Heart disease risk assessment method fusing common disease features and deep learning algorithm

Abstract

The invention discloses a heart disease risk assessment method fusing common disease features and a deep learning algorithm, and the method comprises the steps: carrying out the missing value filling and data cleaning of an original data set, and carrying out the data preprocessing through a standardization method; in combination with expert knowledge, important features significantly related to the coronary heart disease are screened out through association rule analysis, clustering analysis and complex network analysis; combining the screened important features with common disease features and patient network features, constructing a prediction model by using multiple machine learning algorithms, evaluating model performance by using cross validation and performance evaluation indexes, and finding out a model with optimal performance; and performing parameter optimization on the optimal model. The method can accurately evaluate the risk that an individual suffers from the coronary heart disease, provides scientific decision support for medical care professionals, helps to realize early intervention and precise treatment, and enables the patient to recover early.

Description

Technical Field

[0001] The present invention relates to the technical field of medical data analysis, and in particular to a heart disease risk assessment method integrating comorbidity characteristics and deep learning algorithms. Background Art

[0002] Heart disease is a common and serious cardiovascular disease in life. Cardiovascular disease is one of the biggest threats to the health of people in my country and even the world. This disease has brought a serious burden to my country's medical system. Therefore, it is necessary to train a heart disease prediction model based on existing medical data to provide health guidance for patients.

[0003] However, in the existing field of heart disease prediction, there is an important but difficult problem to solve: it is difficult to accurately extract features related to heart disease risk from large-scale physical examination data and build an interpretable heart disease prediction model. Summary of the invention

[0004] The purpose of the present invention is to provide a heart disease risk assessment method that integrates comorbidity characteristics and deep learning algorithms. This method can accurately assess an individual's risk of developing coronary heart disease, provide scientific decision support for healthcare professionals, help achieve early intervention and precise treatment, and enable patients to recover as soon as possible.

[0005] In order to achieve the above object, the present invention provides a method for assessing the risk of heart disease by integrating comorbidity characteristics and deep learning algorithms, the method comprising:

[0006] Fill missing values ​​and clean the original data set, and use the Z-score standardization method for data preprocessing; subtract the mean of each feature value from the feature and divide it by its standard deviation to convert it to a distribution with a mean of 0 and a standard deviation of 1;

[0007] Combined with expert knowledge, association rule analysis was used to explore the comorbidity relationship between chronic diseases. The Apriori algorithm was used and the thresholds of support, confidence and lift were set to identify comorbidity pairs significantly associated with coronary heart disease, and new comorbidity features were constructed accordingly. At the same time, K-means cluster analysis was used to group patients according to comorbidity features to generate clustering features that reflect the similarity of patients' health status. Patient networks were constructed and complex network analysis techniques were used to extract network features.

[0008] The selected important features were combined with comorbidity characteristics and patient network characteristics, and a variety of machine learning algorithms were used, including support vector machine SVM, logistic regression LR, K nearest neighbor KNN, decision tree DT, random forest RF, gradient boosting tree LightGBM, XGBoost and convolutional neural network CNN to build a coronary heart disease prediction model; and cross-validation and performance evaluation indicators were used to evaluate the model performance to find the best performing model; at the same time, the parameters of the optimal model were tuned.

[0009] Preferably, the screened important features significantly associated with coronary heart disease include basic features, comorbidity features and patient network features.

[0010] Preferably, the basic characteristics include demographic information and physical measurements.

[0011] Preferably, demographic information includes age and drinking history, and physical measurements include height, weight, body mass index, systolic blood pressure, and diastolic blood pressure.

[0012] Preferably, comorbidity features are extracted through association rule analysis and cluster analysis.

[0013] Preferably, the comorbidity feature includes a plurality of comorbidity pair features: coronary heart disease and hypertension, diabetes and hypertension, dyslipidemia and hypertension.

[0014] Preferably, the patient network features include multiple indicators: degree centrality, closeness centrality, betweenness centrality, eigenvector centrality and clustering coefficient.

[0015] Preferably, the performance evaluation indicators include accuracy, precision, recall, F1-Score and AUC value.

[0016] Preferably, parameter tuning for the optimal model includes grid search combined with cross-validation to optimize model performance.

[0017] According to the above technical solution, the present invention first integrates the data, fills in missing values ​​and cleans the data; secondly, extracts multi-dimensional features through comorbidity analysis and patient network construction, and integrates them into the deep learning model; performs parameter tuning during model training, and uses cross-validation and performance indicators for performance evaluation. In this way, the comorbidity analysis and patient network characteristics are effectively combined, which can significantly improve the accuracy of coronary heart disease risk prediction, provide a scientific basis for clinical diagnosis and early intervention, and contribute to the prevention, treatment and management of cardiovascular diseases. At the same time, this method can be further promoted and applied to other chronic disease risk prediction fields, providing new ideas and tools for disease prevention and control.

[0018] Other features and advantages of the present invention will be described in detail in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the following specific embodiments, they are used to explain the present invention but do not constitute a limitation of the present invention. In the accompanying drawings:

[0020] Figure 1 is a flowchart of the implementation of the heart disease risk assessment method integrating comorbidity features and deep learning algorithms provided by the present invention;

[0021] Figure 2 is a flow chart of a method for assessing heart disease risk by integrating comorbidity features and a deep learning algorithm provided by the present invention. DETAILED DESCRIPTION

[0022] The specific implementation of the present invention is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the present invention, and is not used to limit the present invention.

[0023] See also Figure 1 and Figure 2 The present invention provides a method for assessing the risk of heart disease by integrating comorbidity features and a deep learning algorithm, the method comprising:

[0024] Fill missing values ​​and clean the original data set, and use the Z-score standardization method for data preprocessing; subtract the mean of each feature value from the feature and divide it by its standard deviation to convert it to a distribution with a mean of 0 and a standard deviation of 1;

[0025] Combined with expert knowledge, association rule analysis was used to explore the comorbidity relationship between chronic diseases. The Apriori algorithm was used and the thresholds of support, confidence and lift were set to identify comorbidity pairs significantly associated with coronary heart disease, and new comorbidity features were constructed accordingly. At the same time, K-means cluster analysis was used to group patients according to comorbidity features to generate clustering features that reflect the similarity of patients' health status. Patient networks were constructed and complex network analysis techniques were used to extract network features.

[0026] The selected important features were combined with comorbidity characteristics and patient network characteristics, and a variety of machine learning algorithms were used, including support vector machine SVM, logistic regression LR, K nearest neighbor KNN, decision tree DT, random forest RF, gradient boosting tree LightGBM, XGBoost and convolutional neural network CNN to build a coronary heart disease prediction model; and cross-validation and performance evaluation indicators were used to evaluate the model performance to find the best performing model; at the same time, the parameters of the optimal model were tuned.

[0027] In this embodiment, the basic characteristics include demographic information and physical measurement indicators. More specifically, preferably, the demographic information includes age and drinking history, and the physical measurement indicators include height, weight, body mass index, systolic blood pressure and diastolic blood pressure.

[0028] Specifically, the comorbidity relationship between diseases is extracted through association rule analysis methods (such as Apriori algorithm), and the comorbidity pair features with high support and high confidence are screened out, such as the comorbidity relationship between coronary heart disease and hypertension, diabetes and hypertension, etc. In addition, network features such as degree centrality, closeness centrality, and betweenness centrality are extracted based on the patient similarity network, thereby effectively reflecting the potential correlation between patients and the characteristics of disease transmission.

[0029] The above training process uses the Adam optimizer to update parameters. The key parameters involved include Epochs (number of iterations), Batch Size, Learning Rate, etc., to optimize the model training effect and prevent overfitting.

[0030] A specific implementation is provided below to illustrate the method:

[0031] S1: Obtain data features including age, height, weight, systolic blood pressure, diastolic blood pressure, total cholesterol, triglycerides, high-density lipoprotein, low-density lipoprotein, albumin and total protein;

[0032] S2: Standardize the acquired data. The standardization formula used is as follows:

[0033]

[0034] Where x is the original data point, M is the median, Q1 and Q3 are the first quartile and the third quartile respectively.

[0035] S3: Extract comorbidity characteristics and network characteristics through association rule analysis and complex network analysis, such as comorbidity pair characteristics (CHD_HYT, DIA_HYT, etc.) and network centrality indicators (degree centrality, eigenvector centrality, etc.);

[0036] S4: Input the feature fused data into the convolutional neural network model, and perform feature extraction and integration through the feature fusion module and attention mechanism;

[0037] S5: During model training, the Adam optimizer was used for parameter tuning, with the learning rate set to 0.001, the number of iterations to 500, and the batch size to 64.

[0038] The implementation of the above experiments showed that after integrating the comorbidity features, the accuracy rate increased by 10 percentage points; and after adding the network features, the accuracy rate further increased by 9 percentage points; when both the comorbidity features and the network features were integrated with the basic features, the accuracy rate increased by about 8 percentage points compared with the fusion of a single feature. At the same time, other evaluation indicators also improved significantly, proving the effectiveness of the comorbidity features and network features.

[0039] The above methods are used to conduct data mining and modeling based on comorbidity analysis and patient network characteristics, which significantly improves the accuracy and reliability of coronary heart disease risk prediction. At the same time, it also significantly improves the accuracy of CHD risk prediction, making the model's accuracy rate reach 94.50%, sensitivity 96.87%, and precision 91.17%. These results are not only reflected in the data, but also have potential application value in clinical practice, providing a new perspective for the early identification and intervention of cardiovascular disease.

[0040] In summary, the present invention first integrates data, fills in missing values ​​and cleans data; secondly, extracts multi-dimensional features through comorbidity analysis and patient network construction, and integrates them into the deep learning model; performs parameter tuning during model training, and uses cross-validation and performance indicators for performance evaluation. In this way, comorbidity analysis and patient network characteristics are effectively combined, which can significantly improve the accuracy of coronary heart disease risk prediction, provide a scientific basis for clinical diagnosis and early intervention, and contribute to the prevention, treatment and management of cardiovascular diseases. At the same time, this method can be further promoted and applied to other chronic disease risk prediction fields, providing new ideas and tools for disease prevention and control.

[0041] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0042] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0043] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0044] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0045] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0046] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0047] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0048] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0049] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A heart disease risk assessment method integrating comorbidity features and deep learning algorithms, characterized in that: The method comprises: Fill missing values ​​and clean the original data set, and use the Z-score standardization method for data preprocessing; subtract the mean of each feature value from the feature and divide it by its standard deviation to convert it to a distribution with a mean of 0 and a standard deviation of 1; Combined with expert knowledge, association rule analysis was used to explore the comorbidity relationship between chronic diseases. The Apriori algorithm was used and the thresholds of support, confidence and lift were set to identify comorbidity pairs significantly associated with coronary heart disease, and new comorbidity features were constructed accordingly. At the same time, K-means cluster analysis was used to group patients according to comorbidity features to generate clustering features that reflect the similarity of patients' health status. Patient networks were constructed and complex network analysis techniques were used to extract network features. The selected important features were combined with comorbidity characteristics and patient network characteristics, and a variety of machine learning algorithms were used, including support vector machine SVM, logistic regression LR, K nearest neighbor KNN, decision tree DT, random forest RF, gradient boosting tree LightGBM, XGBoost and convolutional neural network CNN to build a coronary heart disease prediction model; and cross-validation and performance evaluation indicators were used to evaluate the model performance to find the best performing model; at the same time, the parameters of the optimal model were tuned.

2. The heart disease risk assessment method integrating comorbidity features and deep learning algorithms according to claim 1, characterized in that: The important characteristics screened out that were significantly associated with CHD included basic characteristics, comorbidity characteristics, and patient network characteristics.

3. The heart disease risk assessment method integrating comorbidity features and deep learning algorithms according to claim 2, characterized in that: The basic characteristics include demographic information and physical measurements.

4. The heart disease risk assessment method integrating comorbidity features and deep learning algorithms according to claim 3, characterized in that: The demographic information included age and drinking history, and the physical measurements included height, weight, body mass index, systolic blood pressure, and diastolic blood pressure.

5. The heart disease risk assessment method integrating comorbidity features and deep learning algorithms according to claim 1, characterized in that: Comorbidity characteristics were extracted through association rule analysis and cluster analysis.

6. The heart disease risk assessment method integrating comorbidity features and deep learning algorithms according to claim 5, characterized in that: The comorbidity characteristics include multiple comorbidity pair characteristics: coronary heart disease and hypertension, diabetes and hypertension, dyslipidemia and hypertension.

7. The heart disease risk assessment method integrating comorbidity features and deep learning algorithms according to claim 1, characterized in that: The patient network characteristics include multiple indicators: degree centrality, closeness centrality, betweenness centrality, eigenvector centrality and clustering coefficient.

8. The heart disease risk assessment method integrating comorbidity features and deep learning algorithms according to claim 1, characterized in that: The performance evaluation indicators include accuracy, precision, recall, F1-Score and AUC value.

9. The heart disease risk assessment method integrating comorbidity features and deep learning algorithms according to claim 1, characterized in that: Parameter tuning for the optimal model includes grid search combined with cross-validation to optimize model performance.

Images