Complex time sequence compound mutation identification method and device based on forward decomposition
By constructing a layer-by-layer forward decomposition recognition model of mean mutation-variability mutation-turn mutation, the interference problem of mutation process recognition in complex time series is solved, and the accurate identification and accuracy of mutation processes in complex time series is achieved.
Patent Information
- Application Number
- CN202411162831.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-23
- Publication Date
- 2024-12-27
AI Technical Summary
The prior art is difficult to clearly characterize the composite mutation process in complex time series, and different categories of mutation processes interfere with each other, resulting in some mutation points being masked or confused.
Using a complex time series complex mutation recognition method based on forward decomposition, the layer-by-layer forward decomposition recognition model of mean mutation-variability mutation-transformation mutation is constructed, and the mutations in complex time series are regarded as the mean mutation and variability mutation. The three types of turning mutations are coupled to form, identify them one by one, strip them down layer by layer, and eliminate mutual interference between multiple mutations. The specific steps include means mutation analysis, variability mutation analysis and transition mutation analysis of complex time series, and subsequent analysis of unaffected sequences are obtained by layer by layer.
Accurate recognition of complex mutation processes in complex time series is achieved, the accuracy of identification of mutation points and mutation processes is improved, the adverse effects of upper mutations on lower mutation analysis is eliminated, and a variety of mutations hidden in the time series are revealed.
Smart Images
Figure CN119207559A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of time series analysis, and in particular relates to a method and device for identifying complex time series composite mutations based on forward decomposition. Background Art
[0002] Mutation analysis is widely used in the analysis and processing of time series. Complex time series are very common in practical applications. Such sequences often have compound mutations of multiple mutations. The mutation processes of various mutations interfere with each other, resulting in some mutation points being obscured or confused. Mutation analysis in the prior art usually performs a holistic analysis of all mutations in the sequence and identifies mutation points in the time series at a single level. It does not consider different types of mutations and is difficult to clearly characterize the mutation process of compound mutations. In order to accurately identify compound mutation points and mutation processes in complex time series, it is necessary to propose a complex time series compound mutation identification method and device based on forward decomposition. Summary of the invention
[0003] The main purpose of the present invention is to overcome the shortcomings and deficiencies of the prior art, and to provide a method and device for identifying complex time series compound mutations based on forward decomposition, which breaks through the thinking framework of single mutation identification, overcomes the mutual interference of mutation processes under different categories, eliminates the adverse effects of the upper-level mutation process on the lower-level mutation analysis, and accurately identifies the compound mutation process of complex time series.
[0004] In order to achieve the above object, the present invention adopts the following technical solutions:
[0005] In a first aspect, the present invention provides a method for identifying complex time series compound mutations based on forward decomposition, comprising the following steps:
[0006] A layer-by-layer forward decomposition and identification model of a composite mutation of mean mutation-rate mutation-turning mutation is constructed; the composite mutation layer-by-layer forward decomposition and identification model regards the mutation in a complex time series as formed by the coupling of mean mutation, rate mutation and turning mutation, identifies the three mutations by category, and decomposes them layer by layer to eliminate mutual interference between multiple mutations;
[0007] The mean mutation process is obtained by performing mean mutation analysis on complex time series;
[0008] The mean mutation process is separated from the original time series to obtain a mean mutation sequence without the influence of mean mutation. The rate mutation analysis is performed on the mean mutation sequence to identify the rate mutation points in the sequence. The rate mutation process is identified to quantify the mutation of the sequence variance state.
[0009] The rate mutation process is separated from the de-mean mutation sequence to obtain the de-mean mutation and rate mutation sequence without the influence of mean mutation and rate mutation. For the de-mean mutation and rate mutation sequence, the turning mutation point is identified.
[0010] As a preferred technical solution, the mean mutation process obtained by performing mean mutation analysis on the complex time series is specifically as follows:
[0011] Perform MK test on complex time series to identify all mutation points in complex time series;
[0012] Perform MTT test on complex time series to examine whether the difference between the mean values of the two sub-time series is significant, and select significant mean mutation points from the MK test mutation points;
[0013] Identify mean mutation processes and quantify the mutation of the mean value of the sequence.
[0014] As a preferred technical solution, the MK test is performed on the complex time series to identify all mutation points in the complex time series, specifically:
[0015] For a complex time series x={x 1 ,x 2 ,…,x n}, construct a set of ordered columns:
[0016]
[0017] in,
[0018]
[0019] Under the assumption that the time series are randomly independent, the statistic is defined:
[0020]
[0021] UF 1 =0, Var(s k ) is the cumulative number s k The mean and variance of UF i is a statistical sequence that obeys the standard normal distribution. Under a given significance level α, the normal distribution table is used to obtain U α If UF i |>U α , it indicates that there is an obvious trend change in the sequence;
[0022] Repeat the above process in reverse order of time series x, and make UB k =-UF k , k=n,n-1,…,1,UB1 =0;
[0023] Confidence interval [-U α ,U α ] inner UF k With UB k The intersection of is the mutation point, a positive value represents a sudden increase, and a negative value represents a sudden decrease.
[0024] As a preferred technical solution, the complex time series is subjected to an MTT test to examine whether the difference between the average values of the two sub-time series before and after is significant, and significant mean mutation points are screened out from the MK test mutation points, specifically:
[0025] Determine the two subsequences x before and after the reference point 1 and x 2 The length of n 1 =n 2 The average values of the two subsequences are and The variance is s 1 2 and 2 2 , define the statistic:
[0026]
[0027] in,
[0028]
[0029] The reference points are set continuously by sliding, and the statistics are calculated according to the above formula to obtain the statistical sequence t i ,i=1,2,…,n-(n 1 +n 2 )+1; given the significance level α, the critical value t is obtained by looking up the t distribution table α , if |t i | <t α , then it is considered that there is no significant difference in the means of the two subsequences before and after the reference point, otherwise it is considered that a mean mutation occurs at the reference point;
[0030] The MTT test was used to analyze whether there were significant mean mutations before and after all the mutation points obtained by the MK test, and the mean mutation points above the significant level were screened out.
[0031] As a preferred technical solution, the Pettitt method is used to perform rate mutation analysis on the mean-devalued mutation sequence to identify the rate mutation points in the sequence, specifically:
[0032] For the sequence y={y 1 ,y 2,…,y n}, construct the order column again:
[0033]
[0034] in,
[0035]
[0036] Among them, the order column s k It is the cumulative number of values at the i-th moment that are greater than or less than the values at the j-th moment, reflecting the sharp change of the sequence from one variance state to another;
[0037] Directly use the order sequence to detect the mutation point of the rate change. If t 0 Always Satisfied
[0038]
[0039] Then t 0 The point is the mutation point of rate of change;
[0040]
[0041] If P ≤ 0.5, the detected mutation point is considered to be statistically significant.
[0042] As a preferred technical solution, the rate mutation process is separated from the mean mutation sequence to obtain the mean mutation and rate mutation sequence without the influence of mean mutation and rate mutation. For the mean mutation and rate mutation sequence, specifically:
[0043] The rate mutation process is separated from the mean-depleted mutation sequence to obtain the mean-depleted mutation and rate mutation sequence;
[0044] Introduce linear semi-truncated polynomials to construct new sets of independent variables and linear regression forms;
[0045] The SP-SRA method was used to perform turning mutation analysis on the mean-degraded mutation and rate-variable mutation sequences to identify the turning mutation points in the sequences.
[0046] As a preferred technical solution, the introduction of linear semi-truncated polynomials to construct a new set of independent variables and linear regression form is specifically:
[0047] For the sequence z={z 1 ,z 2 ,…,z n}, suppose that in this time series t c The point is a mutation point, which means that t c Before and after, two subsequences t 1=1,2,…,t c and t 1 =t c+1 ,t c+2 …,n have significant differences in their changing trends. Based on the above assumptions, n-1 points are inserted into the time interval [1,n] of the sequence, and the time variable t is divided into n small intervals, recorded as (1,2], (2,3], …, (n-1,n]; in each interval, a set of linear functions is used to describe the changes of the sequence z:
[0048]
[0049] Introduce linear semi-truncated polynomials:
[0050]
[0051] Then we have:
[0052]
[0053] The above process transforms the selection of turning points in the time series into the selection of independent variables that have a significant impact on the dependent variable, thereby solving the problem of significance testing for multiple selected mutation points;
[0054] Construct a new set of independent variables, let
[0055] Z 1 =t
[0056] Z 2 =(t-1) +
[0057] Z 3 =(t-2) +
[0058] …
[0059] Z n =(t-(n-1)) +
[0060] The linear regression form is obtained:
[0061]
[0062] In a second aspect, the present invention provides a complex time series composite mutation identification system based on forward decomposition, a model building module, a mean mutation module, a rate mutation module and a turning mutation module;
[0063] The model building module is used to build a layer-by-layer forward decomposition and identification model of a composite mutation of mean mutation-rate mutation-turning mutation; the composite mutation layer-by-layer forward decomposition and identification model regards the mutation in a complex time series as being formed by the coupling of mean mutation, rate mutation and turning mutation, and identifies the three mutations by category, peels off and decomposes them layer by layer to eliminate mutual interference between multiple mutations;
[0064] The mean mutation module is used to perform mean mutation analysis on complex time series to obtain a mean mutation process;
[0065] The rate mutation module is used to separate the mean mutation process from the original time series to obtain a mean mutation sequence without the influence of mean mutation, perform rate mutation analysis on the mean mutation sequence, identify the rate mutation points in the sequence; identify the rate mutation process, and quantify the mutation of the sequence variance state;
[0066] The turning mutation module is used to separate the rate mutation process from the de-mean mutation sequence, obtain the de-mean mutation and rate mutation sequence without the influence of mean mutation and rate mutation, and identify the turning mutation point for the de-mean mutation and rate mutation sequence.
[0067] In a third aspect, the present invention provides an electronic device, the electronic device comprising:
[0068] at least one processor; and,
[0069] a memory communicatively connected to the at least one processor; wherein,
[0070] The memory stores computer program instructions that can be executed by the at least one processor, and the computer program instructions are executed by the at least one processor so that the at least one processor can execute the complex time series composite mutation identification method based on forward decomposition.
[0071] In a fourth aspect, the present invention provides a computer-readable storage medium storing a program, which, when executed by a processor, implements the complex time series composite mutation identification method based on forward decomposition.
[0072] Compared with the prior art, the present invention has the following advantages and beneficial effects:
[0073] 1. The present invention provides a complex time series composite mutation identification method and device based on forward decomposition. For three different types of mutations, namely, mean mutation, rate mutation, and turning mutation, a composite mutation layer-by-layer decomposition and identification model of mean mutation-rate mutation-turning mutation is constructed; mean mutation analysis is performed on the complex time series to obtain the mean mutation process; decomposition and stripping are performed to obtain the mean mutation sequence, and rate mutation analysis is performed to obtain the rate mutation process; further decomposition and stripping are performed to obtain the mean mutation and rate mutation sequence, and turning mutation analysis is performed to obtain the turning mutation process. The upper layer mutation process is decomposed and stripped, the mutation points of the layer are identified, and the mutation process of the layer is analyzed layer by layer, so as to gradually eliminate the influence of the upper layer mutation on the lower layer mutation analysis and reveal the various mutations hidden in the time series. The hydrological and meteorological time series analysis of the present invention provides a four-layer systematic framework, which can decompose and identify the mutation process of the complex time series layer by layer while considering the composite mutation, and improve its recognition accuracy.
[0074] 2. The present invention constructs a layer-by-layer forward decomposition and identification model of a composite mutation of mean mutation-rate mutation-turning mutation, which regards the mutation in a complex time series as being formed by the coupling of mean mutation, rate mutation and turning mutation, thereby improving the accuracy of identifying mutation points and mutation processes.
[0075] 3. The present invention uses the MK-MTT method to perform mean mutation analysis on complex time series. The MK method is simple to calculate, but it is not suitable for sequences with multiple mutation points and lacks specificity for mean mutation. The MTT method makes up for these two points well. The coupling of the two methods can accurately identify all mean mutation points.
[0076] 4. The present invention decomposes and strips the mean mutation process from the original time series to obtain a de-mean mutation sequence, and uses the Pettitt method to perform rate mutation analysis on the de-mean mutation sequence. The de-mean mutation process converts the time series to the same mean level, which can make the subsequent rate mutation analysis more refined.
[0077] 5. The present invention decomposes and strips the rate mutation process from the mean-removed mutation sequence to obtain the mean-removed mutation and rate mutation sequence, and uses the SP-SRA method to perform turning mutation analysis on the mean-removed mutation and rate mutation sequence to achieve layer-by-layer decomposition and identification of multiple types of mutations. This method constructs a new set of independent variables by introducing a linear semi-truncated polynomial, and uses stepwise regression to automatically select the optimal segmentation points that pass the statistical significance test, converting the identification problem of turning mutation points into the identification problem of independent variables that significantly affect the dependent variable, which can not only identify multiple turning mutation points in the time series, but also improve the reliability of the detected turning mutation points. BRIEF DESCRIPTION OF THE DRAWINGS
[0078] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0079] Figure 1 The present invention is a flowchart of a complex time series composite mutation identification method based on forward decomposition according to an embodiment of the present invention.
[0080] Figure 2 Flow chart of step S2 of an embodiment of the present invention.
[0081] Figure 3 Flow chart of step S3 of the embodiment of the present invention.
[0082] Figure 4 This is a flow chart of step S4 of an embodiment of the present invention.
[0083] Figure 5 The block diagram of the complex time series composite mutation identification system based on forward decomposition according to an embodiment of the present invention.
[0084] Figure 6 4 is a structural diagram of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION
[0085] In order to enable those skilled in the art to better understand the present application, the technical solutions 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, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present application.
[0086] Reference to "embodiments" in this application means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described in this application may be combined with other embodiments.
[0087] See also Figure 1 This embodiment proposes a complex time series composite mutation identification method and device based on forward decomposition, including the following steps:
[0088] Step S1: Construct a layer-by-layer forward decomposition recognition model of a compound mutation of mean mutation, rate mutation, and turning mutation
[0089] The present application regards mutations in complex time series as formed by the coupling of mean mutations, rate mutations, and turning point mutations. Therefore, the present application constructs a composite mutation recognition model that can identify category by category, decompose layer by layer, and analyze item by item.
[0090] It can be understood that the layer-by-layer decomposition and identification model of the composite mutation of mean mutation-rate mutation-turning mutation can identify the three mutations category by category, decompose them layer by layer, eliminate the mutual interference between multiple mutations, and thus accurately identify various mutation points and mutation processes in complex time series.
[0091] Step S2: Use MK-MTT method to perform mean mutation analysis on complex time series
[0092] The mean value is a basic characteristic quantity that reflects the state of the data, and the mean mutation describes the process of a sequence changing sharply from one mean value to another, which is characterized by the discontinuity of the sequence.
[0093] It can be understood that the main idea of the MK-MTT method is to identify all mutation points in complex time series through the MK test (Man-Kendall test), and then verify whether there is a significant mean mutation before and after the mutation point through the moving T test, screen out the mean mutation points that meet the requirements, construct the mean mutation process line, and eliminate the impact of the mean mutation on subsequent analysis.
[0094] For further information, see Figure 2 , the step S2 comprises the following steps:
[0095] Step S2a: Perform MK test on the complex time series to identify all mean mutation points in the series.
[0096] In one embodiment, for a complex time series x={x 1 ,x 2 ,…,x n}, construct a set of ordered columns:
[0097]
[0098] in,
[0099]
[0100] Under the assumption that the time series are randomly independent, the statistic is defined:
[0101]
[0102] UF 1 =0, Var(s k ) is the cumulative number s k The mean and variance of UF i is a statistical sequence that obeys the standard normal distribution. Under a given significance level α, the normal distribution table is used to obtain U α If UF i |>U α , it indicates that there is an obvious trend change in the sequence.
[0103] Repeat the above process in reverse order of time series x, and make UB k =-UF k , k=n,n-1,…,1,UB 1 =0.
[0104] Confidence interval [-U α ,U α ] inner UF k With UB k The intersection of is the mutation point, a positive value represents a sudden increase, and a negative value represents a sudden decrease.
[0105] Step S2b: Perform an MTT test on the complex time series to examine whether the difference between the mean values of the two subsequences is significant, and select significant mean mutation points from the MK test mutation points.
[0106] Artificially determine the two subsequences x before and after the reference point 1 and x 2 The length is usually the same as n 1 =n 2 The average values of the two subsequences are and The variance is s 1 2 and 2 2 , define the statistic:
[0107]
[0108] in,
[0109]
[0110] The reference points are set continuously by sliding, and the statistical quantities are calculated in sequence according to the above formula to obtain the statistical quantity sequence t i ,i=1,2,…,n-(n 1 +n 2 )+1. Given a significant level α, the critical value t is obtained by looking up the t distribution table α , if |t i | <t α, then it is considered that there is no significant difference in the means of the two subsequences before and after the reference point, otherwise it is considered that a mean mutation occurs at the reference point.
[0111] By using the MTT test to analyze whether there is a significant mean mutation before and after all the mutation points obtained by the MK test method, the mean mutation points above the significant level can be screened out.
[0112] Step S2c: Identify the mean mutation process and quantify the mutation of the sequence mean value.
[0113] It can be understood that according to the above mean mutation point m 1 ,m 2 …m o , o is the number of mean mutation points, dividing the original time series into m 1 ,m 2 …m o ,m o+1 There are (o+1) subsequences in total. Calculate the mean E corresponding to each subsequence. 1 ,E 2 ,…,E o ,E o+1 , the mean sequence corresponding to the n sample points constitutes the mean mutation process line.
[0114] Step S3: decompose and remove the mean mutation process from the original time series to obtain a de-mean mutation sequence, and use the Pettitt method to perform rate mutation analysis on the de-mean mutation sequence.
[0115] Rate mutation refers to a process in which a sequence changes sharply from one variance state to another, but the mean value does not change significantly. Since the mean mutation process will have an adverse effect on the identification of rate mutation, it is necessary to remove the mean mutation process from the original time series to eliminate the impact of mean value changes on rate mutation analysis.
[0116] For further information, see Figure 2 , the step S3 comprises the following steps:
[0117] Step S3a: remove the mean mutation process from the original time series to obtain a de-mean mutation sequence.
[0118] In one embodiment, the mean mutation process can be stripped from the original time series by subtracting the corresponding mean mutation process line from the original time series to obtain the mean mutation sequence y={y 1 ,y 2 ,…,y n}.
[0119] Step S3b, using the Pettitt method to perform rate mutation analysis on the mean-devalued mutation sequence to identify rate mutation points in the sequence.
[0120] Furthermore, for the sequence y={y 1 ,y 2 ,…,y n}, construct an ordered sequence of the form S1a:
[0121]
[0122] in,
[0123]
[0124] Different from the MK test used in mean mutation analysis, the order column s k It is the cumulative number of values at the i-th moment that are greater than or less than the values at the j-th moment. This can well reflect the sharp change of the sequence from one variance state to another, and is therefore used in rate mutation analysis.
[0125] Directly use the order sequence to detect the mutation point of the rate change. If t 0 Always Satisfied
[0126]
[0127] Then t 0 The point where the rate of change changes suddenly.
[0128]
[0129] If P ≤ 0.5, the detected mutation point is considered to be statistically significant.
[0130] Step S3c, identifying the rate mutation process and quantifying the mutation of the sequence variance state.
[0131] According to the above mutation point m 1 ,m 2 …m p , p is the number of mutation points, and the mean mutation sequence is divided into m 1 ,m 2 …m o ,m p+1 There are (p+1) subsequences in total. Calculate the mean Var corresponding to each subsequence. 1 ,Var 2 ,…,Var p ,Var p+1 , the variance sequence corresponding to the n sample points constitutes the rate mutation process line.
[0132] Step S4, decomposing and stripping the rate mutation process from the mean-removed mutation sequence to obtain the mean-removed mutation and rate mutation sequence, and using the SP-SRA method to perform turning mutation analysis on the mean-removed mutation and rate mutation sequence.
[0133] A turning mutation refers to a sequence that continuously decreases (increases) over a period of time, and then suddenly starts to increase (decrease) continuously at a certain point. The mean mutation process has been stripped away in S2, but the rate mutation process will also have an adverse effect on the identification of turning mutations, eliminating the influence of mean and variance changes on turning mutation analysis.
[0134] Linear step polynomial and stepwise regression analysis were introduced to detect sequence transition mutations.
[0135] For further information, see Figure 3 , the step S4 comprises the following steps:
[0136] Step S4a, separating the rate mutation process from the mean-removed mutation sequence to obtain the mean-removed mutation and rate mutation sequence.
[0137] In one embodiment, the rate mutation process can be separated from the original time series by subtracting the corresponding rate mutation process line from the mean mutation sequence, and the mean mutation and rate mutation sequence z={z 1 ,z 2 ,…,z n}.
[0138] Step S4b: Introduce linear semi-truncated polynomials to construct a new set of independent variables and linear regression form
[0139] Furthermore, for the sequence z={z 1 ,z 2 ,…,z n}, suppose that in this time series t c The point is a mutation point, which means that t c Before and after, two subsequences t 1 =1,2,…,t c and t 1 =t c+1 ,t c+2 …,n have significant differences in their changing trends. Based on the above assumptions, n-1 points are inserted into the time interval [1,n] of the sequence, and the time variable t is divided into n small intervals, recorded as (1,2], (2,3], …, (n-1,n]. In each interval, a set of linear functions are used to describe the changes of the sequence z:
[0140]
[0141] Introduce linear semi-truncated polynomials:
[0142]
[0143] Then there is
[0144]
[0145] The above process transforms the selection of turning points in the time series into the selection of independent variables that have a significant impact on the dependent variable, thereby solving the problem of significance testing for multiple selected mutation points.
[0146] Construct a new set of independent variables, let
[0147] Z 1 =t
[0148] Z 2 =(t-1) +
[0149] Z 3 =(t-2) +
[0150] …
[0151] Z n =(t-(n-1)) +
[0152] The linear regression form is obtained:
[0153]
[0154] Step S4c, using the SP-SRA method to perform turning mutation analysis on the mean-degraded mutation and rate-variable mutation sequences to identify turning mutation points in the sequences.
[0155] Furthermore, the stepwise regression technique is introduced to calculate Z under the condition that the regression coefficients of adjacent independent variables have different signs. i By screening, the optimal point value can be automatically screened out. By analyzing the selected variables, the turning points in the mean mutation and rate mutation sequences can be obtained.
[0156] The present application overcomes the interference of different types of mutations on the identification of mutation processes in complex time series. It can perform the next layer of mutation analysis on the basis of peeling off the known mutation processes layer by layer, and perform hierarchical identification of mutation points and mutation processes in complex time series, thereby improving the accuracy of mutation analysis while ensuring optimal layer-by-layer decomposition.
[0157] It should be noted that, for the sake of convenience, the aforementioned method embodiments are all expressed as a series of action combinations, but those skilled in the art should know that the present invention is not limited to the described order of actions, because according to the present invention, certain steps can be performed in other orders or simultaneously.
[0158] Based on the same idea as the complex time series composite mutation identification method based on forward decomposition in the above embodiment, the present invention also provides a complex time series composite mutation identification system based on forward decomposition, which can be used to execute the above complex time series composite mutation identification method based on forward decomposition. For ease of explanation, the structural schematic diagram of the embodiment of the complex time series composite mutation identification system based on forward decomposition only shows the parts related to the embodiment of the present invention. Those skilled in the art can understand that the illustrated structure does not constitute a limitation on the device, and may include more or fewer components than shown in the diagram, or combine certain components, or arrange the components differently.
[0159] See also Figure 5 In another embodiment of the present application, a complex time series composite mutation identification system 100 based on forward decomposition is provided, the system comprising a model building module, a mean mutation module, a variability mutation module and a turning mutation module;
[0160] The model building module is used to build a layer-by-layer forward decomposition and identification model of a composite mutation of mean mutation-rate mutation-turning mutation; the composite mutation layer-by-layer forward decomposition and identification model regards the mutation in a complex time series as being formed by the coupling of mean mutation, rate mutation and turning mutation, and identifies the three mutations by category, peels off and decomposes them layer by layer to eliminate mutual interference between multiple mutations;
[0161] The mean mutation module is used to perform mean mutation analysis on complex time series to obtain a mean mutation process;
[0162] The rate mutation module is used to separate the mean mutation process from the original time series to obtain a mean mutation sequence without the influence of mean mutation, perform rate mutation analysis on the mean mutation sequence, identify the rate mutation points in the sequence; identify the rate mutation process, and quantify the mutation of the sequence variance state;
[0163] The turning mutation module is used to separate the rate mutation process from the de-mean mutation sequence, obtain the de-mean mutation and rate mutation sequence without the influence of mean mutation and rate mutation, and identify the turning mutation point for the de-mean mutation and rate mutation sequence.
[0164] It should be noted that the complex time series compound mutation identification system based on forward decomposition of the present invention corresponds one to one with the complex time series compound mutation identification method based on forward decomposition of the present invention. The technical features and beneficial effects described in the embodiment of the complex time series compound mutation identification method based on forward decomposition are applicable to the embodiment of the complex time series compound mutation identification based on forward decomposition. For specific contents, please refer to the description in the embodiment of the method of the present invention, which will not be repeated here. This is hereby declared.
[0165] In addition, in the implementation of the complex time series composite mutation identification system based on forward decomposition in the above-mentioned embodiment, the logical division of each program module is only an example. In actual applications, the above-mentioned functions can be assigned to different program modules as needed, for example, for the convenience of corresponding hardware configuration requirements or software implementation. That is, the internal structure of the complex time series composite mutation identification system based on forward decomposition is divided into different program modules to complete all or part of the functions described above.
[0166] See also Figure 6 In one embodiment, an electronic device for implementing a complex time series compound mutation identification method based on forward decomposition is provided. The electronic device 200 may include a first processor 201, a first memory 202 and a bus, and may also include a computer program stored in the first memory 202 and executable on the first processor 201, such as a complex time series compound mutation identification program 203 based on forward decomposition.
[0167] The first memory 202 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, mobile hard disk, multimedia card, card-type memory (for example: SD or DX memory, etc.), magnetic memory, disk, optical disk, etc. In some embodiments, the first memory 202 can be an internal storage unit of the electronic device 200, such as a mobile hard disk of the electronic device 200. In other embodiments, the first memory 202 can also be an external storage device of the electronic device 200, such as a plug-in mobile hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device 200. Further, the first memory 202 can also include both an internal storage unit of the electronic device 200 and an external storage device. The first memory 202 can not only be used to store application software and various types of data installed in the electronic device 200, such as the code of the complex time series composite mutation identification program 203 based on forward decomposition, but also can be used to temporarily store data that has been output or is to be output.
[0168] In some embodiments, the first processor 201 may be composed of an integrated circuit, for example, a single packaged integrated circuit, or a plurality of packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and combinations of various control chips, etc. The first processor 201 is the control core (Control Unit) of the electronic device, and uses various interfaces and lines to connect various components of the entire electronic device, and executes various functions and processes data of the electronic device 200 by running or executing programs or modules stored in the first memory 202, and calling data stored in the first memory 202.
[0169] Figure 6 Only an electronic device with components is shown, and those skilled in the art will understand that Figure 6 The structure shown does not constitute a limitation on the electronic device 200 , and may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.
[0170] The complex time series composite mutation identification program 203 based on forward decomposition stored in the first memory 202 in the electronic device 200 is a combination of multiple instructions. When running in the first processor 201, it can achieve:
[0171] A layer-by-layer forward decomposition and identification model of a composite mutation of mean mutation-rate mutation-turning mutation is constructed; the composite mutation layer-by-layer forward decomposition and identification model regards the mutation in a complex time series as formed by the coupling of mean mutation, rate mutation and turning mutation, identifies the three mutations by category, and decomposes them layer by layer to eliminate mutual interference between multiple mutations;
[0172] The mean mutation process is obtained by performing mean mutation analysis on complex time series;
[0173] The mean mutation process is separated from the original time series to obtain a mean mutation sequence without the influence of mean mutation. The rate mutation analysis is performed on the mean mutation sequence to identify the rate mutation points in the sequence. The rate mutation process is identified to quantify the mutation of the sequence variance state.
[0174] The rate mutation process is separated from the de-mean mutation sequence to obtain the de-mean mutation and rate mutation sequence without the influence of mean mutation and rate mutation. For the de-mean mutation and rate mutation sequence, the turning mutation point is identified.
[0175] Furthermore, if the module / unit integrated in the electronic device 200 is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a non-volatile computer-readable storage medium. The computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, and a read-only memory (ROM).
[0176] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
[0177] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0178] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.
Claims
1. A complex time series compound mutation identification method based on forward decomposition, characterized by: The steps include: A layer-by-layer forward decomposition and identification model of a composite mutation of mean mutation-rate mutation-turning mutation is constructed; the composite mutation layer-by-layer forward decomposition and identification model regards the mutation in a complex time series as formed by the coupling of mean mutation, rate mutation and turning mutation, identifies the three mutations by category, and decomposes them layer by layer to eliminate mutual interference between multiple mutations; The mean mutation process is obtained by performing mean mutation analysis on complex time series; The mean mutation process is separated from the original time series to obtain a mean mutation sequence without the influence of mean mutation. The rate mutation analysis is performed on the mean mutation sequence to identify the rate mutation points in the sequence. The rate mutation process is identified to quantify the mutation of the sequence variance state. The rate mutation process is separated from the de-mean mutation sequence to obtain the de-mean mutation and rate mutation sequence without the influence of mean mutation and rate mutation. For the de-mean mutation and rate mutation sequence, the turning point is identified.
2. The complex time series composite mutation identification method based on forward decomposition according to claim 1 is characterized in that: The mean mutation process obtained by performing mean mutation analysis on the complex time series is specifically as follows: Perform MK test on complex time series to identify all mutation points in complex time series; Perform MTT test on complex time series to examine whether the difference between the mean values of the two sub-time series is significant, and select significant mean mutation points from the MK test mutation points; Identify mean mutation processes and quantify the mutation of the mean value of the sequence.
3. The complex time series composite mutation identification method based on forward decomposition according to claim 2 is characterized in that: The MK test is performed on the complex time series to identify all mutation points in the complex time series, specifically: For a complex time series x with n samples, x={x1,x2,…,x n }, construct a set of ordered columns: in, Under the assumption that the time series are randomly independent, the statistic is defined: in Var(s k ) is the cumulative number s k The mean and variance of UF i is a statistical sequence that obeys the standard normal distribution. Under a given significance level α, the normal distribution table is used to obtain U α If UF i |>U α , it indicates that there is an obvious trend change in the sequence; Repeat the above process in reverse order of time series x, and make UB k =-UF k , k=n,n-1,…,1, UB1=0; Confidence interval [-U α ,U α ] inner UF k With UB k The intersection of is the mutation point, a positive value represents a sudden increase, and a negative value represents a sudden decrease.
4. The complex time series composite mutation identification method based on forward decomposition according to claim 2 is characterized in that: The MTT test is performed on the complex time series to examine whether the difference between the average values of the two sub-time series before and after is significant, and the significant mean mutation points are screened out from the MK test mutation points, specifically: Determine the lengths of the two subsequences x1 and x2 before and after the reference point, take the same length n1 = n2, and the average values of the two subsequences are and The variance is s1 2 and s2 2 , define the statistic: in, The reference points are set continuously by sliding, and the statistics are calculated according to the above formula to obtain the statistical sequence t i , i = 1, 2, ..., n-(n1 + n2) + 1; given a significant level α, look up the t distribution table to get the critical value t α , if |t i | <t α , then it is considered that there is no significant difference in the means of the two subsequences before and after the reference point, otherwise it is considered that a mean mutation occurs at the reference point; The MTT test was used to analyze whether there were significant mean mutations before and after all the mutation points obtained by the MK test, and the mean mutation points above the significant level were screened out.
5. The complex time series composite mutation identification method based on forward decomposition according to claim 1 is characterized in that: The Pettitt method was used to perform rate mutation analysis on the mean-devalued mutation sequence to identify the rate mutation points in the sequence, specifically: For the sequence y={y1,y2,…,y n }, construct the order column again: in, Among them, the order column s k It is the cumulative number of values at the i-th moment that are greater than or less than the values at the j-th moment, reflecting the sharp change of the sequence from one variance state to another; Directly use the order sequence to detect the mutation point of the rate of change. If the moment t0 satisfies Then t0 is the sudden change point of rate of change; If P ≤ 0.5, the detected mutation point is considered to be statistically significant.
6. The complex time series composite mutation identification method based on forward decomposition according to claim 1 is characterized in that: The rate mutation process is separated from the mean mutation sequence to obtain the mean mutation and rate mutation sequence without the influence of mean mutation and rate mutation. For the mean mutation and rate mutation sequence, the specific results are: The rate mutation process is separated from the mean-depleted mutation sequence to obtain the mean-depleted mutation and rate mutation sequence; Introduce linear semi-truncated polynomials to construct new sets of independent variables and linear regression forms; The SP-SRA method was used to perform turning mutation analysis on the mean-degraded mutation and rate-variable mutation sequences to identify the turning mutation points in the sequences.
7. The complex time series composite mutation identification method based on forward decomposition according to claim 6 is characterized in that: The introduction of linear semi-truncated polynomials to construct a new set of independent variables and linear regression form is specifically: For the sequence z={z1,z2,…,z n }, suppose that in this time series t c The point is a mutation point, which means that t c Before and after, two subsequences t1=1,2,…,t c and t1=t c+1 ,t c+2 …,n have significant differences in their changing trends. Based on the above assumptions, n-1 points are inserted into the time interval [1,n] of the sequence, and the time variable t is divided into n small intervals, recorded as (1,2], (2,3], …, (n-1,n]; in each interval, a set of linear functions is used to describe the changes of the sequence z: Introduce linear semi-truncated polynomials: Then we have: The above process transforms the selection of turning points in the time series into the selection of independent variables that have a significant impact on the dependent variable, thereby solving the problem of significance testing for multiple selected mutation points; Construct a new set of independent variables, let Z1=t Z2=(t-1) + Z3=(t-2) + … Z n =(t-(n-1)) + The linear regression form is obtained:
8. The complex time series composite mutation identification system based on forward decomposition according to claim 1 is characterized in that: Model building module, mean mutation module, rate mutation module and transition mutation module; The model building module is used to build a layer-by-layer forward decomposition and identification model of a composite mutation of mean mutation-rate mutation-turning mutation; the composite mutation layer-by-layer forward decomposition and identification model regards the mutation in a complex time series as being formed by the coupling of mean mutation, rate mutation and turning mutation, and identifies the three mutations by category, peels off and decomposes them layer by layer to eliminate mutual interference between multiple mutations; The mean mutation module is used to perform mean mutation analysis on complex time series to obtain a mean mutation process; The rate mutation module is used to separate the mean mutation process from the original time series to obtain a mean mutation sequence without the influence of mean mutation, perform rate mutation analysis on the mean mutation sequence, identify the rate mutation points in the sequence; identify the rate mutation process, and quantify the mutation of the sequence variance state; The turning mutation module is used to separate the rate mutation process from the de-mean mutation sequence, obtain the de-mean mutation and rate mutation sequence without the influence of mean mutation and rate mutation, and identify the turning mutation point for the de-mean mutation and rate mutation sequence.
9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores computer program instructions that can be executed by the at least one processor, and the computer program instructions are executed by the at least one processor so that the at least one processor can execute the complex time series composite mutation identification method based on forward decomposition as described in any one of claims 1-7.
10. A computer-readable storage medium storing a program, characterized in that: When the program is executed by a processor, the complex time series composite mutation identification method based on forward decomposition according to any one of claims 1 to 7 is implemented.