Coronary heart disease prediction method fusing cardiopulmonary motion parameter time sequence information
By integrating static parameters and time-series information from cardiopulmonary exercise testing and using deep learning models to analyze patient data, the accuracy and early prediction problems of existing coronary heart disease diagnostic methods have been solved, achieving more efficient coronary heart disease prediction and early warning.
Patent Information
- Application Number
- CN202411014378.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-26
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-07-26
AI Technical Summary
Existing methods for diagnosing coronary heart disease rely on high-radiation and high-cost imaging technologies, lack precise early prediction, and there are homogeneous differences in the interpretation of cardiopulmonary exercise test results in primary hospitals.
By integrating static parameters and temporal information from cardiopulmonary exercise testing, and using deep learning models such as Long Short-Term Memory (LSTM) networks to analyze patients' multidimensional health data, a coronary heart disease prediction model is established. Combined with real-time data monitoring and parameter changes, multidimensional and accurate coronary heart disease prediction is achieved.
It improves the accuracy and timeliness of early prediction of coronary heart disease, better captures dynamic changes in patients, reduces error accumulation, improves prediction accuracy, and supports early treatment.
Smart Images

Figure CN118969244A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of biotechnology, and in particular to a method for predicting coronary heart disease that integrates cardiopulmonary exercise parameter time series information. Background Art
[0002] Coronary Heart Disease (CHD) is a heart disease caused by myocardial ischemia, hypoxia or necrosis due to coronary atherosclerosis, and is one of the major diseases that threaten human health. Traditional methods for diagnosing CHD rely on coronary artery computed tomography or coronary angiography, which cause certain radiation or trauma to the human body and have high medical costs.
[0003] Existing prediction or diagnosis models mostly focus on a wide range of cardiovascular diseases, lacking accurate early prediction of coronary heart disease. At the same time, previous prediction methods mostly focused on the selection and application of traditional clinical indicators and parameters, lacking the comprehensiveness and accuracy of the disease characteristics of patients with coronary heart disease (CN118098575A, CN115831374A, CN115040156A).
[0004] Cardiopulmonary exercise test (CPET) is an objective, quantitative, non-invasive examination method that has been gradually applied to the early diagnosis and prognostic risk stratification of coronary heart disease. CPET examination results contain rich time series information, and diagnosis is often assisted by manual interpretation of its changes. However, the learning cost of interpreting CPET results is high, and the interpretation level of grassroots hospitals varies, especially the understanding of image changes is poorly homogeneous among medical institutions. Therefore, the applicant intends to establish a cardiopulmonary exercise prediction model for patients with coronary heart disease based on static parameters of cardiopulmonary exercise and the integration of time series information, using deep learning and time series information analysis technology of artificial intelligence. Summary of the invention
[0005] In order to overcome the shortcomings of the prior art, the present invention provides a coronary heart disease prediction model that integrates static parameters and time series information. By analyzing the patient's continuous health data, it is intended to combine traditional clinical indicators with cardiopulmonary exercise test parameters, especially time series information parameters, and use the fusion of cardiopulmonary exercise static parameters and time series information to predict coronary heart disease. Based on the characteristics of real-time data monitoring and parameter changes during exercise, the occurrence of coronary heart disease can be predicted more multi-dimensionally and accurately, which can better capture the dynamic changes of patients during exercise, improve the prediction accuracy, help to take treatment measures in time, and ensure the physical and mental health of patients. Improve the accuracy and timeliness of early prediction of coronary heart disease.
[0006] The present invention is achieved by at least one of the following technical solutions.
[0007] A method for predicting coronary heart disease by integrating cardiopulmonary exercise parameter time series information comprises the following steps:
[0008] S1. Collect multi-dimensional health data of patients;
[0009] S2. Preprocessing the multi-dimensional health data;
[0010] S3, extract features from the preprocessed data;
[0011] S4, using deep learning model to perform time series analysis on the extracted features;
[0012] S5. Input the time series analysis results and static features into the classification model to predict coronary heart disease, output the prediction results and generate relevant health reports.
[0013] Furthermore, in step S1, the multi-dimensional health data includes demographic information, basic medical information, static parameters and timing information of the target subject, and the static parameters and timing information of the cardiopulmonary exercise test include ventilation function, gas exchange, exercise tolerance and circulatory function.
[0014] Furthermore, the preprocessing includes data cleaning, default value processing, and normalization operations.
[0015] Furthermore, the extracted features include time domain features, frequency domain features and other features that may be related to coronary heart disease.
[0016] Furthermore, the deep learning model is a long short-term memory network LSTM combined with patient time series information to predict diseases.
[0017] Furthermore, the long short-term memory network LSTM includes an input gate, a forget gate and an output gate, and the calculation formula of the forget gate is as follows:
[0018] f t =σ(W f .[h t-1 ,x t ]+b f )
[0019] Among them, f t Represents the weight of forgetting the cell state, W f is the weight matrix of the forget gate, b f is the bias term, x t is the input of the current time step, h t-1 is the hidden state of the previous time step, σ is the sigmoid function, which maps the input value to a probability value between 0 and 1;
[0020] The calculation formula of the input gate is as follows:
[0021] i t =σ(W i .[h t-1 ,x t ]+b i )
[0022] where i t Represents the weight for updating the cell state, W i is the weight matrix of the input gate, b i is the bias term, x t is the input of the current time step, h t-1 is the hidden state of the previous time step, σ is the sigmoid function, which maps the input value to a probability value between 0 and 1;
[0023] LSTM calculates candidate cell states, which represent the impact of the new input at the current time step on the cell state. The calculation formula is as follows
[0024]
[0025] Among them, W c is the weight matrix of the candidate cell state, b c is the bias term, x t is the input of the current time step, h t-1 is the hidden state of the previous time step, and tanh is the hyperbolic tangent function, which maps the input value to a value between -1 and 1;
[0026] The cell state of LSTM will be updated and passed to the next time step. At each time step t, the update formula of the cell state is as follows:
[0027]
[0028] in, is the candidate cell state of the current time step, indicating the impact of the new input of the current time step on the cell state; C t is the candidate cell state for the next time step;
[0029] The output gate determines the output information o through the sigmoid function t , at each time step t, the output gate is calculated as follows:
[0030] o t =σ(W o .[h t-1 ,x t ]+b o )
[0031] Among them, W ois the weight matrix of the output gate, b o is the bias term, x t is the input of the current time step, h t-1 is the hidden state of the previous time step, σ is the sigmoid function, which maps the input value to a probability value between 0 and 1;
[0032] LSTM will transform the cell state C t Through the tanh function, the hidden state h of the current time step is obtained t :
[0033] h t =o t *tanh(C t )
[0034] Here, tanh is the hyperbolic tangent function, which maps the input value to a value between -1 and 1.
[0035] Furthermore, the long short-term memory network LSTM adopts step-by-step prediction in the time series prediction task, and the specific steps are as follows:
[0036] 1) Initial input:
[0037] Input the data of the historical time step into the long short-term memory network LSTM to obtain the initial hidden state and cell state, ensuring that the data is arranged in chronological order;
[0038] The initial input data is usually a fixed-length time window containing historical data of multiple time steps;
[0039] 2) Step-by-step prediction:
[0040] For each time step, the data of the current time step and the hidden state of the previous time step are input into the long short-term memory network LSTM, and the output result of the current time step and the new hidden state and cell state are obtained through the long short-term memory network LSTM, and the long short-term memory network LSTM outputs the patient's time series information;
[0041] 3) Iteration process:
[0042] Use the output result of the current time step as the input of the next time step, and repeat step 2) until the specified number of times is reached;
[0043] 4) Prediction strategy:
[0044] True value input: After each prediction, the prediction of whether it is coronary heart disease is compared with the actual case data to evaluate the prediction error; choose to use the true value as the input of the next time step to reduce error accumulation;
[0045] Prediction value input: Use the long short-term memory network LSTM to predict whether it is coronary heart disease;
[0046] Use the cell state updated in this time step and the time series features corresponding to the next time step as the input of the next time step. This method is suitable for situations where long-time step prediction is required, but it may cause errors to accumulate gradually.
[0047] 5) Error assessment and adjustment:
[0048] After each prediction, the prediction result is compared with the true value, the prediction error is evaluated, and according to the error size, the LSTM input of the long short-term memory network is adjusted or the LSTM model of the long short-term memory network is retrained.
[0049] Furthermore, the classification model includes multiple layers of fully connected layers, each neuron in the fully connected layer is connected to all neurons in the previous layer, and each connection has a weight for adjusting the intensity of information transmission.
[0050] Furthermore, the fully connected layer consists of two parts: parameter initialization and forward propagation;
[0051] Parameter initialization: When initializing the weight matrix, use random initialization, and when initializing the bias vector, use the zero vector for initialization;
[0052] Forward propagation: The fully connected layer establishes a fully connected relationship between all neurons in the previous layer and all neurons in the current layer, and transmits and transforms information through the weight matrix.
[0053] Furthermore, binary cross entropy is used as the loss function for back propagation to update the parameters of the long short-term memory network LSTM.
[0054] A system for implementing the coronary heart disease prediction method integrating cardiopulmonary exercise parameter time series information comprises:
[0055] Data collection module: collects multi-dimensional health data of patients, including demographic information of target subjects, basic medical information, static parameters and time series information of cardiopulmonary exercise test, including ventilation function, gas exchange, exercise tolerance and circulatory function, etc.
[0056] Data preprocessing module: preprocess the collected multi-dimensional health data, including data cleaning, default value processing, normalization and other operations.
[0057] Feature extraction module: extracts features from the preprocessed time series data, including time domain features, frequency domain features and other features that may be related to coronary heart disease.
[0058] Time series analysis module: Use LSTM to perform time series analysis on the extracted features to mine the time series information in the data.
[0059] Disease prediction module: The time series analysis results and static features are input into the fully connected layer of the classification module to make prediction decisions on coronary heart disease.
[0060] Result output module: outputs the prediction results of coronary heart disease and generates relevant health reports.
[0061] Compared with the prior art, the present invention has the following beneficial effects:
[0062] The present invention obtains the demographic data, basic medical information and cardiopulmonary exercise data of the target object, wherein the demographic data includes age, gender and body mass index, and the basic medical information includes whether hypertension and diabetes are combined. By fusing static parameters and time series information to predict coronary heart disease, the present invention can better capture the dynamic changes in the patient's health data and improve the prediction accuracy. At the same time, the present invention can achieve early warning, help to take timely treatment measures, and ensure the physical and mental health of the patient. BRIEF DESCRIPTION OF THE DRAWINGS
[0063] Figure 1 is a flow chart of coronary heart disease prediction incorporating time series information in an embodiment;
[0064] Figure 2 is a schematic diagram of the structure of a timing analysis module in an embodiment;
[0065] Figure 3 It is a structural diagram of a classification model of an embodiment. DETAILED DESCRIPTION
[0066] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.
[0067] like Figure 1 to Figure 3 As shown, a method for predicting coronary heart disease by integrating cardiopulmonary exercise parameter time series information in this embodiment includes the following steps:
[0068] S1. Use the data collection module to collect multi-dimensional health data of patients;
[0069] As an embodiment, the multi-dimensional health data includes the target subject's demographic information, basic medical information, static parameters and timing information of a cardiopulmonary exercise test, wherein the static parameters and timing information of the cardiopulmonary exercise test include ventilation function, gas exchange, exercise tolerance and circulatory function.
[0070] 1) The demographic data and basic medical information of the target subjects are collected through the electronic medical record system, and the patient's personal information is kept confidential.
[0071] 2) Cardiopulmonary exercise test (CPET) parameter acquisition:
[0072] ① Experimental equipment: Use cardiopulmonary exercise test equipment, including power bike, respiratory gas analyzer, electrocardiograph, electronic blood pressure monitor, pulse oximeter, etc.
[0073] ② Experimental process: The subjects performed cycling exercise with gradually increasing intensity within a certain period of time, during which various physiological index data were collected in real time. The data included but were not limited to exercise capacity: respiratory exchange rate, oxygen uptake, exercise load time, maximum load; cardiovascular function: exercise electrocardiogram changes, heart rate, blood pressure, oxygen pulse; ventilation and perfusion: respiratory reserve, maximum ventilation, carbon dioxide emission, etc.
[0074] S2. The data preprocessing module preprocesses the health data; the preprocessing includes operations such as data cleaning, default value processing, and normalization.
[0075] 1) Time series structure of data:
[0076] ① Definition of time step: During the cardiopulmonary exercise test, data collection is performed at fixed time intervals. For CPET data, data is collected every 10 seconds, so each time step corresponds to one second of data.
[0077] ②Data preprocessing: Preprocess the collected data, including data cleaning, noise filtering, missing value filling, etc., to ensure data quality.
[0078] 2) Construction of historical time step data:
[0079] ① Time step length: Select an appropriate time step length based on actual conditions. Collect data for the first 15 minutes at a rate of 10 seconds, for a total of 90 time steps.
[0080] ②Data arrangement: Arrange the data of each time step in chronological order to form time series data.
[0081] S3. The feature extraction module extracts features from the preprocessed data, and extracts features that can reflect the patient's cardiovascular condition, such as oxygen pulse, oxygen uptake / power; the extracted features include time domain features and other features that may be related to coronary heart disease.
[0082] 1) Data preprocessing:
[0083] The collected cardiopulmonary exercise test (CPET) data were cleaned to remove noise and outliers to ensure data quality. The objects with more than 20% missing variables were removed. Missing values were filled using linear interpolation and mean filling methods.
[0084] 2) Time domain feature extraction: Calculate the data of each time step and extract the time domain features. Including but not limited to:
[0085] Oxygen pulse: Calculated as the ratio of oxygen uptake (VO2) and heart rate (HR).
[0086] Oxygen uptake / power ratio (VO2 / Watt): calculated by the ratio of oxygen uptake (VO2) and exercise power.
[0087] Heart Rate (HR): Direct reading of the number of heart beats per minute through electrocardiogram data.
[0088] S4, the timing analysis module uses the deep learning model to perform timing analysis on the extracted features;
[0089] As an embodiment, the deep learning model is a long short-term memory network (Long Short-Term Memory, LSTM), a special recurrent neural network (RNN) structure, which is usually used to process time series and extract time series information. Therefore, LSTM is a neural network model in the field of deep learning.
[0090] In deep learning, LSTM is widely used in tasks that require processing long-term dependencies, such as speech recognition, natural language processing, time series prediction, etc. Compared with the traditional RNN structure, LSTM can better capture and utilize long-term dependencies in time series data by introducing a gating unit mechanism, thereby improving the performance and generalization ability of the model.
[0091] LSTM sets two key variables:
[0092] ① The hidden state h is mainly responsible for memorizing short-term information, especially the information of the current time step;
[0093] ②Cell state C is mainly responsible for memorizing long-term information.
[0094] LSTM has three key gating units, namely input gate, forget gate and output gate. These gating units control the flow of information through learning, helping LSTM networks better handle long-term dependencies, such as Figure 2 shown.
[0095] The calculation formulas for the three LSTM gates are as follows:
[0096] The forget gate determines which information in the cell state should be forgotten or discarded. It contains a sigmoid activation function that outputs a value between 0 and 1 based on the input data and the hidden state of the previous time step. Through multiplication operations, the forget gate controls the degree of retention of historical information in the cell state and determines which information will persist in the cell state. The calculation formula of the forget gate is as follows:
[0097] f t =σ(W f .[h t-1 ,x t ]+b f )
[0098] Among them, f t Represents the weight of forgetting the cell state, W f is the weight matrix of the forget gate, b f is the bias term, x t is the input of the current time step, h t-1 is the hidden state of the previous time step, σ is the sigmoid function, which maps the input value to a probability value between 0 and 1;
[0099] The input gate controls the degree of influence of the new input information on the current cell state. It contains a sigmoid activation function that generates a value between 0 and 1, indicating which values in each cell state should be updated. By using multiplication operations, the input gate determines which parts of the cell state to add the new input information to and passes the result to the cell state. The calculation formula of the input gate is as follows:
[0100] i t =σ(W i .[h t-1 ,x t ]+b i )
[0101] where i t Represents the weight for updating the cell state, W i is the weight matrix of the input gate, b i is the bias term, x t is the input of the current time step, h t-1 is the hidden state of the previous time step, σ is the sigmoid function, which maps the input value to a probability value between 0 and 1;
[0102] Output gate, the output gate controls the hidden state of the current moment and how the cell state at the next moment affects the final output. It contains a sigmoid activation function to determine which parts of the output state will be activated. The output gate also contains a tanh activation function to generate a value between -1 and 1, representing the candidate value of the cell state at the current time step. Through the multiplication operation, the output gate sends the filtered and updated information in the cell state to the hidden state of the next time step, and based on this hidden state h t Generate the final output. The calculation formula of the output gate is as follows:
[0103] o t =σ(W o .[h t-1 ,x t ]+b o )
[0104] h t =o t *tanh(C t )
[0105] Among them, W o is the weight matrix of the output gate, b o is the bias term, x t is the input of the current time step, h t-1 is the hidden state of the previous time step, σ is the sigmoid function, which maps the input value to a probability value between 0 and 1; tanh is the hyperbolic tangent function, which maps the input value to a value between -1 and 1. LSTM usually performs predictions step by step in time series prediction tasks. The specific steps are as follows:
[0106] 1) Initial input:
[0107] Input the data of historical time steps in step S2 and step S3 into the LSTM model to obtain the initial hidden state and cell state, ensuring that the data is arranged in chronological order.
[0108] The initial input data is usually a fixed-length time window containing historical data of multiple time steps.
[0109] 2) Step-by-step prediction:
[0110] For each time step, the data of the current time step and the hidden state of the previous time step are input into the LSTM network. The output result of the current time step and the new hidden state and cell state are obtained through the calculation of the LSTM network. The output of the LSTM network can be a predicted value (such as the value of the next time step) or multiple predicted values (such as the values of multiple future time steps).
[0111] 3) Iteration process:
[0112] The output result of the current time step is used as the input of the next time step, and the above process is repeated until the predicted time range is reached or the stopping condition is met.
[0113] 4) Prediction strategy:
[0114] Training phase: input the preprocessed cardiopulmonary exercise test (CPET) data into the LSTM model, input the true value into the long short-term memory network LSTM as the input of the next time step, and make predictions step by step: after each prediction, compare the prediction of coronary heart disease with the actual case data to evaluate the prediction error. Choose to use the true value as the input of the next time step to reduce error accumulation.
[0115] Prediction stage: input the predicted value into the long short-term memory network LSTM: use the predicted value as the input of the next time step, and continue to predict the subsequent time steps. Record the results of the prediction of coronary heart disease at each step until the prediction of all time steps is completed.
[0116] 5) Error assessment and adjustment:
[0117] After each prediction, the prediction result is compared with the true value to evaluate the prediction error. Depending on the size of the error, the model input can be adjusted or the model can be retrained. Common error evaluation indicators include mean square error (MSE), root mean square error (RMSE), mean absolute error (MAE), etc.
[0118] Long-term forecast considerations
[0119] Error accumulation:
[0120] As the forecast time step increases, the forecast error may gradually accumulate, resulting in a decrease in forecast accuracy. Therefore, when making long-term forecasts, special attention should be paid to the error accumulation problem.
[0121] Model retraining:
[0122] In order to improve the accuracy of long-term predictions, the model can be retrained regularly using the latest actual data to update the model parameters so that it can better adapt to data changes.
[0123] Dynamic Adjustment:
[0124] The input sequence and model parameters can be dynamically adjusted according to the prediction results to improve the flexibility and accuracy of the prediction. For example, the autoregressive model is combined with LSTM to enhance the prediction ability of the model.
[0125] S5. Input the time series analysis results and static cardiopulmonary characteristics into the classification model to predict coronary heart disease; the result output module outputs the prediction results and generates relevant health reports.
[0126] The classification model predicts coronary heart disease in patients based on the results of time series analysis and static parameters. The classification model uses multiple layers of fully connected layers (Fully Connected Layer), also known as densely connected layers (Dense Layer), which is a basic layer type in deep learning neural networks. Each neuron in the fully connected layer is connected to all neurons in the previous layer, and each connection has a weight to adjust the strength of information transmission.
[0127] The implementation of the fully connected layer mainly includes two parts: parameter initialization and forward propagation.
[0128] Parameter initialization: The fully connected layer is composed of a weight matrix and a bias vector. When initializing the weight matrix, random initialization is usually used to make full use of the network's adaptive ability. When initializing the bias vector, a zero vector is usually used for initialization.
[0129] Forward propagation: The fully connected layer establishes a fully connected relationship between all neurons in the previous layer and all neurons in the current layer, and transmits and converts information through the weight matrix. The specific implementation process is as follows:
[0130] 1) Input data: Input data timing analysis results and preprocessed static features, usually a batch of feature vectors or feature tensors, each sample is represented as a vector. Pull the output tensor of the previous layer into one dimension, that is, construct the input vector. Merge the feature vector containing timing information and the static feature vector into a one-dimensional vector.
[0131] 2) Weight initialization: Each neuron is connected to each neuron in the previous layer, and each connection has a weight. In a fully connected layer, these weights are usually initialized, for example, by performing matrix multiplication of the input vector with the weight matrix using a random initialization method.
[0132] 3) Add the bias vector to the result of the matrix multiplication.
[0133] 4) The result of adding the bias vector is input into the activation function for processing to obtain the output result of the current layer.
[0134] 5) Repeat: If there are more fully connected layers, repeat the above steps. The output of each layer will be used as the input of the next layer until it reaches the last layer or the output layer.
[0135] 6) Loss calculation: After the output layer, the loss between the predicted value and the true label is usually calculated to measure the accuracy of the model prediction.
[0136] 7) Back-propagation and optimization: Use the back-propagation algorithm to calculate the gradient of the loss function with respect to each parameter, and then use an optimization algorithm (such as stochastic gradient descent) to update the weights to minimize the loss function.
[0137] Sigmoid activation function: Among them, x represents the final output of the model
[0138] Binary cross entropy loss function:
[0139]
[0140] Among them, n represents the number of predicted samples; y i Represents the sample label, when the patient has coronary heart disease, y i is 1, otherwise it is 0; p(y i =1) represents the probability that the model predicts that the patient has coronary heart disease.
[0141] Missing Value Imputation: handles missing values in relevant data and fills in missing values using appropriate methods, such as mean, median, mode filling, etc. Missing value processing helps improve prediction performance.
[0142] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well.
Claims
1. A method for predicting coronary heart disease by integrating the time series information of cardiopulmonary exercise parameters, characterized in that: The following steps are involved: S1. Collect multi-dimensional health data of patients; S2. Preprocessing the multi-dimensional health data; S3, extract features from the preprocessed data; S4, using deep learning model to perform time series analysis on the extracted features; S5. Input the time series analysis results and static features into the classification model to predict coronary heart disease, output the prediction results and generate relevant health reports.
2. A method for predicting coronary heart disease by integrating cardiopulmonary exercise parameter time series information according to claim 1, characterized in that: In step S1, the multi-dimensional health data includes the demographic information, basic medical information, static parameters and timing information of the cardiopulmonary exercise test of the target subject, and the static parameters and timing information of the cardiopulmonary exercise test include ventilation function, gas exchange, exercise tolerance and circulatory function.
3. According to the method for predicting coronary heart disease integrating cardiopulmonary exercise parameter time series information as claimed in claim 1, in said step S2, it is characterized in that: The preprocessing includes data cleaning, default value processing, and normalization operations.
4. According to the method for predicting coronary heart disease integrating cardiopulmonary exercise parameter time series information as claimed in claim 1, in said step S3, it is characterized in that: The extracted features include time domain features, frequency domain features and other features that may be related to coronary heart disease.
5. The method for predicting coronary heart disease by integrating the time series information of cardiopulmonary exercise parameters according to claim 1, characterized in that: The deep learning model is a long short-term memory network (LSTM) combined with patient time series information to predict diseases.
6. A method for predicting coronary heart disease integrating cardiopulmonary exercise parameter time series information according to claim 5, characterized in that: The long short-term memory network LSTM includes an input gate, a forget gate and an output gate. The calculation formula of the forget gate is as follows: f t =σ(W f ·[h t-1 ,x t ]+b f ) Among them, f t Represents the weight of forgetting the cell state, W f is the weight matrix of the forget gate, b f is the bias term, x t is the input of the current time step, h t-1 is the hidden state of the previous time step, σ is the sigmoid function, which maps the input value to a probability value between 0 and 1; The calculation formula of the input gate is as follows: i t =σ(W i .[h t-1 ,x t ]+b i ) where i t Represents the weight for updating the cell state, W i is the weight matrix of the input gate, b i is the bias term, x t is the input of the current time step, h t-1 is the hidden state of the previous time step, σ is the sigmoid function, which maps the input value to a probability value between 0 and 1; LSTM calculates candidate cell states, which represent the impact of the new input at the current time step on the cell state. The calculation formula is as follows Among them, W c is the weight matrix of the candidate cell state, b c is the bias term, x t is the input of the current time step, h t-1 is the hidden state of the previous time step, and tanh is the hyperbolic tangent function, which maps the input value to a value between -1 and 1; The cell state of LSTM will be updated and passed to the next time step. At each time step t, the update formula of the cell state is as follows: in, is the candidate cell state of the current time step, indicating the impact of the new input of the current time step on the cell state; C t is the candidate cell state for the next time step; The output gate determines the output information o through the sigmoid function t , at each time step t, the output gate is calculated as follows: the t =σ(W o .[h t-1 ,x t ]+b o ) Among them, W o is the weight matrix of the output gate, b o is the bias term, x t is the input of the current time step, h t-1 is the hidden state of the previous time step, σ is the sigmoid function, which maps the input value to a probability value between 0 and 1; LSTM will transform the cell state C t Through the tanh function, the hidden state h of the current time step is obtained t : h t =o t *tanh(C t ) Here, tanh is the hyperbolic tangent function, which maps the input value to a value between -1 and 1.
7. A method for predicting coronary heart disease by integrating cardiopulmonary exercise parameter time series information according to claim 6, characterized in that: The long short-term memory network LSTM adopts step-by-step prediction in the time series prediction task. The specific steps are as follows: 1) Initial input: Input the data of the historical time step into the long short-term memory network LSTM to obtain the initial hidden state and cell state, ensuring that the data is arranged in chronological order; The initial input data is usually a fixed-length time window containing historical data of multiple time steps; 2) Step-by-step prediction: For each time step, the data of the current time step and the hidden state of the previous time step are input into the long short-term memory network LSTM, and the output result of the current time step and the new hidden state and cell state are obtained through the long short-term memory network LSTM, and the long short-term memory network LSTM outputs the patient's time series information; 3) Iteration process: Use the output result of the current time step as the input of the next time step, and repeat step 2) until the specified number of times is reached; 4) Prediction strategy: True value input: After each prediction, the prediction of whether it is coronary heart disease is compared with the actual case data to evaluate the prediction error; choose to use the true value as the input of the next time step to reduce error accumulation; Prediction value input: Use the long short-term memory network LSTM to predict whether it is coronary heart disease; Use the cell state updated in this time step and the time series features corresponding to the next time step as the input of the next time step. This method is suitable for situations where long-time step prediction is required, but it may cause errors to accumulate gradually. 5) Error assessment and adjustment: After each prediction, the prediction result is compared with the true value, the prediction error is evaluated, and according to the error size, the LSTM input of the long short-term memory network is adjusted or the LSTM model of the long short-term memory network is retrained.
8. A method for predicting coronary heart disease by integrating time series information of cardiopulmonary exercise parameters according to claim 7, characterized in that: The classification model includes multiple layers of fully connected layers, each neuron in the fully connected layer is connected to all neurons in the previous layer, and each connection has a weight for adjusting the intensity of information transmission.
9. A method for predicting coronary heart disease by integrating time series information of cardiopulmonary exercise parameters according to claim 9, characterized in that: The fully connected layer consists of two parts: parameter initialization and forward propagation; Parameter initialization: When initializing the weight matrix, use random initialization, and when initializing the bias vector, use the zero vector for initialization; Forward propagation: The fully connected layer establishes a fully connected relationship between all neurons in the previous layer and all neurons in the current layer, and transmits and transforms information through the weight matrix.
10. A method for predicting coronary heart disease by integrating time series information of cardiopulmonary exercise parameters according to claim 7, characterized in that: Use binary cross entropy as the loss function for back propagation to update the parameters of the long short-term memory network LSTM.
Citation Information
Patent Citations
Coronary heart disease assessment method and system based on electrocardio and heart sound signal coupling analysis
CN115553784A
Multi-parameter cardiac function monitoring system and method
CN117598674A
Cardiopulmonary function diagnosis system based on medical priori knowledge
CN117731248A
Cited By
Intelligent state monitoring system and method for patient after radiotherapy
CN119541867A
Risk prediction method for malnutrition of gastric cancer surgery patient
CN119889706A
Method for realizing exercise health management based on individualized cardiopulmonary function digital twinborn model
CN120164618A
A method for realizing exercise health management based on individualized cardiopulmonary function digital twin model
CN120164618B
Auxiliary diagnosis method and system based on chest image analysis
CN120260876A
