Cardiovascular event occurrence probability prediction method and system based on ambulatory blood pressure monitoring

Abstract

The invention discloses a cardiovascular event occurrence probability prediction method and system based on ambulatory blood pressure monitoring, and relates to the technical field of artificial intelligence. According to the method, a feature data set is constructed according to ambulatory blood pressure sequence features and personnel information features; inputting the feature data set into a trained ambulatory blood pressure encoder, and outputting a hidden feature matrix; according to a preset moment, dividing the hidden feature matrix into a first half-day hidden feature and a second half-day hidden feature; based on the first half day hidden feature and the second half day hidden feature, obtaining a morning period hidden feature and a night period hidden feature; the morning period hidden features and the night period hidden features are input into a feature fusion and occurrence probability output module, cardiovascular event occurrence probabilities corresponding to the morning period and the night period are obtained, and the feature fusion and occurrence probability output module is constructed based on a cross attention mechanism. By implementing the technical scheme of the invention, the occurrence probability of the cardiovascular event corresponding to the early morning time period and the night time period can be accurately obtained.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and specifically to a method and system for predicting the probability of occurrence of cardiovascular events based on dynamic blood pressure monitoring. Background Art

[0002] Cardiovascular events (CVE) refer to serious adverse events caused by heart and blood vessel diseases, such as heart attack, stroke, sudden cardiac death, etc. Cardiovascular events are very harmful. They are one of the leading causes of death worldwide, causing about 18 million deaths each year, one-third of which occur in people under the age of 70. In addition, cardiovascular events can cause disability and seriously reduce the quality of life of patients.

[0003] In the research related to cardiovascular events, there is a lack of direct warning of the probability of cardiovascular events, especially the lack of warning of the probability of cardiovascular events at specific times such as early morning and night. This also makes it impossible for people at high risk of cardiovascular events to receive more comprehensive probability warnings and timely rescue measures.

[0004] Therefore, how to accurately obtain the probability of cardiovascular events corresponding to the early morning and evening hours has become a technical problem that needs to be solved urgently. Summary of the invention

[0005] The present application provides a method and system for predicting the probability of occurrence of cardiovascular events based on dynamic blood pressure monitoring, which can accurately obtain the probability of occurrence of cardiovascular events corresponding to the early morning and evening time periods.

[0006] In a first aspect, the present application provides a method for predicting the probability of occurrence of cardiovascular events based on dynamic blood pressure monitoring, the method comprising: obtaining dynamic blood pressure sequence characteristics and personnel information characteristics of a target person, and constructing a feature data set based on the dynamic blood pressure sequence characteristics and personnel information characteristics; Inputting the feature data set into a trained dynamic blood pressure encoder and outputting a latent feature matrix; According to a preset time, the latent feature matrix is ​​divided into the first half of the day latent features and the second half of the day latent features; Based on the first half-day hidden features and the second half-day hidden features, obtain the early morning time hidden features and the evening time hidden features corresponding to the target person; The latent features of the early morning period and the latent features of the evening period are input into a preset feature fusion and occurrence probability output module to obtain the probability of occurrence of cardiovascular events corresponding to the target person in the early morning period and the evening period, wherein the preset feature fusion and occurrence probability output module is constructed based on a cross-attention mechanism.

[0007] Optionally, before inputting the early morning period latent features and the evening period latent features into a preset feature fusion and occurrence probability output module to obtain the corresponding cardiovascular event occurrence probability of the target person in the early morning period and the evening period, the method further includes: Acquire historical dynamic blood pressure sequence features and historical personnel information features of historical target personnel, and construct a training feature data set based on the historical dynamic blood pressure sequence features and historical personnel information features; Based on the time series model or the non-time series model, a dynamic blood pressure encoder is constructed; Input the training feature data set into the occurrence probability prediction model framework for training to obtain an initial occurrence probability prediction model. The occurrence probability prediction model framework includes the dynamic blood pressure encoder, a preset morning encoder, a preset night encoder, a preset feature fusion and occurrence probability output module; Evaluate the initial occurrence probability prediction model using a cross-validation method to obtain a performance indicator of the initial occurrence probability prediction model; the performance indicator includes a cross-validation result; Based on the cross-validation result, the hyperparameters of the initial occurrence probability prediction model are adjusted using a hyperparameter tuning method to obtain the preset occurrence probability prediction model.

[0008] Optionally, the probability of occurrence of a cardiovascular event includes a first probability of occurrence of a cardiovascular event and a second probability of occurrence of a cardiovascular event; the step of inputting the latent features of the early morning period and the latent features of the evening period into a preset feature fusion and occurrence probability output module to obtain the corresponding probability of occurrence of a cardiovascular event of the target person in the early morning period and the evening period specifically includes: Inputting the early morning period hidden feature and the evening period hidden feature into the preset feature fusion and occurrence probability output module to obtain a first splicing feature and a second splicing feature; Based on the first splicing feature, obtaining the probability of occurrence of the first cardiovascular event corresponding to the target person in the early morning period; Based on the second splicing feature, the probability of occurrence of the second cardiovascular event corresponding to the target person during the night time period is obtained.

[0009] Optionally, the preset feature fusion and occurrence probability output module includes a first encoder, a second encoder, a first fully connected layer, a second fully connected layer, a first cross attention module, and a second cross attention module; the early morning period latent feature and the evening period latent feature are input into the preset feature fusion and occurrence probability output module to obtain a first splicing feature and a second splicing feature, specifically including: Inputting the early morning period latent feature into the first encoder to obtain a first feature; Inputting the night-time latent feature into the second encoder to obtain a second feature; Inputting the first feature into the first fully connected layer to obtain a first intermediate feature; Inputting the second feature into the second fully connected layer to obtain a second intermediate feature; Inputting the first intermediate feature and the second feature into the first cross attention module to obtain a first attention feature; Inputting the second intermediate feature and the first feature into the second cross attention module to obtain a second attention feature; Concatenate the first feature and the first attention feature to obtain a first concatenated feature; The second feature and the second attention feature are concatenated to obtain a second concatenated feature.

[0010] Optionally, obtaining the early morning time period hidden features and the evening time period hidden features corresponding to the target person based on the first half of the day hidden features and the second half of the day hidden features specifically includes: Input the first half of the day latent features into a preset morning encoder to obtain the awakening time of the target person, and input the second half of the day latent features into a preset night encoder to obtain the sleeping time of the target person; According to the awakening time, the early morning period latent features are extracted from the first half of the day latent features, and according to the sleeping time, the evening period latent features are extracted from the second half of the day latent features.

[0011] Optionally, extracting the early morning period latent features from the first half of the day latent features according to the awakening time, and extracting the evening period latent features from the second half of the day latent features according to the sleeping time specifically includes: Based on the awakening moment, determining a first time window corresponding to the awakening moment, wherein the awakening moment is a center moment of the first time window; Extracting the intermediate features of the first half of the day corresponding to the first time window from the latent features of the first half of the day; Using the intermediate features of the first half of the day as the latent features of the early morning period; Based on the sleeping time, determining a second time window corresponding to the sleeping time, wherein the sleeping time is the center time of the second time window; Extracting the second half-day intermediate features corresponding to the second time window from the second half-day latent features; The intermediate features of the second half of the day are used as the latent features of the night time period.

[0012] Optionally, the dynamic blood pressure encoder is any one of a timing model or a non-timing model.

[0013] Optionally, constructing a feature data set according to the dynamic blood pressure sequence features and personnel information features specifically includes: Based on a preset assessment standard, determining a first disease label and a second disease label corresponding to the target person; Constructing a basic data set based on the dynamic blood pressure sequence feature, the personnel information feature, the first disease label, and the second disease label; The basic data set is filled with null values ​​to obtain the feature data set.

[0014] Optionally, after obtaining the corresponding cardiovascular event occurrence probabilities of the target person in the early morning and evening periods, the method further includes: When the probability of the cardiovascular event occurring is greater than the preset probability of the disease occurring, an early warning prompt message is output to the physician terminal.

[0015] In a second aspect of the present application, a cardiovascular event occurrence probability prediction system based on dynamic blood pressure monitoring is provided, the system comprising: a data acquisition module, a feature acquisition module, and a feature fusion and output module; The data acquisition module is used to obtain dynamic blood pressure sequence characteristics and personnel information characteristics of the target person, and construct a feature data set based on the dynamic blood pressure sequence characteristics and personnel information characteristics; The feature acquisition module is used to input the feature data set into the trained dynamic blood pressure encoder and output a latent feature matrix; divide the latent feature matrix into the first half of the day and the second half of the day according to the preset time; based on the first half of the day and the second half of the day, obtain the early morning time period latent features and the evening time period latent features corresponding to the target person; The feature fusion and output module is used to input the latent features of the early morning period and the latent features of the evening period into a preset feature fusion and occurrence probability output module to obtain the corresponding probability of cardiovascular events of the target person in the early morning period and the evening period, wherein the preset feature fusion and occurrence probability output module is constructed based on a cross-attention mechanism.

[0016] Optionally, the system further comprises: an early warning module; The early warning module is used to output early warning information to the physician terminal when the probability of the cardiovascular event is greater than the preset probability of the disease.

[0017] In the third aspect of the present application, an electronic device is provided, including a processor, a memory, a user interface and a network interface, the memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device executes any method as described in the first aspect of the present application.

[0018] In a fourth aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program that can be loaded by a processor and execute the method as described in any one of the first aspects of the present application.

[0019] In summary, one or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: The present application provides a method for early warning of the probability of occurrence of cardiovascular events at specific times such as early morning and evening based on dynamic blood pressure. It can predict whether a patient is at risk of adverse cardiovascular events in the early morning and at night, and remind doctors of abnormal blood pressure in patients in the early morning and at night. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 This is one of the flow charts of a method for predicting the probability of occurrence of cardiovascular events based on dynamic blood pressure monitoring provided in an embodiment of the present application; Figure 2 It is a macroscopic framework schematic diagram of the probability prediction of cardiovascular events provided in the embodiment of the present application; Figure 3 This is the second flow chart of a method for predicting the probability of occurrence of cardiovascular events based on dynamic blood pressure monitoring provided in an embodiment of the present application; Figure 4 This is a flowchart of a method for predicting the probability of occurrence of cardiovascular events based on dynamic blood pressure monitoring provided in an embodiment of the present application; Figure 5 This is a fourth flow chart of a method for predicting the probability of occurrence of cardiovascular events based on dynamic blood pressure monitoring provided in an embodiment of the present application; Figure 6 This is a flowchart of a method for predicting the probability of occurrence of cardiovascular events based on dynamic blood pressure monitoring provided in an embodiment of the present application; Figure 7 It is a macroscopic framework diagram of a feature fusion and occurrence probability output module provided in an embodiment of the present application; Figure 8 It is a structural schematic diagram of a cardiovascular event occurrence probability prediction system based on dynamic blood pressure monitoring provided in an embodiment of the present application; Fig. 9 It is a structural schematic diagram of an electronic device disclosed in an embodiment of the present application.

[0021] Explanation of the accompanying drawings: 1. Data acquisition module; 2. Feature acquisition module; 3. Feature fusion and output module; 4. Occurrence probability prediction model building module; 5. Early warning module; 900. Electronic device; 901. Processor; 902. Communication bus; 903. User interface; 904. Network interface; 905. Memory. DETAILED DESCRIPTION

[0022] In order to enable technicians in this field to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments.

[0023] In the description of the embodiments of the present application, words such as "for example" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "for example" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "for example" or "for example" is intended to present related concepts in a specific way.

[0024] In the description of the embodiments of the present application, the meaning of the term "multiple" refers to two or more. For example, multiple systems refer to two or more systems, and multiple screen terminals refer to two or more screen terminals. In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the indicated technical features. Thus, the features defined as "first" and "second" can explicitly or implicitly include one or more of the features. The terms "include", "comprise", "have" and their variations all mean "including but not limited to", unless otherwise specifically emphasized.

[0025] The present application provides a method for predicting the probability of occurrence of cardiovascular events based on dynamic blood pressure monitoring, referring to Figure 1 , which shows one of the flow diagrams of a method for predicting the probability of occurrence of cardiovascular events based on dynamic blood pressure monitoring provided in an embodiment of the present application. The method includes steps S1-S5, which are as follows: Based on the description of steps S1-S5, in order to facilitate the understanding of the overall solution, refer to Figure 2 , which shows a macroscopic framework schematic diagram of the probability of occurrence of cardiovascular events provided in an embodiment of the present application.

[0026] Step S1: Acquire the dynamic blood pressure sequence features and personnel information features of the target person, and construct a feature data set based on the dynamic blood pressure sequence features and personnel information features.

[0027] Specifically, in the process of implementing step S1, the dynamic blood pressure sequence characteristics and personnel information characteristics of the target person are first determined as basic data. The feature data set includes the dynamic blood pressure sequence characteristics and personnel information characteristics of the target person. The dynamic blood pressure sequence characteristics include blood pressure and heart rate, and the personnel information characteristics include age, gender, weight, and medical history. These data include physiological parameters such as blood pressure and heart rate, as well as basic information such as age, gender, weight, and medical history, and are key factors in assessing the probability of occurrence of cardiovascular diseases.

[0028] By implementing step S1, a comprehensive, accurate and complete feature dataset can be constructed, which provides a solid foundation for the prediction of the probability of subsequent cardiovascular events.

[0029] It should be noted that the specific implementation of step S1 will be described in detail in subsequent embodiments.

[0030] Step S2: Input the feature data set into the trained dynamic blood pressure encoder and output the latent feature matrix.

[0031] Specifically, in step S2, the preprocessed feature data set is input into the trained dynamic blood pressure encoder. It should be noted that the dynamic blood pressure encoder is any one of a time series model and a non-time series model.

[0032] The purpose of this step is to transform the raw physiological and personal information data into a latent feature matrix that is more conducive to subsequent analysis. The dynamic blood pressure encoder was selected for this transformation based on its ability to capture complex patterns and associations in time series data.

[0033] When implementing this step, you first need to determine the specific type of dynamic blood pressure encoder. Time series models such as long short-term memory networks (LSTM) or recurrent neural networks (RNN) are particularly suitable for processing time series data because they can capture the time dependencies and dynamic changes in the data. Non-time series models, such as multi-layer perceptrons (MLP) or Transformers, although they do not directly deal with time series dependencies, can still capture time series features through appropriate data preprocessing (such as integration of time windows). This choice will be determined according to the specific application scenario and data characteristics.

[0034] The dynamic blood pressure encoder outputs a latent feature matrix by learning the patterns in the data set. These latent features comprehensively reflect the dynamic changes in the patient's blood pressure and their association with personal basic information. In this way, the raw data is converted into a form that is more suitable for machine learning analysis and probability assessment. The output latent feature matrix provides detailed, encoded information for subsequent steps, which is the basis for accurate prediction of the probability of cardiovascular events.

[0035] The effect of this implementation is that it can greatly improve the ability to express data, converting simple physiological and personal information into complex data forms that can reflect the deep changes in the patient's health status. This not only improves the accuracy of the model's prediction of the probability of future cardiovascular events, but also provides the possibility for more detailed probability assessment, thereby supporting medical decision-making and the implementation of early intervention measures.

[0036] Step S3: According to the preset time, the latent feature matrix is ​​divided into the first half of the day latent features and the second half of the day latent features.

[0037] Specifically, in step S3, the latent feature matrix output by the dynamic blood pressure encoder is divided into the first half of the day latent features and the second half of the day latent features according to the preset time. This step is to divide the key time period related to the biological rhythm from the continuous 24-hour data, so as to conduct more accurate health status analysis and occurrence probability prediction.

[0038] The necessity of step S3 is that the physiological state of the human body will have significant periodic changes during the day, and these changes are closely related to the probability of cardiovascular events. For example, the incidence of cardiovascular events is usually higher in the early morning, which is related to the changes in hormone levels in the body when waking up in the morning. Therefore, dividing the latent feature matrix into the first half of the day and the second half of the day features helps to capture the changes in the probability of occurrence caused by these biological rhythms.

[0039] Therefore, when implementing step S3, it is first necessary to determine the middle time, usually the 12-hour point, and divide the 24-hour data into the first half of the day from midnight to noon, and the second half of the day from noon to midnight of the next day. Such a division allows the data to be better matched with the patient's daily activities and physiological cycles in the model analysis.

[0040] Step S4: Based on the latent features of the first half of the day and the latent features of the second half of the day, the early morning period latent features and the evening period latent features corresponding to the target person are obtained.

[0041] Specifically, in step S4, the segmented first half of the day latent features and the second half of the day latent features are processed by the preset morning encoder and night encoder to further determine the personalized awakening time and sleeping time of the target person, and extract the corresponding morning period latent features and night period latent features accordingly. The implementation of this step is crucial for accurately assessing the probability of cardiovascular events, because it allows the model to pay special attention to the peak physiological activity periods of individuals, which are often closely related to the increased probability of cardiovascular events.

[0042] The necessity of step S4 is that, in general, the probability of cardiovascular events such as myocardial infarction increases significantly within a few hours after waking up in the morning, and the hours during sleep and after falling asleep at night are also critical periods. By accurately identifying and analyzing the physiological data of these specific time periods, these potential probability periods can be predicted and prevented more effectively.

[0043] It should be noted that the specific implementation of step S4 will be described in detail in subsequent embodiments.

[0044] Step S5: Input the latent features of the early morning period and the latent features of the evening period into a preset feature fusion and occurrence probability output module to obtain the corresponding probability of cardiovascular events for the target person in the early morning period and the evening period, wherein the preset feature fusion and occurrence probability output module is constructed based on a cross-attention mechanism.

[0045] Specifically, in step S5, the latent features of the early morning period and the latent features of the evening period obtained from the previous step are input into the preset feature fusion and occurrence probability output module, which is built based on the cross-attention mechanism to obtain the probability of cardiovascular events of the target person in the early morning period and the evening period. This step is the core of the entire early warning system because it directly outputs the final prediction results, which are used to evaluate and possibly trigger preventive measures.

[0046] The necessity of step S5 is that the prediction of the probability of cardiovascular events needs to take into account time correlation and individual differences, especially in different stages of daily life, such as early morning awakening and nighttime sleep. By analyzing data in specific time periods, high probability periods can be identified more accurately, thereby achieving more effective prevention and intervention. The cross-attention mechanism plays a key role here, which improves the accuracy and relevance of the prediction by strengthening the model's understanding of the differences and correlations between early morning and evening data.

[0047] It should be noted that the specific implementation of step S5 will be described in detail in subsequent embodiments.

[0048] In a possible implementation, before step S5, the method further includes the following steps: The historical dynamic blood pressure sequence features and historical personnel information features of the historical target personnel are obtained, and a training feature data set is constructed based on the historical dynamic blood pressure sequence features and historical personnel information features.

[0049] Specifically, in this step, the task is to obtain the historical dynamic blood pressure sequence characteristics and historical personnel information characteristics of the historical target personnel, and build a training feature data set based on these historical data. This step is to ensure that the cardiovascular event warning system can effectively learn and build models based on historical data, thereby improving the accuracy and reliability of the prediction model.

[0050] The specific method of obtaining the historical dynamic blood pressure sequence characteristics and historical personnel information characteristics of the historical target person, and constructing a training feature data set based on the historical dynamic blood pressure sequence characteristics and historical personnel information characteristics is consistent with the implementation method of step S1 in the aforementioned embodiment, so it will not be elaborated here.

[0051] Based on the timing model or the non-timing model, a dynamic blood pressure encoder is constructed.

[0052] Specifically, in this step, the task is to build an ambulatory blood pressure encoder based on a time series model or a non-time series model. The purpose of this step is to create a tool that can effectively extract features from ambulatory blood pressure data and related health information to support subsequent cardiovascular event probability analysis and prediction.

[0053] In the specific operation, first select a suitable model architecture based on the characteristics of the collected historical data set. If the data shows obvious time dependence, such as the periodic changes of blood pressure and heart rate over time, it would be more appropriate to use a time series model such as LSTM or RNN; if the focus is on extracting features from the data at each time point, MLP or CNN may be selected. After that, the model is designed and trained, including setting the number of network layers, the number of neurons, the activation function, and the optimization algorithm. The design and training of the above model are determined according to the actual application scenario needs and are existing technologies. During the model training process, the encoder is trained using the historical feature dataset, and the parameters are adjusted to achieve the best learning effect to obtain a trained dynamic blood pressure encoder.

[0054] The training feature data set is input into the occurrence probability prediction model framework for training to obtain an initial occurrence probability prediction model. The occurrence probability prediction model framework includes a dynamic blood pressure encoder, a preset morning encoder, a preset night encoder, a preset feature fusion and occurrence probability output module.

[0055] Specifically, in this step, the task is to input the training feature data set into the probability prediction model framework for training, which includes a dynamic blood pressure encoder, a preset morning encoder, a preset night encoder, and a preset feature fusion and probability output module. The purpose of this step is to create a comprehensive prediction model that can accurately predict the probability of cardiovascular events based on the data features obtained from different encoders and modules.

[0056] In the specific operation, first of all, it is necessary to ensure that all encoders and modules have been designed and optimized according to the requirements of pre-training. Then, the training feature data set that integrates dynamic blood pressure features, basic information features of personnel, and other relevant health indicators is input into the model framework. This step involves all aspects of model training, including selecting appropriate loss functions and optimization algorithms, setting reasonable training cycles, and making necessary hyperparameter adjustments to ensure that the model can effectively learn and capture key patterns and relationships in the data.

[0057] Through in-depth learning of the training feature data set, the prediction model is able to establish complex data associations and pattern recognition, which are indispensable for probability assessment and prediction. In addition, by combining the outputs of multiple encoders and feature fusion modules, the model can provide a comprehensive prediction of the probability of cardiovascular events, which is much more accurate than a single data source or simple analysis method.

[0058] The initial occurrence probability prediction model is evaluated by a cross-validation method to obtain performance indicators of the initial occurrence probability prediction model; the performance indicators include cross-validation results.

[0059] Specifically, in the field of machine learning and data science, evaluating the performance of the model and verifying its predictive ability are crucial links. Through cross-validation, the model is trained and tested on multiple data subsets, which can effectively evaluate its stability and accuracy and avoid data contingency and overfitting. Cross-validation can provide a comprehensive evaluation to ensure that the model can achieve the expected results in practical applications.

[0060] Therefore, in this step, the task is to evaluate the initial occurrence probability prediction model using the cross-validation method to obtain the performance indicators of the initial occurrence probability prediction model. This step is to ensure that the model has good generalization ability on different data samples and verify its effect in predicting the probability of cardiovascular events.

[0061] In specific operations, first set up an appropriate cross-validation scheme, such as k-fold cross-validation, in which the dataset is split into k smaller subsets. The model is trained on k-1 subsets and tested on the remaining subset. This process is repeated k times, and a different subset is selected as the test set each time to evaluate the performance of the model. After each test, record the model's performance indicators, such as accuracy, recall, and area under the ROC curve (AUC).

[0062] Based on the cross-validation results, the hyperparameter tuning method is used to adjust the hyperparameters of the initial occurrence probability prediction model to obtain the preset occurrence probability prediction model.

[0063] Specifically, hyperparameter tuning is an important part of the machine learning model development process. By adjusting the model's hyperparameters, such as learning rate, number of neural network layers, number of nodes, batch size, etc., the model's learning ability and final performance can be significantly affected. Appropriate hyperparameter settings can help the model learn from data more effectively, avoid overfitting or underfitting, and thus improve the performance and stability of the model in practical applications.

[0064] Therefore, in this step, the task is to adjust the initial occurrence probability prediction model using the hyperparameter tuning method based on the cross-validation results to obtain an optimized occurrence probability prediction model. This step is to ensure that the model can perform its tasks with the best configuration and improve the accuracy and efficiency of the prediction.

[0065] In specific operations, the current configuration of the model is first evaluated based on the performance indicators obtained during the cross-validation process. Then, an appropriate hyperparameter tuning strategy is selected, such as grid search, random search, or Bayesian optimization-based methods, to systematically explore the effects of different hyperparameter combinations. These methods search within a predefined hyperparameter space to find the parameter combination that maximizes the model performance indicators (such as accuracy, AUC, etc.).

[0066] After hyperparameter tuning, the model usually needs to be retrained on the same training data set to verify the effect of the new parameter settings. The optimized model is once again evaluated for performance to ensure that the adjustments made have indeed improved the model's predictive ability.

[0067] In a possible implementation, the probability of a cardiovascular event occurring includes a first probability of a cardiovascular event occurring and a second probability of a cardiovascular event occurring. Figure 5 , which shows a fourth flow chart of a method for predicting the probability of occurrence of cardiovascular events based on dynamic blood pressure monitoring provided in an embodiment of the present application. Step S5 specifically includes steps S51-S53: Step S51: inputting the latent features of the early morning period and the latent features of the evening period into a preset feature fusion and occurrence probability output module to obtain a first splicing feature and a second splicing feature.

[0068] Specifically, in step S51, the task is to input the latent features of the early morning period and the latent features of the evening period into the preset feature fusion and occurrence probability output module, and then obtain the first splicing feature and the second splicing feature. This step is to effectively integrate and analyze the key features extracted from different time periods to predict the probability of cardiovascular events in the target person in the early morning and evening periods.

[0069] By combining and strengthening the features extracted from the early morning and evening hours, the model can more comprehensively assess the probability of cardiovascular events in patients during these critical hours. This approach not only improves the overall accuracy of the probability prediction, but also makes the prediction results more detailed.

[0070] It should be noted that the specific implementation of step S51 will be described in detail in subsequent embodiments.

[0071] Step S52: Based on the first splicing feature, the probability of occurrence of the first cardiovascular event corresponding to the target person in the early morning period is obtained.

[0072] Specifically, in the specific operation, the first spliced ​​feature obtained by feature fusion and processing is first input into a preset occurrence probability assessment model. This model usually contains machine learning algorithms, such as logistic regression, support vector machine, or deep learning network, which can predict the probability of the first cardiovascular event based on the input features. The model analyzes the information in the first spliced ​​feature, such as patterns and trends extracted from physiological data, and important physiological signals obtained from the cross-attention mechanism, to calculate the probability of cardiovascular events in the early morning period.

[0073] Step S53: Based on the second splicing feature, the probability of occurrence of a second cardiovascular event corresponding to the target person during the night time period is obtained.

[0074] Specifically, the second concatenated features obtained by processing the feature fusion and occurrence probability output module are first used. These features integrate the information extracted and enhanced from the night encoder and related fully connected layers and cross-attention modules. Then, these features are input into a pre-trained prediction model, which may use machine learning algorithms such as random forests, gradient boosting machines, or deep learning networks to predict the probability of a second cardiovascular event. In this process, the model analyzes the patterns and correlations in these features, such as heart rate variability, blood pressure fluctuations, etc., to predict the possibility of nocturnal cardiovascular events.

[0075] In a possible implementation, the preset feature fusion and occurrence probability output module includes a first encoder, a second encoder, a first fully connected layer, a second fully connected layer, a first cross attention module, and a second cross attention module. Figure 6 , which shows a flowchart of a method for predicting the probability of occurrence of cardiovascular events based on dynamic blood pressure monitoring provided in an embodiment of the present application. Step S51 specifically includes steps S511-S518: Based on the description of steps S511-S518, in order to facilitate the understanding of the overall solution, refer to Figure 7 , which shows a macroscopic framework schematic diagram of a feature fusion and occurrence probability output module provided in an embodiment of the present application.

[0076] Step S511: input the latent features of the early morning period into the first encoder to obtain the first features.

[0077] Specifically, in step S511, the task is to input the latent features of the early morning period into the first encoder to obtain the first features. This step is implemented to extract deeper and more detailed information through advanced coding processing, which is crucial for accurately predicting the probability of cardiovascular events in the early morning period.

[0078] In the specific operation, it is first necessary to ensure that the latent features in the early morning period accurately reflect the physiological data collected within two hours before and after waking up, such as time series data of heart rate and blood pressure. These data are input to the first encoder, which usually adopts advanced neural network structures such as deep learning models, including but not limited to convolutional neural networks (CNN) or recurrent neural networks (RNN). These models are good at processing time series data and extracting complex features, so that the first encoder can extract more detailed and key information from the original latent features to form the first feature.

[0079] The first encoder processes the input latent features through advanced algorithms and is able to generate first features that contain deeper information, which more finely reflects the physiological state and potential probability of occurrence in the early morning. For example, by identifying specific heart rate change patterns or abnormal blood pressure increases, the model can more accurately predict the probability of cardiovascular events that may occur in the early morning.

[0080] Step S512: Input the latent features of the night time period into the second encoder to obtain the second features.

[0081] Specifically, in step S512, the task is to input the latent features of the night time period into the second encoder to obtain the second features. The purpose of this step is to use advanced coding technology to further process and refine the latent features of the night time period to enhance the model's ability to identify and predict the probability factors of cardiovascular events at night.

[0082] In the specific operation, we first need to ensure that the latent features of the night time period accurately reflect the relevant physiological data collected within two hours before and after falling asleep. These data are input to the second encoder, which usually adopts advanced neural network structures such as deep learning models, including but not limited to convolutional neural networks (CNN) or recurrent neural networks (RNN). These models are good at processing time series data and extracting complex features, so that the second encoder can extract more detailed and key information from the original latent features to form the second features.

[0083] The second encoder not only improves the depth of data processing, but also enhances the model's ability to identify potential factors that may affect the occurrence of nocturnal cardiovascular events. For example, by identifying specific patterns of heart rate slowing or abnormal blood pressure reduction, the second feature can help the model more accurately predict the probability of cardiovascular events that may occur at night.

[0084] Step S513: Input the first feature into the first fully connected layer to obtain the first intermediate feature.

[0085] Specifically, in step S513, the task is to input the first feature obtained from the first encoder into the first fully connected layer to obtain the first intermediate feature. The purpose of this step is to further process and refine the features of the early morning period, enhance the model's ability to understand and process these data, and thus improve the accuracy of the prediction of the probability of cardiovascular events.

[0086] In the specific operation, first, the first feature, that is, the feature set processed by the first encoder, is input to the first fully connected layer. This fully connected layer consists of multiple neurons, each of which can learn different aspects and correlations in the data and optimize the output by adjusting weights and biases. Set a suitable activation function (such as ReLU or Sigmoid) to ensure nonlinear feature conversion and better learning effect.

[0087] The processing of the first fully connected layer not only enhances the expressiveness of the features, but also improves the sophistication and accuracy of the prediction model in analyzing and evaluating the probability of cardiovascular events in the early morning. For example, it can help the model identify specific physiological change patterns, such as the rapid rise in blood pressure upon awakening, which are key factors in determining the probability of cardiovascular events. Further feature processing through the fully connected layer significantly improves the model's predictive ability, enabling it to more accurately identify which patients face a higher probability of cardiovascular events during the early morning hours.

[0088] Step S514: Input the second feature into the second fully connected layer to obtain a second intermediate feature.

[0089] Specifically, in step S514, the task is to input the second feature obtained from the second encoder into the second fully connected layer to obtain the second intermediate feature. This step is intended to further process and refine the features of the night time period, and to improve the accuracy of predicting the probability of cardiovascular events at night by enhancing the model's ability to process these data.

[0090] In the specific operation, first, the second feature, which is the feature set of the night time processed by the second encoder, is input into the second fully connected layer. This layer consists of a series of neurons, each of which can learn different features and patterns in the data. In this process, choosing the right activation function (such as ReLU or Sigmoid) is the key, which helps to achieve nonlinear feature conversion and optimize the learning effect.

[0091] The second fully connected layer not only improves the expressiveness of features, but also enhances the precision and accuracy of the prediction model in processing the probability of cardiovascular events at night. For example, this layer can help the model better understand and identify patterns such as slowing heart rate or falling blood pressure that are unique to nighttime, which may be a sign of an impending cardiovascular event. Through this layer of efficient feature processing, the model's ability to predict the probability of cardiovascular events during the nighttime period is significantly improved.

[0092] Step S515: Input the first intermediate feature and the second feature into the first cross-attention module to obtain the first attention feature.

[0093] Specifically, in step S515, the task is to input the first intermediate feature and the second feature into the first cross-attention module to obtain the first attention feature. The execution of this step is crucial because it uses the cross-attention mechanism to enhance the model's ability to integrate and analyze features from different time periods (such as early morning and night), which is crucial to improving the overall accuracy and detail of the prediction of the probability of cardiovascular events.

[0094] In the specific operation, first, the first intermediate feature (output from the processed data in the early morning period) and the second feature (output directly from the night encoder) are input into the first cross attention module. The first cross attention module automatically learns which feature combinations are most informative for the prediction model by calculating the relationship weights between features. In this process, attention mechanisms such as those in Transformer may be used, which adjust and optimize the flow and processing of information by calculating the relationship between Query, Key, and Value.

[0095] The first cross-attention module can more accurately map the complex physiological patterns that affect the probability of cardiovascular events by fine-tuning the interactions between features. For example, this module can identify the impact of a rapid rise in heart rate in the morning on nighttime blood pressure fluctuations, or vice versa, thereby providing key physiological clues for early warning of cardiovascular events. In addition, through this fine-grained feature integration and analysis, the model's predictive ability is significantly improved.

[0096] Step S516: Input the second intermediate feature and the first feature into the second cross-attention module to obtain the second attention feature.

[0097] Specifically, in step S516, the task is to input the second intermediate feature and the first feature into the second cross-attention module to obtain the second attention feature. This step is performed to enhance the model's ability to analyze and integrate features from different time periods using the cross-attention mechanism, especially combining nighttime data with morning data, thereby improving the comprehensiveness and accuracy of cardiovascular event probability prediction.

[0098] In the specific operation, first, the second intermediate feature (output from the night-time data processing) and the first feature (output from the early morning encoder) are input into the second cross-attention module. Similarly, the second cross-attention module automatically learns which feature combinations are most informative for the prediction model by calculating the relationship weights between features. In this process, attention mechanisms such as those in Transformer may be used, which adjust and optimize the flow and processing of information by calculating the relationship between Query, Key, and Value.

[0099] The second cross-attention module enables the model to explore and utilize the potential connection between night and morning data in more detail, improving the understanding of the relationship between cardiovascular probability factors such as sleep quality and morning blood pressure peak. This in-depth feature integration significantly improves the prediction accuracy of the model.

[0100] Step S517: concatenate the first feature and the first attention feature to obtain a first concatenated feature.

[0101] Specifically, in step S517, the task is to concatenate the first feature and the first attention feature to obtain the first concatenated feature. This step is implemented to comprehensively utilize the finely processed morning time period data and the related features enhanced by the cross-attention module, thereby improving the prediction accuracy of the probability of cardiovascular events in the morning.

[0102] In the prediction of cardiovascular events, combining features from different processing stages is a key strategy to improve the accuracy of the prediction model. Concatenating the first feature and the first attention feature can fuse the information directly extracted from the original data and the information emphasized by the cross-attention mechanism. This method can more comprehensively reflect the cardiovascular health status in the early morning hours, especially those subtle physiological changes that may indicate the probability of occurrence.

[0103] In the specific operation, first, the first feature (generated directly by the first encoder, reflecting the processing results based on the original data) and the first attention feature (generated by the first cross-attention module, highlighting the relevance and importance of features in other time periods) are input into a concatenation operation in parallel. This operation simply combines the two sets of feature vectors into a longer vector, the first concatenated feature, element by element, thereby ensuring that the information extracted from the two different processing flows can be fully utilized in subsequent model training and prediction.

[0104] By splicing the first feature and the first attention feature, the first spliced ​​feature provides a richer and more detailed data basis for the model, enhancing the model's ability to capture the probability factors of cardiovascular disease in the early morning. This feature fusion strategy enables the prediction model to more accurately identify the high probability of occurrence in the early morning, such as abnormal blood pressure rise or heart rate increase, thereby providing more effective warnings.

[0105] Step S518: concatenate the second feature and the second attention feature to obtain a second concatenated feature.

[0106] Specifically, in step S518, the task is to splice the second feature and the second attention feature to generate a second spliced ​​feature. This step is implemented to effectively integrate the direct features of the night time period and the correlation features strengthened by cross-attention, so as to improve the accuracy and comprehensiveness of the model's prediction of the probability of cardiovascular events at night.

[0107] In the prediction of cardiovascular events, combining features from different processing stages is a key strategy to improve the accuracy of prediction models. Concatenating the second feature and the second attention feature can fuse the information directly extracted from the original data and the information emphasized by the cross-attention mechanism. This method can more comprehensively reflect the cardiovascular health status during the night, especially those subtle physiological changes that may indicate the probability of occurrence.

[0108] In the specific operation, similar to step S517, first, the second feature (generated by the second encoder, reflecting the basic physiological data processing results of the night period) and the second attention feature (generated by the second cross-attention module, emphasizing the important connection and difference between the night and morning features) are input into a concatenation operation in parallel. This operation simply combines the two sets of feature vectors into a longer vector, the second concatenated feature, by element, thereby ensuring that the information extracted from the two different processing flows can be fully utilized in subsequent model training and prediction.

[0109] By splicing the second feature and the second attention feature, the second spliced ​​feature provides a richer and more detailed data basis for the model, enhancing the model's ability to capture the probability factors of cardiovascular disease during the night. This feature fusion strategy enables the prediction model to more accurately identify the high probability of occurrence at night, such as abnormal heart rate or sudden changes in blood pressure during sleep, thereby providing more effective warnings.

[0110] In one possible implementation, refer to Figure 3 , which shows a second flow chart of a method for predicting the probability of occurrence of cardiovascular events based on dynamic blood pressure monitoring provided in an embodiment of the present application. Step S4 specifically includes steps S41-S42: Step S41: input the latent features of the first half of the day into a preset morning encoder to obtain the awakening time of the target person, and input the latent features of the second half of the day into a preset night encoder to obtain the sleeping time of the target person.

[0111] Specifically, studies have shown that the occurrence of cardiovascular events such as myocardial infarction often peaks within a few hours after waking up in the morning, and the probability of occurrence also increases within a few hours after falling asleep. By accurately determining the moment of waking up and falling asleep, we can more effectively focus on these key periods and conduct more in-depth physiological data analysis, thereby effectively predicting and managing the probability of occurrence.

[0112] In step S41, the first half of the day latent features are input into a preset morning encoder to determine the target person's awakening time, and the second half of the day latent features are input into a preset night encoder to determine the target person's sleeping time. The purpose of this step is to accurately identify the key transition points in the individual's biorhythm, that is, the transition from the sleeping state to the awakening state and the transition from the awakening state to the sleeping state, which are closely related to the change in the probability of cardiovascular health.

[0113] In specific operations, we first need to use specially designed encoders to process the latent features of the first half of the day and the second half of the day. These encoders may be time series models based on deep learning, such as recurrent neural networks (RNNs) or long short-term memory networks (LSTMs), which can learn time-dependent patterns and features from continuous time series data. The morning encoder is specially adjusted to recognize physiological activity patterns in the morning period, such as changes in blood pressure and heart rate, while the night encoder focuses on changes in physiological activities in the evening.

[0114] Through the analysis of these encoders, the moment of awakening and the moment of falling asleep can be accurately marked, which is usually done by identifying significant points of changes in physiological parameters, such as increased blood pressure or increased heart rate. After determining these moments, the physiological data before and after these moments can be further analyzed to assess the probability of cardiovascular events.

[0115] Step S42: according to the awakening time, the early morning period latent features are extracted from the latent features of the first half of the day, and according to the sleeping time, the latent features of the night period are extracted from the latent features of the second half of the day.

[0116] Specifically, accurately identifying and utilizing physiological data related to wakefulness and sleep can significantly improve the accuracy of cardiovascular disease prediction. Cardiovascular events, especially myocardial infarction and sudden cardiac death, are often significantly correlated with daily physiological rhythms, especially the physiological changes when the human body transitions from sleep to wakefulness.

[0117] In step S42, the latent features of the first half of the day and the second half of the day are deeply mined through detailed analysis and feature extraction, and the information of the awakening time and the sleeping time is accurately located and utilized, and then the corresponding early morning time latent features and evening time latent features are extracted from the latent feature matrix according to these key moments. This step is a crucial part of the entire early warning system for individualized probability assessment, because it allows the model to more specifically analyze the time period most relevant to the probability of cardiovascular events.

[0118] In one possible implementation, refer to Figure 4 , which shows a third flow chart of a method for predicting the probability of occurrence of cardiovascular events based on dynamic blood pressure monitoring provided in an embodiment of the present application. Step S42 specifically includes steps S421-S426: Step S421: Based on the awakening moment, determine a first time window corresponding to the awakening moment, wherein the awakening moment is the center moment of the first time window.

[0119] Specifically, the human body undergoes a series of significant physiological changes during the awakening moment, especially during the transition from sleep to wakefulness, such as increased heart rate and blood pressure. These changes are closely related to the increased probability of cardiovascular events such as myocardial infarction.

[0120] In the specific operation, the awakening time determined by the morning encoder is first used as the center point. For example, if the encoder analyzes and identifies the patient's awakening time as 6 am, the first time window will be set from 4 am to 8 am. The selection of this time window is based on the significant changes in physiological activities before and after awakening, which are key factors in assessing the probability of cardiovascular events.

[0121] The effect of step S421 is that it provides a highly targeted analysis window, allowing the model to focus on key physiological data before and after the awakening moment. By conducting in-depth analysis of the data during this time period, physiological markers associated with a high probability of cardiovascular disease, such as a sharp change in blood pressure or an irregular pattern of heart rate, can be accurately identified.

[0122] Step S422: extracting the intermediate features of the first half of the day corresponding to the first time window from the latent features of the first half of the day.

[0123] Specifically, in step S422, the task is to extract features corresponding to the first time window (the awakening moment and the two hours before and after) from the latent features of the first half of the day, and these features are defined as the latent features of the early morning period. The implementation of this step is crucial for in-depth analysis and utilization of physiological data related to awakening, because these data usually cover the physiological changes that are most relevant to the increased probability of cardiovascular events.

[0124] In the specific operation, the awakening time is first determined, for example, 6 am, and the first time window is determined from 4 am to 8 am. Then, the data in this time window is accurately selected from the latent feature data of the first half of the day processed by the dynamic blood pressure encoder. This includes cutting out the part corresponding to 4 am to 8 am from the latent feature matrix to ensure that this part of the data contains all the physiological changes related to awakening.

[0125] Step S423: Use the intermediate features of the first half of the day as latent features of the early morning period.

[0126] Specifically, in step S423, the first half-day intermediate features extracted from step S422 are set as the early morning period latent features. This step is implemented to transform the specific data analysis into a specific feature set that can be used for further prediction of the probability of occurrence, which is a key link in achieving accurate probability assessment in the cardiovascular event early warning system.

[0127] In the specific operation, we first ensure that the data segments before and after awakening have been correctly extracted from the latent features of the first half of the day, which usually involves identifying and segmenting the period from 4 am to 8 am in the time series data. These data are marked and stored as latent features of the early morning period, which contain the dynamic changes of physiological parameters such as heart rate and blood pressure during the awakening process. Such processing not only ensures the specificity and relevance of the data, but also provides the model with highly concentrated occurrence probability prediction factors.

[0128] Step S424: Based on the sleep onset time, determine a second time window corresponding to the sleep onset time, wherein the sleep onset time is the center time of the second time window.

[0129] Specifically, the study shows that the human body also undergoes a series of important physiological changes during the nighttime sleep phase, which may be closely related to cardiovascular health. For example, a decrease in heart rate and blood pressure usually occurs during deep sleep, but any abnormal pattern of changes, such as tachycardia or blood pressure fluctuations at night, may indicate an increased probability of cardiovascular events. By accurately defining and monitoring data from this critical time window, early warning signs of nocturnal cardiovascular events can be better identified.

[0130] In step S424, based on the sleep onset time determined from the night encoder, a second time window corresponding to the sleep onset time is set, and usually this time window is centered on the sleep onset time and includes data for two hours before and after sleep onset. This step is to accurately capture physiological data related to sleep onset, which is crucial in the assessment of the probability of cardiovascular events, especially in assessing the potential probability of nocturnal cardiovascular events.

[0131] In specific operation, once the night encoder identifies the patient's sleep time, for example 10 pm, the second time window is set to 8 pm to 12 midnight. Within this time window, the relevant physiological data segments are accurately extracted from the latent features of the second half of the day. This involves screening the time series data to ensure that the selected data segments can accurately cover the physiological activities before and after falling asleep.

[0132] The effect of step S424 is that by focusing on analyzing key physiological changes before and after falling asleep, the probability of cardiovascular events at night can be effectively predicted and managed. This method allows the model to gain in-depth understanding of changes in cardiovascular status at night, providing a scientific basis for medical providers to implement targeted interventions, especially for patients who may have a high probability of cardiovascular events at night.

[0133] Step S425: extracting the intermediate features of the second half of the day corresponding to the second time window from the latent features of the second half of the day.

[0134] Specifically, in step S425, the task is to extract features corresponding to the second time window (the moment of falling asleep and two hours before and after) from the latent features of the second half of the day, and these features are defined as the latent features of the night period. This step is performed to ensure that the physiological changes associated with falling asleep can be accurately captured and analyzed, which are crucial for the assessment of the probability of cardiovascular events at night.

[0135] In specific operations, once the patient's sleep time is determined through the night encoder, for example, 10 pm, the second time window is set to 8 pm to 12 midnight. Within this time window, the relevant physiological data segments are accurately selected and extracted from the latent features of the second half of the day. This involves filtering out data from a specific time range from the time series to ensure that the selected data segments can accurately cover the physiological activities before and after falling asleep, especially focusing on key indicators such as heart rate drop and blood pressure changes.

[0136] Step S426: Use the middle features of the second half of the day as latent features of the night period.

[0137] In step S426, the task is to set the second half of the day intermediate features extracted from step S425 as the night time hidden features. This step is to use a specific set of physiological data to further analyze and predict the probability of cardiovascular events at night. The determination and application of the night time hidden features are crucial because they focus on key physiological indicators that may affect cardiovascular health at night.

[0138] In the specific operation, step S425 has ensured that the data segment from 8 pm to 12 midnight is accurately extracted from the latent features of the second half of the day. This data segment is now formally defined as the latent feature of the night time period, which includes the change data of heart rate, blood pressure, etc. before and after falling asleep. This process ensures the concentration and relevance of the data, and provides a scientific and accurate data basis for the evaluation of the probability of cardiovascular disease at night.

[0139] In a possible implementation, step S1 specifically includes the following steps: Based on the preset assessment criteria, the first disease label and the second disease label corresponding to the target person are determined.

[0140] Specifically, the accurate definition of disease labels is fundamental and core to any medical prediction model, as these labels are directly related to the accuracy of model training and the reliability of the final prediction. By clarifying disease labels, accurate predictions for specific cardiovascular events can be ensured during model development and subsequent analysis.

[0141] In specific operations, it is necessary to first set the assessment criteria for cardiovascular events by collaborating with medical experts or referring to the latest clinical guidelines. These criteria are usually based on a series of clinical parameters, such as electrocardiogram changes, blood pressure levels, cholesterol levels, past medical history, etc. For each participating patient, a comprehensive health examination and data collection are carried out according to these preset assessment criteria.

[0142] Next, one or more disease labels are assigned to each patient based on the collected data, such as electrocardiogram results, blood test results, etc. For example, if a patient's electrocardiogram shows typical signs of myocardial infarction, he / she will be labeled as a high-probability individual for myocardial infarction (the first disease label). Similarly, if the patient has a history of sudden cardiac death or has factors that lead to the probability of the disease, they will also be labeled accordingly.

[0143] The effect of this step is to provide accurate target variables, namely disease labels, for subsequent data processing and model training, which helps the model learn and predict the probability of cardiovascular events more accurately. The disease labels determined in this way make the entire early warning system more targeted.

[0144] A basic data set is constructed based on dynamic blood pressure sequence features, personnel information features, first disease labels, and second disease labels.

[0145] Specifically, the basic data set is constructed by combining dynamic blood pressure sequence features, personnel information features, and disease labels. This data integration provides comprehensive input for the dynamic blood pressure encoder, which helps to accurately extract latent features in subsequent steps. The basic data set is established to ensure the integrity and multidimensionality of the input data, so that the occurrence probability prediction model can capture the relationship between complex physiological and personal information.

[0146] This integrated approach ensures that the interconnectedness and complexity of the data are fully utilized, providing enough information during model training to learn how to distinguish different health conditions and predict future cardiovascular events.

[0147] In the specific operation, the collected dynamic blood pressure sequence features and personnel information features of each patient are first integrated into a data framework. Dynamic blood pressure sequence features may include but are not limited to blood pressure and heart rate change data within 24 hours, while personnel information features include basic and medical background information such as age, gender, weight, and medical history. After these data are integrated together, the first disease label and the second disease label are added to each patient. Next, the necessary data preprocessing work is performed on the integrated data set, such as filling missing values, standardization, and regularization to ensure the consistency and applicability of the data. The construction of this basic data set involves not only the collection and integration of data, but also the cleaning and formatting of the data to make it suitable for advanced analysis and machine learning processing.

[0148] The effect of this step is to provide a rich, structured and high-quality input data source for the cardiovascular disease prediction model. In this way, the prediction model can more effectively identify the key factors that affect cardiovascular events. In addition, this comprehensive data processing method enhances the representativeness of the data and the generalization ability of the prediction model, so that the model can not only be applied to the current research group, but also provide reliable predictions in a wider population.

[0149] Fill the basic data set with empty values ​​to obtain the feature data set.

[0150] Specifically, during the data collection process, missing data is often encountered. Missing data may be caused by various reasons, such as measurement errors, missing data entry, or sensor failure. Unhandled null values ​​will affect the quality and accuracy of data analysis, and may cause the model to learn wrong information, thus affecting the final prediction results. Therefore, properly handling these null values ​​is crucial to ensure the effectiveness of data analysis and subsequent model training.

[0151] In the specific operation, first evaluate the null values ​​in the basic data set, identify which features have null values ​​and the distribution of null values. Choose the appropriate filling method according to different types of features: For null values ​​in dynamic blood pressure series features, consider using model fitting interpolation methods, such as time series forecasting, interpolation techniques, or filling based on the average value of adjacent time points, which helps to maintain the continuity and consistency of the data in the time series.

[0152] For null values ​​in personnel information features, such as age and weight, the median or mode can be used to fill them. This method is simple and effective, and can reduce the impact of outliers on the overall distribution of the data set.

[0153] After these filling processes, the completeness of the data set will be significantly improved, creating a solid foundation for subsequent data analysis and model training. Through such data preprocessing, it is ensured that the data of all patients are complete, so that the model can learn on the basis of complete data, improving the accuracy and robustness of the prediction model.

[0154] In a possible implementation, after step S5, the method further includes the following steps: When the probability of the cardiovascular event occurring is greater than the preset probability of the disease occurring, an early warning prompt message is output to the physician terminal.

[0155] Specifically, in this step, the task is to output warning information to the physician terminal when the probability of cardiovascular events in the target person is determined to be greater than the preset threshold. This step is implemented to ensure that any potential probability of cardiovascular events can be identified and handled by medical professionals in a timely manner, so that appropriate intervention measures can be taken to reduce the patient's health probability.

[0156] In specific operation, firstly, the probability of cardiovascular events in the morning and evening for each patient is evaluated based on the trained probability prediction model. These probabilities are obtained from the final output of the model, including the first and second cardiovascular event probabilities calculated by steps S51 to S53. Then, these probabilities are compared with the preset probability thresholds. If any probability exceeds the threshold, the system will automatically generate an early warning message and send the warning to the physician terminal through the medical information system, such as the hospital's electronic health record system or mobile medical application.

[0157] Reference Figure 8 , which shows a schematic diagram of the structure of a cardiovascular event occurrence probability prediction system based on dynamic blood pressure monitoring provided in an embodiment of the present application, the system includes: a data acquisition module 1, a feature acquisition module 2 and a feature fusion and occurrence probability output module 3; Data acquisition module 1, used to obtain dynamic blood pressure sequence characteristics and personnel information characteristics of the target person, and construct a feature data set based on the dynamic blood pressure sequence characteristics and personnel information characteristics; Feature acquisition module 2 is used to input the feature data set into the trained dynamic blood pressure encoder and output the latent feature matrix; divide the latent feature matrix into the first half of the day and the second half of the day according to the preset time; based on the first half of the day and the second half of the day, obtain the early morning period latent features and the evening period latent features corresponding to the target person; The feature fusion and output module 3 is used to input the latent features of the early morning period and the latent features of the latent features of the latent night period into the preset feature fusion and occurrence probability output module to obtain the corresponding probability of cardiovascular events of the target person in the early morning period and the latent features of the latent night period, wherein the preset feature fusion and output module 3 is constructed based on the cross-attention mechanism.

[0158] In a possible implementation, the system also includes: an occurrence probability prediction model construction module 4; the occurrence probability prediction model construction module 4 is used to obtain the historical dynamic blood pressure sequence characteristics and historical personnel information characteristics of the historical target personnel, and construct a training feature data set based on the historical dynamic blood pressure sequence characteristics and historical personnel information characteristics; based on the time series model or the non-time series model, a dynamic blood pressure encoder is constructed; the training feature data set is input into the occurrence probability prediction model framework for training to obtain an initial occurrence probability prediction model. The occurrence probability prediction model framework includes a dynamic blood pressure encoder, a preset morning encoder, a preset night encoder, a preset feature fusion and occurrence probability output module; the initial occurrence probability prediction model is evaluated by a cross-validation method to obtain the performance index of the initial occurrence probability prediction model; the performance index includes a cross-validation result; based on the cross-validation result, the hyperparameter tuning method is used to adjust the hyperparameters of the initial occurrence probability prediction model to obtain a preset occurrence probability prediction model.

[0159] In a possible implementation, the feature fusion and output module 3 is also used to input the latent features of the early morning period and the latent features of the evening period into a preset feature fusion and occurrence probability output module to obtain a first splicing feature and a second splicing feature; based on the first splicing feature, the probability of occurrence of a first cardiovascular event corresponding to the target person in the early morning period is obtained; based on the second splicing feature, the probability of occurrence of a second cardiovascular event corresponding to the target person in the evening period is obtained.

[0160] In a possible implementation, the feature fusion and output module 3 is also used to input the latent feature of the early morning period into the first encoder to obtain the first feature; input the latent feature of the evening period into the second encoder to obtain the second feature; input the first feature into the first fully connected layer to obtain the first intermediate feature; input the second feature into the second fully connected layer to obtain the second intermediate feature; input the first intermediate feature and the second feature into the first cross-attention module to obtain the first attention feature; input the second intermediate feature and the first feature into the second cross-attention module to obtain the second attention feature; splice the first feature and the first attention feature to obtain the first spliced ​​feature; splice the second feature and the second attention feature to obtain the second spliced ​​feature.

[0161] In a possible implementation, the feature acquisition module 2 is also used to input the latent features of the first half of the day into a preset morning encoder to obtain the awakening time of the target person, and input the latent features of the second half of the day into a preset night encoder to obtain the sleeping time of the target person; according to the awakening time, the early morning period latent features are extracted from the latent features of the first half of the day, and according to the sleeping time, the night period latent features are extracted from the latent features of the second half of the day.

[0162] In a possible implementation, the feature acquisition module 2 is further used to determine, based on the awakening moment, a first time window corresponding to the awakening moment, wherein the awakening moment is the center moment of the first time window; extract the first half-day middle feature corresponding to the first time window from the first half-day latent feature; use the first half-day middle feature as the early morning time period latent feature; determine, based on the sleeping moment, a second time window corresponding to the sleeping moment, wherein the sleeping moment is the center moment of the second time window; extract the second half-day middle feature corresponding to the second time window from the second half-day latent feature; use the second half-day middle feature as the night time period latent feature In a possible implementation, the data acquisition module 1 is also used to determine the first disease label and the second disease label corresponding to the target person based on a preset assessment standard; construct a basic data set based on dynamic blood pressure sequence characteristics, personnel information characteristics, the first disease label and the second disease label; and fill the basic data set with null values ​​to obtain a feature data set.

[0163] In a possible implementation, the system further includes: an early warning module 5; the early warning module 5 is used to output early warning information to the physician terminal when the probability of a cardiovascular event is greater than a preset probability of a disease.

[0164] It should be noted that: when the device provided in the above embodiment realizes its function, only the division of the above functional modules is used as an example. In actual application, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the device and method embodiments provided in the above embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0165] The present application also discloses an electronic device. Fig. 9 , Fig. 9 The electronic device 900 may include: at least one processor 901 , at least one network interface 904 , a user interface 903 , a memory 905 , and at least one communication bus 902 .

[0166] The communication bus 902 is used to realize the connection and communication between these components.

[0167] The user interface 903 may include a display screen (Display) and a camera (Camera). Optionally, the user interface 903 may also include a standard wired interface and a wireless interface.

[0168] The network interface 904 may optionally include a standard wired interface or a wireless interface (such as a WI-FI interface).

[0169] Among them, the processor 901 may include one or more processing cores. The processor 901 uses various interfaces and lines to connect various parts in the entire server, and executes various functions of the server and processes data by running or executing instructions, programs, code sets or instruction sets stored in the memory 905, and calling data stored in the memory 905. Optionally, the processor 901 can be implemented in at least one hardware form of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), and programmable logic array (Programmable Logic Array, PLA). The processor 901 can integrate one or a combination of a central processing unit (Central Processing Unit, CPU), a graphics processor (Graphics Processing Unit, GPU) and a modem. Among them, the CPU mainly processes the operating system, user interface and application programs; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 901, and it can be implemented separately through a chip.

[0170] Among them, the memory 905 may include a random access memory (Random Access Memory, RAM) and may also include a read-only memory (Read-Only Memory). Optionally, the memory 905 includes a non-transitory computer-readable storage medium. The memory 905 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 905 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned method embodiments, etc.; the data storage area may store data involved in the above-mentioned method embodiments, etc. The memory 905 may optionally be at least one storage device located away from the aforementioned processor 901. Refer to Fig. 9 , the memory 905 as a computer-readable storage medium may include an operating system, a network communication module, a user interface module and an application program.

[0171] exist Fig. 9In the electronic device 900 shown, the user interface 903 is mainly used to provide an input interface for the user and obtain the data input by the user; and the processor 901 can be used to call an application stored in the memory 905, and when executed by one or more processors 901, the electronic device 900 executes one or more methods in the above-mentioned embodiments. It should be noted that for the aforementioned method embodiments, for the sake of simple description, they are all expressed as a series of action combinations, but those skilled in the art should know that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required for the present application.

[0172] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0173] In the several embodiments provided in this application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0174] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0175] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0176] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a memory and includes several instructions for a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned memory includes: various media that can store program codes, such as USB flash drives, mobile hard drives, magnetic disks or optical disks.

[0177] The above are only exemplary embodiments of the present disclosure and cannot be used to limit the scope of the present disclosure. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the disclosure of the specification and the truth of practice, those skilled in the art will easily think of other embodiments of the present disclosure.

[0178] This application is intended to cover any variation, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the art not described in the present disclosure. The description and examples are to be regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. A method for predicting the probability of occurrence of cardiovascular events based on dynamic blood pressure monitoring, characterized in that: The method comprises: Acquire dynamic blood pressure sequence features and personnel information features of the target person, and construct a feature data set based on the dynamic blood pressure sequence features and personnel information features; Inputting the feature data set into a trained dynamic blood pressure encoder and outputting a latent feature matrix; According to a preset time, the latent feature matrix is ​​divided into the first half of the day latent features and the second half of the day latent features; Based on the first half-day hidden features and the second half-day hidden features, obtain the early morning time hidden features and the evening time hidden features corresponding to the target person; The latent features of the early morning period and the latent features of the evening period are input into a preset feature fusion and occurrence probability output module to obtain the probability of occurrence of cardiovascular events corresponding to the target person in the early morning period and the evening period, wherein the preset feature fusion and occurrence probability output module is constructed based on a cross-attention mechanism.

2. The method according to claim 1, characterized in that Before inputting the latent features of the early morning period and the latent features of the evening period into a preset feature fusion and occurrence probability output module to obtain the corresponding probability of occurrence of cardiovascular events of the target person in the early morning period and the evening period, the method further includes: Acquire historical dynamic blood pressure sequence features and historical personnel information features of historical target personnel, and construct a training feature data set based on the historical dynamic blood pressure sequence features and historical personnel information features; Based on the time series model or the non-time series model, a dynamic blood pressure encoder is constructed; Inputting the training feature data set into the occurrence probability prediction model framework for training to obtain an initial occurrence probability prediction model, wherein the occurrence probability prediction model framework includes the dynamic blood pressure encoder, a preset morning encoder, a preset night encoder, a preset feature fusion and occurrence probability output module; Evaluate the initial occurrence probability prediction model using a cross-validation method to obtain a performance indicator of the initial occurrence probability prediction model; the performance indicator includes a cross-validation result; Based on the cross-validation result, the hyperparameters of the initial occurrence probability prediction model are adjusted using a hyperparameter tuning method to obtain the preset occurrence probability prediction model.

3. The method according to claim 1, characterized in that The probability of occurrence of a cardiovascular event includes a first probability of occurrence of a cardiovascular event and a second probability of occurrence of a cardiovascular event; the step of inputting the latent features of the early morning period and the latent features of the evening period into a preset feature fusion and occurrence probability output module to obtain the corresponding probability of occurrence of a cardiovascular event of the target person in the early morning period and the evening period specifically includes: Inputting the early morning period hidden feature and the evening period hidden feature into the preset feature fusion and occurrence probability output module to obtain a first splicing feature and a second splicing feature; Based on the first splicing feature, obtaining the probability of occurrence of the first cardiovascular event corresponding to the target person in the early morning period; Based on the second splicing feature, the probability of occurrence of the second cardiovascular event corresponding to the target person during the night time period is obtained.

4. The method according to claim 3, characterized in that The preset feature fusion and occurrence probability output module includes a first encoder, a second encoder, a first fully connected layer, a second fully connected layer, a first cross attention module, and a second cross attention module; the early morning period latent feature and the evening period latent feature are input into the preset feature fusion and occurrence probability output module to obtain a first splicing feature and a second splicing feature, specifically including: Inputting the early morning period latent feature into the first encoder to obtain a first feature; Inputting the night-time latent feature into the second encoder to obtain a second feature; Inputting the first feature into the first fully connected layer to obtain a first intermediate feature; Inputting the second feature into the second fully connected layer to obtain a second intermediate feature; Inputting the first intermediate feature and the second feature into the first cross attention module to obtain a first attention feature; Inputting the second intermediate feature and the first feature into the second cross attention module to obtain a second attention feature; Concatenate the first feature and the first attention feature to obtain a first concatenated feature; The second feature and the second attention feature are concatenated to obtain a second concatenated feature.

5. The method according to claim 1, characterized in that: The step of obtaining the early morning hidden features and the evening hidden features corresponding to the target person based on the first half of the day hidden features and the second half of the day hidden features specifically includes: Input the first half of the day latent features into a preset morning encoder to obtain the awakening time of the target person, and input the second half of the day latent features into a preset night encoder to obtain the sleeping time of the target person; According to the awakening time, the early morning period latent features are extracted from the first half of the day latent features, and according to the sleeping time, the evening period latent features are extracted from the second half of the day latent features.

6. The method according to claim 5, characterized in that The step of extracting the early morning period latent features from the first half of the day latent features according to the awakening time, and extracting the evening period latent features from the second half of the day latent features according to the sleeping time, specifically includes: Based on the awakening moment, determining a first time window corresponding to the awakening moment, wherein the awakening moment is a center moment of the first time window; Extracting the intermediate features of the first half of the day corresponding to the first time window from the latent features of the first half of the day; Using the intermediate features of the first half of the day as the latent features of the early morning period; Based on the sleeping time, determining a second time window corresponding to the sleeping time, wherein the sleeping time is the center time of the second time window; Extracting the second half-day intermediate features corresponding to the second time window from the second half-day latent features; The intermediate features of the second half of the day are used as the latent features of the night time period.

7. The method according to claim 1, characterized in that The dynamic blood pressure encoder is any one of a time series model and a non-time series model.

8. The method according to claim 1, characterized in that The step of constructing a feature data set according to the dynamic blood pressure sequence features and the personnel information features specifically includes: Based on a preset assessment standard, determining a first disease label and a second disease label corresponding to the target person; Constructing a basic data set based on the dynamic blood pressure sequence feature, the personnel information feature, the first disease label, and the second disease label; The basic data set is filled with null values ​​to obtain the feature data set.

9. The method according to claim 1, characterized in that: After obtaining the corresponding probability of occurrence of cardiovascular events of the target person in the early morning and evening periods, the method further includes: When the probability of the cardiovascular event occurring is greater than the preset probability of the disease occurring, an early warning prompt message is output to the physician terminal.

10. An auxiliary warning system for cardiovascular event warning, characterized in that: The system includes: a data acquisition module, a feature acquisition module and a feature fusion and output module; The data acquisition module is used to obtain dynamic blood pressure sequence characteristics and personnel information characteristics of the target person, and construct a feature data set based on the dynamic blood pressure sequence characteristics and personnel information characteristics; The feature acquisition module is used to input the feature data set into the trained dynamic blood pressure encoder and output a latent feature matrix; divide the latent feature matrix into the first half of the day and the second half of the day according to the preset time; based on the first half of the day and the second half of the day, obtain the early morning time period latent features and the evening time period latent features corresponding to the target person; The feature fusion and output module is used to input the latent features of the early morning period and the latent features of the evening period into a preset feature fusion and occurrence probability output module to obtain the corresponding probability of cardiovascular events of the target person in the early morning period and the evening period, wherein the preset feature fusion and occurrence probability output module is constructed based on a cross-attention mechanism.

11. The system according to claim 10, characterized in that The system also includes: an early warning module; The early warning module is used to output early warning information to the physician terminal when the probability of the cardiovascular event is greater than the preset probability of the disease.

12. An electronic device, characterized in that: The electronic device (900) comprises a processor (901), a memory (905), a user interface (903) and a network interface (904), wherein the memory (905) is used to store instructions, the user interface (903) and the network interface (904) are used to communicate with other devices, and the processor (901) is used to execute the instructions stored in the memory (905) so that the electronic device (900) executes the method according to any one of claims 1 to 9.

13. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed, the method according to any one of claims 1 to 9 is executed.

Images