A medical data processing method and system for inspection and examination

Abstract

The present invention relates to the technical field of medical data processing, and particularly relates to a medical data processing method and system for inspection and examination. First, the present invention determines the search radius of the antigen immune response spectrum according to the data values and eigenvalue of each principal component of each antigen immune response spectrum; obtains the need for radius adjustment of the graph to be analyzed according to the difference between the graph to be analyzed and the overall representativeness of all antigen immune response spectra in the local range; adjusts the search radius according to the need for radius adjustment to obtain the adjusted search radius of the graph to be analyzed; and then performs DBSCAN clustering on all antigen immune response spectra to obtain each antigen immune response spectrum cluster. The present invention reasonably sets the adjusted search radius of the antigen immune response spectrum, can better capture the key information in the antigen immune response spectrum, and improves the accuracy of the clustering result of the antigen immune response spectrum.

Description

Technical Field

[0001] The present invention relates to the technical field of medical data processing, and particularly relates to a medical data processing method and system for inspection and examination. Background Art

[0002] Medical data for inspection and examination refers to data on various physical indicators of patients obtained through means such as clinical laboratory tests, medical imaging examinations, and bioelectric measurements, which is of great significance for assisting treatment plans. As the result of one of the detection means, the antigen immune response spectrum is a graph showing the change of antigen reaction intensity with antigen concentration measured by experimental methods. The antigen immune response spectrum can intuitively display the characteristics of the immune response, including affinity, specificity, and sensitivity, etc., which is crucial for the medical evaluation and treatment of all patients. Classifying the antigen immune response spectra of all patients can provide a basis for doctors to formulate personalized treatment plans.

[0003] The DBSCAN algorithm is a widely used clustering algorithm and has been widely applied in the task of classifying antigen immune response spectra. When the prior art uses the DBSCAN algorithm to classify all antigen immune response spectra, it often uses a fixed search radius for clustering, without fully considering the feature representativeness of the antigen immune response spectra, and it is difficult to accurately capture the key information in the antigen immune response spectra, resulting in inaccurate clustering results of the antigen immune response spectra. Summary of the Invention

[0004] In order to solve the technical problem of inaccurate clustering of antigen immune response spectra in the prior art, the purpose of the present invention is to provide a medical data processing method and system for inspection and examination, and the specific technical solutions adopted are as follows:

[0005] A medical data processing method for inspection and examination, the method includes:

[0006] Obtain the antigen signal data of each antigen immune response spectrum and all patient characteristic indicators;

[0007] Perform dimensionality reduction processing according to the antigen signal data of all the antigen immune response spectra to obtain the data values and eigenvalue of each principal component of each antigen immune response spectrum; determine the search radius of the antigen immune response spectrum according to the data values and eigenvalue of each principal component of each antigen immune response spectrum;

[0008] Take any one of the antigen immune response spectra as the graph to be analyzed, and take any one of the patient characteristic indicators as the indicator to be analyzed; according to the deviation from the central tendency of the indicator to be analyzed in the graph to be analyzed, obtain the characteristic representativeness of the indicator to be analyzed in the graph to be analyzed; according to the correlation between the data values of all the antigen immune response spectra corresponding to the indicator to be analyzed and the principal components, obtain the influence measure of the indicator to be analyzed; according to the characteristic representativeness of all the patient characteristic indicators in the graph to be analyzed and the influence measure, obtain the overall representativeness of the graph to be analyzed; according to the difference between the graph to be analyzed and the overall representativeness of all the antigen immune response spectra in its local range, obtain the need for radius adjustment of the graph to be analyzed; adjust the search radius according to the need for radius adjustment to obtain the adjusted search radius of the graph to be analyzed.

[0009] Perform DBSCAN clustering on all the antigen immune response spectra according to the adjusted search radius of each antigen immune response spectrum to obtain each antigen immune response spectrum cluster.

[0010] Furthermore, the method for obtaining the data values and eigenvalues of the principal components includes:

[0011] Using the PCA method, perform dimensionality reduction processing on the antigen signal data of all the antigen immune response spectra to obtain the data values and eigenvalues of each principal component of each antigen immune response spectrum.

[0012] Furthermore, the method for obtaining the search radius includes:

[0013] Obtain the distance metric value according to the distance metric value formula, and the distance metric value formula includes:

[0014] ; where is the distance metric value between the th antigen immune response spectrum and the th antigen immune response spectrum; is the eigenvalue of the th principal component; is the data value of the th principal component of the th antigen immune response spectrum; is the data value of the th principal component of the th antigen immune response spectrum; is the total number of all principal components;

[0015] Using the k-distance analysis method, determine the search radius of the antigen immune response spectrum according to the distance metric value between every two antigen immune response spectra.

[0016] Furthermore, the method for obtaining the feature representativeness includes:

[0017] According to all antigen immune response spectra corresponding to the index to be analyzed, obtain the normal distribution mean; calculate the absolute value of the difference between the index to be analyzed of the graph to be analyzed and the normal distribution mean and perform a negative correlation mapping to obtain the feature representativeness of the index to be analyzed of the graph to be analyzed.

[0018] Furthermore, the method for obtaining the influence measure includes:

[0019] Label all the antigen immune response spectra; in the order of the labels, sequentially count the data values of the principal components corresponding to all the antigen immune response spectra to obtain a sequence of data values of the principal components; in the order of the labels, sequentially count the index to be analyzed corresponding to all the antigen immune response spectra to obtain a sequence of the index to be analyzed.

[0020] Obtain the influence measure according to the influence measure formula, and the influence measure formula includes:

[0021] ; where is the influence measure of the index to be analyzed; is the eigenvalue of the th principal component; is the sum value of the eigenvalues of all the principal components; is the Pearson correlation coefficient between the sequence of the index to be analyzed and the sequence of data values of the th principal component; is the total number of all the principal components; is the absolute value symbol.

[0022] Furthermore, the formula for obtaining the overall representativeness includes:

[0023] ; where is the overall representativeness of the graph to be analyzed; is the influence measure corresponding to the th patient characteristic index; is the cumulative value of the influence measures corresponding to all the patient characteristic indexes; is the feature representativeness of the th patient characteristic index of the graph to be analyzed; is the total number of all the patient characteristic indexes; is the preset denominator adjustment value.

[0024] Furthermore, the method for obtaining the radius adjustment requirement includes:

[0025] In the principal component space, each antigen immune response spectrum within the search radius of the graph to be analyzed is used as each neighborhood analysis graph of the graph to be analyzed;

[0026] Relative typicality is obtained according to a relative typicality formula, wherein the relative typicality formula includes:

[0027] ;in, The graph to be analyzed Relative to neighborhood analysis graph the relative typicality of The graph to be analyzed No. Neighborhood analysis diagram; For the The eigenvalues of the principal components; is the sum of the eigenvalues of all principal components; is the total number of all principal components; For the The principal components correspond to all eigenvalues, and the graph to be analyzed No. The total number of all eigenvalues with the same value corresponding to the principal component; For the The principal components correspond to all eigenvalues, and the neighborhood analysis diagram No. The total number of all eigenvalues with the same value corresponding to the principal component; The graph to be analyzed overall representativeness of Neighborhood analysis graph overall representativeness of Adjust the value for the preset denominator;

[0028] The relative typicality of the image to be analyzed relative to all neighborhood analysis images and the overall representativeness of the image to be analyzed are forwardly integrated to obtain the radius adjustment requirement of the image to be analyzed.

[0029] Furthermore, the method for obtaining the radius adjustment requirement includes:

[0030] The radius adjustment requirement is obtained according to the radius adjustment requirement formula, wherein the radius adjustment requirement formula includes:

[0031] ;in, The graph to be analyzed The radius adjustment needs to be made; The graph to be analyzed Relative to neighborhood analysis graph the relative typicality of The graph to be analyzed The nth neighborhood analysis graph; is the graph to be analyzed the total number of all neighborhood analysis graphs; is the overall representativeness of the graph to be analyzed ; is the normalization function.

[0032] Furthermore, the method for obtaining the adjusted search radius includes:

[0033] Calculate the product of the radius adjustment requirement of the graph to be analyzed and the preset adjustment value as the target adjustment value of the graph to be analyzed; take the product of the target adjustment value and the search radius as the adjusted search radius of the graph to be analyzed.

[0034] The present invention provides a medical data processing system for inspection and examination, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the medical data processing method for inspection and examination are implemented.

[0035] The present invention has the following beneficial effects:

[0036] Since antigen signal data usually has high dimensionality, direct analysis is relatively complex and requires dimensionality reduction processing to obtain the data values and eigenvalues of the main components of each antigen immune response spectrum; the main components represent the dimensions with more immune response information. Further, based on the data values and eigenvalues of the main components of each antigen immune response spectrum, the search radius of the antigen immune response spectrum is initially determined. Reasonably determining the search radius is a key step in subsequent adjustment and optimization of the clustering results. Select any antigen immune response spectrum as the graph to be analyzed and select a patient characteristic index as the index to be analyzed. Analyze the proximity to the central tendency of the index to be analyzed in the graph to be analyzed, and reflect the representativeness of the index to be analyzed in the graph to be analyzed through feature representativeness; by analyzing the correlation between the index to be analyzed and the data values of all antigen immune response spectra on the main components, obtain the influence measure of the index to be analyzed. The influence measure reflects the importance of the index to be analyzed for the antigen immune response spectrum. Combine the feature representativeness of the graph to be analyzed on all patient characteristic indices and the influence measures of these indices to obtain the overall representativeness of the graph to be analyzed. The overall representativeness comprehensively evaluates the overall representativeness degree of the graph to be analyzed. Compare the difference between the graph to be analyzed and its overall representativeness with all antigen immune response spectra within a local range, and calculate the radius adjustment requirement of the graph to be analyzed. The greater the difference, the higher the adjustment requirement may be, which means that a larger search radius needs to be allocated to the graph to be analyzed to capture more similar points. According to the calculated radius adjustment requirement, adjust the original search radius of the graph to be analyzed, and reasonably set the adjusted search radius to improve the accuracy of clustering. Brief Description of the Drawings

[0037] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0038] Figure 1 It is a flowchart of a medical data processing method for inspection and examination provided by an embodiment of the present invention;

[0039] Figure 2 It is a structural diagram of a medical data processing system for inspection and examination provided by an embodiment of the present invention. Detailed Embodiments

[0040] To further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following, in combination with the drawings and preferred embodiments, details the specific embodiments, structures, features and effects of a medical data processing method and system for inspection and examination proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0041] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0042] The following specifically describes the specific solutions of a medical data processing method and system for inspection and examination provided by the present invention in conjunction with the drawings.

[0043] An embodiment of the present invention provides a medical data processing method and system for inspection and examination. Please refer to Figure 1 , which shows a flowchart of a medical data processing method for inspection and examination provided by an embodiment of the present invention. The method includes the following steps:

[0044] Step S1: Obtain the antigen signal data of each antigen immune response spectrum and all patient characteristic indicators.

[0045] To perform accurate clustering subsequently, first obtain the antigen signal data of each antigen immune response spectrum and all patient characteristic indicators.

[0046] The data acquisition process must strictly follow the principle of user authorization to ensure the legal acquisition of antigen signal data of each antigen immune response spectrum and all patient characteristic indicators from the database. The specific acquisition process includes:

[0047] The antigen immune response spectra and all patient characteristic indicators of each patient are stored in the database; for the antigen immune response spectrum of each patient, the antigen signal data is extracted.

[0048] As the result of one of the detection methods, the antigen immune response spectrum is a graph that measures the change in antigen reaction intensity with antigen concentration through experimental methods. The horizontal axis of the antigen immune response spectrum represents the antigen concentration, and the vertical axis represents the antigen reaction intensity; the antigen signal data is a sequence of antigen reaction intensities statistically arranged in the order of antigen concentration, which clearly shows the characteristics of the change in antigen reaction intensity with antigen concentration.

[0049] All patient characteristic indicators of the patients are collected in the database. For example, patient characteristic indicators include: height, weight, age, blood pressure, etc.; all patient characteristic indicators of the patients are statistically analyzed as the corresponding all patient characteristic indicators of the antigen immune response spectrum of the patients, for subsequent in-depth understanding of the patient background information of the antigen immune response spectrum.

[0050] In the present invention, the data collection is authorized by the user, does not violate relevant laws and regulations, and does not violate public order and good customs. For the convenience of calculation, all index data involved in the operation in the embodiments of the present invention have undergone data preprocessing, thereby eliminating the influence of dimension. The specific means of eliminating the influence of dimension are well-known technical means to those skilled in the art and will not be limited here. It should be noted that the present invention aims at an antigen immune response spectrum containing one antigen signal data, the number of antigen reaction intensities included in the antigen signal data of all antigen immune response spectra is the same, and at the same time, the number of all patient characteristic indicators of all antigen immune response spectra is consistent.

[0051] Step S2: According to the antigen signal data of all antigen immune response spectra, perform dimensionality reduction processing to obtain the data values and eigenvalues of each principal component of each antigen immune response spectrum; according to the data values and eigenvalues of each principal component of each antigen immune response spectrum, determine the search radius of the antigen immune response spectrum.

[0052] Since antigen signal data usually has high dimensionality and is relatively complex to analyze directly, dimensionality reduction processing is required to obtain the data values and eigenvalues of each principal component of each antigen immune response spectrum; the principal components represent the dimensions with more immune response information. Further, based on the data values and eigenvalues of each principal component of each antigen immune response spectrum, the search radius of the antigen immune response spectrum is initially determined. Reasonably determining the search radius is a key step in subsequent adjustment and optimization of the clustering results.

[0053] Due to the high-dimensionality of the antigen signal data, direct analysis is complex and it is difficult to capture the main information. Therefore, dimensionality reduction processing is required to obtain the data values and eigenvalue of each principal component of the antigen immune response spectrum. Preferably, in one embodiment of the present invention, the method for obtaining the data values and eigenvalue of the principal component includes:

[0054] Using the PCA method, perform dimensionality reduction processing on the antigen signal data of all antigen immune response spectra to obtain the data values and eigenvalue of each principal component of each antigen immune response spectrum.

[0055] It should be noted that the PCA method (Principal Component Analysis) is a commonly used dimensionality reduction technique. Here, only a brief description is given of using the PCA method to perform dimensionality reduction processing on the antigen signal data of all antigen immune response spectra to obtain the data values and eigenvalue of each principal component of each antigen immune response spectrum:

[0056] First, organize the antigen signal data of all antigen immune response spectra into a two-dimensional data matrix. In this matrix, each row represents an antigen immune response spectrum, and each column represents the antigen reaction intensity corresponding to an antigen concentration. Next, calculate the covariance matrix of this data matrix to capture the correlation between the monitoring values at different antigen concentration points. Subsequently, perform eigenvalue decomposition on the covariance matrix. This step will generate eigenvalues and corresponding eigenvectors. The magnitude of the eigenvalue represents the amount of variance explained by each principal component, that is, the amount of information. The eigenvector defines the direction of the principal component, that is, the projection direction of the data in the principal component space. Based on the magnitude of the eigenvalue, select the first preset number of principal components with the largest variance. These principal components will be used for subsequent dimensionality reduction processing. Using the eigenvectors corresponding to these selected principal components, perform a linear transformation on the original data. This transformation process projects the original data into the principal component space, thereby obtaining the dimensionality-reduced data. In the dimensionality-reduced data matrix, each row still represents an antigen immune response spectrum, but each column now represents the coordinate value of the antigen immune response spectrum in the principal component space, that is, the data value of the principal component. At the same time, the eigenvalue of the principal component of each antigen immune response spectrum is also obtained during the eigenvalue decomposition process. These eigenvalues reflect the importance of each principal component in the dataset. In the present invention, the preset number is 3, that is, the total number of all principal components in the present invention is 3, and the implementer can set it according to the implementation scenario.

[0057] It should be noted that the subsequent clustering is performed in the principal component space containing all antigen immune response spectra. Here, a brief description of the principal component space is given: Each axis of the principal component space corresponds to each principal component, and each antigen immune response spectrum is represented as a data point in the principal component space. Specifically, the data values of each antigen immune response spectrum on each principal component are used as its coordinates in the principal component space. In this way, each antigen immune response spectrum is mapped into a multi-dimensional space composed of principal components to form a data point, that is, the principal component space containing all antigen immune response spectra is obtained. Among them, the principal component space is a multi-dimensional space, and each of its axes corresponds to a principal component. In the present invention, the total number of principal components is set to 3, that is, the principal component space is a three-dimensional space.

[0058] Considering that in the DBSCAN algorithm, the search radius refers to defining the neighborhood range around a point, and the search radius is a core parameter affecting the clustering effect of the DBSCAN algorithm. By explicitly calculating the distance metric value between two antigen immune response spectra, the search radius can be set more accurately, thereby ensuring the accuracy of the initially determined search radius. Preferably, in an embodiment of the present invention, the method for obtaining the search radius includes:

[0059] Obtaining the distance metric value according to the distance metric value formula, and the distance metric value formula includes:

[0060] ; where is the distance metric value between the th antigen immune response spectrum and the th antigen immune response spectrum; is the eigenvalue of the th principal component; is the data value of the th antigen immune response spectrum on the th principal component; is the data value of the th antigen immune response spectrum on the th principal component; is the total number of all principal components. It should be noted that the data values of different antigen immune response spectra in the principal component space are not necessarily the same, while the eigenvalues of the principal components used in the principal component analysis of different antigen immune response spectra are the same.

[0061] In the formula, reflects the importance of the th principal component. The larger the eigenvalue, the more information the principal component contains, the stronger the ability to explain the data, and the higher the importance. When calculating the distance, the important principal components will be given greater weights. reflects the th antigen immune response spectrum and the The degree of difference of an antigen immune response spectrum in the th principal component; the distance metric value comprehensively reflects the difference between two antigen immune response spectra in the principal component space.

[0062] Using the k-distance analysis method, according to the distance metric values of every two antigen immune response spectra, determine the search radius of the antigen immune response spectra.

[0063] It should be noted that the k-distance analysis method is a well-known prior art to those skilled in the art, and only a simple description is given here: for each antigen immune response spectrum, calculate the distance metric values between it and all other antigen immune response spectra, and sort them according to the magnitude of the distance metric values. Then, select the distance of the k-th nearest antigen immune response spectrum as the search radius of this antigen immune response spectrum. Among them, k is a preset parameter, and in the present invention, the value is 5, which is usually set according to the characteristics of the data and the purpose of analysis.

[0064] Considering that in the DBSCAN clustering algorithm, the setting of the search radius has a crucial impact on the clustering result. If the search radius is too small, it may lead to the inability to effectively identify representative data points as core points, thereby hindering the correct formation of clusters. This is because a smaller search radius limits the connectivity between data points, causing some points that should belong to the same cluster to be unable to be grouped together due to excessive distance. On the contrary, if the search radius is set too large, although it can enhance the connectivity between data points, it may also cause some non-representative data points to be misidentified as core points. In this case, the clustering result may become too broad, including some points that do not belong to the cluster, thereby reducing the representativeness and accuracy of the classification result. An overly large search radius makes the boundary of the cluster blurred and difficult to accurately describe the characteristics of the cluster. To solve this problem, the present invention adjusts the search radius, and the core idea of the adjustment method is: for data points that are more representative in the principal component space, a larger clustering search radius should be set in order to more accurately capture the similar points around it and form a more compact and representative cluster. The search radius represents the range of other antigen immune response spectra that are similar to a given antigen immune response spectrum in this principal component space. Specifically, if the search radius of an antigen immune response spectrum is small, then in this principal component space, there are fewer antigen immune response spectra similar to it; conversely, if the search radius of an antigen immune response spectrum is large, then in this principal component space, there are more antigen immune response spectra similar to it.

[0065] Step S3: Take any antigen immune response spectrum as the graph to be analyzed, and take any patient characteristic index as the index to be analyzed; according to the deviation from the central tendency of the index to be analyzed in the graph to be analyzed, obtain the characteristic representativeness of the index to be analyzed in the graph to be analyzed; according to the correlation between the data values of the index to be analyzed corresponding to all antigen immune response spectra and the principal components, obtain the influence measure of the index to be analyzed; according to the characteristic representativeness and influence measure of all patient characteristic indexes in the graph to be analyzed, obtain the overall representativeness of the graph to be analyzed; according to the difference between the graph to be analyzed and the overall representativeness of all antigen immune response spectra in its local range, obtain the need for radius adjustment of the graph to be analyzed; adjust the search radius according to the need for radius adjustment to obtain the adjusted search radius of the graph to be analyzed.

[0066] Select any antigen immune response spectrum as the graph to be analyzed and select a patient characteristic index as the index to be analyzed. Analyze the proximity to the central tendency of the index to be analyzed in the graph to be analyzed, and reflect the representativeness of the index to be analyzed in the graph to be analyzed through the characteristic representativeness; by analyzing the correlation between the index to be analyzed and the data values of all antigen immune response spectra on the principal components, obtain the influence measure of the index to be analyzed, and the influence measure reflects the importance of the index to be analyzed for the antigen immune response spectrum. Combine the characteristic representativeness of the graph to be analyzed on all patient characteristic indexes and the influence measures of these indexes to obtain the overall representativeness of the graph to be analyzed, and the overall representativeness comprehensively evaluates the overall representativeness degree of the graph to be analyzed. Compare the difference between the graph to be analyzed and the overall representativeness of all antigen immune response spectra in its local range, and calculate the need for radius adjustment of the graph to be analyzed. The greater the difference, the higher the adjustment need, which means that a larger search radius needs to be allocated for the graph to be analyzed to capture more similar points. According to the calculated need for radius adjustment, adjust the original search radius of the graph to be analyzed, and reasonably set the adjusted search radius to improve the accuracy of clustering.

[0067] To analyze the representativeness of the index to be analyzed in the graph to be analyzed, preferably, in an embodiment of the present invention, the method for obtaining the characteristic representativeness includes:

[0068] According to the index to be analyzed corresponding to all antigen immune response spectra, obtain the normal distribution mean value; calculate the absolute value of the difference between the index to be analyzed in the graph to be analyzed and the normal distribution mean value and perform a negative correlation mapping to obtain the characteristic representativeness of the index to be analyzed in the graph to be analyzed. It should be noted that the negative correlation mapping is a well-known technical means in the art, and the negative correlation mapping can be in the form of inverse proportion or negative exponential power, which is not limited here.

[0069] For the above steps, since the sample size of the antigen immune response spectra involved in the present invention is large enough, the distribution of these antigen immune response spectra on the corresponding indexes to be analyzed presents the characteristics of a normal distribution. The mean of the normal distribution, as the central tendency of the index to be analyzed, can reflect the typical values of most antigen immune response spectra on the index to be analyzed. First, calculate the mean of the normal distribution based on the values of all antigen immune response spectra on the index to be analyzed. This mean of the normal distribution represents the central position of the index to be analyzed among all antigen immune response spectra, that is, the most typical value. Next, for the value of the index to be analyzed of each antigen immune response spectrum, calculate the absolute value of the difference between it and the mean of the normal distribution. This difference reflects the degree to which the antigen immune response spectrum deviates from the central tendency on the index to be analyzed. In order to convert the absolute value of the difference into characteristic representativeness, a negative correlation mapping is performed. Specifically, the larger the absolute value of the difference, the less typical the antigen immune response spectrum is on the index to be analyzed, that is, the more it deviates from the central tendency, so its characteristic representativeness is smaller.

[0070] In order to quantify the importance of the index to be analyzed for the antigen immune response spectrum, preferably, in an embodiment of the present invention, the method for obtaining the influence measure includes:

[0071] Label all antigen immune response spectra; in the order of the labels, sequentially count the data values of the corresponding principal components of all antigen immune response spectra to obtain a data value sequence of the principal components; in the order of the labels, sequentially count the indexes to be analyzed of all antigen immune response spectra to obtain a sequence of the indexes to be analyzed.

[0072] For the above steps, label all antigen immune response spectra one by one. This step is to ensure the orderliness and accuracy of subsequent data processing. Each antigen immune response spectrum is assigned a unique identifier for easy tracking and statistics. The data value sequence of the principal components reflects the change direction of the main immune response information volume and is a direct manifestation of the result of principal component analysis. The sequence of the indexes to be analyzed reflects the change direction of the patient characteristics of the antigen immune response spectrum.

[0073] Obtain the influence measure according to the influence measure formula, and the influence measure formula includes:

[0074] ; where is the influence measure of the index to be analyzed; is the eigenvalue of the th principal component; is the sum of the eigenvalues of all principal components; is the Pearson correlation coefficient between the sequence of the indexes to be analyzed and the data value sequence of the th principal component; is the absolute value symbol. It should be noted that the Pearson correlation coefficient is an existing technology well-known to those skilled in the art and will not be elaborated here.

[0075] In the formula, is used to perform normalization on The larger is, the greater the information content of the principal component, and the stronger the representativeness of the original data. When calculating the influence degree, the principal component with a larger eigenvalue should have a greater weight; represents the linear correlation degree between the sequence of the index to be analyzed and the data value sequence of the th principal component. The larger the value, the stronger the linear relationship between the index to be analyzed and the principal component. Therefore, the influence degree of this principal component on the index to be analyzed is also greater. The influence measure can quantitatively evaluate the importance of each index to be analyzed against the antigen immune response spectrum.

[0076] In order to comprehensively evaluate the overall representativeness of the graph to be analyzed, preferably, in an embodiment of the present invention, the acquisition formula for the overall representativeness includes:

[0077] ; where is the overall representativeness of the graph to be analyzed; is the influence measure corresponding to the th patient characteristic index; is the cumulative value of the influence measures corresponding to all patient characteristic indexes; is the characteristic representativeness of the th patient characteristic index of the graph to be analyzed; is the total number of all patient characteristic indexes; is the preset denominator adjustment value.

[0078] In the formula, is the cumulative value of the influence measures corresponding to all patient characteristic indexes, providing a normalization benchmark. Using to perform normalization on to obtain ; The importance of each patient characteristic index in the antigen immune response spectrum is measured by , and then by combining the importance of each index with its own characteristic representativeness, the overall representativeness of the graph to be analyzed is comprehensively evaluated. The larger the overall representativeness, the more the graph to be analyzed can represent the patient characteristics as a whole, and the greater the information content it contains.

[0079] Preferably, in an embodiment of the present invention, the method for obtaining the need for radius adjustment includes:

[0080] In the principal component space, each antigen immune response spectrum in the search radius of the graph to be analyzed is used as each neighborhood analysis graph of the graph to be analyzed;

[0081] In order to quantify the representativeness of the graph to be analyzed relative to its neighborhood analysis graph, the relative typicality is obtained according to the relative typicality formula. The relative typicality formula includes:

[0082] ;in, The graph to be analyzed Relative to neighborhood analysis graph the relative typicality of The graph to be analyzed No. Neighborhood analysis diagram; For the The eigenvalues of the principal components; is the sum of the eigenvalues of all principal components; is the total number of all principal components; For the The principal components correspond to all eigenvalues, and the graph to be analyzed No. The total number of all eigenvalues with the same value corresponding to the principal component; For the The principal components correspond to all eigenvalues, and the neighborhood analysis diagram No. The total number of all eigenvalues with the same value corresponding to the principal component; The graph to be analyzed overall representativeness of Neighborhood analysis graph overall representativeness of The default denominator adjustment value is 0.01 to prevent the denominator from being zero. Implementers can set this value based on the implementation scenario.

[0083] Where, The larger the eigenvalue is, the greater the information content of the principal component is and the more representative it is of the original data. In the calculation, the principal component with a larger eigenvalue should have a greater weight. It reflects the representativeness of the principal component corresponding eigenvalues of the graph to be analyzed relative to its neighborhood analysis graph. The larger the value, the more the principal component corresponding eigenvalues of the graph to be analyzed relative to its neighborhood analysis graph appear, which means that the representativeness of the graph to be analyzed on this principal component is stronger. It reflects the representativeness of the graph to be analyzed relative to the neighborhood analysis graph. The larger the value, the stronger the representativeness of the graph to be analyzed relative to its neighborhood analysis graph. Relative typicality quantifies the representativeness of the graph to be analyzed relative to its neighborhood analysis graph.

[0084] It should be noted that, in other embodiments of the present invention, the relative typicality calculation formula also includes: . That is, directly through Calculation .

[0085] Forwardly fuse the relative typicality of the graph to be analyzed with respect to all neighborhood analysis graphs and the overall representativeness of the graph to be analyzed to obtain the radius adjustment requirement degree of the graph to be analyzed. It should be noted that forward fusion is an existing technology well-known to those skilled in the art, and forward fusion can adopt simple product, arithmetic mean or other suitable fusion methods. In an embodiment of the present invention, the radius adjustment requirement degree formula includes:

[0086] ; where is the radius adjustment requirement degree of the graph to be analyzed ; is the relative typicality of the graph to be analyzed with respect to the neighborhood analysis graph ; is the th neighborhood analysis graph of the graph to be analyzed; is the total number of all neighborhood analysis graphs of the graph to be analyzed ; is the overall representativeness of the graph to be analyzed ; is the normalization function. It should be noted that the method of adopting normalization is: use the norm normalization function for normalization to limit the numerical range between 0 and 1. Among them, normalization is a technical means well-known to those skilled in the art, and the choice of the normalization function can be linear normalization or standard normalization, etc., and the specific normalization method is not limited here.

[0087] In the formula, reflects the overall performance of the relative typicality of the graph to be analyzed with respect to all neighborhood analysis graphs, and in combination with , a quantified radius adjustment requirement degree is obtained, which characterizes the comprehensive representativeness of the graph to be analyzed.

[0088] Preferably, in an embodiment of the present invention, the method for obtaining the adjusted search radius includes:

[0089] Calculate the product of the radius adjustment requirement degree of the graph to be analyzed and the preset adjustment value as the target adjustment value of the graph to be analyzed; take the product of the target adjustment value and the search radius as the adjusted search radius of the graph to be analyzed. In an embodiment of the present invention, the preset adjustment value is 1.4, and the implementer can set it according to the implementation scenario by himself.

[0090] For the above steps, the radius adjustment requirement takes into account the relative typicality of the graph to be analyzed and its neighborhood analysis graph, as well as the overall representativeness of the graph to be analyzed itself, so as to accurately reflect the importance of each graph in the clustering process. Calculate the product of the radius adjustment requirement of the graph to be analyzed and the preset adjustment value as the target adjustment value of the graph to be analyzed; take the product of the target adjustment value and the search radius as the adjusted search radius of the graph to be analyzed. By assigning a larger adjusted search radius to the more representative antigen immune response spectra, the key information in the antigen immune response spectra can be captured more accurately, thus effectively reducing the possibility of misclassification. This helps to form tighter and more clearly structured clusters, thereby improving the accuracy of clustering.

[0091] Step S4: Perform DBSCAN clustering on all antigen immune response spectra according to the adjusted search radius of the antigen immune response spectra to obtain each antigen immune response spectrum cluster.

[0092] By reasonably setting the adjusted search radius of the antigen immune response spectra, the key information in the antigen immune response spectra can be better captured, improving the accuracy of the clustering results of the antigen immune response spectra.

[0093] It should be noted that DBSCAN density clustering is a well-known technical means in the art and will not be elaborated here. Only a brief method for performing DBSCAN clustering on all antigen immune response spectra according to the adjusted search radius of the antigen immune response spectra to obtain each antigen immune response spectrum cluster is described:

[0094] In the principal component space, first calculate the distance metric value between every two antigen immune response spectra, which reflects the relative spatial distance between every two antigen immune response spectra. Subsequently, in combination with the adjusted search radius of each antigen immune response spectrum, which is set in the present invention and used to control the neighborhood range in the clustering process, perform DBSCAN clustering on all antigen immune response spectra to obtain each antigen immune response spectrum cluster. The antigen immune response spectrum clusters can better capture the key information in the antigen immune response spectra, improving the accuracy of the clustering results of the antigen immune response spectra.

[0095] The present invention also proposes a medical data processing system for examination and inspection. Please refer to Figure 2 , which shows the structure diagram of a medical data processing system for examination and inspection provided by an embodiment of the present invention. The system includes: a data acquisition module 101, a search radius analysis module 102, a search radius adjustment module 103, and a clustering module 104.

[0096] The data acquisition module 101 is used to acquire the antigen signal data of each antigen immune response spectrum and all patient characteristic indicators.

[0097] The search radius analysis module 102 is used to perform dimensionality reduction processing based on the antigen signal data of all antigen immune response spectra, and obtain the data values and eigenvalues of each principal component of each antigen immune response spectrum; determine the search radius of the antigen immune response spectrum according to the data values and eigenvalues of each principal component of each antigen immune response spectrum.

[0098] The search radius adjustment module 103 is used to take any antigen immune response spectrum as the graph to be analyzed, and any patient characteristic index as the index to be analyzed; obtain the characteristic representativeness of the index to be analyzed according to the deviation trend of the index to be analyzed of the graph to be analyzed from the center; obtain the influence measure of the index to be analyzed according to the correlation between the data values of the index to be analyzed corresponding to all antigen immune response spectra and the principal components; obtain the overall representativeness of the graph to be analyzed according to the characteristic representativeness and influence measure of all patient characteristic indices of the graph to be analyzed; obtain the need for radius adjustment of the graph to be analyzed according to the difference between the graph to be analyzed and its overall representativeness in the local range; adjust the search radius according to the need for radius adjustment to obtain the adjusted search radius of the graph to be analyzed.

[0099] The clustering module 104 is used to perform DBSCAN clustering on all antigen immune response spectra according to the adjusted search radius of each antigen immune response spectrum, and obtain each antigen immune response spectrum cluster.

[0100] It should be noted that: for the system provided in the above embodiments, only the above division of each functional module is used for illustration. In actual applications, the above functions can be allocated to different functional modules according to needs, that is, the internal structure of the computer device is divided into different functional modules to complete all or part of the functions described above. In addition, a medical data processing system for inspection and a medical data processing method embodiment provided in the above embodiments belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0101] In summary, the embodiments of the present invention provide a medical data processing method and system for inspection and examination. First, according to the data values and eigenvalue of each principal component of each antigen immune response spectrum, the search radius of the antigen immune response spectrum is determined; according to the feature representativeness and influence metric of all patient characteristic indicators of the graph to be analyzed, the overall representativeness of the graph to be analyzed is obtained; according to the difference between the graph to be analyzed and the overall representativeness of all antigen immune response spectra in the local range, the need for radius adjustment of the graph to be analyzed is obtained; the search radius is adjusted according to the need for radius adjustment to obtain the adjusted search radius of the graph to be analyzed; and then DBSCAN clustering is performed on all antigen immune response spectra to obtain each antigen immune response spectrum cluster. The present invention reasonably sets the adjusted search radius of the antigen immune response spectrum, can better capture the key information in the antigen immune response spectrum, and improve the accuracy of the clustering result of the antigen immune response spectrum.

[0102] It should be noted that the above sequence of the embodiments of the present invention is only for description and does not represent the advantages or disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the particular order or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0103] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between each embodiment can be referred to each other. The key point of each embodiment is to illustrate the differences from other embodiments.

Claims

1. A medical data processing method for inspection and examination, characterized in that The method includes: Obtaining antigen signal data and all patient characteristic indicators of each antigen immune response spectrum; Performing dimensionality reduction processing based on the antigen signal data of all the antigen immune response spectra to obtain the data values and eigenvalues of each principal component of each antigen immune response spectrum; determining the search radius of each antigen immune response spectrum according to the data values and eigenvalues of each principal component of each antigen immune response spectrum; Taking any one of the antigen immune response spectra as the graph to be analyzed and any one of the patient characteristic indicators as the indicator to be analyzed; obtaining the characteristic representativeness of the indicator to be analyzed of the graph to be analyzed according to the deviation from the central tendency of the indicator to be analyzed of the graph to be analyzed; obtaining the influence measure of the indicator to be analyzed according to the correlation between all the antigen immune response spectra corresponding to the indicator to be analyzed and the data values of the principal components; obtaining the overall representativeness of the graph to be analyzed according to the characteristic representativeness of all the patient characteristic indicators of the graph to be analyzed and the influence measure; obtaining the radius adjustment requirement of the graph to be analyzed according to the difference between the graph to be analyzed and its overall representativeness among all the antigen immune response spectra in the local range; adjusting the search radius according to the radius adjustment requirement to obtain the adjusted search radius of the graph to be analyzed; the radius adjustment requirement characterizes the comprehensive representativeness of the graph to be analyzed; Performing DBSCAN clustering on all the antigen immune response spectra according to the adjusted search radius of each antigen immune response spectrum to obtain each antigen immune response spectrum cluster; The method for obtaining the characteristic representativeness includes: Obtaining the normal distribution mean according to all the antigen immune response spectra corresponding to the indicator to be analyzed; calculating the absolute value of the difference between the indicator to be analyzed of the graph to be analyzed and the normal distribution mean and performing negative correlation mapping to obtain the characteristic representativeness of the indicator to be analyzed of the graph to be analyzed; The method for obtaining the influence measure includes: Numbering all the antigen immune response spectra; sequentially counting the data values of the principal components corresponding to all the antigen immune response spectra in the order of the numbers to obtain the data value sequence of the principal components; sequentially counting the indicator to be analyzed corresponding to all the antigen immune response spectra in the order of the numbers to obtain the indicator sequence of the indicator to be analyzed; Obtaining the influence measure according to the influence measure formula, and the influence measure formula includes: ; wherein, is the influence measure of the index to be analyzed; is the th eigenvalue of the principal component; is the sum of the eigenvalues of all principal components; is the Pearson correlation coefficient between the sequence of the index to be analyzed and the data value sequence of the th principal component; is the total number of all principal components; is the absolute value symbol; The formula for obtaining the overall representativeness includes: ; wherein, is the overall representativeness of the graph to be analyzed; is the influence measure corresponding to the th patient characteristic index; is the cumulative value of the influence measures corresponding to all patient characteristic indexes; is the characteristic representativeness of the th patient characteristic index of the graph to be analyzed; is the total number of all patient characteristic indexes; is a preset denominator adjustment value.

2. The medical data processing method for inspection and examination according to claim 1, wherein The method for obtaining the data values and eigenvalues of the principal components includes: Using the PCA method to perform dimensionality reduction processing on the antigen signal data of all the antigen immune response spectra to obtain the data values and eigenvalues of each principal component of each antigen immune response spectrum.

3. The medical data processing method for inspection according to claim 1, wherein The method for obtaining the search radius includes: Obtaining the distance measure value according to the distance measure value formula, and the distance measure value formula includes: ; wherein, is the distance metric value between the th antigen immune response spectrum and the th antigen immune response spectrum; is the eigenvalue of the th principal component; is the data value of the th principal component of the th antigen immune response spectrum; is the data value of the th principal component of the th antigen immune response spectrum; is the total number of all principal components; Using the k-distance analysis method to determine the search radius of the antigen immune response spectrum according to the distance measure value between every two antigen immune response spectra.

4. The medical data processing method for inspection and examination according to claim 1, characterized in that, The method for obtaining the radius adjustment requirement includes: In the principal component space, each antigen immune response spectrum within the search radius of the graph to be analyzed is used as each neighborhood analysis graph of the graph to be analyzed; The relative typicality is obtained according to the relative typicality formula, and the relative typicality formula includes: ; wherein, is the graph to be analyzed relative to the neighborhood analysis graph of relative typicality; is the graph to be analyzed the th neighborhood analysis graph; is the eigenvalue of the th principal component; is the sum value of the eigenvalues of all principal components; is the total number of all principal components; is in the th principal component corresponding to all eigenvalues, and is the total number of all eigenvalues with the same numerical value as the eigenvalue corresponding to the th principal component of the graph to be analyzed ; the th is in the th principal component corresponding to all eigenvalues, and is the total number of all eigenvalues with the same numerical value as the eigenvalue corresponding to the th principal component of the neighborhood analysis graph ; the th is the overall representativeness of the graph to be analyzed ; is the overall representativeness of the neighborhood analysis graph ; is the preset denominator adjustment value; The need degree for radius adjustment of the graph to be analyzed is obtained by fusing the relative typicality of the graph to be analyzed with respect to all neighborhood analysis graphs and the overall representativeness of the graph to be analyzed in a positive direction.

5. The medical data processing method for inspection according to claim 4, wherein The method for obtaining the need degree for radius adjustment includes: The need degree for radius adjustment is obtained according to the need degree for radius adjustment formula, and the need degree for radius adjustment formula includes: ; wherein, is the required degree of radius adjustment of the graph to be analyzed ; is the graph to be analyzed with respect to the neighborhood analysis graph relative typicality; is the graph to be analyzed the th neighborhood analysis graph; is the graph to be analyzed the total number of all neighborhood analysis graphs of the graph to be analyzed; is the overall representativeness of the graph to be analyzed ; is the normalization function.

6. The medical data processing method for inspection according to claim 1, wherein The method for obtaining the adjusted search radius includes: Calculating the product of the need degree for radius adjustment of the graph to be analyzed and a preset adjustment value as the target adjustment value of the graph to be analyzed; taking the product of the target adjustment value and the search radius as the adjusted search radius of the graph to be analyzed.

7. A medical data processing system for inspection and examination, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the medical data processing method for inspection and examination according to any one of claims 1 to 6.

Images