A chest pain intelligent triage method and system

Abstract

The present invention discloses a method and system for intelligent chest pain triage, relating to the field of intelligent medical triage. The method comprises collecting patient condition information and chest vibration signals, standardizing the collected data, and extracting a multimodal feature set; inputting the multimodal feature set into a chest pain knowledge graph, mapping semantic feature vectors using the BERT semantic embedding model, and filtering out graph subgraphs; calculating etiology likelihood scores based on the graph subgraphs through nonlinear feature association analysis, and optimizing associations using a Bayesian network to obtain a ranked list of etiologies and an inference path; using a Q-learning reinforcement learning algorithm, combined with chest vibration signals, to weight the ranked list of etiologies and update the inference path; and generating triage priorities and recommendations based on the adjusted ranked list of etiologies and the inference path. By combining multimodal data with the Q-learning reinforcement learning algorithm, the present invention achieves accurate triage priority sorting and personalized triage recommendation generation.

Description

Technical Field

[0001] The present invention relates to the field of intelligent medical triage, and in particular to an intelligent chest pain triage method and system. Background Art

[0002] Chest pain is a common clinical emergency symptom with numerous potential causes, including but not limited to heart disease, lung disease, and the digestive and musculoskeletal systems. Traditional chest pain triage methods rely primarily on physicians' clinical experience and routine physical examinations. However, due to the complexity and similarities of the causes of chest pain, traditional methods may not be able to quickly and accurately diagnose the cause in some cases. In recent years, with the rapid development of artificial intelligence technology, intelligent chest pain triage methods have emerged. In particular, intelligent triage systems that combine machine learning, big data analysis, and knowledge graphs have gradually become an important auxiliary tool in the diagnosis and treatment of chest pain.

[0003] However, the existing intelligent triage technology for chest pain still has certain shortcomings in many aspects. First, the existing intelligent triage systems mostly rely on a single data source, such as clinical examinations, medical history records or basic physiological data, but these data often cannot fully and accurately reflect the patient's symptoms and signs, resulting in limited diagnostic accuracy. Secondly, traditional chest pain triage systems based on rules or classical machine learning models often lack a deep understanding of the complex relationships between causes of disease, resulting in limitations in cause sorting and reasoning paths, especially in complex situations where multiple causes exist at the same time or change alternately, the system may not be able to make reasonable judgments. In addition, although the collection and analysis of vibration signals have been applied in some fields, how to effectively integrate them into the chest pain triage process and optimize the reasoning process through intelligent algorithms is still a problem that needs to be solved. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides an intelligent chest pain triage method to solve the problems of insufficient data diversity and limitations in cause sorting in existing chest pain triage methods.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides an intelligent chest pain triage method, which includes collecting patient condition information and chest vibration signals, standardizing the collected data, and extracting a multimodal feature set; inputting the multimodal feature set into a chest pain knowledge graph, mapping semantic feature vectors through a BERT semantic embedding model, and screening out graph subgraphs; based on the graph subgraphs, calculating the cause possibility score through nonlinear feature association analysis, and using a Bayesian network to optimize the association to obtain a cause ranking list and reasoning path; using a Q-learning reinforcement learning algorithm, combined with chest vibration signals, to adjust the weights of the cause ranking list and update the reasoning path; based on the adjusted cause ranking list and reasoning path, generating triage priorities and triage suggestions.

[0008] As a preferred solution of the intelligent chest pain triage method described in the present invention, the patient's condition information includes patient personal information, main complaint symptom description and core physical sign data.

[0009] As a preferred embodiment of the intelligent chest pain triage method of the present invention, the steps of collecting patient condition information and chest vibration signals, standardizing the collected data, and extracting a multimodal feature set are as follows:

[0010] Collect patient personal information through self-service terminals and patient-side mobile devices;

[0011] Through text input and doctor inquiries, the patient's subjective feelings about current chest pain are obtained and the description of the main complaints and symptoms is summarized;

[0012] Collect patients' core vital signs data in real time through medical equipment;

[0013] Use chest vibration detection equipment to non-invasively collect the patient's chest vibration signal;

[0014] The patient's personal information, main complaint symptom description and core vital sign data are combined into the patient's condition information, and the patient's condition information and chest vibration signal are sent to the data processing terminal via wireless transmission;

[0015] Preprocess and normalize the patient's condition information and chest vibration signals;

[0016] Multimodal features are extracted from the processed patient condition information and chest vibration signals to generate a multimodal feature set.

[0017] As a preferred solution of the chest pain intelligent triage method of the present invention, wherein: the multimodal feature set is input into the chest pain knowledge graph, the semantic feature vector is mapped through the BERT semantic embedding model, and the graph subgraph is screened out. The specific steps are:

[0018] Build a chest pain knowledge graph based on medical literature, historical cases, and clinical data;

[0019] The multimodal feature set is input into the chest pain knowledge graph and mapped to a high-dimensional vector space using the BERT semantic embedding model to generate a semantic feature vector.

[0020] Calculate the cosine similarity between the semantic feature vector and the nodes in the chest pain knowledge graph. The expression is:

[0021]

[0022] Among them, S il is the semantic feature vector V i and chest pain knowledge graph node N l The cosine similarity between i is the i-th semantic feature vector, N l is the lth chest pain knowledge graph node, ∥V i ∥ is the modulus length of the semantic feature vector, ∥N l ∥ is the modulus of the chest pain knowledge graph node, i is the index coefficient of the number of semantic feature vectors, l is the index coefficient of the number of chest pain knowledge graph nodes, and N is the number of chest pain knowledge graph nodes;

[0023] Subgraph nodes are filtered according to cosine similarity, and the associated paths between subgraph nodes are retained to generate a graph subgraph.

[0024] As a preferred solution of the chest pain intelligent triage method of the present invention, wherein: based on the graph subgraph, the cause possibility score is calculated by nonlinear feature association analysis, and the correlation degree is optimized using the Bayesian network to obtain the cause ranking list and reasoning path. The specific steps are:

[0025] In the generated graph subgraph, the multimodal interaction relationship between the semantic feature vector and the cause node is analyzed to calculate the cause possibility score, which is expressed as:

[0026]

[0027] Among them, P k is the probability score of the cause, k is the index of the number of cause nodes, the integral interval [0,1] is the standardized time range, n is the total number of features in the multimodal feature set, α i is a single feature weight coefficient, exp is an exponential function with the natural constant e as the base, ∥V i -N k ∥ is the Euclidean distance, N k is the kth causal node in the graph subgraph, β is the multi-feature interaction weight coefficient, m is the number of features involved in the multi-feature interaction calculation, j is the index coefficient of m, V j is the jth semantic feature vector participating in the multi-feature interaction calculation, and dt is the integral identifier;

[0028] Based on the Bayesian network model, the feature weights are adjusted to optimize the correlation between the semantic feature vector and the cause node. The expression is:

[0029]

[0030] in, is the updated single feature weight coefficient, is the prior single feature weight coefficient, L(V i |N k ) is the semantic feature vector V i At the cause node N k Likelihood value under the condition;

[0031] Arrange the cause nodes in descending order according to the cause possibility score to generate a cause ranking list;

[0032] In the generated graph subgraph, the associated path of the highest-scoring cause node in the cause ranking list is traced to generate an inference path.

[0033] As a preferred embodiment of the intelligent chest pain triage method of the present invention, the Q-learning reinforcement learning algorithm is used to adjust the weight of the cause ranking list and update the reasoning path in combination with the chest vibration signal. The specific steps are as follows:

[0034] Through fast Fourier transform and wavelet transform, the time-frequency characteristics of the chest vibration signal are extracted to form the vibration signal feature vector;

[0035] Using the Q-learning reinforcement learning algorithm, the weight of the cause node is adjusted based on the vibration signal feature vector. The expression is:

[0036]

[0037] Among them, Q(s t ,a t ) is the current state s t and action a t Q value of state s t is the vibration signal feature vector, action a t is the weight adjustment of the cause node, η is the learning rate, and the reward r t is the accuracy evaluation of the cause ranking list inference result, γ is the discount factor, is the maximum Q value of all actions in the next state;

[0038] Based on the adjusted causal node weight, the reasoning path is updated, and the expression is:

[0039]

[0040] Among them, P is the updated reasoning path, p is the number of cause nodes in the cause sorting list, is the weight of the causal node after being updated through reinforcement learning.

[0041] As a preferred embodiment of the chest pain intelligent triage method of the present invention, wherein: the triage priority and triage suggestion are generated based on the adjusted cause ranking list and reasoning path, the specific steps are:

[0042] Based on the adjusted cause ranking list and combined with the chest pain knowledge graph, the priority score of each cause is calculated;

[0043] Generate a personalized triage recommendation matrix based on the type of causal node and the depth of the reasoning path;

[0044] The priority score and personalized triage recommendation matrix are integrated to generate the triage report.

[0045] In a second aspect, the present invention provides an intelligent chest pain triage system, comprising a condition data acquisition module, a knowledge graph screening module, a cause ranking reasoning module, a vibration signal adjustment module and a triage suggestion generation module; the condition data acquisition module is used to collect patient condition information and chest vibration signals, standardize the collected data, and extract a multimodal feature set; the knowledge graph screening module is used to input the multimodal feature set into the chest pain knowledge graph, map the semantic feature vector through the BERT semantic embedding model, and filter out the graph subgraph; the cause ranking reasoning module is used to calculate the cause possibility score based on the graph subgraph through nonlinear feature association analysis, and use the Bayesian network to optimize the association to obtain a cause ranking list and reasoning path; the vibration signal adjustment module is used to use the Q-learning reinforcement learning algorithm, combined with the chest vibration signal, to adjust the weight of the cause ranking list and update the reasoning path; the triage suggestion generation module is used to generate triage priorities and triage suggestions based on the adjusted cause ranking list and reasoning path.

[0046] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the intelligent chest pain triage method as described in the first aspect of the present invention is implemented.

[0047] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the intelligent chest pain triage method as described in the first aspect of the present invention is implemented.

[0048] The beneficial effects of the present invention are as follows: by collecting the patient's condition information and chest vibration signals, and performing standardization processing and multimodal feature extraction on the data, it is possible to comprehensively obtain the patient's condition information and provide a data basis with strong consistency and high comparability for subsequent analysis. By inputting these multimodal features into the chest pain knowledge graph and mapping them to a high-dimensional semantic space through the BERT semantic embedding model, it is possible to accurately screen out the graph subgraphs most relevant to the patient's condition, thereby achieving personalized etiology reasoning. Through nonlinear feature association analysis and Bayesian network optimization, the accuracy of the etiology possibility score is further improved, and the reasoning path is optimized. Combined with the Q-learning reinforcement learning algorithm, the weights of the etiology ranking list can be adjusted in real time, and the reasoning path can be continuously optimized to enhance the adaptive ability. Based on the adjusted etiology ranking and reasoning path, priority scores and personalized triage recommendations are generated to provide clinical scientific and efficient triage decision support, thereby improving triage efficiency, accuracy and utilization of medical resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0050] Figure 1 This is a flow chart of the intelligent chest pain triage method in Example 1.

[0051] Figure 2 This is a module diagram of the chest pain intelligent triage system in Example 1. DETAILED DESCRIPTION

[0052] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0053] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0054] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0055] Example 1, with reference to Figure 1 and Figure 2 , which is the first embodiment of the present invention, provides a method for intelligent chest pain triage, comprising the following steps:

[0056] S1: Collect patient condition information and chest vibration signals, standardize the collected data, and extract multimodal feature sets.

[0057] Specifically, the following steps are included:

[0058] S1.1: Collect patient personal information through self-service terminals and patient-side mobile devices.

[0059] Specifically, patient personal information includes:

[0060] Demographic data such as name, gender, age, weight, height, etc.;

[0061] Past medical history: such as hypertension, diabetes, coronary heart disease, chronic lung disease, etc.;

[0062] Family history: such as whether there is a family history of aortic dissection, coronary artery disease, or hyperlipidemia;

[0063] Lifestyle habits: smoking, drinking, exercise, etc.

[0064] The above information is obtained with the user's consent and is used for legal purposes.

[0065] S1.2: Obtain the patient's subjective experience of current chest pain through text input and doctor's inquiries, and summarize the description of the main symptoms.

[0066] Specifically, the main symptom description includes:

[0067] Nature of chest pain: such as pressure, stabbing pain, dull pain;

[0068] Chest pain location: such as behind the sternum, left chest, right chest;

[0069] Duration of chest pain: such as intermittent or continuous;

[0070] Aggravating or relieving factors: such as aggravation after activity and relief after rest;

[0071] Are there other symptoms: such as difficulty breathing, sweating, nausea, dizziness, radiating pain in the left arm, etc.

[0072] S1.3: Collect patients’ core vital signs data in real time through medical devices.

[0073] Specifically, the heart rate is recorded by an electrocardiogram monitoring device, the blood pressure is measured by an electronic sphygmomanometer, the blood oxygen saturation is collected by a fingertip pulse oximeter, the body temperature is measured by an infrared thermometer or a contact thermometer, and the patient's respiratory rate is collected by a chest sensor or manual counting.

[0074] S1.4: Use chest vibration detection equipment to non-invasively collect the patient's chest vibration signal.

[0075] Specifically, chest heart sound sensors (such as accelerometers or microphones) are used to collect acoustic vibrations when the heart valves close and the myocardium contracts; lung sound sensors are used to collect airflow characteristics in chest vibrations to detect abnormalities such as dry rales, wet rales, and wheezing; and high-sensitivity vibration sensors (such as skin surface accelerometers) are used to collect chest wall tremors, capture low-frequency tremor signals, and diagnose causes such as aortic dissection.

[0076] S1.5: Combine the patient's personal information, chief symptom description, and core vital sign data into patient condition information, and send the patient's condition information and chest vibration signal to a data processing terminal via wireless transmission.

[0077] It should be understood that the chest vibration signal is sent to the data processing terminal via wireless transmission (such as Bluetooth or Wi-Fi), and the data is stored in real time in the database of the hospital information platform or telemedicine platform for subsequent analysis.

[0078] S1.5.1: Patient medical information includes patient personal information, description of main symptoms and core vital signs data.

[0079] S1.6: Preprocess and normalize the patient's condition information and chest vibration signals.

[0080] Specifically, the collected patient information and chest vibration signal data are preprocessed to remove invalid or noisy data. The patient information data is checked for completeness (e.g., whether key fields are missing) and correctness (e.g., whether the age range is reasonable). A filtering algorithm is used to remove background noise and extract valid heart sound, lung sound, and tremor data.

[0081] Data from different sources were standardized to ensure that the eigenvalues ​​were within the same scale. The patient's text description was converted into structured data, for example, "chest pressure" was assigned a classification number of 1, and "radiating pain in the left arm" was assigned a classification number of 2. Numerical data such as heart rate and blood pressure were normalized to a range of 0-1 to facilitate subsequent algorithm processing. Amplitude and frequency were normalized, and the main characteristic peaks of the heart sound signal, the spectral distribution characteristics of the lung sound signal, and the low-frequency components of the chest wall tremor signal were extracted.

[0082] S1.7: Extract multimodal features from the processed patient condition information and chest vibration signal to generate a multimodal feature set.

[0083] Specifically, the multimodal feature set includes the following:

[0084] Symptom characteristics: including the nature of chest pain, duration, and associated symptoms;

[0085] Physical characteristics: including heart rate, blood pressure, blood oxygen saturation and other physiological parameters;

[0086] Chest vibration signal characteristics: including amplitude, frequency, and power distribution characteristics of heart sounds, lung sounds, and chest wall tremors;

[0087] Feature format: stored in structured table or vector form as input data for the next step of knowledge graph analysis.

[0088] Preferably, by integrating the patient's personal information, main symptoms, core vital sign data and chest vibration signals, and after standardized processing and multimodal feature extraction, a comprehensive, accurate and efficient intelligent triage method for chest pain is provided. Through automated data collection and non-invasive chest vibration signal detection, the limitation of traditional triage relying on a single physiological indicator is overcome, which can effectively improve the accuracy of diagnosis, especially for the identification of complex diseases such as aortic dissection. In addition, the standardized processing of data ensures the high quality and consistency of information, allowing for more accurate correlation analysis of different types of data, thereby providing a solid foundation for the intelligent triage of chest pain and improving the efficiency and reliability of the triage process.

[0089] S2: Input the multimodal feature set into the chest pain knowledge graph, map the semantic feature vectors through the BERT semantic embedding model, and filter out the graph subgraphs.

[0090] Specifically, the following steps are included:

[0091] S2.1: Construct a chest pain knowledge graph based on medical literature, historical cases, and clinical data.

[0092] Specifically, the chest pain knowledge graph is a multimodal, multi-layered structured data system used to represent the relationships between chest pain-related symptoms, signs, test results, medical knowledge, causes, and treatment pathways. Its nodes (entities) and edges (relationships) form a network that supports semantic reasoning, path analysis, and dynamic updates.

[0093] S2.1.1: Collect chest pain knowledge graph data.

[0094] Specifically, obtain chest pain-related papers, case analyses, and guidelines from authoritative medical databases (such as PubMed, UpToDate, and Medline); obtain standardized medical terminology data from SNOMED-CT (medical terminology collection) to build a unified entity representation; and obtain chest pain-related disease codes and classifications from ICD-10 / ICD-11 (International Classification of Diseases).

[0095] The mechanisms and diagnostic and treatment pathways of chest pain etiology were extracted from clinical guidelines and medical literature.

[0096] Data related to symptoms, examinations, diagnosis, and treatment of patients with chest pain were extracted from historical cases and clinical data, including examination data such as electrocardiogram, chest CT, cardiac ultrasound, blood test results, and the patient's medical history.

[0097] By communicating with clinicians, cardiovascular experts or emergency department experts, the medical experience and diagnosis and treatment logic that are difficult to extract directly from the data in the knowledge graph can be supplemented.

[0098] S2.1.2: Define and extract chest pain knowledge graph entities (nodes) from chest pain knowledge graph data.

[0099] Specifically, the chest pain knowledge graph entities include:

[0100] Symptom entity: nature of chest pain (pressure, tingling, burning), duration, location (retrosternal, left chest), and radiation site (left shoulder, mandible).

[0101] Physical signs: abnormal heart rate (tachycardia / bradycardia), abnormal blood pressure, respiratory rate, chest wall tremor.

[0102] Examination entities: electrocardiogram (ST segment elevation, T wave inversion), blood tests (troponin, D-dimer), imaging examinations (chest CT, coronary angiography).

[0103] Etiological entities: Acute coronary syndrome (ACS), aortic dissection, pulmonary embolism, pericarditis, gastroesophageal reflux disease (GERD).

[0104] Treatment entities: drugs (aspirin, clopidogrel), surgery (PCI, coronary artery bypass grafting), supportive care (oxygen therapy, analgesia).

[0105] S2.1.3: Define and extract chest pain knowledge graph relationships (edges) from chest pain knowledge graph data.

[0106] Specifically, define the semantic relationship types between entities and extract specific relationships from the data:

[0107] Symptom-cause relationship: nature of chest pain → cause (such as pressure → acute coronary syndrome).

[0108] Symptom-examination relationship: nature of chest pain → recommended examination (e.g., chest pain with dyspnea → lung CT).

[0109] Examination-cause relationship: examination results → cause (such as ST segment elevation → acute myocardial infarction).

[0110] Cause-treatment relationship: cause → treatment plan (such as aortic dissection → emergency surgery).

[0111] Etiology-complication relationship: cause → complication (such as acute coronary syndrome → cardiogenic shock).

[0112] S2.1.3: Organize the extracted chest pain knowledge graph entities and relationships into a graph structure.

[0113] Specifically, triples are used to represent the knowledge graph, for example:

[0114] <Chest pain with dyspnea, recommended examination, chest CT>

[0115] ST-segment elevation, etiology, acute myocardial infarction

[0116] Acute Coronary Syndrome, Treatment Options, PCI

[0117] S2.1.4: Define knowledge graph inference rules.

[0118] Specifically, by combining medical guidelines with expert knowledge, inference rules are defined. For example, if the symptoms are chest pain accompanied by dyspnea, and the examination results show pulmonary embolism, the cause is inferred to be pulmonary embolism; if a blood test shows elevated troponin and elevated ST segment on the electrocardiogram, the cause is inferred to be acute myocardial infarction.

[0119] S2.2: Input the multimodal feature set into the chest pain knowledge graph and map it to a high-dimensional vector space using the BERT semantic embedding model to generate a semantic feature vector.

[0120] It's important to note that the BERT model uses a self-attention mechanism to learn the contextual relationships between features, mapping the multimodal feature set into a high-dimensional vector space to form a semantic feature vector. BERT's bidirectional encoding capability captures complex relationships between features, ensuring that the contextual information of each input feature is fully considered, laying the foundation for subsequent semantic analysis.

[0121] S2.3: Calculate the cosine similarity between the semantic feature vector and the nodes in the chest pain knowledge graph. The expression is:

[0122]

[0123] Among them, S ilis the semantic feature vector V i and chest pain knowledge graph node N l The cosine similarity between i is the i-th semantic feature vector, N l is the lth chest pain knowledge graph node, ∥V i ∥ is the modulus length of the semantic feature vector, ∥N l ∥ is the modulus of the chest pain knowledge graph node, i is the index coefficient of the number of semantic feature vectors, l is the index coefficient of the number of chest pain knowledge graph nodes, and N is the chest pain knowledge graph node.

[0124] Specifically, cosine similarity measures the angular difference between two vectors. Values ​​closer to 1 indicate greater similarity, while values ​​closer to 0 indicate lower similarity. This process quantifies the similarity between each multimodal feature and nodes in the knowledge graph, such as different causes, symptoms, and pathological mechanisms, providing a basis for subsequent screening of the most relevant nodes (i.e., subgraphs).

[0125] S2.4: Filter subgraph nodes based on cosine similarity and retain the associated paths between subgraph nodes to generate a graph subgraph.

[0126] Specifically, the cosine similarities between all chest pain knowledge graph nodes and semantic feature vectors are sorted, and nodes with higher similarity are selected. By setting a similarity threshold, nodes with lower similarity are filtered out, thereby reducing the interference of irrelevant information on subsequent analysis. The association paths between the filtered nodes are further retained. These association paths describe the connection between different causes, symptoms or physiological mechanisms. The retention of association paths can provide important structural information for reasoning and cause association analysis. Ultimately, the generated graph subgraph contains the nodes most relevant to the current patient's condition and the association relationships between them, which can provide an accurate semantic framework for subsequent nonlinear feature association analysis and reasoning processes.

[0127] Preferably, by constructing a chest pain knowledge graph and combining it with the BERT model for semantic mapping, the patient's condition can be accurately analyzed and rapid diagnosis can be assisted. By extracting multi-level chest pain-related knowledge from medical literature, clinical data, and expert experience, the knowledge graph provides a complete information architecture for the relationship between symptoms, signs, causes, examinations, and treatments. The BERT model uses contextual semantic understanding to map multimodal feature data to a high-dimensional space and calculates the similarity with the knowledge graph nodes, thereby accurately screening out nodes that are highly relevant to the patient's condition and optimizing the diagnostic path. The generated graph subgraph can provide accurate and personalized support for clinical decision-making, improve the efficiency and accuracy of diagnosis and treatment, and reduce misdiagnosis.

[0128] S3: Based on the graph subgraph, the cause possibility score is calculated through nonlinear feature association analysis, and the correlation degree is optimized using the Bayesian network to obtain the cause ranking list and reasoning path.

[0129] Specifically, the following steps are included:

[0130] S3.1: In the generated graph subgraph, analyze the multimodal interaction relationship between the semantic feature vector and the cause node, and calculate the cause possibility score, which is expressed as:

[0131]

[0132] Among them, P k is the probability score of the cause, k is the index of the number of cause nodes, the integral interval [0,1] is the standardized time range, n is the total number of features in the multimodal feature set, α i is a single feature weight coefficient, exp is an exponential function with the natural constant e as the base, ∥V i -N k ∥ is the Euclidean distance, N k is the kth causal node in the graph subgraph, β is the multi-feature interaction weight coefficient, m is the number of features involved in the multi-feature interaction calculation, j is the index coefficient of m, V j is the jth semantic feature vector participating in the multi-feature interaction calculation, and dt is the integral identifier.

[0133] Preferably, for each causal node N k , first through the Euclidean distance ∥V i -N k ∥Calculate multimodal features V i The similarity between the cause node and the cause node is used to obtain the contribution of each feature to the cause. At the same time, the combined function of sine and cosine sin(πV j ) and cos(πN k ) to model the interaction effects between different features. For example, the nature and location of chest pain may have stronger nonlinear relationships with certain causes (such as acute coronary syndrome), and these relationships are captured by Weighted calculation.

[0134] S3.2: Adjust the feature weights based on the Bayesian network model to optimize the correlation between the semantic feature vector and the cause node. The expression is:

[0135]

[0136] in, is the updated single feature weight coefficient, is the prior single feature weight coefficient, L(V i |N k) is the semantic feature vector V i At the cause node N k Likelihood value under the condition.

[0137] Preferably, the weight of each feature is optimized using a Bayesian network model. The Bayesian network can combine prior knowledge and observed evidence to infer a more accurate relationship between the feature and the causal node. The updated weight This reflects the change in the relative importance of features in the diagnosis process. Through this update, the association between causal nodes and features is more accurate and reliable.

[0138] S3.3: Arrange the cause nodes in descending order according to the cause likelihood scores to generate a ranked cause list.

[0139] Specifically, according to the causal probability score P of each causal node k , sorting all causal nodes from high to low by score. Nodes with higher scores are more likely to match the patient's condition and are therefore placed higher in the list. The sorted causal nodes are then ranked by likelihood score to generate a ranked list of causal factors, which serves as the basis for subsequent diagnosis and reasoning.

[0140] S3.4: In the generated graph subgraph, trace the associated path of the highest-scoring cause node in the cause ranking list to generate an inference path.

[0141] Specifically, based on the ranked list of causes, the highest-scoring cause node is selected and its associated paths are traced from the knowledge graph. These associated paths represent the logical connections between the cause node and other related nodes (such as symptoms, signs, and test results). By tracing the associated paths of the cause nodes, potential causes and possible treatment options are gradually derived. This path provides structured support for subsequent diagnostic decisions and guides clinicians in developing personalized treatment plans.

[0142] Optimally, the combination of nonlinear feature association analysis and Bayesian network optimization effectively improves the accuracy and personalization of chest pain etiology diagnosis. By calculating the similarity between multimodal features and etiology nodes and incorporating nonlinear interaction effects, complex etiology associations can be accurately captured. Bayesian networks are used to dynamically adjust feature weights and optimize the correlation between features and causes, thereby improving diagnostic reliability. The calculation of etiology probability scores and the generation of ranked lists of causes help identify the most relevant causes and, by tracing the inference path, provide clinicians with decision support for personalized treatment plans.

[0143] S4: Use the Q-learning reinforcement learning algorithm, combined with chest vibration signals, to adjust the weights of the cause ranking list and update the reasoning path.

[0144] Specifically, the following steps are included:

[0145] S4.1: Extract the time-frequency features of the chest vibration signal through fast Fourier transform and wavelet transform to form a vibration signal feature vector.

[0146] Specifically, a fast Fourier transform (FFT) is applied to the chest vibration signal, converting it from the time domain to the frequency domain. This allows the spectral characteristics of the chest vibration signal to be determined, including key features such as peak frequency, amplitude, and power spectral density. Wavelet transforms are then used to analyze the local time-frequency characteristics of the chest vibration signal, particularly the instantaneous variations within different frequency ranges. Time-frequency features related to heart sounds, lung sounds, and chest wall tremor are extracted, such as the instantaneous frequency and vibration amplitude within the time window.

[0147] The time-frequency features extracted by FFT and wavelet transform are integrated into a high-dimensional feature vector to represent the chest vibration signal. This high-dimensional feature vector contains information such as the chest vibration signal's spectral distribution, instantaneous frequency, and amplitude changes, reflecting the key characteristics of the chest vibration signal.

[0148] S4.2: Use the Q-learning reinforcement learning algorithm to adjust the cause node weight based on the vibration signal feature vector. The expression is:

[0149]

[0150] Among them, Q(s t ,a t ) is the current state s t and action a t Q value of state s t is the vibration signal feature vector, action a t is the weight adjustment of the cause node, η is the learning rate, and the reward r t is the accuracy evaluation of the cause ranking list inference result, γ is the discount factor, is the maximum Q value of all actions in the next state.

[0151] Preferably, the reward is calculated by comparing the difference between the result of the current reasoning path and the actual cause. If the result of the reasoning path closely matches the actual cause, the reward is large; if the match is low, the reward is small or negative. In each iteration, the Q value is adjusted based on the feedback reward information, gradually optimizing the weight of the cause node.

[0152] S4.3: Based on the adjusted causal node weights, update the inference path, which is expressed as:

[0153]

[0154] Among them, P is the updated reasoning path, p is the number of cause nodes in the cause sorting list, is the weight of the causal node after being updated through reinforcement learning.

[0155] It should be understood that the reasoning path describes the logical path from current symptoms and signs to the potential causes through the relationships between cause nodes. The updated reasoning path reflects the relative importance of each cause node in the overall reasoning process, making it easier to deduce the most relevant causes.

[0156] Optimally, combining the time-frequency feature extraction of chest vibration signals with the Q-learning reinforcement learning algorithm can continuously optimize the weights of causal nodes and the inference path, improving the accuracy of cause sorting and diagnostic efficiency. The time-frequency features extracted through fast Fourier transform and wavelet transform can precisely capture key pathological information in chest vibration signals. Q-learning uses a feedback mechanism to adjust the weights of causal nodes and optimize the inference path, providing more personalized and accurate cause analysis.

[0157] S5: Generate triage priorities and triage recommendations based on the adjusted ranked list of causes and reasoning pathways.

[0158] Specifically, the following steps are included:

[0159] S5.1: Based on the adjusted ranked list of causes and the chest pain knowledge graph, calculate the priority score for each cause.

[0160] Specifically, the priority score of each cause is calculated based on the adjusted cause ranking list and the association weight of the cause node in the reasoning path. The expression is:

[0161]

[0162] Among them, R k Score the priority of the cause, P k is the probability score of the cause, E k is the medical severity of the cause, with a value range of [1,5], for example, aortic dissection is 5 and GERD is 1; (provided by the severity label of the cause in the knowledge graph, for example, aortic dissection > acute coronary syndrome > gastroesophageal reflux disease), D k,z is the cause support strength of the zth signal feature in the chest vibration signal, with a value range of [0,1], Z is the total number of chest vibration signal features, C k is the weight of the impact of medical resource constraints on the cause of disease, with a value range of [0,1]. For example, when emergency resources are tight, the priority of non-critical causes of disease will be lowered.

[0163] R kThe higher the value, the more urgent the cause of the disease is, and triage should be prioritized. k A negative value indicates that the cause does not require immediate treatment or may be a misdiagnosis.

[0164] S5.2: Generate a personalized triage recommendation matrix based on the type of causal node and the depth of the reasoning path.

[0165] Specifically, specific triage recommendations are generated based on the type of causal node and the depth of the reasoning path.

[0166] The type of a cause node refers to the category or nature of each cause node. Depending on different medical fields and pathological characteristics, cause nodes can belong to different types, for example:

[0167] Acute causes (such as myocardial infarction, aortic dissection, etc.): These causes usually require urgent treatment and high-priority triage.

[0168] Chronic causes (such as chronic bronchitis, gastroesophageal reflux disease, etc.): These causes are not urgent, but still require timely diagnosis and treatment.

[0169] Functional diseases (such as anxiety disorders, mild depression, etc.): This type of cause is not directly life-threatening, but still affects the patient's health and quality of life.

[0170] Non-critical causes: such as some mild pathological symptoms, which may be related to other factors but do not require emergency treatment.

[0171] The depth of an inference path refers to the length of the path from current symptoms and signs, through the inference process, to the potential cause. A deeper path indicates that the inference process involves more causal nodes or more complex relationships, which may require more examinations or professional judgment. The depth of an inference path can be understood in the following ways:

[0172] Shallow reasoning pathways: When the reasoning process points directly to one or a few causes, the reasoning pathway is shallow. This is associated with a few common causes and can usually be diagnosed with basic testing.

[0173] Deep reasoning path: When the reasoning process involves multiple causal nodes or a complex relationship network, the reasoning path is deep. In this case, more examinations and expert diagnosis are needed to confirm the cause.

[0174] The depth of the inference pathway helps determine the triage complexity of the patient. Deeper pathways may require multiple tests or referrals to specialists to ensure diagnostic accuracy.

[0175] Triage recommendations include the following:

[0176] Recommended examinations: Based on the reasoning path of the cause node, relevant examination nodes are extracted to generate specific examination recommendations (such as blood tests, electrocardiograms, chest CT, etc.).

[0177] Referral department: Patients are triaged to appropriate departments (e.g., cardiology, emergency medicine, respiratory medicine) based on the type of etiology node.

[0178] Treatment time limit: Based on the priority of the cause, give clear treatment time recommendations (such as immediate examination, follow-up visit within 24 hours).

[0179] A personalized triage recommendation matrix is ​​generated by combining the cause priority scores and reasoning paths. Each row of the matrix corresponds to a cause, and the columns represent the examination, department, and time limit recommendations. The rows of the matrix are sorted in descending order to select the examination, department, and treatment recommendation with the highest score.

[0180] S5.3: Integrate the priority score and personalized triage recommendation matrix to generate a triage report.

[0181] Specifically, the triage priority and triage recommendations are integrated to generate a triage report, which includes the following:

[0182] Priority diagnosis cause: Displays the emergency cause and its probability score;

[0183] Recommended examinations: List the examination items in order of priority (such as chest CT, electrocardiogram, D-dimer test);

[0184] Referral department: clearly indicate the recommended referral department (e.g., cardiology department, emergency department);

[0185] Treatment time: Suggested treatment time (such as immediate treatment or follow-up within 48 hours).

[0186] The optimal approach is to quantify the urgency of the cause through a priority scoring formula based on the adjusted cause ranking list and chest pain knowledge graph. This, combined with the reasoning path, dynamically generates a personalized triage recommendation matrix, which is then integrated into a triage report, effectively improving the accuracy and efficiency of triage. This approach not only prioritizes high-risk causes and optimizes the allocation of medical resources, but also improves doctor-patient communication and the patient experience through clear examination recommendations, referral departments, and treatment timelines. It also provides support for clinical decision-making and training, demonstrating its significant clinical application value.

[0187] This embodiment also provides a chest pain intelligent triage system, including:

[0188] The condition data acquisition module is used to collect patient condition information and chest vibration signals, standardize the collected data, and extract multimodal feature sets;

[0189] The knowledge graph screening module is used to input the multimodal feature set into the chest pain knowledge graph, map the semantic feature vectors through the BERT semantic embedding model, and filter out the graph subgraphs;

[0190] The cause ranking reasoning module is used to calculate the cause possibility score based on the graph subgraph through nonlinear feature association analysis, and use the Bayesian network to optimize the association degree to obtain the cause ranking list and reasoning path;

[0191] The vibration signal adjustment module uses the Q-learning reinforcement learning algorithm to combine chest vibration signals to adjust the weights of the cause ranking list and update the inference path;

[0192] The triage recommendation generation module is used to generate triage priorities and triage recommendations based on the adjusted cause ranking list and reasoning path.

[0193] This embodiment also provides a computer device suitable for the case of an intelligent chest pain triage method, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the intelligent chest pain triage method proposed in the above embodiment.

[0194] The computer device may be a terminal, comprising a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner may be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a button, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse.

[0195] This embodiment also provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for intelligent chest pain triage proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, disk or optical disk.

[0196] In summary, the present invention collects patient condition information and chest vibration signals, standardizes the data, and extracts multimodal features, thereby comprehensively acquiring patient condition information and providing a highly consistent and comparable data foundation for subsequent analysis. By inputting these multimodal features into the chest pain knowledge graph and mapping them to a high-dimensional semantic space using the BERT semantic embedding model, the graph subgraphs most relevant to the patient's condition can be accurately screened, thereby enabling personalized etiology reasoning. Through nonlinear feature association analysis and Bayesian network optimization, the accuracy of the etiology probability score is further improved, and the reasoning path is optimized. Combined with the Q-learning reinforcement learning algorithm, the weights of the etiology ranking list can be adjusted in real time, and the reasoning path can be continuously optimized to enhance adaptive capabilities. Based on the adjusted etiology ranking and reasoning path, priority scores and personalized triage recommendations are generated, providing scientific and efficient triage decision support for clinicians, thereby improving triage efficiency, accuracy, and the utilization of medical resources.

[0197] Example 2, referring to Table 1, is the second embodiment of the present invention. In order to further verify the technical solution of the present invention, experimental simulation data of the intelligent triage method for chest pain are provided.

[0198] Thirty patients with chest pain were selected as research subjects. Fifteen of them were diagnosed using existing chest pain triage technology (the control group), and the other 15 were diagnosed using the intelligent chest pain triage method presented in this paper (the experimental group). The experimental process included collecting patient information, chest vibration signals, data standardization, multimodal feature extraction, knowledge graph reasoning, and etiology ranking, and generating triage recommendations based on priority.

[0199] First, demographic data (such as name, gender, age, and weight), medical history, presenting symptoms (such as the nature and duration of chest pain and associated symptoms), and core physiological signs (such as heart rate, blood pressure, blood oxygen saturation, and respiratory rate) were collected from patients in the experimental and control groups. The experimental group also used a chest vibration sensor to collect chest vibration signals, including heart sound signals (amplitude and frequency), lung sound signals (spectral distribution), and chest wall tremor signals (low-frequency tremor characteristics).

[0200] Next, the multimodal data from the experimental group was normalized, using filtering algorithms to remove noise and extract signal features (such as peak frequency and power density). Text descriptions were converted into structured data, and numerical data was normalized to a range of 0-1 to ensure uniformity and efficient processing.

[0201] Secondly, the multimodal features of the experimental group were input into the chest pain knowledge graph. The BERT model was used to generate semantic feature vectors, calculate similarities with the cause nodes, and filter out subgraphs. Nonlinear feature association analysis and Bayesian network optimization were combined to generate a ranked list of causes. The Q-learning reinforcement learning algorithm was then used to adjust the cause weights and optimize the inference path.

[0202] Finally, triage recommendations were generated for patients in the experimental group based on the etiology priority score, including recommended tests (such as blood tests, chest CT, electrocardiogram, etc.), referral departments (such as cardiology, emergency department), and treatment timelines (such as immediate examination or follow-up within 24 hours). The control group used traditional physician experience combined with a single physiological indicator for triage.

[0203] The details are shown in Table 1 below:

[0204] Table 1 Experimental data comparison record

[0205]

[0206] By analyzing the three sets of experimental data, we can clearly see the technical advantages of the present invention:

[0207] The average diagnostic accuracy in the experimental group was 96.9%, compared to 85.2% in the control group. The highest diagnostic accuracy in the experimental group was 97.2% (experimental group 3), while the highest in the control group was only 91.4% (control group 3). The experimental group used multimodal feature extraction and knowledge graph reasoning to accurately identify the cause of disease from multidimensional data, improving diagnostic accuracy.

[0208] The average diagnosis time for the experimental group was 13 minutes, while that for the control group was 28.6 minutes. The fastest diagnosis time in the experimental group was 10 minutes (experimental group 1), while the fastest in the control group was 25 minutes (control group 3). The automated diagnostic process of this invention effectively reduces manual intervention, improves diagnostic efficiency, and saves patients valuable treatment time.

[0209] The experimental group's average recognition rate for high-risk causes was 96.6%, higher than the control group's 65.6%. The experimental group was able to accurately identify critical causes such as aortic dissection, with a minimum recognition rate of 96.0% (experimental group 2), while the control group's highest recognition rate was only 61.0% (control group 2). This demonstrates the excellent performance of the present invention in diagnosing critical causes.

[0210] The experimental group also had superior patient satisfaction scores and missed diagnosis rates compared to the control group. The lowest patient satisfaction score in the experimental group reached 9.2 (experimental group 2), while the control group had a score of only 7 (control group 1). The highest missed diagnosis rate in the experimental group was only 4% (experimental group 2), while the control group reached 20% (control group 1), reflecting the comprehensive advantages of the present invention in diagnostic efficiency, accuracy, and patient experience.

[0211] Through comparative analysis of three groups of data, the present invention has shown advantages in key indicators such as diagnostic accuracy, diagnostic efficiency, high-risk cause identification rate, patient satisfaction and missed diagnosis rate, achieving more efficient, more accurate and safer intelligent triage of chest pain, and demonstrating certain clinical application value.

[0212] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for intelligent chest pain triage, characterized by: include, Collect patient condition information and chest vibration signals, standardize the collected data, and extract multimodal feature sets; The multimodal feature set is input into the chest pain knowledge graph, and the semantic feature vector is mapped through the BERT semantic embedding model to filter out the graph subgraph. The specific steps are as follows: Build a chest pain knowledge graph based on medical literature, historical cases, and clinical data; The multimodal feature set is input into the chest pain knowledge graph and mapped to a high-dimensional vector space using the BERT semantic embedding model to generate a semantic feature vector. Calculate the cosine similarity between the semantic feature vector and the nodes in the chest pain knowledge graph; Filter subgraph nodes based on cosine similarity and retain the associated paths between subgraph nodes to generate a graph subgraph; Based on the graph subgraph, the cause probability score is calculated through nonlinear feature association analysis, and the correlation degree is optimized using Bayesian network to obtain the cause ranking list and reasoning path; Using the Q-learning reinforcement learning algorithm, combined with chest vibration signals, the cause ranking list is weighted and the reasoning path is updated. The specific steps are: Through fast Fourier transform and wavelet transform, the time-frequency characteristics of the chest vibration signal are extracted to form the vibration signal feature vector; Using the Q-learning reinforcement learning algorithm, the weight of the cause node is adjusted based on the vibration signal feature vector; Based on the adjusted causal node weights, the reasoning path is updated; Based on the adjusted ranked list of etiologies and reasoning pathways, triage priorities and triage recommendations are generated.

2. The intelligent chest pain triage method according to claim 1, wherein: The patient's condition information includes the patient's personal information, description of the main symptoms and core physical sign data.

3. The intelligent chest pain triage method according to claim 2, wherein: The specific steps of collecting patient condition information and chest vibration signals, standardizing the collected data, and extracting a multimodal feature set are as follows: Collect patient personal information through self-service terminals and patient-side mobile devices; Through text input and doctor inquiries, the patient's subjective feelings about current chest pain are obtained and the description of the main complaints and symptoms is summarized; Collect patients' core vital signs data in real time through medical equipment; Use chest vibration detection equipment to non-invasively collect the patient's chest vibration signal; The patient's personal information, main complaint symptom description and core vital sign data are combined into the patient's condition information, and the patient's condition information and chest vibration signal are sent to the data processing terminal via wireless transmission; Preprocess and normalize the patient's condition information and chest vibration signals in the data processing terminal; Multimodal features are extracted from the processed patient condition information and chest vibration signals to generate a multimodal feature set.

4. The intelligent chest pain triage method according to claim 3, wherein: The cosine similarity between the calculated semantic feature vector and the nodes in the chest pain knowledge graph is expressed as: ； in, is the semantic feature vector and chest pain knowledge graph node The cosine similarity between For the semantic feature vectors, For the chest pain knowledge graph node, is the modulus length of the semantic feature vector, is the node module length of the chest pain knowledge graph, is the index coefficient of the number of semantic feature vectors, is the index coefficient of the number of nodes in the chest pain knowledge graph, It is a node in the chest pain knowledge graph.

5. The intelligent chest pain triage method according to claim 4, characterized in that: Based on the graph subgraph, the cause possibility score is calculated through nonlinear feature association analysis, and the correlation degree is optimized using the Bayesian network to obtain the cause ranking list and reasoning path. The specific steps are: In the generated graph subgraph, the multimodal interaction relationship between the semantic feature vector and the cause node is analyzed to calculate the cause possibility score, which is expressed as: ； in, The probability score of the cause, is the index of the number of cause nodes, the integral interval is the standardized time frame, is the total number of features in the multimodal feature set, is the single feature weight coefficient, For the natural constant The exponential function with base , is the Euclidean distance, is the first causal nodes, is the multi-feature interaction weight coefficient, is the number of features involved in multi-feature interaction calculation, for The index coefficient of For the semantic feature vectors that participate in multi-feature interaction calculations, is the integral identifier; Based on the Bayesian network model, the feature weights are adjusted to optimize the correlation between the semantic feature vector and the cause node. The expression is: ； in, is the updated single feature weight coefficient, is the prior single feature weight coefficient, is the semantic feature vector At the etiology node Likelihood value under the condition; Arrange the cause nodes in descending order according to the cause possibility score to generate a cause ranking list; In the generated graph subgraph, the associated path of the highest-scoring cause node in the cause ranking list is traced to generate an inference path.

6. The intelligent chest pain triage method according to claim 5, characterized in that: The Q-learning reinforcement learning algorithm is used to adjust the cause node weight based on the vibration signal feature vector, and the expression is: ； in, Current status and actions of Value, Status is the vibration signal feature vector, action is the weight adjustment of the cause node, is the learning rate, reward To evaluate the accuracy of the inference results of the ranked list of causes, is the discount factor, is the maximum of all actions in the next state value; Based on the adjusted causal node weight, the reasoning path is updated, and the expression is: ； in, is the updated reasoning path, is the number of cause nodes in the cause sorting list, is the weight of the causal node after being updated through reinforcement learning.

7. The intelligent chest pain triage method according to claim 6, characterized in that: The steps of generating triage priorities and triage suggestions based on the adjusted cause ranking list and reasoning path are as follows: Based on the adjusted cause ranking list and combined with the chest pain knowledge graph, the priority score of each cause is calculated; Generate a personalized triage recommendation matrix based on the type of causal node and the depth of the reasoning path; The priority score and personalized triage recommendation matrix are integrated to generate the triage report.

8. An intelligent chest pain triage system, based on the intelligent chest pain triage method according to any one of claims 1 to 7, characterized in that: It includes a disease data collection module, a knowledge graph screening module, a cause ranking and reasoning module, a vibration signal adjustment module, and a triage suggestion generation module; The condition data acquisition module is used to collect patient condition information and chest vibration signals, standardize the collected data, and extract a multimodal feature set; The knowledge graph screening module is used to input the multimodal feature set into the chest pain knowledge graph, map the semantic feature vectors through the BERT semantic embedding model, and screen out the graph subgraphs; The etiology ranking reasoning module is used to calculate the etiology possibility score based on the graph subgraph through nonlinear feature association analysis, and use the Bayesian network to optimize the association degree to obtain the etiology ranking list and reasoning path; The vibration signal adjustment module is used to use the Q-learning reinforcement learning algorithm, combined with the chest vibration signal, to adjust the weight of the cause ranking list and update the reasoning path; The triage suggestion generating module is used to generate triage priorities and triage suggestions based on the adjusted cause ranking list and reasoning path.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the chest pain intelligent triage method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the chest pain intelligent triage method according to any one of claims 1 to 7 are implemented.

Images