A metastatic breast cancer data preprocessing method

By obtaining the iterative scale target degree of breast cancer data, screening the optimal iterative scale concentrated noise, solving the problem of difficult to determine the location of noise signals in the prior art, and improving the preprocessing accuracy and noise removal effect of breast cancer data.

CN118709126BActive Publication Date: 2024-11-22FIRST PEOPLES HOSPITAL OF YUNNAN PROVINCE
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411201001.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-29
Publication Date
2024-11-22
Estimated Expiration
2044-08-29

AI Technical Summary

Technical Problem

When the existing time-scale decomposition algorithms process metastatic breast cancer data, the location of the noise signal is difficult to determine, resulting in high noise processing complexity and affecting the data preprocessing accuracy.

Method used

By obtaining the original signal target degree at each iteration scale, filter out the optimal iteration scale, concentrate all complex noise components into the first two IMF component signals, reduce the complexity of noise signal processing, and obtain denoising through the optimal iteration scale protein signal sequence in breast cancer patients.

Benefits of technology

It improves the preprocessing accuracy of metastatic breast cancer data, simplifies the noise signal processing process, clarifies the noise location, and enhances the signal denoising effect.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118709126B_ABST
    Figure CN118709126B_ABST
Patent Text Reader

Abstract

The present invention relates to a metastatic breast cancer data preprocessing method, belonging to the technical field of data processing, and the present invention comprises four main steps: obtaining a protein signal time series of a breast cancer patient; obtaining all IMF component signals of an original signal under different iteration scales, obtaining a distribution feature extraction factor of each IMF component signal and a contribution rate to the original signal; obtaining the correlation of the convergent decomposition component singularity of each IMF component signal according to the decomposition component singularity of each IMF component signal; obtaining an optimal iteration scale according to the original signal target degree under each iteration scale, and obtaining a protein signal sequence of a breast cancer patient after denoising through the optimal iteration scale. The present invention can obtain the original signal target degree under each iteration scale, so that all complex noise components are concentrated in the first two IMF component signals, reduce the complexity of noise signal processing, and improve the preprocessing accuracy of metastatic breast cancer data.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of data processing, and in particular, relates to a metastatic breast cancer data preprocessing method. Background Art

[0002] Metastatic breast cancer refers to a situation in which cancer cells spread from a primary tumor to other parts of the body. Data sets related to breast cancer may include multiple types of information, of which signal data is a part. Signal data includes biological signals related to tumors, such as protein signals. Various factors in the experiment or measurement process may cause random or systematic errors, that is, noise, which in turn will mask the true biological signals. The quality of the data and the accuracy of the analysis can be improved by removing the noise. Due to the complexity of the frequency of the noise signal, the protein signal can be analyzed using the intrinsic time decomposition (ITD) algorithm.

[0003] The existing time scale decomposition algorithm selects the iterative scale of the decomposition iterative time scale based on adjacent extreme points, which causes noise of different frequencies to appear in the component signal, making it difficult to determine the location of the signal noise, thereby increasing the complexity of subsequent signal denoising operations. Summary of the invention

[0004] In order to overcome the problems existing in the background technology, the present invention provides a metastatic breast cancer data preprocessing method, which can obtain the target degree of the original signal at each iteration scale, so that all complex noise components are concentrated in the first two IMF component signals, reduce the complexity of noise signal processing, and improve the preprocessing accuracy of metastatic breast cancer data.

[0005] To achieve the above object, the present invention is implemented through the following technical solutions:

[0006] The metastatic breast cancer data preprocessing method comprises the following steps:

[0007] S1, obtaining the protein content of breast cancer patient tissue samples to form a protein signal time series, recorded as the original signal;

[0008] S2, decomposing the original signal at different iteration time scales, obtaining all IMF component signals of the original signal at different iteration scales, and obtaining the distribution feature extraction factor of each IMF component signal, and obtaining the contribution rate of each IMF component signal to the original signal according to the difference in the distribution feature extraction factor between each IMF component signal and the original signal;

[0009] S3, dividing each IMF component signal into component signal segments, and obtaining the decomposition component singularity of each IMF component signal according to the change of the signal amplitude in all component signal segments; and obtaining the correlation of the convergent decomposition component singularity of each IMF component signal according to the decomposition component singularity of each IMF component signal;

[0010] S4, according to the contribution rate of each IMF component signal of the original signal to the original signal at each iterative time scale and the correlation with the singleness of the convergent decomposition component, the target degree of the original signal at each iterative scale is obtained; according to the size of the target degree of the original signal, the optimal iterative scale among all iterative scales is screened out; and the protein signal sequence of the breast cancer patient after denoising is obtained through the optimal iterative scale.

[0011] Further, according to the distribution of the spectrum signal of each IMF component signal, the distribution feature extraction factor of each IMF component signal is obtained. The specific method is: each IMF component signal is divided into several single-peak signals, and the amplitude and bandwidth of each single-peak signal are obtained; then The calculation method of the distribution feature extraction factor of the IMF component signal is:

[0012] ;

[0013] In the formula, Indicates The distribution feature extraction factor of the IMF component signal; Indicates The variance of the amplitude of all single-peak signals of the IMF component signal; Indicates The number of all single-peak signals of the IMF component signal; Indicates The first IMF component signal The bandwidth of a single peak signal.

[0014] Furthermore, the specific method for obtaining the amplitude and bandwidth of each single-peak signal is as follows: using fast Fourier transform to obtain the spectrum signals of all IMF component signals of the original signal, taking all the minimum values ​​of the spectrum signals of each IMF component signal as segmentation points, taking the spectrum signal segment between any two adjacent segmentation points as a single-peak signal of the IMF component signal, and then dividing each IMF component signal into several single-peak signals; the ordinate of the maximum point in each single-peak signal of each IMF component signal is taken as the amplitude of each single-peak signal of each IMF component signal; taking the ordinate of the maximum point in each single-peak signal of each IMF component signal as the amplitude of each single-peak signal of each IMF component signal; taking the The first IMF component signal The product of the amplitude of the single-peak signal and 0.5 is recorded as the first height, and the horizontal line passing through the first height is named the first height horizontal line. The two intersection points of the first single-peak signal and the first height horizontal line are taken as the two dividing points of the single-peak signal. The absolute value of the difference between the horizontal coordinates of the two dividing points of the single peak signal is taken as the The bandwidth of a single peak signal.

[0015] Furthermore, the contribution rate of each IMF component signal to the original signal is obtained by: The absolute value of the difference between any two adjacent extreme points of the IMF component signal is taken as the The absolute value of the difference between the horizontal coordinates of any two adjacent extreme points of the original signal is taken as a time difference of the original signal. The calculation method of the contribution rate of each IMF component signal to the original signal is:

[0016]

[0017] In the formula, Indicates The contribution rate of the IMF component signal to the original signal; Indicates The distribution feature extraction factor of the IMF component signal; Represents the distribution feature extraction factor of the original signal; Indicates The total number of all time differences of the IMF component signals; Represents the total number of all time differences of the original signal; Indicates The first IMF component signal time difference; The original signal time difference; Indicates The maximum value of the protein signal of the IMF component signal; Indicates The minimum protein signal of the IMF component signal; The minimum value of protein signal representing the raw signal; The protein signal maximum value represents the raw signal.

[0018] Furthermore, the method for dividing each IMF component signal into several component signal segments is to use all extreme points of each IMF component signal as segmentation points, and use the signal segment between any two adjacent segmentation points as a component signal segment of the IMF component signal, thereby dividing each IMF component signal into several component signal segments.

[0019] Further, the method for obtaining the uniqueness of the decomposition components of each IMF component signal is: obtain the first The first IMF component signal The characteristic frequency of the component signal segment; The calculation method of the unity of the decomposition components of the IMF component signal is:

[0020]

[0021] In the formula, Indicates The decomposition component singularity of the IMF component signal; Indicates The total number of all component signal segments of an IMF component signal; Indicates The first IMF component signal The characteristic frequency of each component signal segment; Represents the total number of all IMF component signals of the original signal; Represents a logarithmic function with a natural constant as base.

[0022] Further, the obtaining The first IMF component signal The method of calculating the characteristic frequency of the component signal segment is as follows: The standard deviation of all protein signals within the component signal segment is taken as The first IMF component signal The characteristic value of the component signal segment; among all IMF component signals, all The first IMF component signal The first IMF component signal with the same eigenvalues The total number of component signal segments is recorded as the first number; the ratio of the first number to the total number of all IMF component signals is recorded as the The first IMF component signal The characteristic frequency of each component signal segment.

[0023] Further, the method for obtaining the correlation of the unity of the convergent decomposition components of each IMF component signal is: obtain the first The convergence data set and the unity data set of the IMF component signal; then The calculation method of the correlation of the unity of the convergent decomposition components of the IMF component signal is:

[0024]

[0025] In the formula, Indicates The correlation of the convergence decomposition components of the IMF component signals; Indicates A data set of convergence amounts of IMF component signals; Represents a single data set; Indicates The covariance between the convergence data set and the unity data set of the IMF component signals; Indicates The standard deviation of the convergence data set of the IMF component signals; Represents the standard deviation of a single data set; Indicates taking the absolute value.

[0026] Further, get the The method of obtaining the convergence data set and the unity data set of the IMF component signal is as follows: The first IMF component signal The upper and lower envelopes of the first component signal segment are obtained, and the average of the upper and lower envelopes is used as the The first IMF component signal The mean envelope of the component signal segment is The first IMF component signal The sum of all ordinates in the component signal segment is taken as the The first IMF component signal The mean envelope convergence value of the component signal segment; The set of mean envelope convergence values ​​of all component signal segments of the IMF component signal is recorded as The convergence data set of the IMF component signals is recorded as the unity data set. The set of the unity of the decomposed components of all IMF component signals is recorded as the unity data set.

[0027] Further, the method for obtaining the target degree of the original signal at each iteration scale is: The upper and lower envelopes of the original signal under the iteration scale are taken as the average of the upper and lower envelopes. The mean envelope of the original signal under the iteration scale is The sum of all ordinates in the original signal under the iteration scale is taken as the The convergence of the mean envelope of the original signal under the iteration scale; The calculation method of the original signal target degree under the iteration scale is:

[0028]

[0029] In the formula, Indicates The original signal target degree at each iteration scale; Indicates The original signal at the iteration scale The contribution rate of the IMF component signal to the original signal; Indicates The original signal at the iteration scale The contribution rate of the IMF component signal to the original signal; Indicates The total number of all IMF component signals of the original signal under the iteration scale; Indicates The original signal at the iteration scale The contribution rate of the IMF component signal to the original signal; Indicates The original signal at the iteration scale The correlation of the convergence decomposition components of the IMF component signals; Indicates The convergence of the mean envelope of the original signal under the iteration scale; It represents the average value of the mean envelope convergence of the original signal under all iteration scales; are preset parameters.

[0030] Beneficial effects of the present invention:

[0031] The present invention obtains the target degree of the original signal at each iteration scale according to the correlation between the contribution rate of each IMF component signal of the original signal at each iteration scale to the original signal and the unity of the convergent decomposition components; uses this as the iteration scale selection of the decomposition iteration time scale, so that all complex noise components are concentrated in the first two IMF component signals, which can better locate the signal noise position and reduce the complexity of noise signal processing; selects the optimal iteration scale according to the size of the target degree of the original signal; obtains the protein signal sequence of the breast cancer patient after denoising through the optimal iteration scale, thereby improving the preprocessing accuracy of metastatic breast cancer data. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a data processing flow chart of the present invention. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solution and beneficial effects of the present invention clearer, the technical solution of the present invention will be described in detail below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, other embodiments obtained by those skilled in the art without creative work are all within the scope of protection of the present invention.

[0034] Example

[0035] The processing steps of the metastatic breast cancer data preprocessing method provided by the present invention are as follows: Figure 1 As shown, specifically:

[0036] Step S1: Obtaining protein signal time series of breast cancer patients.

[0037] It should be noted that the specific content of metastatic breast cancer data is a data set related to breast cancer, including clinical data, pathological data, molecular biology data, medical imaging data, protein signal data, etc.; this embodiment mainly describes the protein signal data of breast cancer patients.

[0038] Specifically, we first need to collect the protein signal time series of breast cancer patients. The specific process is as follows:

[0039] All protein signals of breast cancer patients within 24 hours are combined into a protein signal time series of breast cancer patients; the protein signal time series of breast cancer patients is recorded as the original signal; the original signal is input into a two-dimensional coordinate system, with the horizontal axis being time and the vertical axis being the protein signal corresponding to each moment.

[0040] The specific process of collecting protein signals from breast cancer patients is as follows: collecting tissue samples from breast cancer patients, performing cell lysis, ultrasonic crushing and centrifugation on the tissue samples from breast cancer patients to obtain the protein content in the tissues of breast cancer patients, and using the protein content in the tissues of breast cancer patients as the protein signals of breast cancer patients.

[0041] So far, the protein signal time sequence of breast cancer patients has been obtained through the above method.

[0042] Step S2: Obtain all IMF component signals of the original signal at different iteration scales, and obtain the distribution feature extraction factor of each IMF component signal and its contribution rate to the original signal.

[0043] It should be noted that the process of obtaining protein signal data sequences of breast cancer patients will be affected by many factors. For example, the electromagnetic pulse of the measuring instrument or the signal of the experimental background will interfere with the detection of protein content. The influence of multiple noise sources causes noise signals of different frequencies to be distributed in the protein signal. In the process of inherent time scale decomposition, due to the small original time decomposition scale, the component signals are too single and the number is large. The noise signals are distributed in different component signals, which makes it difficult to determine the position of the noise signal, thereby increasing the complexity of noise signal processing.

[0044] Specifically, the original signal is decomposed into different iterative time scales using the intrinsic time scale decomposition algorithm to obtain all IMF component signals of the original signal at any iterative scale.

[0045] Since the IMF component signals are obtained by decomposing the original signals, the horizontal axis of each IMF component signal is time, and the vertical axis is the protein signal corresponding to each moment. All IMF component signals are recorded as the first IMF component signal to the Nth IMF component signal in the order of protein signals from large to small, and N represents the total number of IMF component signals.

[0046] Among them, the intrinsic time scale decomposition algorithm is a prior art and will not be described in detail in this embodiment.

[0047] (1) Obtain the distribution feature extraction factor of each IMF component signal

[0048] Specifically, the fast Fourier transform (FFT) is used to obtain the spectrum signals of all IMF component signals of the original signal. The horizontal axis of the spectrum signal is the frequency, and the vertical axis is the amplitude corresponding to each frequency. The spectrum signal records the information of the protein content of each IMF component signal of the original signal at different frequencies.

[0049] Among them, fast Fourier transform is a prior art and will not be described in detail in this embodiment.

[0050] All the minimum values ​​of the spectrum signal of each IMF component signal are taken as the segmentation points, and the spectrum signal segment between any two adjacent segmentation points is taken as a single-peak signal of the IMF component signal, and then each IMF component signal is divided into several single-peak signals; the ordinate of the maximum point in each single-peak signal of each IMF component signal is taken as the amplitude of each single-peak signal of each IMF component signal; The first IMF component signal The product of the amplitude of the single-peak signal and 0.5 is recorded as the first height, and the horizontal line passing through the first height is named the first height horizontal line. The two intersection points of the first single-peak signal and the first height horizontal line are taken as the two dividing points of the single-peak signal. The absolute value of the difference between the horizontal coordinates of the two dividing points of the single peak signal is taken as the The bandwidth of a single-peak signal; The calculation method of the distribution feature extraction factor of the IMF component signal is:

[0051]

[0052] In the formula, Indicates The distribution feature extraction factor of the IMF component signal; Indicates The variance of the amplitude of all single-peak signals of the IMF component signal; Indicates The number of all single-peak signals of the IMF component signal; Indicates The first IMF component signal The bandwidth of a single peak signal.

[0053] It should be noted that Indicates the overall fluctuation degree of the component signal. The larger the value, the greater the fluctuation degree of the component signal. The component signal may contain signal components of different frequencies. The smaller it is, the higher the frequency uniformity of the component signal is. The number of unimodal signals cannot be used as the only clue to determine the frequency uniformity of the component signal, so the bandwidth of the unimodal signal is introduced.

[0054] At this point, the distribution feature extraction factor of each IMF component signal is obtained.

[0055] (2) Obtain the contribution rate of each IMF component signal to the original signal.

[0056] It should be noted that in the process of inherent time scale decomposition, different component signals represent different time decomposition scales, that is, different component signals represent signal components of a certain frequency in the original signal; in order to effectively process the noise signal later, it is necessary to make the component signal not only maintain its original singleness, but also contain more components in the original signal, that is, the component signal must also have a greater contribution rate to the original signal.

[0057] Specifically, the The absolute value of the difference between any two adjacent extreme points of the IMF component signal is taken as the The absolute value of the difference between the horizontal coordinates of any two adjacent extreme points of the original signal is taken as a time difference of the original signal; then The calculation method of the contribution rate of each IMF component signal to the original signal is:

[0058]

[0059] In the formula, Indicates The contribution rate of the IMF component signal to the original signal; Indicates The distribution feature extraction factor of the IMF component signal; Represents the distribution feature extraction factor of the original signal; Indicates The total number of all time differences of the IMF component signals; Represents the total number of all time differences of the original signal; Indicates The first IMF component signal time difference; The original signal time difference; Indicates The maximum value of the protein signal of the IMF component signal; Indicates The minimum protein signal of the IMF component signal; The minimum value of protein signal representing the raw signal; The protein signal maximum value represents the raw signal.

[0060] It should be noted that the calculation method of the distribution feature extraction factor of the original signal is the same as that of the first The calculation method of the distribution feature extraction factor of each IMF component signal is consistent; Indicates the difference between the overall distribution characteristics of the component signal and the overall distribution of the original signal. The larger the value, the higher the similarity between the distribution characteristics of the component signal and the original signal, that is, the more signal components in the original signal the component signal contains; It indicates the characteristic difference between the component signal and the original signal. The larger the value, the higher the contribution rate of the distribution characteristics of the component signal to the distribution characteristics of the original signal.

[0061] So far, the contribution rate of each IMF component signal to the original signal is obtained through the above method.

[0062] Step S3: according to the decomposition component singularity of each IMF component signal, obtain the correlation of the convergent decomposition component singularity of each IMF component signal.

[0063] All extreme value points of the original signal are taken as segmentation points, and the signal segment between any two adjacent segmentation points is taken as an original signal segment of the original signal, thereby dividing the original signal into a plurality of original signal segments; all extreme value points of each IMF component signal are taken as segmentation points, and the signal segment between any two adjacent segmentation points is taken as a component signal segment of the IMF component signal, thereby dividing each IMF component signal into a plurality of component signal segments; wherein each original signal segment of the original signal corresponds one-to-one to the serial numbers of the component signal segments of all the IMF component signals.

[0064] For The first IMF component signal component signal segment, The standard deviation of all protein signals within the component signal segment is taken as The first IMF component signal The characteristic value of the component signal segment; among all IMF component signals, all The first IMF component signal The first IMF component signal with the same eigenvalues The total number of component signal segments is recorded as the first number; the ratio of the first number to the total number of all IMF component signals is recorded as the The first IMF component signal The characteristic frequency of the component signal segment; The calculation method of the unity of the decomposition components of the IMF component signal is:

[0065]

[0066] In the formula, Indicates The decomposition component singularity of the IMF component signal; Indicates The total number of all component signal segments of an IMF component signal; Indicates The first IMF component signal The characteristic frequency of each component signal segment; Represents the total number of all IMF component signals of the original signal; Represents a logarithmic function with a natural constant as base.

[0067] It should be noted that Indicates The information entropy between the eigenvalues ​​of different component signal segments in the IMF component signal is smaller, indicating that the The fewer the types of information components contained in the first IMF component signal, the The better the decomposition of the IMF component signal, the larger the entropy value, the better the The more complex the signal components in the IMF component signal, the less uniform the decomposition components will be.

[0068] At this point, the decomposition component unity of each IMF component signal is obtained.

[0069] Get the The first IMF component signal The upper and lower envelopes of the first component signal segment are obtained, and the average of the upper and lower envelopes is used as the The first IMF component signal The mean envelope of the component signal segment is The first IMF component signal The sum of all ordinates in the component signal segment is taken as the The first IMF component signal The mean envelope convergence value of the component signal segment; The set of mean envelope convergence values ​​of all component signal segments of the IMF component signal is recorded as The convergence data set of the IMF component signals; the set of the decomposition components of all IMF component signals is composed of unity, which is recorded as the unity data set; then The calculation method of the correlation of the unity of the convergent decomposition components of the IMF component signal is:

[0070]

[0071] In the formula, Indicates The correlation of the convergence decomposition components of the IMF component signals; Indicates A data set of convergence amounts of IMF component signals; Represents a single data set; Indicates The covariance between the convergence data set and the unity data set of the IMF component signals; Indicates The standard deviation of the convergence data set of the IMF component signals; Represents the standard deviation of a single data set; Indicates taking the absolute value.

[0072] At this point, the correlation of the unity of the converged decomposition components of each IMF component signal is obtained.

[0073] Step S4: according to the target degree of the original signal under each iteration scale, the optimal iteration scale is obtained; and the protein signal sequence of the breast cancer patient after denoising is obtained through the optimal iteration scale.

[0074] Preset a parameter , wherein this embodiment is based on This example is described as an example, and this embodiment is not specifically limited. It depends on the specific implementation situation.

[0075] Get the The upper and lower envelopes of the original signal under the iteration scale are taken as the average of the upper and lower envelopes. The mean envelope of the original signal under the iteration scale is The sum of all ordinates in the original signal under the iteration scale is taken as the The convergence of the mean envelope of the original signal under the iteration scale; The calculation method of the original signal target degree under the iteration scale is:

[0076]

[0077] In the formula, Indicates The original signal target degree at each iteration scale; Indicates The original signal at the iteration scale The contribution rate of the IMF component signal to the original signal; Indicates The original signal at the iteration scale The contribution rate of the IMF component signal to the original signal; Indicates The total number of all IMF component signals of the original signal under the iteration scale; Indicates The original signal at the iteration scale The contribution rate of the IMF component signal to the original signal; Indicates The original signal at the iteration scale The correlation of the convergence decomposition components of the IMF component signals; Indicates The convergence of the mean envelope of the original signal under the iteration scale; It represents the average value of the mean envelope convergence of the original signal under all iteration scales; are preset parameters.

[0078] It should be noted that The larger the In the iterative scale decomposition process, the stronger the correlation between the envelope convergence of different IMF component signals and the unity of component signals, the stronger the correlation between the envelope convergence of different IMF component signals and the unity of component signals. The greater the convergence of the original signal envelope at the iteration scale, the The worse the unity of the IMF component signal of the original signal at each iteration scale, the better the denoising effect; Indicates The complexity of the iteration scale. The higher the complexity, the The worse the unity of the IMF component signal of the original signal at each iteration scale, the better the denoising effect.

[0079] At this point, the target level of the original signal at each iteration scale is obtained.

[0080] As a preferred solution, the iteration scale when the target degree of the original signal is the largest is selected as the optimal iteration scale; the original signal is subjected to the inherent time scale decomposition method through the optimal iteration scale to obtain all the IMF component signals of the original signal, the first and second IMF component signals of the original signal are removed, and all the remaining IMF component signals are merged and reconstructed to obtain the reconstructed original signal, and the reconstructed original signal is used as the protein signal sequence of the breast cancer patient after denoising.

[0081] Finally, it should be noted that the above preferred embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail through the above preferred embodiments, those skilled in the art should understand that various changes can be made in form and details without departing from the scope defined by the claims of the present invention.

Claims

1. A metastatic breast cancer data preprocessing method, characterized in that: The following steps are involved: S1, obtaining the protein content of breast cancer patient tissue samples to form a protein signal time series, recorded as the original signal; S2, decomposing the original signal at different iteration time scales, obtaining all IMF component signals of the original signal at different iteration scales, and obtaining the distribution feature extraction factor of each IMF component signal, and obtaining the contribution rate of each IMF component signal to the original signal according to the difference in the distribution feature extraction factor between each IMF component signal and the original signal; S3, dividing each IMF component signal into component signal segments, and obtaining the decomposition component singularity of each IMF component signal according to the change of the signal amplitude in all component signal segments; and obtaining the correlation of the convergent decomposition component singularity of each IMF component signal according to the decomposition component singularity of each IMF component signal; S4, according to the correlation between the contribution rate of each IMF component signal of the original signal to the original signal at each iteration time scale and the unity of the converged decomposition components, the target degree of the original signal at each iteration scale is obtained; according to the size of the target degree of the original signal, the optimal iteration scale among all iteration scales is screened out; Obtain the denoised protein signal sequence of breast cancer patients through the optimal iterative scale; Among them, in step S2, According to the distribution of the spectrum signal of each IMF component signal, the distribution feature extraction factor of each IMF component signal is obtained. The specific method is to divide each IMF component signal into several single-peak signals and obtain the amplitude and bandwidth of each single-peak signal; then the calculation method of the distribution feature extraction factor of the i-th IMF component signal is: In the formula, w i represents the distribution feature extraction factor of the i-th IMF component signal; represents the variance of the amplitude of all single-peak signals of the i-th IMF component signal; m i represents the number of all single-peak signals of the i-th IMF component signal; G i,n represents the bandwidth of the nth single-peak signal of the i-th IMF component signal; The specific method for obtaining the amplitude and bandwidth of each single-peak signal is to use fast Fourier transform to obtain the spectrum signals of all IMF component signals of the original signal, take all the minimum values ​​of the spectrum signals of each IMF component signal as segmentation points, take the spectrum signal segment between any two adjacent segmentation points as a single-peak signal of the IMF component signal, and then divide each IMF component signal into several single-peak signals; the ordinate of the maximum point in each single-peak signal of each IMF component signal is taken as the amplitude of each single-peak signal of each IMF component signal; the product of the amplitude of the j-th single-peak signal of the ith IMF component signal and 0.5 is recorded as the first height, the horizontal line passing through the first height is named the first height horizontal line, the two intersection points of the j-th single-peak signal and the first height horizontal line are taken as the two dividing points of the single-peak signal, and the absolute value of the difference between the horizontal coordinates of the two dividing points of the j-th single-peak signal is taken as the bandwidth of the j-th single-peak signal; The method for obtaining the contribution rate of each IMF component signal to the original signal is to use the absolute value of the difference between the horizontal coordinates of any two adjacent extreme points of the i-th IMF component signal as a time difference of the i-th IMF component signal; the absolute value of the difference between the horizontal coordinates of any two adjacent extreme points of the original signal is used as a time difference of the original signal, and the calculation method of the contribution rate of the i-th IMF component signal to the original signal is: In the formula, p j represents the contribution rate of the i-th IMF component signal to the original signal; w j represents the distribution feature extraction factor of the i-th IMF component signal; W represents the distribution feature extraction factor of the original signal; A j represents the total number of all time differences of the i-th IMF component signal; M represents the total number of all time differences of the original signal; t a represents the ath time difference of the i-th IMF component signal; T e represents the e-th time difference of the original signal; q j,max represents the maximum value of the protein signal of the i-th IMF component signal; q j ,min represents the minimum value of the protein signal of the i-th IMF component signal; Q min The protein signal minimum value representing the original signal; Q max The maximum value of protein signal represents the raw signal; In step S4, the method for obtaining the target degree of the original signal at each iteration scale is: The upper envelope and the lower envelope of the original signal at the l-th iteration scale are obtained, and the mean of the upper envelope and the lower envelope is used as the mean envelope of the original signal at the l-th iteration scale. The sum of all ordinates in the original signal at the l-th iteration scale is used as the convergence amount of the mean envelope of the original signal at the l-th iteration scale. Then the calculation method of the target degree of the original signal at the l-th iteration scale is: In the formula, R l represents the original signal target degree at the lth iteration scale; p l,1 represents the contribution rate of the i-th IMF component signal of the original signal to the original signal at the l-th iteration scale; p l , 2 represents the contribution rate of the second IMF component signal of the original signal to the original signal at the lth iteration scale; N l represents the total number of all IMF component signals of the original signal at the lth iteration scale; p d represents the contribution rate of the dth IMF component signal of the original signal at the lth iteration scale to the original signal; ρli represents the correlation of the unity of the convergent decomposition components of the ith IMF component signal of the original signal at the lth iteration scale; K l It represents the convergence of the mean envelope of the original signal at the lth iteration scale; It represents the average value of the mean envelope convergence of the original signal under all iteration scales; L is a preset parameter.

2. The metastatic breast cancer data preprocessing method according to claim 1, characterized in that: The method for dividing each IMF component signal into several component signal segments is to use all extreme value points of each IMF component signal as segmentation points, and use the signal segment between any two adjacent segmentation points as a component signal segment of the IMF component signal, thereby dividing each IMF component signal into several component signal segments.

3. The metastatic breast cancer data preprocessing method according to claim 1, characterized in that: The method to obtain the unity of the decomposition components of each IMF component signal is, Obtain the characteristic frequency of the rth component signal segment of the i-th IMF component signal; then the calculation method for the unity of the decomposition components of the i-th IMF component signal is: In the formula, c i Indicates the uniqueness of the decomposition components of the i-th IMF component signal; B i represents the i-th IMF component signal The total number of all component signal segments of the signal; i,r represents the characteristic frequency of the rth component signal segment of the i-th IMF component signal; N represents the total number of all IMF component signals of the original signal; ln() represents a logarithmic function with a natural constant as the base.

4. The metastatic breast cancer data preprocessing method according to claim 3, characterized in that: The method for obtaining the characteristic frequency of the rth component signal segment of the i-th IMF component signal is: The standard deviation of all protein signals in the rth component signal segment is taken as the characteristic value of the rth component signal segment of the i-th IMF component signal; Among all IMF component signals, the total number of r-th component signal segments of all IMF component signals having the same characteristic value as the r-th component signal segment of the i-th IMF component signal is recorded as the first number; and the ratio of the first number to the total number of all IMF component signals is taken as the characteristic frequency of the r-th component signal segment of the i-th IMF component signal.

5. The metastatic breast cancer data preprocessing method according to claim 1, characterized in that: The method to obtain the correlation of the unity of the convergent decomposition components of each IMF component signal is, Obtain the convergence data set and the unity data set of the i-th IMF component signal; then the calculation method of the correlation of the unity of the convergence decomposition components of the i-th IMF component signal is: In the formula, ρ i represents the correlation of the unity of the convergent decomposition components of the i-th IMF component signal; K i represents the convergence data set of the i-th IMF component signal; C represents the singleness data set; Cov(K i , C) represents the covariance between the convergence data set and the unity data set of the i-th IMF component signal; σ(K i ) represents the standard deviation of the convergence data set of the i-th IMF component signal; σ(C) represents the standard deviation of the unity data set; | | represents the absolute value.

6. The metastatic breast cancer data preprocessing method according to claim 5, characterized in that: The method for obtaining the convergence data set and the unity data set of the i-th IMF component signal is: The upper envelope and the lower envelope of the rth component signal segment of the ith IMF component signal are obtained, and the average of the upper envelope and the lower envelope is used as the mean envelope of the rth component signal segment of the ith IMF component signal; the sum of all ordinates in the rth component signal segment of the ith IMF component signal is used as the mean envelope convergence amount of the rth component signal segment of the ith IMF component signal; the set consisting of the mean envelope convergence amounts of all component signal segments of the ith IMF component signal is recorded as the convergence amount data set of the ith IMF component signal; the set consisting of the unity of the decomposition components of all IMF component signals is recorded as the unity data set.