Intelligent dynamic early warning model construction method and system for acute exacerbation of chronic obstructive pulmonary disease

Abstract

The invention relates to a method and system for constructing an intelligent dynamic early warning model for acute exacerbation of chronic obstructive pulmonary disease, and the method comprises the steps: carrying out the cleaning, sorting, structuring and standardization of collected acute exacerbation data of chronic obstructive pulmonary disease, and obtaining the basic data of acute exacerbation of chronic obstructive pulmonary disease, which is larger in quantity and higher in quality and can be directly used for scientific research. Then, an efficient and practical chronic obstructive pulmonary acute exacerbation risk prediction model is constructed by using a big data technology and an artificial intelligence method, and high-risk factors influencing chronic obstructive pulmonary acute exacerbation are found out from numerous risk factors, so that the patient detection cost is reduced, the clinical operability is improved, and effective guidance is provided for early treatment of chronic obstructive pulmonary acute exacerbation.

Description

Technical Field

[0001] The present invention relates to the field of data science and technology, and in particular to a method and system for constructing an intelligent dynamic early warning model for acute exacerbation of chronic obstructive pulmonary disease. Background Art

[0002] Establishing a database of patients with acute exacerbations of chronic obstructive pulmonary disease (AECOPD) is fundamental to developing a risk prediction model for AECOPD using artificial intelligence and big data mining methods. Currently, few large-scale, prospective AECOPD cohort studies have been conducted, and establishing a database of patients with AECOPD is a challenging task.

[0003] Since the sample size of the COPD acute exacerbation patient database is large and there are many risk factors, directly constructing a prediction model for prediction is not effective. It is difficult to effectively screen the influencing indicators of COPD acute exacerbation, and thus it is difficult to find high-risk influencing factors for COPD acute exacerbation. Summary of the Invention

[0004] The technical problem to be solved by the present invention is how to effectively screen the influencing indicators of acute exacerbation of COPD and find the high-risk influencing factors for the occurrence of acute exacerbation of COPD. In order to overcome the defects of the above-mentioned prior art, the present invention provides a method and system for constructing an intelligent dynamic early warning model for acute exacerbation of COPD, including a method for constructing an intelligent dynamic early warning model for acute exacerbation of COPD and an intelligent dynamic early warning system for acute exacerbation of COPD.

[0005] The present invention provides a method for constructing an intelligent dynamic early warning model for acute exacerbation of chronic obstructive pulmonary disease, comprising the following steps:

[0006] S1: Obtain risk factor information for multiple patients with acute exacerbation of COPD from a medical diagnosis database, and use statistical analysis methods to correct incomplete, abnormal, similar, repeated, and inconsistent risk factor information;

[0007] S2: Post-structuring the risk factor information restored in step S1 using a machine learning algorithm, and then performing de-privacy and encryption processing by constructing a desensitization algorithm intermediate mapping table to obtain a standardized data set;

[0008] S3: Screening potential risk factors for acute exacerbation of COPD from the standardized data set using statistical analysis methods, and obtaining a prediction model based on these risk factors;

[0009] S4: The prediction model is used to predict each COPD patient to obtain the prediction results and the high-risk factor data that the hospital regularly detects for each COPD patient to obtain the real-time and accurate COPD acute exacerbation risk probability value and corresponding risk level of each COPD patient, and the results are visualized.

[0010] The method for constructing an intelligent dynamic early warning model for acute exacerbations of COPD disclosed in the present invention addresses the technical problems of the present invention. By executing step S1, the collected data on acute exacerbations of COPD are cleaned, sorted, structured, and standardized to obtain a large amount of high-quality basic data on acute exacerbations of COPD that can be directly used for scientific research. Subsequently, through steps S2 and S3, an efficient and practical risk prediction model for acute exacerbations of COPD is constructed using big data technology and artificial intelligence methods. The high-risk factors that affect the occurrence of acute exacerbations of COPD are identified from among the numerous risk factors, thereby reducing patient testing costs, improving clinical operability, and providing effective guidance for the early treatment of acute exacerbations of COPD.

[0011] In a possible implementation, in step S1, the process of repairing incomplete parts, abnormal content, similar duplicate content, and inconsistent data of the risk factor information using a statistical analysis method is as follows:

[0012] Repairing the incomplete parts of the risk factor information by interpolation filling, variance analysis filling or mean filling method;

[0013] Identify and repair abnormal content of the risk factor information through mean comparison, correlation comparison method or interval boundary method;

[0014] Deleting similar and repeated content of the risk factor information through a data cleaning method based on a permutation and merging algorithm, and merging through a data cleaning technique based on an N-Gram algorithm;

[0015] Repairing inconsistent data of the risk factor information by sequentially extending a micro-function dependency identification method and an automatic repair method based on confidence and entropy;

[0016] Compared with the existing technology, the above technical solution can repair the incompleteness, anomaly, similar duplication and inconsistency of the risk factor information obtained, thereby ensuring the integrity of the data to facilitate structured processing of some data.

[0017] In one possible implementation, the risk factor information includes clinical characteristic indicators, biomarker indicators, and personal clinical survey data; this solution can ensure comprehensive control of risk factors for COPD patients and can ensure accurate predictions.

[0018] In a possible implementation, step S2 includes the following steps:

[0019] S21: performing post-structuring processing on the symptom information in the risk factor information restored in step S1 by using a natural language processing method or a conditional random field model processing method based on a bidirectional encoding representation of a converter to obtain structured data;

[0020] S22: Associating the structured data with the patient primary index as a unique identifier, and performing de-privacy and encryption processing by constructing a desensitization algorithm intermediate mapping table to obtain the standardized data set;

[0021] This solution can achieve data structuring processing for projects with relatively strong textual information, and build standardized data sets to facilitate model training and result prediction while ensuring de-identification.

[0022] In a possible implementation, the disease information includes medical records, examination reports, and imaging images of patients with acute exacerbation of COPD.

[0023] In a possible implementation, step S3 includes the following steps:

[0024] S31: performing a descriptive analysis on the standardized data set to characterize the statistical characteristics of different patients with acute exacerbation of COPD and obtain descriptive results;

[0025] S32: Using the severity of COPD exacerbation as the dependent variable, perform a single factor preliminary screening of the risk of COPD exacerbation on multiple indicators in the risk factor information describing the results using statistical analysis methods to obtain potential risk factors for COPD exacerbation;

[0026] S33: selecting a deep neural network, and optimizing parameters of the deep neural network using an optimization algorithm based on the risk factors to obtain the prediction model;

[0027] This solution can achieve the acquisition and parameter optimization of the prediction model, ensuring the accuracy of the model's prediction results.

[0028] In a possible implementation, step S33 includes the following steps:

[0029] S331: Construct a multi-layer deep neural network model for predicting the risk of acute exacerbation of COPD, and use sensitivity, specificity, and area under the receiver operating characteristic (ROC) curve to evaluate the model under different parameters to determine the optimal risk prediction model.

[0030] S332: Using the optimal risk prediction model to classify risk levels according to the distribution of risk probabilities;

[0031] S333: The risk factors after dimensionality reduction are used as input variables, the severity level of acute exacerbation of COPD is used as the output variable, and the activation function adopts non-smooth ReLU and LeakyReLU functions to obtain a preliminary optimal risk prediction model;

[0032] S334: Solving the model parameters of the preliminary optimal risk prediction model through an optimization algorithm to obtain the prediction model.

[0033] In a possible implementation, step S4 includes the following steps:

[0034] S41: Based on the prediction results of the prediction model for each COPD patient, combined with the high-risk factor data regularly tested by the hospital for each COPD patient, obtain a real-time and accurate COPD acute exacerbation risk probability value and corresponding risk level for each COPD patient;

[0035] S42: Based on the historical risk probability data obtained for each COPD patient over multiple times, using time series technology to predict the risk probability value of the COPD patient suffering from an acute exacerbation of COPD in the future time period, and obtain the changing trend of the risk probability of each COPD patient and the corresponding risk level;

[0036] S43: Based on the results obtained in step S42, interpretable techniques including local interpretable unknown model interpretation technology, hierarchical correlation transfer, and Shapley superposition interpretation are used to provide a human-understandable visual explanation of the prediction results for each COPD patient;

[0037] This solution can ensure the visualization of prediction results and provide a visual basis for disease diagnosis or development trend prediction.

[0038] In a possible embodiment, the use of time series technology in step S42 to predict the risk probability value of the COPD patient suffering from an acute exacerbation of COPD in a future time period is to use a fractional-order cumulative grey prediction model to predict the risk probability value of the COPD patient suffering from an acute exacerbation of COPD in a future time period; this solution fully takes into account the characteristic of a relatively small sample size, and the use of this technical solution can ensure the accuracy of the risk probability value of an acute exacerbation of COPD.

[0039] Another technical solution of the present invention is to provide an intelligent dynamic early warning system for acute exacerbation of chronic obstructive pulmonary disease, including:

[0040] a collection module configured to obtain risk factor information of multiple patients with acute exacerbation of COPD from a medical diagnosis database, and to correct incomplete, abnormal, similar, repeated, and inconsistent risk factor information using a statistical analysis method;

[0041] A standard module is configured to perform post-structuring processing on the risk factor information repaired by the collection module through a machine learning algorithm, and then perform de-privacy and encryption processing by constructing an intermediate mapping table of a desensitization algorithm to obtain a standardized data set;

[0042] an extraction module configured to screen out potential risk factors for acute exacerbation of COPD from the standardized data set using a statistical analysis method;

[0043] A prediction module is configured to obtain a severity level of an acute exacerbation of COPD based on the risk factors obtained by the extraction module;

[0044] The visualization module is configured to visualize the real-time and accurate COPD acute exacerbation risk probability value and corresponding risk level of each COPD patient;

[0045] The intelligent dynamic early warning system for acute exacerbations of COPD disclosed in this invention addresses the technical problems of the present invention by providing a collection module to clean, organize, structure, and standardize the collected data on acute exacerbations of COPD, thereby obtaining a large amount of high-quality basic data on acute exacerbations of COPD that can be directly used for scientific research. By providing a standard module and an extraction module to identify high-risk factors affecting acute exacerbations of COPD from among numerous risk factors, patient testing costs are reduced. Prediction results are obtained by providing a predictive model and visualization module, thereby improving clinical operability and providing effective guidance for the early treatment of acute exacerbations of COPD. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 This is a flow chart of a method for constructing an intelligent dynamic early warning model for acute exacerbation of chronic obstructive pulmonary disease disclosed in Example 1 of this application;

[0047] Figure 2 This is a flowchart corresponding to step S2 disclosed in Example 1 of this application;

[0048] Figure 3 This is a flowchart corresponding to step S3 disclosed in Example 1 of this application;

[0049] Figure 4 This is a flowchart corresponding to step S33 disclosed in the first embodiment of the present application;

[0050] Figure 5 This is a flowchart corresponding to step S4 disclosed in Example 1 of this application;

[0051] Figure 6 This is a structural schematic diagram of an intelligent dynamic early warning system for acute exacerbation of chronic obstructive pulmonary disease disclosed in Example 2 of this application. DETAILED DESCRIPTION

[0052] First, those skilled in the art should understand that these embodiments are merely used to explain the technical principles of the embodiments of the present application and are not intended to limit the scope of protection of the embodiments of the present application. Those skilled in the art may adjust them as needed to suit specific application scenarios.

[0053] In the description of the embodiments of the present application, it should be noted that, unless otherwise clearly specified and limited, the terms "electrical connection" and "electrical connection relationship" should be understood in a broad sense, that is, referring to a connection method with an electrical relationship, for example, it can be a circuit connection through a wire, or it can be an electrical connection through a radio signal channel (channel), or a combination of the two. In addition, "electrical connection" and "electrical connection relationship" can be based on a mechanical connection (such as a wire set in a connecting key); it can be a direct connection or an indirect connection through an intermediate medium. For those of ordinary skill in the art, the specific meanings of the above terms in the embodiments of the present application can be understood according to specific circumstances.

[0054] Two embodiments will be used below to further illustrate the present application in detail in conjunction with the accompanying drawings and specific embodiments.

[0055] Example 1:

[0056] See also Figures 1 to 5 As shown, the embodiment of the present application discloses a method for constructing an intelligent dynamic early warning model for acute exacerbation of chronic obstructive pulmonary disease, see Figure 1 As shown, the method includes the following steps:

[0057] S1: Obtain risk factor information for multiple patients with acute exacerbation of COPD from the medical diagnosis database, and use statistical analysis methods to correct incomplete, abnormal, similar, repeated, and inconsistent risk factor information.

[0058] Specifically, in step S1 of this embodiment, the process of repairing incomplete parts, abnormal contents, similar duplicate contents and inconsistent data of risk factor information using statistical analysis methods is as follows: (1) Repairing incomplete parts of risk factor information by interpolation filling, variance analysis filling or mean filling method. (2) Identifying and repairing abnormal contents of risk factor information by average comparison, correlation comparison method or interval boundary method. (3) Deleting similar duplicate contents of risk factor information by data cleaning method based on permutation merging algorithm; merging by data cleaning technology based on N-Gram algorithm. (4) Repairing inconsistent data of risk factor information by extended micro-function dependency identification method and automatic repair method based on confidence and entropy.

[0059] The risk factor information in this embodiment includes clinical characteristic indicators, biomarker indicators, and personal clinical survey data. Specifically, the clinical characteristic indicator data used include: clinical characteristic indicators: respiratory symptoms, non-respiratory symptoms, degree of airflow limitation, exhaled nitric oxide, comorbidities, and lung imaging; biomarker indicators include BMI index and eosinophils; and personal clinical survey data includes fibrinogen, air pollution, occupational dust, medical records, examination reports, and imaging images.

[0060] S2: The risk factor information repaired in step S1 is post-structured through a machine learning algorithm, and then de-privacy and encryption are performed by constructing a desensitizing algorithm intermediate mapping table to obtain a standardized data set.

[0061] See also Figure 2 As shown, in this embodiment, step S2 includes the following steps:

[0062] S21: Post-structuring the symptom information in the risk factor information repaired in step S1 by natural language processing or a conditional random field model based on bidirectional encoding representation of a transformer to obtain structured data. The symptom information includes medical records, examination reports, and imaging images of patients with acute exacerbation of COPD.

[0063] S22: Structured data is associated using the patient's primary index as a unique identifier, and de-privacy and encryption are performed by constructing an intermediate mapping table of a desensitizing algorithm to obtain a standardized data set.

[0064] S3: Use statistical analysis methods to screen out potential risk factors for acute exacerbation of COPD from standardized data sets, and obtain a prediction model based on these risk factors.

[0065] See also Figure 3 As shown, in this embodiment, step S3 includes the following steps:

[0066] S31: Perform descriptive analysis on the standardized data set to characterize the statistical characteristics of the data of different patients with acute exacerbation of COPD and obtain descriptive results.

[0067] S32: Taking the severity of COPD exacerbation as the dependent variable, statistical analysis methods are used to conduct a single factor initial screening of the risk of COPD exacerbation for multiple indicators in the risk factor information describing the results, and to obtain potential risk factors for COPD exacerbation.

[0068] S33: Select a deep neural network, and optimize the parameters of the deep neural network through an optimization algorithm based on risk factors to obtain a prediction model.

[0069] See also Figure 5 As shown, in this embodiment, step S33 includes the following steps:

[0070] S331: Construct a multi-layer deep neural network model for predicting the risk of acute exacerbation of COPD, and use sensitivity, specificity, and area under the receiver operating characteristic (ROC) curve (AUC) indicators to evaluate the model under different parameters to determine the optimal risk prediction model.

[0071] S332: Using the optimal risk prediction model, classify the risk levels according to the distribution of risk probabilities.

[0072] S333: Use risk factors after dimensionality reduction as input variables and the severity of COPD exacerbations as output variables. Use non-smooth ReLU and LeakyReLU functions as activation functions to obtain a preliminary optimal risk prediction model.

[0073] S334: Solve the model parameters of the preliminary optimal risk prediction model through the optimization algorithm to obtain the prediction model.

[0074] In step S334, the process of solving the DNN model with a non-smooth activation function through the optimization algorithm is as follows: (1) two regularization terms are introduced into the objective function of the optimization problem corresponding to the preliminary optimal risk prediction model, and the problem is equivalently converted into a problem with a convex feasible set and Slater type constraints; (2) a piecewise smooth function is constructed to approximate the ReLU activation function, so that the objective function becomes a semi-smooth function; (3) the optimization problem is solved using the augmented Lagrangian method based on sparse semi-smooth Newton, because this algorithm is currently the most effective algorithm for solving large-scale optimization problems.

[0075] S4: The prediction model is used to predict each COPD patient to obtain the prediction results and the high-risk factor data that the hospital regularly tests for each COPD patient to obtain the real-time and accurate COPD acute exacerbation risk probability value and corresponding risk level of each COPD patient, and the results are visualized.

[0076] See also Figure 5 As shown, in this embodiment, step S4 includes the following steps:

[0077] S41: Based on the prediction results of the prediction model for each COPD patient, combined with the high-risk factor data regularly tested by the hospital for each COPD patient, the real-time and accurate risk probability value of acute exacerbation of COPD and the corresponding risk level of each COPD patient are obtained.

[0078] S42: Based on the historical risk probability data obtained from each COPD patient multiple times, time series technology is used to predict the risk probability value of the COPD patient suffering from acute exacerbation of COPD in the future time period, and the changing trend of the risk probability of each COPD patient and the corresponding risk level are obtained.

[0079] Among them, the use of time series technology in step S42 to predict the risk probability value of the COPD patient suffering from an acute exacerbation of COPD in the future time period is to use the fractional-order cumulative grey prediction model to predict the risk probability value of the COPD patient suffering from an acute exacerbation of COPD in the future time period. The specific process is: the first-order cumulative generation operator and the first-order cumulative generation operator in the grey model GM(1,1) constructed based on the historical data of risk probabilities obtained multiple times for each COPD patient are generalized to fractional-order cumulative generation operators and fractional-order cumulative generation operators, thereby improving the modeling accuracy and prediction accuracy of the grey model. At the same time, in order to obtain the optimal nonlinear parameters of the fractional-order cumulative grey prediction model, it is necessary to solve a nonlinear optimization problem with equality constraints. A variety of intelligent optimization algorithms such as the whale optimization algorithm (WOA), the particle swarm optimization algorithm (PSO), and the grey wolf optimization algorithm (GWO) can be used for calculation and comparison and selection.

[0080] S43: Based on the results obtained in step S42, interpretable technologies including local interpretable unknown model interpretation technology, hierarchical correlation transfer and Shapley superposition interpretation are used to provide human-understandable visual explanations for the prediction results of each COPD patient.

[0081] The method for constructing an intelligent dynamic early warning model for acute exacerbations of COPD disclosed in this embodiment cleans, organizes, structures, and standardizes the collected data on acute exacerbations of COPD by executing step S1, thereby obtaining a large amount of high-quality basic data on acute exacerbations of COPD that can be directly used for scientific research. Subsequently, through steps S2 and S3, an efficient and practical risk prediction model for acute exacerbations of COPD is constructed using big data technology and artificial intelligence methods. The high-risk factors that affect the occurrence of acute exacerbations of COPD are identified from among the numerous risk factors, thereby reducing patient testing costs, improving clinical operability, and providing effective guidance for the early treatment of acute exacerbations of COPD.

[0082] Example 2:

[0083] See also Figure 6 As shown, this embodiment discloses an intelligent dynamic early warning system for acute exacerbation of COPD on the basis of embodiment one, including a collection module, a standard module, an extraction module, a prediction module and a visualization module, wherein the standard module is electrically connected to the collection module, the extraction module is electrically connected to the standard module, the prediction module is electrically connected to the extraction module, and the visualization module is electrically connected to the prediction module.

[0084] In this system, the collection module is configured to obtain risk factor information of multiple patients with acute exacerbation of COPD from the medical diagnosis database, and to repair the incomplete, abnormal, similar, repeated and inconsistent risk factor information using statistical analysis methods; the standard module is configured to perform post-structuring processing on the risk factor information repaired by the collection module through a machine learning algorithm, and then perform de-privacy and encryption processing by constructing a desensitizing algorithm intermediate mapping table to obtain a standardized data set; the extraction module is configured to screen out potential risk factors for acute exacerbation of COPD from the standardized data set using statistical analysis methods; the prediction module is configured to obtain the severity level of acute exacerbation of COPD based on the risk factors obtained by the extraction module; the visualization module is configured to visualize the real-time and accurate risk probability value of acute exacerbation of COPD and the corresponding risk level of each COPD patient.

[0085] The intelligent dynamic early warning system for acute exacerbations of COPD disclosed in this embodiment addresses the technical issues of the present invention by providing a collection module to clean, organize, structure, and standardize the collected data on acute exacerbations of COPD, thereby obtaining a large amount of high-quality basic data on acute exacerbations of COPD that can be directly used for scientific research. By providing a standard module and an extraction module to identify high-risk factors affecting acute exacerbations of COPD from among numerous risk factors, patient testing costs are reduced. Prediction modules and visualization modules are provided to obtain prediction results, thereby improving clinical operability and providing effective guidance for the early treatment of acute exacerbations of COPD.

[0086] In the description of the embodiments of the present application, it should be noted that in the description of the present application, terms such as "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and does not indicate or imply that the device or component must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present application.

[0087] In the description of the present application, the description with reference to the terms "one embodiment", "some embodiments", "in the present embodiment", "specific example", or "some examples" means that the specific features, mechanisms, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, mechanisms, materials or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and the features of different embodiments or examples, unless they are contradictory.

[0088] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A method for constructing an intelligent dynamic early warning model for acute exacerbation of chronic obstructive pulmonary disease, characterized in that: The steps include: S1: Obtain risk factor information of multiple patients with acute exacerbation of COPD from the medical diagnosis database, and use statistical analysis methods to repair the incomplete, abnormal, similar, repeated and inconsistent risk factor information; S2: Post-structuring the risk factor information repaired in step S1 by using a machine learning algorithm, and then de-privacy and encryption processing are performed by constructing a desensitization algorithm intermediate mapping table to obtain a standardized data set; S3: Screening out potential risk factors for acute exacerbation of COPD from the standardized data set by statistical analysis methods, and obtaining a prediction model based on the risk factors; S4: The prediction model is used to predict each COPD patient to obtain the prediction results and the high-risk factor data regularly tested by the hospital for each COPD patient to obtain the real-time and accurate risk probability value of acute exacerbation of COPD and the corresponding risk level of each COPD patient, and the results are visualized.

2. The method for constructing an intelligent dynamic early warning model for acute exacerbation of COPD according to claim 1, characterized in that: In step S1, the process of repairing the incomplete parts, abnormal contents, similar repeated contents and inconsistent data of the risk factor information by using the statistical analysis method is as follows: Repairing the incomplete parts of the risk factor information by interpolation filling, variance analysis filling or mean filling method; Identify and repair abnormal content of the risk factor information by using mean comparison, correlation comparison method or interval boundary method; Deleting similar and repeated contents of the risk factor information by using a data cleaning method based on an arrangement and merging algorithm, and merging by using a data cleaning technique based on an N-Gram algorithm; The inconsistent data of the risk factor information is repaired by successively extending the micro-function dependency identification method and the automatic repair method based on confidence and entropy.

3. The method for constructing an intelligent dynamic early warning model for acute exacerbation of COPD according to claim 2, characterized in that: The risk factor information includes clinical characteristic indicators, biomarker indicators, and personal clinical investigation data.

4. The method for constructing an intelligent dynamic early warning model for acute exacerbation of COPD according to any one of claims 1 to 3, characterized in that: The step S2 comprises the following steps: S21: performing post-structural processing on the symptom information in the risk factor information repaired in step S1 by natural language processing or conditional random field model processing based on bidirectional encoding representation of a converter to obtain structured data; S22: The structured data is associated with each other by using the patient primary index as a unique identifier, and de-privacy and encryption processing is performed by constructing an intermediate mapping table of a desensitizing algorithm to obtain the standardized data set.

5. The method for constructing an intelligent dynamic early warning model for acute exacerbation of COPD according to claim 4, characterized in that: The disease information includes medical records, examination reports, and imaging images of patients with acute exacerbation of COPD.

6. The method for constructing an intelligent dynamic early warning model for acute exacerbation of COPD according to claim 5, characterized in that: The step S3 comprises the following steps: S31: performing descriptive analysis on the standardized data set to characterize the statistical characteristics of data of different patients with acute exacerbation of COPD and obtain descriptive results; S32: Taking the severity level of COPD acute exacerbation as the dependent variable, using statistical analysis methods, a single factor preliminary screening of the risk of COPD acute exacerbation is performed on multiple indicators in the risk factor information describing the results to obtain potential risk factors for COPD acute exacerbation; S33: Select a deep neural network, and optimize the parameters of the deep neural network through an optimization algorithm based on the risk factors to obtain the prediction model.

7. The method for constructing an intelligent dynamic early warning model for acute exacerbation of COPD according to claim 6, characterized in that: The step S33 comprises the following steps: S331: Construct a multi-layer deep neural network model for the risk prediction of acute exacerbation of COPD, and use sensitivity, specificity, and area under the ROC curve to evaluate the model under different parameters to determine the optimal risk prediction model; S332: using the optimal risk prediction model to classify risk levels according to the distribution of risk probabilities; S333: The risk factors are used as input variables after dimension reduction, and the severity level of acute exacerbation of COPD is used as output variable, and the activation function adopts non-smooth ReLU and LeakyReLU functions to obtain a preliminary optimal risk prediction model; S334: Solving the model parameters of the preliminary optimal risk prediction model through an optimization algorithm to obtain the prediction model.

8. The method for constructing an intelligent dynamic early warning model for acute exacerbation of COPD according to claim 7, characterized in that: The step S4 comprises the following steps: S41: Based on the prediction results of each COPD patient by the prediction model, combined with the high-risk factor data regularly tested by the hospital for each COPD patient, obtain the real-time and accurate risk probability value of acute exacerbation of COPD and the corresponding risk level of each COPD patient; S42: Based on the historical risk probability data obtained for each COPD patient for multiple times, the risk probability value of acute exacerbation of COPD for the COPD patient in the future time period is predicted using time series technology to obtain the change trend of the risk probability of each COPD patient and the corresponding risk level; S43: Based on the results obtained in step S42, interpretable techniques including locally interpretable unknown model interpretation technology, hierarchical correlation transfer and Shapley superposition interpretation are used to provide a human-understandable visual explanation of the prediction results for each COPD patient.

9. The method for constructing an intelligent dynamic early warning model for acute exacerbation of COPD according to claim 8, characterized in that: The step S42 of using time series technology to predict the risk probability value of acute exacerbation of COPD in the COPD patient in the future time period is to use a fractional order cumulative grey prediction model to predict the risk probability value of acute exacerbation of COPD in the COPD patient in the future time period.

10. An intelligent dynamic early warning system for acute exacerbation of chronic obstructive pulmonary disease, characterized in that: The method for constructing an intelligent dynamic early warning model for acute exacerbation of chronic obstructive pulmonary disease according to any one of claims 1 to 9 comprises establishing an electrical connection relationship: A collection module is configured to obtain risk factor information of multiple patients with acute exacerbation of COPD from a medical diagnosis database, and to repair the incomplete, abnormal, similar, repeated and inconsistent risk factor information by using a statistical analysis method; A standard module is configured to perform post-structural processing on the risk factor information repaired by the collection module through a machine learning algorithm, and then perform de-privacy and encryption processing by constructing an intermediate mapping table of a desensitization algorithm to obtain a standardized data set; An extraction module is configured to screen out potential risk factors for acute exacerbation of COPD from the standardized data set by a statistical analysis method; A prediction module, configured to obtain a severity level of acute exacerbation of COPD based on the risk factors obtained by the extraction module; The visualization module is configured to visualize the real-time and accurate risk probability value of acute exacerbation of COPD and the corresponding risk level of each COPD patient.

Images