the sliding window is used to determine the number of data selected continuously from the target data sequence each time, and then based on the target data sequence, the sliding window is slid M times from front to back to obtain M sub-data sequences, and then the data in the middle position of each sub-data sequence is used as the target data of the sub-data sequence, and then the correction data of each sub-data sequence is determined according to the sliding window, wherein the correction data is used to reduce the volatility between multiple data in each sub-data sequence, and finally the target data of each sub-data sequence is replaced by the sub-data sequence. The corrected data of the data series is obtained, and the target data series is synchronously updated according to the target data replacement result of each sub-time series, that is, the processing method of dynamic window smoothing data is used to achieve the purpose of improving the accuracy and real-time performance of financial time series data analysis, thereby realizing the technical effect of accelerating transaction decisions and reducing system risks, and further solving the problem that the financial data in the form of time series in the prior art is mainly analyzed manually, resulting in low analysis efficiency, and thus in some transaction scenarios with strong timeliness, it is easy to have problems due to untimely determination of analysis results, which may cause systemic risks of financial transactions and cause financial losses to the transaction parties. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0027] Figure 1 is a flowchart of an optional method for processing time series data according to an embodiment of the present application;

[0028] Figure 2 is a schematic diagram of an optional method for processing time series data according to an embodiment of the present application;

[0029] Figure 3 It is a schematic diagram of an optional time series data processing device according to an embodiment of the present application. DETAILED DESCRIPTION

[0030] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0031] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0032] It should be noted that the collected information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for display, data for analysis, etc.) involved in this application are information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of relevant data are in compliance with relevant laws, regulations and standards, necessary confidentiality measures are taken, and public order and good customs are not violated, and corresponding operation entrances are provided for users to choose to authorize or refuse. For example, an interface is set up between this system and relevant users or institutions to provide users with corresponding operation entrances for users to choose to agree or refuse the results of automated decision-making; if the user chooses to refuse, the expert decision-making process will be entered.

[0033] It should be noted that an intelligent processing system can be used as the execution subject of the time series data processing method of the embodiment of the present application. It is understandable that the time series data processing method provided in the embodiment of the present application can also be executed by other systems or devices, and the embodiment of the present application does not specifically limit this.

[0034] According to an embodiment of the present application, a method embodiment of a method for processing time series data is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0035] Figure 1 is a flowchart of an optional method for processing time series data according to an embodiment of the present application. Figure 1 As shown, the method comprises the following steps:

[0036] Step S101, setting a sliding window for a target data sequence.

[0037] In step S101, the target data sequence includes N data collected at N moments, where N is an integer greater than 1, and the sliding window is used to determine the number of data continuously selected from the target data sequence each time.

[0038] Optionally, the target data sequence is a time series, in which data is arranged in chronological order, such as stock prices and exchange rates in financial markets, and the sequence contains N data points, which are collected at N different time points.

[0039] Optionally, a sliding window is a tool for calculating local features of a data sequence. It defines a fixed-size data range for continuous selection of data from the sequence. By setting a sliding window, the amount of data analyzed each time can be controlled, which is particularly useful when processing large amounts of time series data. The size of the window directly affects the accuracy and efficiency of the calculation.

[0040] Step S102: based on the target data sequence, slide the sliding window M times from front to back to obtain M sub-data sequences.

[0041] Optionally, the intelligent processing system applies a sliding window to the target data sequence, starting from the beginning of the sequence and moving forward according to a certain step size until the end of the sequence. M is the number of slides, which is also the number of sub-data sequences generated, which depends on the length of the sequence and the size of the window.

[0042] Optionally, the sub-data sequence refers to a data set within each sliding window forming a sub-data sequence, which is a local view of the original sequence for subsequent analysis and processing.

[0043] Step S103: taking the data at the middle position in each sub-data sequence as the target data of the sub-data sequence.

[0044] Optionally, in each sub-data sequence, the intelligent processing system uses the data in the middle position of each sub-data sequence as a key data point representing the sequence, that is, the target data.

[0045] Step S104, determining the correction data of each sub-data sequence according to the sliding window.

[0046] In step S104, the correction data is used to reduce the fluctuation rate between the multiple data in each sub-data sequence.

[0047] Optionally, the modified data is obtained by adjusting or calculating the data in the sub-data sequence, with the purpose of reducing the fluctuation between the data, making it smoother and easier to analyze or predict. For example, it can be determined by weighted average or other statistical methods.

[0048] Step S105: replace the target data of each sub-data sequence with the revised data of the sub-data sequence, and perform synchronous data update on the target data sequence according to the target data replacement result of each sub-time series.

[0049] Optionally, the intelligent processing system updates the target data of each subsequence to its revised data, which is essentially extracting and optimizing the features of the entire data sequence. The updated data sequence is more stable, less volatile, and more suitable for long-term trend analysis or short-term prediction.

[0050] Optionally, the intelligent system can use a Gaussian filter algorithm to process time series data. It should be noted that the core of the Gaussian filter algorithm is to use a Gaussian function as a weight distribution to perform a weighted average on each point in the data sequence to reduce the impact of data fluctuations and market noise while retaining the overall trend and important features of the data. Specifically for financial data processing, the workflow of the Gaussian filter algorithm includes window setting, weight calculation, weighted average, and moving window.

[0051] It can be seen from the contents of step S101 to step S105 that in the present application, a sliding window is first set for the target data sequence, wherein the target data sequence includes N data collected at N moments, N is an integer greater than 1, and the sliding window is used to determine the number of data continuously selected from the target data sequence each time, and then based on the target data sequence, the sliding window is slid M times from front to back to obtain M sub-data sequences, and then the data in the middle position of each sub-data sequence is used as the target data of the sub-data sequence, and then the correction data of each sub-data sequence is determined according to the sliding window, wherein the correction data is used to reduce the volatility between multiple data in each sub-data sequence, and finally each sub-data sequence is The target data of each sub-time series is replaced by the corrected data of the sub-data series, and the target data series is updated synchronously according to the target data replacement result of each sub-time series, that is, the processing method of dynamic window smoothing data is used to achieve the purpose of improving the accuracy and real-time performance of financial time series data analysis, thereby achieving the technical effect of accelerating transaction decisions and reducing system risks, and further solving the problem that in the prior art, financial data in the form of time series are mainly analyzed manually, resulting in low analysis efficiency, and thus in some transaction scenarios with strong timeliness, the problem of untimely determination of analysis results is prone to occur, which may cause systemic risks of financial transactions and cause financial losses to the transaction parties.

[0052] In an optional embodiment, the intelligent processing system first calculates the standard deviation of N data in the target data sequence, and then determines the sliding window according to the standard deviation.

[0053] Optionally, the intelligent processing system first performs a statistical analysis on the target data sequence and calculates the standard deviation of N data in the sequence. The standard deviation is an important indicator for measuring the degree of data dispersion. It reflects the difference between each data point in the data sequence and the sequence average. After calculating the standard deviation, the intelligent processing system dynamically determines the size of the sliding window according to the size of the standard deviation. The size of the sliding window is proportional to the degree of data fluctuation, that is, the greater the data fluctuation, the larger the size of the sliding window, which is conducive to ensuring the accuracy and stability of the analysis while maintaining the flexibility of data processing. Determining the size of the sliding window by the standard deviation can more accurately capture the local features in the data sequence and reduce the analysis errors that may be caused by the fixed window size.

[0054] From the above content, it can be seen that the intelligent processing system can automatically adjust the size of the sliding window according to the volatility characteristics of the target data sequence, which not only improves the flexibility of data processing, but also ensures the accuracy and stability of the analysis results. When processing data sequences with high volatility and timeliness such as financial data, the dynamic sliding window can more effectively capture the dynamic changes of the sequence and reduce short-term fluctuations or long-term trends that may be ignored due to the fixed analysis window, thereby improving the efficiency and reliability of financial data analysis. This method is particularly suitable for trading scenarios that require rapid response to market changes and reduce analysis errors. It can significantly reduce the systemic risks caused by untimely data analysis and reduce potential capital losses.

[0055] In an optional embodiment, the intelligent processing system first obtains the fluctuation range and fluctuation frequency of N data in the target data sequence, and then adjusts the standard deviation according to the fluctuation range and fluctuation frequency of the N data, and uses the adjustment result as a sliding window.

[0056] Optionally, the intelligent processing system first analyzes the target data sequence to obtain the fluctuation range and fluctuation frequency of N data points in the sequence, wherein the fluctuation range refers to the difference between the maximum and minimum values ​​of the data points in the sequence, and the fluctuation frequency reflects the frequency of changes in the data points. This analysis provides a basis for the subsequent adjustment of the sliding window size. Then, based on the obtained fluctuation range and fluctuation frequency, the intelligent processing system adjusts the standard deviation. The standard deviation is a statistical indicator that measures the degree of difference between data points in the data sequence and their average value. The purpose of adjusting the standard deviation is to better reflect the fluctuation characteristics of the data sequence and ensure that the size of the sliding window can adapt to the dynamic changes of the data, so as to more effectively suppress fluctuations in subsequent analysis, improve data stability and prediction accuracy, and then use the adjusted standard deviation value as the basis for determining the size of the sliding window. The size of the sliding window is directly related to the amount of data analyzed each time. A smaller window can respond to data changes more quickly, but may be affected by local fluctuations; a larger window can better capture long-term trends, but the response speed is slower. Finally, through dynamic adjustment based on the standard deviation, the size of the sliding window can find the best balance between real-time and stability.

[0057] From the above content, it can be seen that the intelligent processing system effectively balances the real-time and stability of data processing by dynamically adjusting the sliding window size and combining the fluctuation range and fluctuation frequency of the target data sequence. This intelligent processing method based on adaptive adjustment of data characteristics greatly improves the accuracy and efficiency of financial data processing, reduces analysis errors caused by data fluctuations, thereby reducing the systemic risk of financial transactions and preventing financial losses caused by untimely analysis.

[0058] In an optional embodiment, the intelligent processing system first determines the number of sliding times of the sliding window according to the total number of data in the sliding window and the target data sequence, and then slides the sliding window M times from front to back based on the target data sequence according to the target sliding strategy and the number of sliding times to obtain M sub-data sequences, wherein M represents the number of sliding times, and the target sliding strategy includes: taking the first element in the target data sequence as the first element of the first sub-data sequence, and determining the target data corresponding to the first sub-data sequence according to the sliding window; calculating the difference between the total number of data in the target data sequence and the sliding window, and determining the target data corresponding to the last sub-data sequence according to the difference.

[0059] Optionally, first, the intelligent processing system determines the number of sliding times of the sliding window according to the total length of the target data sequence and the determined sliding window size. This step ensures that the sliding window can completely cover the entire data sequence, so that every part of the sequence can be comprehensively analyzed. Next, the intelligent processing system moves the sliding window according to a preset target sliding strategy, starting from the first data point of the sequence, and the sliding window moves backward one step each time until the end of the sequence.

[0060] Optionally, the target sliding strategy specifically includes: the intelligent processing system takes the first element in the target data sequence as the first element of the first sub-data sequence, and then determines the data in the sliding window as the first sub-data sequence. According to the size of the sliding window, the data in the middle position is selected as the target data, which is based on the assumption that the central data can represent the data characteristics in the window. Then, the difference between the total number of data N in the target data sequence and the size of the sliding window is calculated. This difference is used to determine the boundary of the last sub-data sequence to ensure that even at the end of the sequence, the sliding window can completely contain enough data points to generate a valid sub-data sequence, thereby avoiding analysis deviations at the tail of the sequence due to insufficient data. Through the above strategy, the system moves the sliding window M times from front to back, and each sliding generates a sub-data sequence.

[0061] From the above content, it can be seen that the intelligent processing system analyzes each local feature in the target data sequence through the sliding window mechanism, ensuring the comprehensiveness and depth of the analysis. The system uses the size of the sliding window and the length of the sequence to determine the number of slides, which ensures that when processing data sequences of any length, the sliding window can cover the entire sequence, avoiding the omission of important information that may be caused by fixed window size or position. In addition, the target sliding strategy ensures that the analysis starts from the starting point of the sequence until the end of the sequence, taking into account the timeliness of the data and maintaining the continuity of the analysis, which is particularly important for scenarios such as financial markets that require real-time response. This method not only improves the efficiency of data processing, but also enhances the accuracy and reliability of the analysis, reduces the systemic risks and financial losses caused by untimely analysis results, and provides strong technical support for the stability and efficiency of financial transactions.

[0062] In an optional embodiment, the intelligent processing system takes each data except the target data in the i-th sub-data sequence as the first data, where i is a positive integer less than or equal to M, and then determines the weight value corresponding to each first data based on the difference between each first data and the target data and the sliding window, wherein the weight value corresponding to each first data is smaller than the weight value corresponding to the target data; finally, the corrected data of each sub-data sequence is determined based on the weight values ​​corresponding to all the data in the i-th sub-data sequence.

[0063] Optionally, in each sub-data sequence (denoted as the i-th sub-sequence, i is a positive integer less than or equal to M), except for the target data located in the middle position, the intelligent processing system defines other data in the sequence as first data, and these first data represent other data points in the sub-data sequence other than the target data. Then the system dynamically determines the weight value of each first data according to the difference between each first data and the target data, and the size of the sliding window. The calculation formula of the weight value takes into account the absolute value of the difference and the range of the sliding window. The data point with a larger difference has a lower weight value. This is because the data point farther away from the target data contributes less to the overall trend of the sequence, and the size of the sliding window affects the smoothness of the weight distribution. Among them, the weight value of the target data is set to the maximum, ensuring that the corrected data can reflect the central trend of the sequence. Finally, the intelligent processing system performs weighted average calculation on these data points (including the target data and the first data) according to their weight values ​​in the i-th sub-data sequence to obtain the corrected data of the i-th sub-data sequence.

[0064] From the above content, it can be seen that the intelligent processing system effectively reduces data volatility and improves the accuracy and stability of data processing by assigning a weight value based on the difference between each data point in the sub-data sequence and the target data, and then performing a weighted average calculation. This weight-based modified data calculation method can more accurately capture the central trend of the sequence and reduce analysis errors caused by data fluctuations. For financial data processing, it can improve the reliability of trading strategies and reduce systemic risks caused by market fluctuations, thereby providing trading parties with more stable and accurate financial data analysis results, which helps to make more accurate investment decisions in high-timeliness trading scenarios and avoid unnecessary capital losses that may be caused by large fluctuations in analysis results. This method is particularly suitable for processing financial data with significant volatility and can significantly improve the efficiency and effectiveness of data processing.

[0065] In an optional embodiment, the intelligent processing system first sums the weight values ​​corresponding to all data in the i-th sub-data sequence to obtain a first value, then calculates the ratio of the weight value corresponding to each data in the i-th sub-data sequence to the first value to obtain a target weight value corresponding to each data in the i-th sub-data sequence, and finally determines the corrected data of each sub-data sequence according to the target weight value corresponding to each data in the i-th sub-data sequence.

[0066] Optionally, the intelligent processing system first calculates the corresponding weight values ​​of all data points in the i-th sub-data sequence, and sums these weight values ​​to obtain a total value, i.e., the first value. Next, the system calculates the ratio of the weight value of each data point in the i-th sub-data sequence to the first value (i.e., the sum of all weight values), thereby obtaining a target weight value corresponding to each data point, wherein the determination of the target weight value is based on the consideration of the internal data distribution of the sub-sequence, which can more accurately reflect the relative position of the data point in the sequence, thereby ensuring the fairness and rationality of the calculation of the corrected data. Finally, the system determines the corrected data of the sub-sequence according to the target weight value corresponding to each data point in the i-th sub-data sequence.

[0067] From the above content, it can be seen that the intelligent processing system can not only conduct a detailed analysis of time series data, but also effectively reduce the volatility of data and improve the accuracy and stability of sequence analysis while ensuring the timeliness of data. Using the weight-based correction data calculation method, the system can more intelligently identify and process key data points in the subsequence, ensuring that the calculation of the correction data takes into account both the intrinsic characteristics of the data and the position and influence of the data in the global sequence. This method performs well in processing highly volatile data such as financial markets, and can significantly reduce analysis errors caused by data fluctuations, reduce the systemic risk of financial transactions, provide more reliable data support for trading parties, and avoid potential financial losses.

[0068] In an optional embodiment, the intelligent processing system performs weighted average calculation on all data in the i-th sub-data sequence according to the target weight value corresponding to each data in the i-th sub-data sequence, and determines the corrected data of the i-th sub-data sequence based on the calculation result.

[0069] Optionally, in the i-th sub-data sequence, the system has calculated the target weight value corresponding to each data point based on the difference between each data and the target data, and the characteristics of the sliding window. The system performs weighted average calculation on all data points in the i-th sub-data sequence, including the target data and the first data. This calculation method ensures that the contribution of each data point to the final result matches its importance, and can more accurately reflect the central trend and characteristics of the sub-data sequence. The result obtained by the above weighted average calculation is the corrected data of the i-th sub-data sequence. The determination of the corrected data means that the system has adjusted the representative value of the sub-sequence according to the target weight value of each data point, so that it can more stably and accurately reflect the dynamic changes of the sequence, rather than just a simple average or median value.

[0070] It should be noted that weighted average is a statistical method in which the contribution of each data point is proportional to its weight value. It multiplies each data point by its corresponding target weight value, then adds all the products and finally divides them by the sum of the target weight values. This method can effectively reduce the data fluctuations in the sequence, making the corrected data better represent the central trend of the subsequence, and improving the stability of data processing and prediction accuracy.

[0071] From the above content, it can be seen that by introducing the concept of target weight value, the intelligent processing system can perform more sophisticated analysis and correction of time series data. The weighted average calculation method not only considers the absolute value of each data point, but also its relative position and importance in the sequence, so as to effectively reduce data volatility and improve the accuracy of data analysis. This method is particularly suitable for the processing of high-volatility data such as financial markets. It can significantly improve the stability and predictive ability of data processing, reduce the systemic risk of financial transactions caused by data analysis errors, and reduce potential capital losses. By dynamically adjusting the weights of the data series and correcting the weighted average, the intelligent processing system can respond to market changes more quickly and accurately, and provide strong data support for trading decisions. At the same time, this method also enhances the robustness of the data processing model, making it more flexible and efficient when processing different types of financial market data.

[0072] In an optional embodiment, when the intelligent processing system detects that the data source has switched or the data collection time exceeds a preset time, the target data sequence is updated by the intelligent processing system, wherein the update operation is used to delete at least one historical data with the earliest collection time in the target data sequence, and add the newly collected data to the target data sequence.

[0073] Optionally, the intelligent processing system continuously monitors the status of the data source, including the data collection rate, the stability of the data source and the change of the data type. When it is detected that the data source switches, that is, the data source changes from one to another, this means that the characteristics, quality and format of the data have changed. The system needs to adjust its processing strategy in time to adapt to the new data source to ensure the continuity and accuracy of the analysis. At the same time, the system will monitor the duration of data collection. When the data collection time exceeds the preset time threshold, that is, a time limit point set by the system, this indicates that the data in the sequence is outdated and can no longer accurately reflect the current market conditions. The preset time is set based on the balance between the timeliness requirements of the data and the system processing capacity to ensure that the data can be updated efficiently and in a timely manner. Once any of the above conditions is triggered, the system immediately performs an update operation on the target data sequence.

[0074] Optionally, the update operation mainly includes two parts: first, the intelligent processing system deletes at least one historical data with the earliest collection time in the sequence. This operation ensures that the data sequence will not become too long due to the accumulation of historical data, affecting the processing efficiency and storage resource usage. Then, the system adds the newly collected data to the target data sequence to update the sequence content and reflect the latest data changes. The execution of the entire update strategy ensures that the target data sequence can reflect the latest data information in real time. At the same time, by deleting outdated data, the flexibility and processing efficiency of the sequence are maintained. This dynamic update mechanism is particularly important for scenarios that require rapid response, such as financial markets. It can reduce analysis deviations caused by outdated data and improve the system's real-time response capabilities and decision-making efficiency.

[0075] From the above content, it can be seen that the intelligent processing system can monitor the data source status and data collection time in real time, ensuring that when the data source switches or the data is outdated, the target data sequence can be updated in time to reflect the latest data characteristics. By deleting outdated historical data and adding the latest collected data, the system not only maintains the timeliness of the data, but also improves the efficiency and accuracy of data processing, and reduces the analysis errors caused by data lag. For trading scenarios that require rapid response, such as financial markets, this mechanism can significantly reduce systemic risks, avoid capital losses, and ensure that trading decisions are based on the most accurate and timely data analysis. In addition, this mechanism also enhances the adaptability and robustness of the system, improves its performance in a variable data source environment, and provides a more reliable real-time data processing solution for financial data analysis.

[0076] In an optional embodiment, Figure 2 is a schematic diagram of an optional method for processing time series data according to an embodiment of the present application, such as Figure 2 As shown in the figure, for the processing of gold quotation data, since the bank's gold price changes in real time, and there will be multiple time periods for switching quotation sources throughout the day, frequent switching of quotation sources will result in price mutations at a certain time or a certain period of time, and due to the large fluctuations in gold prices, improper handling will cause serious financial losses to the bank. For the scenario of quotation source switching, two methods are proposed to smooth the gold price. One is to perform price smoothing when the quotation source is switched (the gold price half an hour before and after the quotation source switching time can be added to the price window, and the price can be smoothed while moving the window in chronological order), and the other is to perform price smoothing on the gold price throughout the day. If you are more sensitive to the gold quotation time, you can choose method one. If you need to obtain a relatively smooth gold price throughout the day, you can choose method two, that is, use the Gaussian filter algorithm to smooth the price sequence. The main steps are the following four steps: define the window size, calculate the weight, weighted average, and move the window.

[0077] The first step is to define the window size: according to the characteristics of the data sequence, the intelligent processing system selects a suitable sliding window size. The choice of window size has a great impact on the results. It directly determines which data points will be included in the smoothing calculation. In this embodiment, the window size is determined by the standard deviation σ. The selection of the σ value needs to find a balance between the smoothing effect and the processing speed. The window size is the continuous gold price range selected before and after. It controls the speed at which the weight changes with distance, thereby affecting the degree of smoothing. The larger the σ value, the slower the weight decreases, and the greater the influence of distant points on the current point, so the smoothing effect is more obvious. Since the purpose of this method is to reduce small price mutations in the price sequence and retain most of the characteristics of the original signal, and the σ value not only affects the smoothness, but also affects the algorithm processing speed, the σ value should not be too large. It is necessary to find a balance between maintaining sufficient smoothness and ensuring processing speed. Assume that the price sequence is [λ1, λ2, λ3, λ4...λ n ], standard deviation σ=3, then the first price to be smoothed is λ4, the selected time window is [λ1, λ2, λ3, λ4, λ5, λ6, λ7], and so on, the last price to be smoothed is λ n-3 , the time window selected is [λ n-6 ,λ n-5 ,λ n-4 ,λ n-3 ,λ n-2 ,λ n-1 ,λ n ], the window needs to be moved n-σ*2 times in total, and the price smoothing operation needs to be performed the same number of times. If the smoothing effect is not obvious or the algorithm efficiency is low, the σ value needs to be adjusted repeatedly until the data meets the preset smoothness requirements.

[0078] The second step is to calculate the weight: In the Gaussian filtering algorithm, the weight is calculated based on the Gaussian function. The Gaussian function provides a natural and effective way to measure the influence of each data point in the window on the center point, and this influence can be expressed as a weight. This influence decreases with the increase of distance, that is, the farther away from the center point, the smaller the weight. The selection and calculation of weights are based on the properties of Gaussian distribution, which mainly depends on two factors: the distance x from the center point and the standard deviation σ. For a one-dimensional Gaussian distribution, the calculation method is shown in formula (1):

[0079]

[0080] In the Gaussian filtering algorithm, the mean μ is a parameter that defines the center position of the Gaussian distribution. Each time the filtering operation is performed, the currently processed data point is taken as the center, that is, the mean μ = 0, which means that the center point is the data point with the largest weight. For each point in the window, the required weight is calculated using the above formula (1) according to its distance x relative to the center point (that is, the index of the array minus the center index).

[0081] The third step is weighted averaging: The intelligent processing system first calculates the original weight of each data point in the window according to the Gaussian function, then calculates the sum W of all these original weights, and divides each weight by this sum W to obtain the normalized weight. This operation ensures that the sum of the weights is 1. The normalization factor is It ensures that the total area of ​​the Gaussian function, i.e. the integral over the entire domain, is equal to 1. The normalization is calculated as follows:

[0082] 1) Assume that the weights calculated by the Gaussian filtering algorithm are ω1, ω2, ω3...ω n ;

[0083] 2) The sum of the unnormalized weights is W = ω1 + ω2 + ω3 + ... + ω n ;

[0084] 3) Each weight after normalization is The sum is

[0085] When using normalized weights to perform weighted averaging on data, the total amount of data can be kept unchanged while ensuring that the weight distribution of each price point is more reasonable, reducing the impact of volatility and noise, and improving the accuracy and stability of the smoothed price series.

[0086] The fourth step is to move the window: considering the real-time nature of gold price data, when the intelligent processing system uses the Gaussian filter algorithm to smooth the gold price sequence, the defined window size needs to be moved in real time to ensure that the window always contains the latest price points. Each time the window moves, the system adds the latest price point to the window and deletes the oldest price point in the window. Then, the price point in the window is used to calculate the smoothed price through the Gaussian filter algorithm and updated to the original price sequence. This process is repeated until all price points are processed and a complete smoothed price sequence is obtained. The movement of the window and the real-time update of the data ensure the timeliness and accuracy of the processing results and adapt to the real-time changing characteristics of financial market data.

[0087] Finally, check the price sequence to determine whether it is necessary to continue updating the price sequence. If necessary, re-execute the above steps 2 to 4. If not, determine whether it meets the standards based on the final updated price sequence. If not, adjust the standard deviation and retrain the price sequence.

[0088] The embodiment of the present application also provides a time series data processing device. It should be noted that the time series data processing device of the embodiment of the present application can be used to execute the time series data processing method provided in the embodiment of the present application. The time series data processing device provided in the embodiment of the present application is introduced below.

[0089] According to an embodiment of the present application, a device for implementing the above-mentioned time series data processing method is also provided. Figure 3 is a schematic diagram of an optional time series data processing device according to an embodiment of the present application, such as Figure 3 As shown, the device includes: a first setting unit 301, a first processing unit 302, a first determining unit 303, a second determining unit 304, and a second processing unit 305.

[0090] Optionally, a first setting unit 301 sets a sliding window for a target data sequence, wherein the target data sequence includes N data collected at N moments, N is an integer greater than 1, and the sliding window is used to determine the number of data selected continuously from the target data sequence each time; a first processing unit 302 slides the sliding window M times from front to back based on the target data sequence to obtain M sub-data sequences; a first determining unit 303 uses the data in the middle position of each sub-data sequence as the target data of the sub-data sequence; a second determining unit 304 determines the corrected data of each sub-data sequence according to the sliding window, wherein the corrected data is used to reduce the volatility between multiple data in each sub-data sequence; a second processing unit 305 replaces the target data of each sub-data sequence with the corrected data of the sub-data sequence, and performs a synchronous data update on the target data sequence according to the target data replacement result of each sub-time series.

[0091] Optionally, the first setting unit 301 includes: a first calculating subunit and a first determining subunit, wherein the first calculating subunit is used to calculate the standard deviation of N data in the target data sequence; and the first determining subunit is used to determine the sliding window according to the standard deviation.

[0092] Optionally, the first determination subunit includes: a first acquisition module and a first determination module. The first acquisition module is used to acquire the fluctuation range and fluctuation frequency of N data in the target data sequence; the first determination module is used to adjust the standard deviation according to the fluctuation range and fluctuation frequency of the N data, and use the adjustment result as the sliding window.

[0093] Optionally, the first processing unit 302 includes: a second determining subunit and a first processing subunit. The second determining subunit is used to determine the number of sliding times of the sliding window according to the total number of data in the sliding window and the target data sequence; the first processing subunit is used to slide the sliding window M times from front to back based on the target data sequence according to the target sliding strategy and the number of sliding times to obtain M sub-data sequences, wherein M represents the number of sliding times, and the target sliding strategy includes: taking the first element in the target data sequence as the first element of the first sub-data sequence, and determining the target data corresponding to the first sub-data sequence according to the sliding window; calculating the difference between the total number of data in the target data sequence and the sliding window, and determining the target data corresponding to the last sub-data sequence according to the difference.

[0094] Optionally, the second determination unit 304 includes: a third determination subunit, a fourth determination subunit, and a fifth determination subunit. The third determination subunit is used to take each data except the target data in the i-th sub-data sequence as the first data, wherein i is a positive integer less than or equal to M; the fourth determination subunit is used to determine the weight value corresponding to each first data according to the difference between each first data and the target data and the sliding window, wherein the weight value corresponding to each first data is less than the weight value corresponding to the target data; the fifth determination subunit is used to determine the corrected data of each sub-data sequence according to the weight values ​​corresponding to all the data in the i-th sub-data sequence.

[0095] Optionally, the fifth determination subunit includes: a first calculation subunit, a second calculation subunit, and a sixth determination subunit. The first calculation subunit is used to sum the weight values ​​corresponding to all the data in the i-th sub-data sequence to obtain a first value; the second calculation subunit is used to calculate the ratio of the weight value corresponding to each data in the i-th sub-data sequence to the first value to obtain a target weight value corresponding to each data in the i-th sub-data sequence; the sixth determination subunit is used to determine the correction data of each sub-data sequence according to the target weight value corresponding to each data in the i-th sub-data sequence.

[0096] Optionally, the sixth determination subunit includes: a second determination module, used to perform weighted average calculation on all data in the i-th sub-data sequence according to the target weight value corresponding to each data in the i-th sub-data sequence, and determine the corrected data of the i-th sub-data sequence based on the calculation result.

[0097] Optionally, the time series data processing device also includes: an updating unit, which is used to update the target data sequence when it is detected that the data source is switched or the data collection time exceeds a preset time, wherein the updating operation is used to delete at least one historical data with the earliest collection time in the target data sequence, and add the newly collected data to the target data sequence.

[0098] According to another aspect of an embodiment of the present application, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored executable program, wherein when the executable program runs, the device where the computer-readable storage medium is located is controlled to execute the above-mentioned method for processing time series data.

[0099] According to another aspect of an embodiment of the present application, there is also provided an electronic device, including: a memory storing an executable program; and a processor for running the program, wherein the above-mentioned method for processing time series data is executed when the program is running.

[0100] According to another aspect of an embodiment of the present application, a computer program product is also provided, including computer instructions, which implement the steps of the above-mentioned method for processing time series data when executed by a processor.

[0101] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0102] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0103] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0104] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0105] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0106] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk and other media that can store program codes.

[0107] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A method for processing time series data, characterized in that: include: A sliding window is set for a target data sequence, wherein the target data sequence includes N data collected at N moments, N is an integer greater than 1, and the sliding window is used to determine the number of data continuously selected from the target data sequence each time; Based on the target data sequence, slide the sliding window M times from front to back to obtain M sub-data sequences; The data in the middle position of each sub-data sequence is used as the target data of the sub-data sequence; Determining correction data for each of the sub-data sequences according to the sliding window, wherein the correction data is used to reduce the volatility between the plurality of data in each of the sub-data sequences; The target data of each sub-data sequence is replaced with the revised data of the sub-data sequence, and the target data sequence is synchronously updated according to the target data replacement result of each sub-time series.

2. The method for processing time series data according to claim 1, characterized in that: Set a sliding window for the target data sequence, including: Calculate the standard deviation of N data in the target data sequence; The sliding window is determined according to the standard deviation.

3. The method for processing time series data according to claim 2, characterized in that: Determining the sliding window according to the standard deviation includes: Obtaining the fluctuation range and fluctuation frequency of N data in the target data sequence; The standard deviation is adjusted according to the fluctuation range and fluctuation frequency of the N data, and the adjustment result is used as the sliding window.

4. The method for processing time series data according to claim 1, characterized in that: Based on the target data sequence, the sliding window is slid M times from front to back to obtain M sub-data sequences, including: Determining the number of sliding times of the sliding window according to the total number of data in the sliding window and the target data sequence; According to the target sliding strategy and the number of sliding times, based on the target data sequence, the sliding window is slid M times from front to back to obtain M sub-data sequences, wherein M represents the number of sliding times. The target sliding strategy includes: taking the first element in the target data sequence as the first element of the first sub-data sequence, and determining the target data corresponding to the first sub-data sequence according to the sliding window; calculating the difference between the total number of data in the target data sequence and the sliding window, and determining the target data corresponding to the last sub-data sequence according to the difference.

5. The method for processing time series data according to claim 1, characterized in that: Determining the correction data of each sub-data sequence according to the sliding window includes: Taking each data except the target data in the i-th sub-data sequence as the first data, where i is a positive integer less than or equal to M; Determine, according to the difference between each first data and the target data and the sliding window, a weight value corresponding to each first data, wherein the weight value corresponding to each first data is smaller than the weight value corresponding to the target data; The corrected data of each sub-data sequence is determined according to the weight values ​​respectively corresponding to all the data in the i-th sub-data sequence.

6. The method for processing time series data according to claim 5, characterized in that: Determining the correction data of each sub-data sequence according to the weight values ​​respectively corresponding to all the data in the i-th sub-data sequence includes: Summing the weight values ​​corresponding to all the data in the i-th sub-data sequence to obtain a first value; Calculate the ratio of the weight value corresponding to each data in the i-th sub-data sequence to the first value to obtain the target weight value corresponding to each data in the i-th sub-data sequence; According to the target weight value corresponding to each data in the i-th sub-data sequence, the corrected data of each sub-data sequence is determined.

7. The method for processing time series data according to claim 6, characterized in that: Determining the corrected data of each sub-data sequence according to the target weight value corresponding to each data in the i-th sub-data sequence includes: According to the target weight value corresponding to each data in the i-th sub-data sequence, a weighted average calculation is performed on all the data in the i-th sub-data sequence, and the calculation result is used to determine the corrected data of the i-th sub-data sequence.

8. The method for processing time series data according to claim 1, characterized in that: The method for processing time series data also includes: When it is detected that the data source has switched or the data collection time exceeds a preset time, an update operation is performed on the target data sequence, wherein the update operation is used to delete at least one historical data with the earliest collection time in the target data sequence, and add the newly collected data to the target data sequence.

9. A time series data processing device, characterized in that: include: A first setting unit is configured to set a sliding window for a target data sequence, wherein the target data sequence includes N data collected at N moments, N is an integer greater than 1, and the sliding window is used to determine the number of data continuously selected from the target data sequence each time; A first processing unit, based on the target data sequence, slides the sliding window from front to back M times to obtain M sub-data sequences; A first determining unit uses the data at the middle position in each sub-data sequence as the target data of the sub-data sequence; a second determining unit, which determines the correction data of each sub-data sequence according to the sliding window, wherein the correction data is used to reduce the volatility between the multiple data in each sub-data sequence; The second processing unit replaces the target data of each sub-data sequence with the revised data of the sub-data sequence, and performs synchronous data update on the target data sequence according to the target data replacement result of each sub-time series.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located executes the method for processing time series data according to any one of claims 1 to 8.

11. An electronic device, characterized in that: It includes one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors execute the time series data processing method described in any one of claims 1 to 8.

12. A computer program product comprising computer instructions, characterized in that: When the computer instructions are executed by a processor, the steps of the time series data processing method described in any one of claims 1 to 8 are implemented.

Images