A method and system for constructing a cardiology nursing knowledge base

Abstract

The present invention relates to the field of medical data processing technology, and in particular to a method and system for constructing a cardiology nursing knowledge base. The system is used to: analyze the mapping relationship between heart state and nursing knowledge based on cardiac pathology data and corresponding cardiac nursing data, and generate a cardiac state nursing relationship feature matrix; perform pathological pattern identification processing of the cardiac state nursing relationship characteristics on the cardiac state nursing relationship feature matrix, and generate a pathological pattern cardiac state nursing relationship feature matrix; design an optimized prediction model for deriving cardiac pathological patterns and nursing information based on the pathological pattern cardiac state nursing relationship feature matrix, and generate an optimized cardiac nursing derivation prediction model; design and process an intelligent feedback engine for the cardiology nursing knowledge base based on the optimized cardiac nursing derivation prediction model, so as to construct an intelligent cardiology nursing knowledge base engine. The present invention realizes the efficient and intelligent establishment of a cardiology nursing knowledge base.

Description

Technical Field

[0001] The present invention relates to the technical field of medical data processing, and in particular to a method and system for constructing a cardiology nursing knowledge base. Background Art

[0002] With the continuous development of modern medical technology, the diagnosis, treatment and nursing of heart diseases involve a large amount of clinical data, pathological information, imaging data and physiological monitoring data. The traditional knowledge base for heart disease nursing relies on manual analysis to provide patients with certain diagnosis and treatment support, but it has certain limitations, such as low information processing efficiency, insufficient data analysis accuracy, and inaccurate medical decision support. In recent years, with the development of technologies such as artificial intelligence, big data, and deep learning, the research and application of the medical field, especially cardiology, have been significantly improved. Through the intelligent analysis and processing of massive medical data, medical researchers can understand the pathology and physiological mechanisms of the heart more comprehensively, and then propose personalized and precise nursing plans. However, the construction system of the existing cardiology nursing knowledge base is usually based on empirical rules or rules, lacking deep learning capabilities based on big data and artificial intelligence technologies; and the correlation between cardiac pathology data and clinical nursing data is insufficient, resulting in the inability to give full play to the characteristics of cardiac nursing-related knowledge data; the problem of accurate matching of cardiac pathological status and nursing intervention measures has not been effectively solved, and the nursing plan often does not match the pathological status, thus affecting the quality of nursing. Summary of the invention

[0003] Based on this, the present invention provides a method and system for constructing a cardiology nursing knowledge base to solve at least one of the above technical problems.

[0004] To achieve the above objectives, a system for constructing a cardiology nursing knowledge base includes the following modules:

[0005] A cardiac pathology diagnosis feature analysis module is used to obtain cardiac pathology data and corresponding cardiac care data from a medical database; perform cardiac pathology data analysis on the cardiac pathology data to generate cardiac pathology analysis data, wherein the cardiac pathology analysis data includes cardiac pathology diagnosis data, cardiac pathology physiological index data, and cardiac pathology imaging data; perform cardiac pathology diagnosis feature extraction based on the cardiac pathology diagnosis data to generate cardiac pathology diagnosis feature data;

[0006] A potential cardiac pathological abnormality feature fuzzy logic analysis module is used to perform fuzzy logic analysis and processing of potential cardiac pathological abnormality features based on cardiac pathological physiological index data and cardiac pathological imaging data to generate potential cardiac pathological abnormality feature fuzzy logic data;

[0007] A cardiac state nursing relationship feature analysis module is used to use cardiac pathology diagnosis feature data and potential cardiac pathology abnormality feature fuzzy logic data as cardiac state entity nodes and cardiac nursing data as cardiac nursing entity nodes, and to perform a mapping relationship analysis between cardiac state and nursing knowledge based on the cardiac state entity nodes and cardiac nursing entity nodes to generate a cardiac state nursing relationship feature matrix; to perform pathological pattern identification processing of cardiac state nursing relationship features on the cardiac state nursing relationship feature matrix to generate a pathological pattern cardiac state nursing relationship feature matrix;

[0008] The cardiology nursing knowledge base construction module is used to design an optimized prediction model for deriving cardiac pathology patterns and nursing information based on the pathology pattern heart state nursing relationship feature matrix, and generate an optimized cardiac nursing derivation prediction model; based on the optimized cardiac nursing derivation prediction model, the intelligent feedback engine design and processing of the cardiology nursing knowledge base is carried out to build an intelligent cardiology nursing knowledge base engine.

[0009] Furthermore, the cardiac pathology diagnosis feature analysis module includes the following functions:

[0010] Obtain cardiac pathology data and corresponding cardiac care data from medical databases;

[0011] Perform data modality type analysis based on cardiac pathology data to generate cardiac pathology modality data;

[0012] Performing cardiac pathology data analysis and processing on the cardiac pathology data according to the cardiac pathology modality data to generate cardiac pathology analysis data;

[0013] Performing text coding processing on the diagnostic data according to the cardiac pathology diagnostic data to generate cardiac pathology diagnostic text coding data;

[0014] The cardiac pathology diagnosis feature extraction process is performed based on the cardiac pathology diagnosis text coding data to generate cardiac pathology diagnosis feature data.

[0015] Furthermore, the potential cardiac pathological abnormality feature fuzzy logic analysis module includes the following functions:

[0016] Analyze and process abnormal physiological characteristics of cardiac pathological state according to cardiac pathological and physiological index data, and generate abnormal physiological index characteristic data of cardiac pathological state;

[0017] Analyze abnormal cardiac pathological physiological characteristics according to cardiac pathological and physiological index data to generate abnormal cardiac pathological physiological characteristic data;

[0018] Performing pathological image segmentation processing on cardiac pathological image data to divide cardiac functional structural regions, generating segmented cardiac pathological image data; performing image abnormality feature analysis on cardiac functional structure based on the segmented cardiac pathological image data, generating cardiac functional structure image abnormality feature data;

[0019] Fuzzy logic conversion processing of potential cardiac pathological abnormality characteristics is performed based on cardiac pathological abnormality physiological characteristic data and cardiac functional structural imaging abnormality characteristic data to generate potential cardiac pathological abnormality characteristic fuzzy logic data.

[0020] Furthermore, the cardiac state nursing relationship characteristic analysis module includes the following functions:

[0021] The cardiac pathology diagnosis feature data and the potential cardiac pathology abnormality feature fuzzy logic data are used as cardiac state entity nodes and the cardiac care data are used as cardiac care entity nodes. The mapping relationship between cardiac state and nursing knowledge is analyzed based on the cardiac state entity nodes and cardiac care entity nodes to generate a cardiac state nursing relationship feature matrix.

[0022] Performing cardiac pathology diagnosis cluster analysis on cardiac pathology diagnosis feature data to generate cardiac pathology diagnosis cluster data;

[0023] Conducting cardiac pathology diagnosis correlation analysis based on cardiac pathology diagnosis clustering data to generate cardiac pathology diagnosis correlation data;

[0024] Perform cardiac pathology pattern analysis based on cardiac pathology diagnosis correlation data to generate cardiac pathology pattern data;

[0025] Based on the cardiac pathological pattern data, the cardiac state nursing relationship feature matrix is ​​processed with the pathological pattern identification of the cardiac state nursing relationship features to generate a pathological pattern cardiac state nursing relationship feature matrix.

[0026] Furthermore, the cardiology nursing knowledge base construction module includes the following functions:

[0027] Performing pathological pattern cardiac nursing associated data analysis on the pathological pattern cardiac state nursing relationship characteristic matrix to generate pathological pattern cardiac nursing associated data;

[0028] Performing cardiac care-specific analysis of the pathological pattern based on the cardiac care-related data of the pathological pattern to generate cardiac care-specific data of the pathological pattern;

[0029] Extracting the common nursing features and differential nursing features of the pathological pattern from the pathological pattern heart nursing specific data, and generating the common nursing feature data of the pathological pattern heart nursing and differential nursing feature data of the pathological pattern heart nursing respectively;

[0030] The intervention contribution of nursing difference characteristics is weighted on the pathological pattern heart nursing difference characteristic data to generate pathological pattern heart nursing difference weighted characteristic data;

[0031] Based on the fuzzy logic data of potential cardiac pathological abnormalities, the influencing factors of cardiac nursing differences in pathological patterns are analyzed to generate data on influencing factors of cardiac nursing differences in pathological patterns;

[0032] Using the data of influencing factors of cardiac nursing differences in pathological patterns, the empowerment characteristic data of cardiac nursing differences in pathological patterns are analyzed to generate intervention characteristic data of cardiac nursing differences in pathological patterns;

[0033] The preset deep learning network model is used to design an optimized prediction model for deriving cardiac pathology patterns and nursing information through the common characteristic data of cardiac nursing of pathology patterns and the differential intervention characteristic data of cardiac nursing of pathology patterns, and an optimized cardiac nursing derivation prediction model is generated;

[0034] Based on the optimized cardiac care derivation prediction model, the intelligent feedback engine design and processing of the cardiology nursing knowledge base is carried out to build an intelligent cardiology nursing knowledge base engine.

[0035] Furthermore, the intervention contribution weighting processing of the nursing difference characteristics of the pathological pattern heart nursing difference characteristics data includes:

[0036] Based on the pathological pattern cardiac nursing associated data, the cardiac nursing intervention contribution weight analysis of each pathological pattern is performed to generate the cardiac nursing intervention contribution weight parameters of the pathological pattern;

[0037] According to the pathological pattern heart nursing intervention contribution weight parameter, the pathological pattern heart nursing difference characteristic data is weighted by the intervention contribution of the nursing difference characteristics to generate the pathological pattern heart nursing difference weighted characteristic data.

[0038] Furthermore, the cardiac care intervention contribution weight analysis of each pathological pattern based on the pathological pattern cardiac care associated data includes:

[0039] Perform independent cardiac care parameter analysis of each pathological pattern according to the cardiac care associated data of the pathological pattern, and generate independent cardiac care parameters of the pathological pattern;

[0040] Performing pathology pattern independent heart care parameter contribution analysis according to the pathology pattern independent heart care parameters to generate pathology pattern independent heart care parameter contribution data;

[0041] According to the independent cardiac nursing parameter contribution data of the pathological pattern, the cardiac nursing intervention contribution weight analysis of each pathological pattern is performed to generate the cardiac nursing intervention contribution weight parameters of the pathological pattern.

[0042] Furthermore, the optimization prediction model design for deriving cardiac pathology patterns and nursing information from a preset deep learning network model using the common characteristic data of cardiac nursing of pathology patterns and the differential intervention characteristic data of cardiac nursing of pathology patterns includes:

[0043] Use the preset deep learning network model to design the model architecture for cardiac care derivation prediction and generate a preliminary cardiac care derivation prediction model;

[0044] Using the common characteristic data of cardiac nursing of pathological patterns as the low-level abstract training set for cardiac nursing derivation, and using the differential intervention characteristic data of cardiac nursing of pathological patterns as the high-level abstract training set for cardiac nursing derivation;

[0045] transferring the cardiac care derivation low-level abstract training set to the preliminary cardiac care derivation prediction model for network-level cardiac care derivation training processing to generate the cardiac care derivation prediction model;

[0046] The cardiac care derivation high-order abstract training set is transferred to the cardiac care derivation prediction model for optimized learning processing of the cardiac care derivation difference characteristics of the network-level pathological pattern to generate an optimized cardiac care derivation prediction model.

[0047] Furthermore, the step of transmitting the cardiac care derivation high-order abstract training set to the cardiac care derivation prediction model for performing the cardiac care derivation difference feature optimization learning process of the network-level pathological pattern includes:

[0048] Using the cardiac care derivation high-level abstract training set, the cardiac care derivation prediction model is analyzed for the cardiac care bias characteristics at the network level to generate the cardiac care bias characteristic data at the network level;

[0049] Performing network-level cardiac care feature reconstruction processing according to the network-level cardiac care bias feature data to generate network-level cardiac care feature reconstruction data;

[0050] Through network-level cardiac care feature reconstruction data, the cardiac care derivation prediction model is optimized and learned based on the cardiac care derivation difference features of each pathological pattern to generate an optimized cardiac care derivation prediction model.

[0051] This specification provides a method for constructing a cardiology nursing knowledge base, which is implemented based on the construction system for the cardiology nursing knowledge base as described above. The method for constructing a cardiology nursing knowledge base includes:

[0052] Step S1: for acquiring cardiac pathology data and corresponding cardiac care data from a medical database; performing cardiac pathology data parsing processing on the cardiac pathology data to generate cardiac pathology analysis data, wherein the cardiac pathology analysis data includes cardiac pathology diagnosis data, cardiac pathology physiological index data and cardiac pathology imaging data; performing cardiac pathology diagnosis feature extraction processing on the cardiac pathology diagnosis data to generate cardiac pathology diagnosis feature data;

[0053] Step S2: performing fuzzy logic analysis and processing of potential cardiac pathological abnormality characteristics based on cardiac pathological and physiological index data and cardiac pathological image data to generate fuzzy logic data of potential cardiac pathological abnormality characteristics;

[0054] Step S3: using cardiac pathology diagnosis feature data and potential cardiac pathology abnormality feature fuzzy logic data as cardiac state entity nodes and cardiac nursing data as cardiac nursing entity nodes, performing a mapping relationship analysis between cardiac state and nursing knowledge according to the cardiac state entity nodes and cardiac nursing entity nodes, and generating a cardiac state nursing relationship feature matrix; performing pathological pattern identification processing of cardiac state nursing relationship features on the cardiac state nursing relationship feature matrix, and generating a pathological pattern cardiac state nursing relationship feature matrix;

[0055] Step S4: Design an optimized prediction model for deriving cardiac pathology patterns and nursing information based on the pathology pattern heart state nursing relationship feature matrix, and generate an optimized cardiac nursing derivation prediction model; design and process an intelligent feedback engine for the cardiology nursing knowledge base based on the optimized cardiac nursing derivation prediction model to build an intelligent cardiology nursing knowledge base engine.

[0056] The beneficial effect of the present application is that the cardiac pathology diagnosis feature analysis module can extract key features from multi-source data and comprehensively analyze cardiac pathology information by acquiring cardiac pathology data and corresponding cardiac care data in the medical database. In the data analysis process, the generated cardiac pathology analysis data not only covers cardiac pathology diagnosis data, but also includes cardiac pathology physiological index data and imaging data. These multi-dimensional data provide a rich and accurate basis for subsequent diagnosis and nursing derivation, effectively improve the understanding of cardiac pathological state, and thus support more accurate disease diagnosis. For example, the cardiac pathology diagnosis feature extraction process can extract key cardiac pathology feature data by deeply analyzing cardiac pathology diagnosis data, providing a solid basis for further nursing intervention and treatment decisions. The efficiency and accuracy of this process enable medical workers to quickly identify potential pathological risks in a large amount of clinical data and improve the early detection rate of diagnosis. The potential cardiac pathology abnormality feature fuzzy logic analysis module can identify potential cardiac pathology abnormality features by combining cardiac pathology physiological index data and imaging data for fuzzy logic analysis. Traditional diagnostic methods often rely on manual experience and rough analysis, while this module uses fuzzy logic technology to handle uncertainty and complexity in data, automatically infer potential abnormal pathological characteristics of the heart, and identify potential lesions or physiological abnormalities that have not yet been clinically manifested. Through intelligent analysis of cardiac pathological data and potential abnormal characteristics, feedback adjustments can be made based on real-time data and patient changes. This dynamic adjustment mechanism makes nursing decisions more flexible and accurate during implementation. With the input of more case data, the system's prediction model will be gradually optimized, further improving the accuracy of nursing decisions. The cardiac state nursing relationship feature analysis module uses cardiac pathological diagnostic feature data and potential cardiac pathological abnormal feature fuzzy logic data as cardiac state entity nodes, and combines cardiac nursing data as cardiac nursing entity nodes to perform a mapping relationship analysis between cardiac state and nursing knowledge, establish a relationship between cardiac pathological state and nursing plan, and thus generate a nursing strategy that is more in line with the patient's specific situation. Accurately identifying nursing needs under different pathological states and converting them into nursing decision support information improves the personalization level of nursing plans for heart patients, so that each patient's nursing plan can be dynamically adjusted based on their pathological state, abnormal characteristics and other clinical data, avoiding errors and inadaptability caused by lack of detailed analysis. Based on the generated cardiac state nursing relationship characteristic matrix, the pathological pattern cardiac state nursing relationship characteristic matrix was generated through pathological pattern identification processing, which further optimized and refined the nursing plan, accurately calibrated the relationship between different pathological patterns and nursing needs, and provided a detailed basis for subsequent nursing prediction and intervention. The key to this processing is that it can accurately classify the nursing needs of heart patients based on historical data and disease patterns.Through this identification processing, the system can not only identify the specific nursing needs of patients according to their different pathological patterns, but also dynamically adjust the nursing plan according to the different pathological patterns, so that the accuracy of nursing decisions is further improved. Based on the characteristic matrix of the nursing relationship between the pathological pattern and the heart state, the cardiology nursing knowledge base construction module designs an optimized prediction model for the derivation of cardiac pathological patterns and nursing information. This optimization model can continuously optimize the derivation process of the nursing plan by analyzing a large amount of historical pathological data, nursing data and feedback information, ensuring that each patient can receive the most scientific and most suitable care for their pathological state. The design of this optimized prediction model, especially the optimization combined with the deep learning network model, enables the system to continuously self-learn and evolve according to clinical data to generate more accurate nursing decisions. In this way, the system can quickly and accurately derive the most appropriate nursing plan under various complex pathological conditions, thereby greatly improving the efficiency and accuracy of nursing decisions. In addition, an intelligent feedback engine is designed based on the optimized nursing derivation prediction model, which can monitor the changes in the patient's condition in real time during the nursing process, and adjust and optimize the nursing plan according to the new data. It not only responds to the changes in the patient's condition in real time, but also helps nursing staff to promptly discover potential nursing problems and intervene, effectively improving the accuracy of nursing work and the treatment effect of patients. Through the design of the intelligent feedback engine, the cardiology nursing knowledge base can continuously optimize and improve itself in practical applications. With the input of more and more clinical data and nursing feedback, the system can not only continuously optimize its nursing derivation prediction model, but also flexibly adjust the nursing strategy according to the specific conditions of different patients. This capability makes the knowledge base not only a static reference tool, but also a dynamic and intelligent decision support system, which can improve its application effect and accuracy in the long-term use process. Through the collaborative work of all the above modules, a personalized nursing plan is tailored for each patient, providing a refined nursing intervention plan. Regardless of the patient's pathological state, the system can accurately derive the nursing plan that best meets the patient's needs based on individualized data analysis, and dynamically adjust it during the implementation process. This is the ability of personalized recommendation and real-time adjustment.

[0057] Therefore, the construction system for cardiology nursing knowledge base of the present invention can construct a cardiology nursing knowledge base based on the deep learning ability of big data and artificial intelligence technology, so as to accurately learn cardiac nursing knowledge, thereby intelligently deriving cardiac nursing information related to cardiac pathology. In addition, the correlation between cardiac pathology data and clinical nursing data is deeply mined to ensure that the characteristics of cardiac nursing related knowledge data are fully utilized. The problem of accurate matching of cardiac pathological status and nursing intervention measures can be effectively solved, so that the nursing plan conforms to the pathological status and ensures the quality of nursing. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 A schematic diagram of module functions of a system for constructing a cardiology nursing knowledge base according to the present invention;

[0059] Figure 2 for Figure 1 Detailed implementation functional flow chart of the cardiology nursing knowledge base construction module;

[0060] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0061] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0062] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0063] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0064] To achieve this, please refer to Figure 1 to Figure 2 The present invention provides a system for constructing a cardiology nursing knowledge base. In the embodiments of the present invention, please refer to Figure 1 FIG. 1 is a schematic diagram of module functions of a system for constructing a cardiology nursing knowledge base according to the present invention. The system for constructing a cardiology nursing knowledge base includes the following modules:

[0065] S1: a cardiac pathology diagnosis feature analysis module, which is used to obtain cardiac pathology data and corresponding cardiac care data from a medical database; perform cardiac pathology data analysis on the cardiac pathology data to generate cardiac pathology analysis data, wherein the cardiac pathology analysis data includes cardiac pathology diagnosis data, cardiac pathology physiological index data, and cardiac pathology imaging data; perform cardiac pathology diagnosis feature extraction on the cardiac pathology diagnosis data to generate cardiac pathology diagnosis feature data;

[0066] In an embodiment of the present invention, by calling a standardized medical database interface, heart pathology-related data are extracted, including but not limited to heart imaging data (such as heart CT, heart MRI images), pathology reports, clinical diagnosis data (such as heart disease type, lesion site, course of disease, pathological characteristics, etc.) and patient-related historical health records. These data will be structured and stored in a prescribed format, and associated according to information such as patient identity, diagnosis and treatment time, and treatment plan to ensure data integrity and accuracy. At the same time, heart care data related to heart pathology data are extracted, including nursing records, heart disease nursing effects, patient recovery status and other information, to provide comprehensive data support for the nursing process. By docking with the electronic medical record system within the hospital, formatted heart pathology data (such as electrocardiogram, imaging data, genomic data, etc.) and corresponding patient heart care data (such as nursing cycle, nursing measures, nursing effect evaluation, etc.) are extracted and stored in a specified database. The data extraction process is completed through SQL query and API interface. The modal type analysis of the heart pathology data extracted from the medical database is performed using a data analysis algorithm. By analyzing various data sources in cardiac pathology data (such as imaging data, clinical diagnosis data, genetic data, etc.), the data is classified, labeled and converted to generate corresponding cardiac pathology modal data. By applying data parsing algorithms, such as convolutional neural network CNN or recurrent neural network RNN, image data is recognized, feature extracted and classified, and key lesion features in cardiac pathology data are identified. At the same time, the text modality data is parsed by natural language processing technology to extract clinical information and disease manifestations that are helpful for pathological analysis, and structured cardiac disease analysis data is generated to facilitate subsequent clinical decision-making and nursing support. Convolutional neural network (CNN) in deep learning is used to train and identify cardiac imaging data, automatically detect lesion areas in cardiac images and generate lesion classification reports. Text data (such as pathology reports) are segmented, and named entity recognition (NER) technology is used to extract key information such as disease name, lesion location, and pathological diagnosis, and finally structured pathology analysis data is formed. Text encoding is performed on cardiac pathology diagnosis data, and the parsed pathological data (such as pathological description and diagnosis results) are converted into a unified encoding format. Text coding uses standardized medical coding systems, such as ICD-10 (International Classification of Disease Code) or CPT (Current Procedural Terminology Code) to ensure compatibility and consistency of data between different systems. The coding results are stored as corresponding diagnostic data tables to support subsequent data analysis and automated processing. Based on the analysis of text coding data, key features in cardiac pathology diagnosis are extracted, such as the severity of the disease, the cause of the disease, and the response to treatment. Through data mining techniques (such as association rule mining, feature selection algorithms, etc.), various features in the diagnostic text are extracted and optimized to generate standardized cardiac pathology diagnostic feature data.These feature data can effectively support the accurate diagnosis of heart disease, the formulation of personalized treatment plans, and the support of nursing decisions. Feature extraction is performed on the text coding data of cardiac pathology diagnosis. Features that have a significant impact on cardiac pathology diagnosis (such as the onset time of heart disease, complications, patient age, etc.) are extracted from the coded data through feature selection algorithms in machine learning (such as random forest or L1 regularized logistic regression), and standardized feature data is generated.

[0067] S2: A fuzzy logic analysis module for potential cardiac pathological abnormality characteristics, which is used to perform fuzzy logic analysis and processing of potential cardiac pathological abnormality characteristics based on cardiac pathological physiological index data and cardiac pathological imaging data, and generate fuzzy logic data of potential cardiac pathological abnormality characteristics;

[0068] In an embodiment of the present invention, abnormal detection and analysis are performed on cardiac pathological and physiological index data (such as blood pressure, heart rate, cardiac output, blood oxygen saturation, etc.) to identify cardiac pathological state characteristics that are different from normal physiological states. Statistical methods or machine learning models (such as support vector machines, decision trees, or clustering algorithms) are used to detect outliers and analyze trends on the data. By comparing with the normal range, it is determined which changes in indicators indicate abnormal cardiac pathological states. By training historical patient data using a support vector machine (SVM) model, physiological characteristics that exceed the normal range are identified and marked as abnormal data. For abnormal indicators, they are classified according to severity and recorded as abnormal physiological index characteristic data of cardiac pathological states for subsequent analysis and early warning. By deeply analyzing the cardiac pathological and physiological index data, abnormal physiological characteristics related to specific pathological conditions are further extracted from them. The data is deeply analyzed using time series analysis, abnormal pattern recognition, feature engineering and other technologies to extract features that may indicate changes in cardiac lesions or pathological states. By using image processing and computer vision technology, the cardiac pathological image data is divided and segmented to accurately identify key areas in the cardiac image, such as the ventricles, atria, coronary arteries and other functional structures. Image segmentation algorithms (such as U-Net or graph convolutional network GCN) are used to process cardiac CT, MRI images or cardiac ultrasound images to accurately divide the various regions of the heart. The segmented image data provides a basis for subsequent functional structure analysis. For cardiac CT image data, the U-Net deep learning network is used to automatically segment the cardiac region. The segmented cardiac pathological image data is further analyzed to identify abnormal features related to the functional structure of the heart (such as ventricular wall thickening, ventricular dilatation, valve abnormalities, etc.). Image analysis methods in computer vision, such as morphological analysis, edge detection, regional growing and other algorithms, are used to extract the geometric features of the functional structure (such as area, thickness, volume, etc.) and compare them with the normal range to generate image abnormality feature data of the functional structure of the heart. By comparing the difference between the patient's ventricular wall thickness and the normal range, it is possible to identify whether there is abnormality of ventricular hypertrophy or ventricular dilatation. If abnormalities are found, these features (such as changes in ventricular volume and excessive ventricular wall thickness) are recorded as abnormal cardiac functional structural imaging feature data. The abnormal cardiac pathological physiological feature data and abnormal cardiac functional structural imaging feature data are comprehensively analyzed through fuzzy logic reasoning methods. The fuzzy logic system converts the abnormal features of cardiac pathological data into fuzzy sets to reason about complex physiological and imaging data, thereby identifying potential abnormal cardiac pathological features. This process can fuzzify the data according to the set rules and thresholds, generate potential abnormal cardiac pathological features fuzzy logic data, and provide accurate prediction and diagnosis support for the cardiology nursing knowledge base.Combined with the data of abnormal cardiac pathology and physiological characteristics (such as abnormal blood pressure and heart rate) and abnormal cardiac functional structural imaging characteristics (such as ventricular dilatation and ventricular wall thickening), fuzzy logic reasoning rules are applied to process the data. Through fuzzy reasoning calculation, fuzzy logic data of potential abnormal cardiac pathology characteristics are obtained, that is, quantitative information of potential abnormal cardiac pathology and ventricular dilatation is converted, and risk scores and early warning information of potential cardiac lesions are output.

[0069] S3: A cardiac state nursing relationship feature analysis module, which is used to use cardiac pathology diagnosis feature data and potential cardiac pathology abnormality feature fuzzy logic data as cardiac state entity nodes and cardiac nursing data as cardiac nursing entity nodes, and to perform a mapping relationship analysis between cardiac state and nursing knowledge based on the cardiac state entity nodes and cardiac nursing entity nodes to generate a cardiac state nursing relationship feature matrix; to perform pathological pattern identification processing of cardiac state nursing relationship features on the cardiac state nursing relationship feature matrix to generate a pathological pattern cardiac state nursing relationship feature matrix;

[0070] In an embodiment of the present invention, cardiac pathological diagnosis feature data and potential cardiac pathological abnormality feature fuzzy logic data are first used as entity nodes of cardiac state and associated with cardiac care data as cardiac care entity nodes. A mapping relationship between cardiac state and cardiac care data is established using a graph database or a relational database. A graph analysis technique or an association rule algorithm is used to analyze the correlation between cardiac pathological state and nursing measures, and a cardiac state nursing relationship feature matrix is ​​generated. The matrix can intuitively display the association between cardiac state and nursing process, and help build a personalized nursing plan. The diagnostic features (such as heart disease type, course of disease, complications, etc.) and potential cardiac pathological abnormality feature fuzzy logic data (such as heart failure risk, myocardial ischemia, etc.) extracted from cardiac pathological data are respectively used as entity nodes of cardiac state; nursing measures (such as drug management, electrocardiogram monitoring, cardiac rehabilitation, etc.) extracted from cardiac care data are used as entity nodes of cardiac care. These nodes are connected through a graph database, and association rule analysis (such as Apriori algorithm) is used to identify the association between cardiac pathological state and nursing measures, and a mapping relationship matrix is ​​generated to represent the relationship strength and frequency between cardiac pathological state and corresponding nursing measures. Cluster analysis methods (such as K-means clustering or hierarchical clustering) are used to classify cardiac pathological diagnostic feature data to identify different types of cardiac pathological states. Each cluster represents a typical cardiac pathological state, and the results of clustering provide a subdivision of the disease for clinical diagnosis. The goal of cluster analysis is to divide patients into multiple groups based on the similarity of cardiac pathological diagnostic features, which is convenient for subsequent cardiac pathological pattern analysis and personalized care plan formulation. The clustering results divide patients into different pathological groups, such as heart failure, coronary heart disease, cardiomyopathy, etc., to generate cardiac pathological diagnostic cluster data for subsequent analysis. According to the cardiac pathological diagnosis cluster data of the previous step, the correlation analysis between cardiac pathological diagnostic features is performed. Statistical analysis methods (such as Pearson correlation coefficient, Spearman rank correlation coefficient or mutual information analysis) are used to quantitatively evaluate the relationship between different cardiac pathological diagnosis clusters to identify which diagnostic features have a significant impact on the judgment of cardiac pathological states. The analysis results help to reveal the internal correlation of cardiac pathological states and provide theoretical support for subsequent pathological pattern analysis and personalized care. The Pearson correlation coefficient method is used to perform correlation analysis on the cardiac pathological diagnosis cluster data. The key features in the cardiac pathology clustering (such as left ventricular ejection fraction, coronary artery obstruction, heart disease history, etc.) were selected to calculate the correlation between these features. By calculating the correlation coefficient, the linear relationship between the pathological features was obtained. For example, it was found that the ventricular ejection fraction was highly correlated with the degree of coronary artery obstruction. These correlation information were organized into cardiac pathology diagnosis correlation data for subsequent analysis.Pattern recognition techniques (such as principal component analysis (PCA), factor analysis, or pattern recognition algorithms based on neural networks) are used to perform pattern recognition on cardiac pathology diagnosis-related data, extract core patterns related to cardiac pathological states, better understand the intrinsic relationship between different pathological characteristics, and establish a quantitative model of cardiac pathological states to identify the common characteristics and laws of different types of cardiac pathological states. Based on cardiac pathological pattern data, the cardiac state nursing relationship feature matrix is ​​identified and processed, and the identified pathological patterns are matched with the corresponding relationships in the cardiac state nursing relationship feature matrix. According to different pathological patterns, the mapping relationship between cardiac pathological states and nursing measures is adjusted to ensure a high degree of matching between nursing measures and cardiac pathological states. The pathological pattern cardiac state nursing relationship feature matrix finally generated can provide personalized nursing plans for different types of cardiac pathological states more accurately. By using the identified cardiac pathological pattern data (such as coronary heart disease, heart failure patterns, etc.), corresponding nursing measures (such as drug treatment, lifestyle adjustment, rehabilitation exercises, etc.) are specified for each pathological pattern.

[0071] S4: Cardiology nursing knowledge base construction module, used to design an optimized prediction model for deriving cardiac pathology patterns and nursing information based on the pathology pattern heart state nursing relationship feature matrix, and generate an optimized cardiac nursing derivation prediction model; based on the optimized cardiac nursing derivation prediction model, an intelligent feedback engine design and processing of the cardiology nursing knowledge base is performed to build an intelligent cardiology nursing knowledge base engine.

[0072] In an embodiment of the present invention, the pathological pattern heart state nursing relationship feature matrix is ​​used as input to perform association analysis between the pathological pattern and the heart nursing measures. An association rule mining algorithm (such as an Apriori algorithm or an FP-growth algorithm) is used to analyze the strong correlation between the pathological pattern and the nursing measures and the frequently occurring nursing measures combination, generate pathological pattern heart nursing association data, discover the common nursing measures pattern under different cardiac pathological states, and further support the formulation of personalized nursing plans. By deeply analyzing the pathological pattern heart nursing association data, the unique nursing measures characteristics under each pathological pattern are identified. Through data analysis methods such as differential analysis, information gain analysis, or feature selection based on a tree model, nursing features closely related to a specific pathological pattern are screened out. The generated pathological pattern heart nursing specific data can reflect the unique or special nursing measures under a specific pathological pattern, which is convenient for customizing personalized nursing interventions. The pathological pattern heart nursing specific data is analyzed to extract the common nursing features and differential nursing features between different pathological patterns. Common features refer to nursing measures that are common under multiple pathological patterns, while differential features refer to unique nursing needs under a specific pathological pattern. Through statistical and comparative analysis of nursing characteristics, the common characteristic data of cardiac nursing in pathological patterns and the differential characteristic data of cardiac nursing in pathological patterns are generated respectively. According to the differential characteristic data of cardiac nursing in pathological patterns, an intervention contribution weight is assigned to each nursing characteristic. Weights are assigned to the differential characteristics of nursing through expert evaluation, regression analysis or machine learning algorithms (such as weighted regression, decision tree, etc.) to quantify the contribution of each characteristic to the treatment effect. Through these weights, the relative importance of different nursing measures can be more accurately reflected in the nursing plans of different pathological patterns, so as to provide more targeted nursing strategies for clinicians. Expert evaluation combined with regression analysis is used to assign weights to the differential characteristics of nursing for heart failure patients (such as fluid management, electrolyte monitoring, etc.). Fuzzy logic data of potential abnormal cardiac pathological characteristics are used to analyze the influencing factors of differential cardiac nursing characteristics. The fuzzy logic system converts potential abnormal cardiac pathological characteristics into fuzzy sets, combines the differential characteristics of nursing, analyzes the influencing factors under the pathological pattern through the inference engine, and generates the influencing factor data of differential cardiac nursing in pathological patterns. This process helps to understand the complex relationship between cardiac nursing measures and pathological characteristics, and provides a basis for personalized care. The fuzzy logic reasoning system was used to analyze the influencing factors of the nursing difference characteristics (such as fluid management and electrolyte balance) of heart failure patients. The influencing factor data of cardiac nursing difference in pathological mode were used to analyze the intervention characteristics of cardiac nursing difference empowerment characteristics. The intervention effect of influencing factors on nursing difference characteristics was analyzed through machine learning methods (such as decision trees, neural networks, etc.), so as to generate the intervention characteristic data of cardiac nursing difference in pathological mode. This step helps to clarify which nursing intervention measures can effectively adjust or improve the nursing difference characteristics under the pathological mode.The weighted feature data of cardiac nursing differences in pathological patterns are analyzed by neural networks to determine the adjustment effects of various intervention features (such as fluid management, electrolyte monitoring, etc.) under different pathological patterns. The preset deep learning networks (such as CNN, RNN, etc.) are trained using the common feature data of cardiac nursing in pathological patterns and the differential intervention feature data of cardiac nursing in pathological patterns to design an optimized cardiac nursing derivation prediction model. The model can intelligently predict the most suitable nursing intervention plan based on the patient's pathological pattern. The model is optimized through cross-validation, hyperparameter adjustment and other technologies to ensure the accuracy and reliability of the prediction results. Based on the optimized cardiac nursing derivation prediction model, an intelligent feedback engine is designed and implemented. The engine can receive clinical data, pathological status and nursing records in real time, predict the nursing intervention measures required by the patient through the model, and feed it back to clinical nursing staff. The feedback engine can also continuously optimize the nursing derivation model based on historical data to ensure the continuous updating and adaptability of the knowledge base. An intelligent feedback engine system is designed to integrate the optimized cardiac nursing derivation prediction model with the hospital information system (HIS). Whenever a new patient is admitted to the hospital, the system automatically inputs the pathological pattern data, derives the most suitable nursing measures through the feedback engine, and provides real-time feedback to the nursing staff. Through the feedback mechanism, not only personalized care is provided to patients, but also the nursing decision model is continuously improved and optimized, forming an intelligent cardiology nursing knowledge base engine.

[0073] Furthermore, the cardiac pathology diagnosis feature analysis module includes the following functions:

[0074] Obtain cardiac pathology data and corresponding cardiac care data from medical databases;

[0075] Perform data modality type analysis based on cardiac pathology data to generate cardiac pathology modality data;

[0076] Performing cardiac pathology data analysis and processing on the cardiac pathology data according to the cardiac pathology modality data to generate cardiac pathology analysis data;

[0077] Performing text coding processing on the diagnostic data according to the cardiac pathology diagnostic data to generate cardiac pathology diagnostic text coding data;

[0078] The cardiac pathology diagnosis feature extraction process is performed based on the cardiac pathology diagnosis text coding data to generate cardiac pathology diagnosis feature data.

[0079] In an embodiment of the present invention, by calling a standardized medical database interface, heart pathology-related data are extracted, including but not limited to heart imaging data (such as heart CT, heart MRI images), pathology reports, clinical diagnosis data (such as heart disease type, lesion site, course of disease, pathological characteristics, etc.) and patient-related historical health records. These data will be structured and stored in a prescribed format, and associated according to information such as patient identity, diagnosis and treatment time, and treatment plan to ensure data integrity and accuracy. At the same time, heart care data related to heart pathology data are extracted, including nursing records, heart disease nursing effects, patient recovery status and other information, to provide comprehensive data support for the nursing process. By docking with the electronic medical record system within the hospital, formatted heart pathology data (such as electrocardiogram, imaging data, genomic data, etc.) and corresponding patient heart care data (such as nursing cycle, nursing measures, nursing effect evaluation, etc.) are extracted and stored in a specified database. The data extraction process is completed through SQL query and API interface. The modal type analysis of the heart pathology data extracted from the medical database is performed using a data analysis algorithm. By analyzing various data sources in cardiac pathology data (such as imaging data, clinical diagnosis data, genetic data, etc.), the data is classified, labeled and converted to generate corresponding cardiac pathology modal data. These modal data include text, image, audio or video formats, and will be encoded according to unified standards for subsequent processing. The extracted cardiac imaging data is analyzed using data processing algorithms (such as clustering analysis algorithms), and cardiac CT images, electrocardiograms and other data are classified and labeled to convert them into a unified image data modality. At the same time, natural language processing technology is applied to text data (such as pathology reports and diagnostic texts) to extract valuable clinical information and convert it into a standardized text format modality. All modal data will be saved as structured data tables and labeled with their data types and sources. The cardiac pathology modal data generated in the previous step is further parsed. By applying data parsing algorithms, such as convolutional neural networks (CNN) or recurrent neural networks (RNN), image data is subjected to image recognition, feature extraction and classification analysis to identify key lesion features in cardiac pathology data. At the same time, the text modal data is parsed using natural language processing technology to extract clinical information and disease manifestations that are helpful for pathological analysis, and structured cardiac disease analysis data is generated to facilitate subsequent clinical decision-making and nursing support. The convolutional neural network (CNN) in deep learning is used to train and identify cardiac imaging data, automatically detect lesion areas in cardiac images and generate lesion classification reports. The text data (such as pathology reports) is segmented, and named entity recognition (NER) technology is used to extract key information such as disease name, lesion location, and pathological diagnosis, and finally structured pathological analysis data is formed. The cardiac pathological diagnosis data is text-encoded, and the parsed pathological data (such as pathological descriptions and diagnostic results) are converted into a unified coding format.Text coding uses standardized medical coding systems, such as ICD-10 (International Classification of Diseases) or CPT (Current Procedural Terminology) to ensure compatibility and consistency of data between different systems. In addition, coding processing also includes standardizing the diagnosis text so that the system can automatically identify and classify information such as disease type, cause, and degree of lesion. For text data extracted from pathology reports, the ICD-10 coding system is used to standardize the coding of pathological diagnoses. First, the disease name (such as myocardial infarction) in the diagnosis result is automatically matched and coded, and then the pathological characteristics (such as lesion location, pathological type, etc.) are converted into standardized coding data. The coding results are stored as corresponding diagnostic data tables to support subsequent data analysis and automated processing. Based on the analysis of text coding data, key features in cardiac pathology diagnosis, such as disease severity, cause of disease, treatment response, etc., are extracted. Through data mining techniques (such as association rule mining, feature selection algorithms, etc.), various features in the diagnosis text are extracted and optimized to generate standardized cardiac pathology diagnosis feature data. These feature data can effectively support the accurate diagnosis of heart disease, the formulation of personalized treatment plans, and the support of nursing decisions. Feature extraction is performed on cardiac pathology diagnosis text coding data. Features that have a significant impact on the diagnosis of cardiac pathology (such as the onset of heart disease, complications, patient age, etc.) are extracted from the coded data through feature selection algorithms in machine learning (such as random forest or L1 regularized logistic regression), and standardized feature data are generated. These feature data will be used as input for subsequent data analysis and decision support, providing a reference for clinical care.

[0080] Furthermore, the potential cardiac pathological abnormality feature fuzzy logic analysis module includes the following functions:

[0081] Analyze and process abnormal physiological characteristics of cardiac pathological state according to cardiac pathological and physiological index data, and generate abnormal physiological index characteristic data of cardiac pathological state;

[0082] Analyze abnormal cardiac pathological physiological characteristics according to cardiac pathological and physiological index data to generate abnormal cardiac pathological physiological characteristic data;

[0083] Performing pathological image segmentation processing on cardiac pathological image data to divide cardiac functional structural regions, generating segmented cardiac pathological image data; performing image abnormality feature analysis on cardiac functional structure based on the segmented cardiac pathological image data, generating cardiac functional structure image abnormality feature data;

[0084] Fuzzy logic conversion processing of potential cardiac pathological abnormality characteristics is performed based on cardiac pathological abnormality physiological characteristic data and cardiac functional structural imaging abnormality characteristic data to generate potential cardiac pathological abnormality characteristic fuzzy logic data.

[0085] In an embodiment of the present invention, abnormal detection and analysis are performed on cardiac pathological and physiological index data (such as blood pressure, heart rate, cardiac output, blood oxygen saturation, etc.) to identify cardiac pathological state characteristics that are different from normal physiological states. Statistical methods or machine learning models (such as support vector machines, decision trees, or clustering algorithms) are used to detect outliers and analyze trends on the data. By comparing with the normal range, it is determined which changes in indicators indicate abnormal cardiac pathological states. The detected abnormal features are extracted and stored to form structured abnormal physiological index feature data for further cardiac pathological analysis and clinical decision support. For cardiac pathological and physiological index data, such as heart rate and blood pressure, the normal value range of each indicator is first determined (for example: heart rate is between 60-100 beats / min and blood pressure is 120 / 80 mmHg). By using a support vector machine (SVM) model to train historical patient data, physiological characteristics that exceed the normal range are identified and marked as abnormal data. For abnormal indicators, they are classified according to severity and recorded as abnormal physiological index feature data of cardiac pathological states for subsequent analysis and early warning. By deeply analyzing the data of cardiac pathological and physiological indicators, abnormal physiological features related to specific pathological conditions are further extracted from them. Time series analysis, abnormal pattern recognition, feature engineering and other technologies are used to deeply analyze the data and extract features that may indicate changes in cardiac lesions or pathological states. These features are organized and output as structured abnormal physiological feature data, including information such as the change trend, frequency, and amplitude of relevant indicators, for subsequent diagnosis and nursing processes. For a patient's continuous electrocardiogram (ECG) signal, the waveform features (for example, the duration and interval of the QRS complex) in the ECG signal are extracted through time series analysis methods (such as spectrum analysis based on Fourier transform), and changes in physiological indicators related to heart disease such as heart rate and QT interval are extracted. If the waveform fluctuations in the ECG exceed the normal range, or the frequency and amplitude change significantly, these abnormal features are identified as abnormal physiological feature data of cardiac pathology and recorded. Image processing and computer vision technology are used to divide and segment cardiac pathological image data, and key areas in cardiac images, such as ventricles, atria, coronary arteries and other functional structures, are accurately identified. Use image segmentation algorithms (such as U-Net or graph convolutional network GCN) to process cardiac CT, MRI images or cardiac ultrasound images to accurately divide the various areas of the heart. The segmented image data provides a basis for subsequent functional structural analysis. For cardiac CT image data, the U-Net deep learning network is used to automatically segment the heart area. The algorithm first preprocesses the image data to eliminate noise and enhance image quality. Then, the functional structures of the heart (such as atria, ventricles, valves, coronary arteries, etc.) are accurately segmented through the trained U-Net model. The segmented image data is output as the coordinates and boundary data of each structural area for subsequent abnormal feature analysis.The segmented cardiac pathological image data are further analyzed to identify abnormal features related to the functional structure of the heart (such as ventricular wall thickening, ventricular dilatation, valve abnormalities, etc.). Image analysis methods in computer vision, such as morphological analysis, edge detection, and region growing algorithms, are used to extract the geometric features of the functional structure (such as area, thickness, volume, etc.) and compare them with the normal range to generate abnormal imaging feature data of the functional structure of the heart. These abnormal feature data can provide a quantitative description of abnormal cardiac pathology and assist clinical diagnosis. For the segmented ventricular region, morphological operations (such as expansion, corrosion, etc.) are applied to the region for edge detection, the ventricular boundary is extracted, and the thickness of the ventricular wall is calculated. By comparing the difference between the patient's ventricular wall thickness and the normal range, it is identified whether there is abnormal ventricular hypertrophy or ventricular dilatation. If abnormalities are found, these features (such as ventricular volume changes, ventricular wall thickness exceeding the standard) are recorded as abnormal imaging feature data of the functional structure of the heart. The fuzzy logic reasoning method is used to comprehensively analyze the abnormal physiological feature data of cardiac pathology and the abnormal imaging feature data of the functional structure of the heart. The fuzzy logic system converts the abnormal characteristics of cardiac pathology data into fuzzy sets to infer complex physiological and imaging data, thereby identifying potential abnormal characteristics of cardiac pathology. This process can fuzzify the data according to the set rules and thresholds to generate fuzzy logic data of potential abnormal characteristics of cardiac pathology, providing accurate prediction and diagnosis support for the cardiology nursing knowledge base. Combining the physiological characteristic data of abnormal cardiac pathology (such as abnormal blood pressure and abnormal heart rate) and the abnormal characteristic data of cardiac functional structural imaging (such as ventricular dilatation and ventricular wall thickening), the fuzzy logic reasoning rules are applied to process the data. The fuzzy logic data of potential abnormal characteristics of cardiac pathology is obtained through fuzzy reasoning calculation, that is, the quantitative information of potential abnormal cardiac pathology and ventricular dilatation is converted, and the risk score and early warning information of potential cardiac lesions are output.

[0086] Furthermore, the cardiac state nursing relationship characteristic analysis module includes the following functions:

[0087] The cardiac pathology diagnosis feature data and the potential cardiac pathology abnormality feature fuzzy logic data are used as cardiac state entity nodes and the cardiac care data are used as cardiac care entity nodes. The mapping relationship between cardiac state and nursing knowledge is analyzed based on the cardiac state entity nodes and cardiac care entity nodes to generate a cardiac state nursing relationship feature matrix.

[0088] Performing cardiac pathology diagnosis cluster analysis on cardiac pathology diagnosis feature data to generate cardiac pathology diagnosis cluster data;

[0089] Conducting cardiac pathology diagnosis correlation analysis based on cardiac pathology diagnosis clustering data to generate cardiac pathology diagnosis correlation data;

[0090] Perform cardiac pathology pattern analysis based on cardiac pathology diagnosis correlation data to generate cardiac pathology pattern data;

[0091] Based on the cardiac pathological pattern data, the cardiac state nursing relationship feature matrix is ​​processed with the pathological pattern identification of the cardiac state nursing relationship features to generate a pathological pattern cardiac state nursing relationship feature matrix.

[0092] In an embodiment of the present invention, cardiac pathological diagnosis feature data and potential cardiac pathological abnormality feature fuzzy logic data are first used as entity nodes of cardiac state and associated with cardiac care data as cardiac care entity nodes. A mapping relationship between cardiac state and cardiac care data is established using a graph database or a relational database. A graph analysis technique or an association rule algorithm is used to analyze the correlation between cardiac pathological state and nursing measures, and a cardiac state nursing relationship feature matrix is ​​generated. The matrix can intuitively display the association between cardiac state and nursing process, and help build a personalized nursing plan. The diagnostic features (such as heart disease type, course of disease, complications, etc.) and potential cardiac pathological abnormality feature fuzzy logic data (such as heart failure risk, myocardial ischemia, etc.) extracted from cardiac pathological data are respectively used as entity nodes of cardiac state; nursing measures (such as drug management, electrocardiogram monitoring, cardiac rehabilitation, etc.) extracted from cardiac care data are used as entity nodes of cardiac care. These nodes are connected through a graph database, and association rule analysis (such as Apriori algorithm) is used to identify the association between cardiac pathological state and nursing measures, and a mapping relationship matrix is ​​generated to represent the relationship strength and frequency between cardiac pathological state and corresponding nursing measures. Cluster analysis methods (such as K-means clustering or hierarchical clustering) are used to classify cardiac pathological diagnostic feature data to identify different types of cardiac pathological states. Each cluster represents a typical cardiac pathological state, and the clustering results provide a subdivision of the disease for clinical diagnosis. The goal of cluster analysis is to divide patients into multiple groups based on the similarity of cardiac pathological diagnostic features, which is convenient for subsequent cardiac pathological pattern analysis and personalized care plan formulation. The K-means clustering algorithm is used to perform cluster analysis on cardiac pathological diagnostic feature data (such as cardiac imaging features, clinical symptoms, blood test results, etc.). First, the number of cluster categories k is determined, and the optimal k value is determined by the elbow rule. Then, the cardiac pathological diagnostic feature data of all patients are normalized and input into the K-means algorithm for cluster analysis. Finally, the clustering results divide the patients into different pathological groups, such as heart failure, coronary heart disease, cardiomyopathy, etc., and generate cardiac pathological diagnostic cluster data for subsequent analysis. According to the cardiac pathological diagnostic cluster data of the previous step, the correlation analysis between cardiac pathological diagnostic features is performed. Statistical analysis methods (such as Pearson correlation coefficient, Spearman rank correlation coefficient or mutual information analysis) are used to quantitatively evaluate the relationship between different cardiac pathology diagnosis clusters to identify which diagnostic features have a significant impact on the judgment of cardiac pathological status. The results of this analysis help to reveal the internal correlation of cardiac pathological status and provide theoretical support for subsequent pathological pattern analysis and personalized care. The Pearson correlation coefficient method is used to perform correlation analysis on cardiac pathology diagnosis cluster data. The key features in the cardiac pathology cluster (such as left ventricular ejection fraction, coronary artery obstruction, history of heart disease, etc.) are selected to calculate the correlation between these features.By calculating the correlation coefficient, the linear relationship between each pathological feature is obtained. For example, it is found that the ventricular ejection fraction is highly correlated with the degree of coronary artery obstruction. These correlation information are organized into cardiac pathology diagnosis correlation data for subsequent analysis. Pattern recognition techniques (such as principal component analysis PCA, factor analysis, or pattern recognition algorithms based on neural networks) are used to perform pattern recognition on cardiac pathology diagnosis correlation data, extract the core patterns related to cardiac pathological states, better understand the intrinsic relationship between different pathological features, and establish a quantitative model of cardiac pathological states to identify the common characteristics and laws of different types of cardiac pathological states. Principal component analysis (PCA) is used to reduce the dimension of cardiac pathology diagnosis correlation data and identify the main characteristic patterns related to cardiac pathological states. First, the correlation data is standardized, and then the PCA algorithm is used to compress the multidimensional features into several main principal components. By analyzing the principal components, the core patterns in cardiac pathology diagnosis, such as the relevant characteristic patterns of coronary heart disease and heart failure, can be identified, and cardiac pathology pattern data can be generated for further analysis. Based on the cardiac pathology pattern data, the cardiac state nursing relationship feature matrix is ​​identified, and the identified pathological patterns are matched with the corresponding relationships in the cardiac state nursing relationship feature matrix. According to different pathological patterns, the mapping relationship between cardiac pathological states and nursing measures is adjusted to ensure a high degree of matching between nursing measures and cardiac pathological states. The finally generated pathological pattern cardiac state nursing relationship feature matrix can provide personalized nursing plans for different types of cardiac pathological states more accurately. By using the identified cardiac pathological pattern data (such as coronary heart disease, heart failure pattern, etc.), corresponding nursing measures (such as drug therapy, lifestyle adjustment, rehabilitation exercise, etc.) are specified for each pathological pattern. Combined with the aforementioned generated cardiac state nursing relationship feature matrix, the pathological pattern and nursing measures are matched through fuzzy logic or rule engine to form a pathological pattern cardiac state nursing relationship feature matrix. For example, the nursing measures corresponding to the coronary heart disease pattern may include drug therapy and regular monitoring, while the heart failure pattern focuses more on fluid management and cardiopulmonary function training.

[0093] Further, as an embodiment of the present invention, refer to Figure 2 As shown, Figure 1 The detailed implementation function flow diagram of the cardiology nursing knowledge base construction module is shown in the figure. In this embodiment, the cardiology nursing knowledge base construction module includes the following functions:

[0094] S401: performing a pathological pattern heart care associated data analysis on the pathological pattern heart state care relationship feature matrix to generate pathological pattern heart care associated data;

[0095] In an embodiment of the present invention, the pathological pattern heart state nursing relationship feature matrix is ​​used as input to perform association analysis between the pathological pattern and the heart nursing measures. An association rule mining algorithm (such as the Apriori algorithm or the FP-growth algorithm) is used to analyze the strong correlation between the pathological pattern and the nursing measures and the frequently occurring nursing measures combination, generate pathological pattern heart nursing association data, discover the common nursing measures pattern under different cardiac pathological states, and further support the formulation of personalized nursing plans. By analyzing the pathological pattern heart state nursing relationship feature matrix, the Apriori algorithm is used to mine the association rules between cardiac pathological states (such as heart failure, coronary heart disease, etc.) and nursing measures (such as drug treatment, monitoring, diet adjustment, etc.). For example, if "heart failure mode" and "diuretic use" and "electrolyte balance monitoring" frequently co-occur in multiple cases, the association rule is generated and recorded as pathological pattern heart nursing association data. This data can be used for subsequent nursing measures optimization.

[0096] S402: performing a pathological pattern heart care specific analysis based on the pathological pattern heart care associated data to generate pathological pattern heart care specific data;

[0097] In an embodiment of the present invention, this step identifies the unique nursing measures characteristics under each pathological mode by deeply analyzing the pathological mode heart care associated data. Through data analysis methods, such as differentiation analysis, information gain analysis, or feature selection based on a tree model, nursing features closely related to specific pathological modes are screened out. The generated pathological mode heart care specific data can reflect the unique or special nursing measures under a specific pathological mode, which is convenient for customizing personalized nursing interventions. The pathological mode heart care associated data are analyzed using an information gain algorithm to identify the unique nursing measures characteristics under each pathological mode. Assuming that in the "acute myocardial infarction" mode, patients often need emergency medication and ECG monitoring, while in the "chronic heart failure" mode, they need more fluid management and electrolyte balance regulation. Through feature selection, specific nursing measures related to each pathological mode are extracted, and pathological mode heart care specific data are generated to provide a basis for the design of subsequent nursing strategies.

[0098] S403: extracting common nursing features and differential nursing features of the pathological pattern from the pathological pattern heart nursing specific data, and generating common nursing feature data of the pathological pattern heart nursing and differential nursing feature data of the pathological pattern heart nursing respectively;

[0099] In an embodiment of the present invention, the pathological pattern heart care specific data is analyzed to extract the common nursing features and differential nursing features between different pathological patterns. Common features refer to nursing measures that are common in multiple pathological patterns, while differential features refer to unique nursing needs under specific pathological patterns. Through statistical and comparative analysis of nursing features (such as cluster analysis, variance analysis, etc.), pathological pattern heart care common feature data and pathological pattern heart care differential feature data are generated respectively. The pathological pattern heart care specific data are divided into multiple categories through cluster analysis, and the common nursing features common to "heart failure" and "coronary heart disease" are identified, such as drug therapy (ACE inhibitors, beta-blockers) and ECG monitoring are common features. At the same time, through variance analysis, it is identified that patients with heart failure need to pay extra attention to differential features such as fluid management and cardiopulmonary function training, while patients with coronary heart disease need more vasodilator drug management. Finally, these two types of features are extracted to generate pathological pattern heart care common feature data and pathological pattern heart care differential feature data.

[0100] S404: performing intervention contribution weighting processing on the pathological pattern heart nursing difference characteristic data of the nursing difference characteristics to generate pathological pattern heart nursing difference weighted characteristic data;

[0101] In an embodiment of the present invention, this step assigns an intervention contribution weight to each nursing feature according to the pathological pattern heart nursing difference feature data. Weights are assigned to nursing difference features through expert evaluation method, regression analysis or machine learning algorithm (such as weighted regression, decision tree, etc.) to quantify the contribution of each feature to the treatment effect. Through these weights, the relative importance of different nursing measures can be more accurately reflected in the nursing plans of different pathological patterns, thereby providing more targeted nursing strategies for the clinic. Expert evaluation method combined with regression analysis is used to assign weights to nursing difference features (such as fluid management, electrolyte monitoring, etc.) of heart failure patients. Assume that the importance of fluid management in the treatment of heart failure is 0.4, the importance of electrolyte monitoring is 0.3, and other nursing measures (such as ECG monitoring) are 0.3. By calculating the contribution value of each feature, the pathological pattern heart nursing difference empowerment feature data is finally generated to provide a scientific basis for clinical decision-making.

[0102] S405: analyzing the influencing factors of cardiac nursing differences in pathological patterns based on the fuzzy logic data of potential cardiac pathological abnormalities, and generating data of influencing factors of cardiac nursing differences in pathological patterns;

[0103] In an embodiment of the present invention, fuzzy logic data of potential cardiac pathological abnormalities are used to analyze the influencing factors of cardiac care difference characteristics. The fuzzy logic system converts potential cardiac pathological abnormalities into fuzzy sets, combines nursing difference characteristics, analyzes the influencing factors in the pathological mode through the reasoning engine, and generates pathological mode cardiac care difference influencing factor data. This process helps to understand the complex relationship between cardiac care measures and pathological characteristics, and provides a basis for personalized care. The fuzzy logic reasoning system is used to analyze the influencing factors of the nursing difference characteristics (such as fluid management and electrolyte balance) of heart failure patients. Assuming that fluid management is affected by factors such as the patient's cardiac function and blood pressure changes, fuzzy logic rules (such as "if cardiac function is reduced and blood pressure is low, the priority of fluid management increases") are used to generate pathological mode cardiac care difference influencing factor data to provide a basis for subsequent nursing decisions.

[0104] S406: using the data of influencing factors of cardiac nursing differences in pathological patterns to analyze the cardiac nursing differences intervention characteristics of pathological patterns on the weighted characteristic data of cardiac nursing differences in pathological patterns, and generating cardiac nursing differences intervention characteristic data of pathological patterns;

[0105] In an embodiment of the present invention, the data of influencing factors of cardiac care differences in pathological patterns are used to perform intervention feature analysis on the weighted feature data of cardiac care differences. The intervention effect of influencing factors on the nursing difference features is analyzed by machine learning methods (such as decision trees, neural networks, etc.), thereby generating intervention feature data of cardiac care differences in pathological patterns. This step helps to clarify which nursing intervention measures can effectively adjust or improve the nursing difference features under the pathological pattern. The weighted feature data of cardiac care differences in pathological patterns are analyzed by neural networks to determine the adjustment effect of each intervention feature (such as fluid management, electrolyte monitoring, etc.) under different pathological patterns. Through data training, the impact of intervention measures on the nursing differences in pathological patterns is obtained, and intervention feature data of cardiac care differences in pathological patterns are generated to provide quantitative support for personalized nursing intervention.

[0106] S407: Designing an optimized prediction model for deriving cardiac pathology patterns and nursing information using the preset deep learning network model through the common characteristic data of cardiac nursing of pathology patterns and the differential intervention characteristic data of cardiac nursing of pathology patterns, and generating an optimized cardiac nursing derivation prediction model;

[0107] In an embodiment of the present invention, the common feature data of cardiac care in pathological patterns and the differential intervention feature data of cardiac care in pathological patterns are used to train a preset deep learning network (such as CNN, RNN, etc.) to design an optimized cardiac care derivation prediction model. The model can intelligently predict the most suitable nursing intervention plan based on the patient's pathological pattern. The optimization of the model ensures the accuracy and reliability of the prediction results through cross-validation, hyperparameter adjustment and other technologies. The common feature data of cardiac care in pathological patterns and the differential intervention feature data of cardiac care in pathological patterns are jointly trained using a convolutional neural network (CNN). The model predicts the most effective combination of nursing measures in each pathological pattern by learning the association between these features. Through model training, the cardiac care derivation prediction model is optimized so that it can automatically generate nursing plans according to the patient's pathological state in practical applications.

[0108] S408: Design and process an intelligent feedback engine for the cardiology nursing knowledge base based on the optimized cardiac nursing derivation prediction model to construct an intelligent cardiology nursing knowledge base engine.

[0109] In an embodiment of the present invention, an intelligent feedback engine is designed and implemented based on an optimized cardiac care derivation prediction model. The engine can receive clinical data, pathological conditions, and nursing records in real time, predict the nursing interventions required by the patient through the model, and feed them back to clinical nursing staff. The feedback engine can also continuously optimize the nursing derivation model based on historical data to ensure the continuous updating and adaptability of the knowledge base. An intelligent feedback engine system is designed to integrate the optimized cardiac care derivation prediction model with the hospital information system (HIS). Whenever a new patient is admitted to the hospital, the system automatically inputs the pathological pattern data, derives the most suitable nursing measures through the feedback engine, and feeds back to the nursing staff in real time. Through the feedback mechanism, not only personalized care is provided to patients, but also the nursing decision-making model is continuously improved and optimized to form an intelligent cardiology nursing knowledge base engine.

[0110] Furthermore, the intervention contribution weighting processing of the nursing difference characteristics of the pathological pattern heart nursing difference characteristics data includes:

[0111] Based on the pathological pattern cardiac nursing associated data, the cardiac nursing intervention contribution weight analysis of each pathological pattern is performed to generate the cardiac nursing intervention contribution weight parameters of the pathological pattern;

[0112] According to the pathological pattern heart nursing intervention contribution weight parameter, the pathological pattern heart nursing difference characteristic data is weighted by the intervention contribution of the nursing difference characteristics to generate the pathological pattern heart nursing difference weighted characteristic data.

[0113] In an embodiment of the present invention, the pathological pattern heart care associated data is analyzed to determine the heart care intervention measures and their influence under each pathological pattern. Statistical analysis methods (such as correlation analysis, regression analysis, etc.) are used to quantify the contribution of different nursing intervention measures to each pathological pattern. For example, for the "acute myocardial infarction" pathological pattern, drug intervention (such as thrombolytic drugs) may have a greater impact on the recovery of the pathological pattern, while for the "chronic heart failure" pathological pattern, fluid management and electrolyte balance intervention play a key role in the patient's clinical outcome. Through these analyses, the contribution weight parameters of the heart care intervention under each pathological pattern are generated. Regression analysis is performed on the pathological pattern heart care associated data. Assuming that the pathological pattern is "coronary heart disease", the nursing measures include drug therapy, ECG monitoring, lifestyle intervention, etc. Through the regression model, the influence coefficient of each intervention measure on the course of coronary heart disease patients is obtained. Assuming that the contribution of drug therapy is 0.45, the contribution of ECG monitoring is 0.35, and the contribution of lifestyle intervention is 0.2, these coefficients are the pathological pattern heart care intervention contribution weight parameters. This weight value can be used for subsequent intervention weighting processing. The intervention contribution weight parameters of cardiac nursing intervention in pathological patterns are used to weight the intervention contribution of cardiac nursing difference feature data of pathological patterns. The specific method is to assign an intervention contribution weight to each nursing feature according to the intervention contribution weight parameters of cardiac nursing intervention in pathological patterns. This process weights different nursing features based on the degree of influence of intervention measures to generate weighted feature data of cardiac nursing difference in pathological patterns. Nursing features with higher intervention weights will receive more attention and priority in subsequent nursing plans. Assuming that for the pathological pattern of "heart failure", the intervention contribution weight parameters obtained through weight analysis are: fluid management is 0.4, electrolyte balance is 0.3, ECG monitoring is 0.2, and exercise regulation is 0.1. According to these weights, weighted feature data of cardiac nursing difference in pathological patterns are weighted. For example, the intervention contribution value of fluid management is 0.4, the intervention contribution value of electrolyte balance is 0.3, and ECG monitoring and exercise regulation are 0.2 and 0.1 respectively. Therefore, fluid management has the greatest contribution to the nursing intervention of patients with heart failure, and it is given a higher priority. The finally generated weighted feature data of cardiac nursing difference in pathological patterns can reflect the difference in intervention effect of each nursing feature in different pathological patterns.

[0114] Furthermore, the cardiac care intervention contribution weight analysis of each pathological pattern based on the pathological pattern cardiac care associated data includes:

[0115] Perform independent cardiac care parameter analysis of each pathological pattern according to the cardiac care associated data of the pathological pattern, and generate independent cardiac care parameters of the pathological pattern;

[0116] Performing pathology pattern independent heart care parameter contribution analysis according to the pathology pattern independent heart care parameters to generate pathology pattern independent heart care parameter contribution data;

[0117] According to the independent cardiac nursing parameter contribution data of the pathological pattern, the cardiac nursing intervention contribution weight analysis of each pathological pattern is performed to generate the cardiac nursing intervention contribution weight parameters of the pathological pattern.

[0118] In an embodiment of the present invention, the pathological pattern heart care associated data is analyzed to extract independent heart care parameters under each pathological pattern. These care parameters are usually independent care measures directly related to the pathological pattern, such as drug use, vital signs monitoring, nutritional support, etc. The independent care parameters related to each pathological pattern are determined by statistical analysis and data mining methods, such as correlation analysis or factor analysis. Factor analysis is performed on the pathological pattern heart care associated data to extract independent care parameters directly related to the "acute myocardial infarction" pattern. Assume that the analysis results show that ECG monitoring, antithrombotic drug use, and pain management are independent care parameters of this pattern. Similarly, for the "chronic heart failure" pattern, parameters such as fluid management, cardiac function monitoring, and nutritional support are obtained through data analysis. These independent care parameters do not overlap with other care measures and can reflect the care needs under a specific pathological pattern alone. For the independent heart care parameters of the pathological pattern extracted in the previous step, a contribution analysis of each independent parameter is performed to calculate the contribution value of each care parameter to the pathological pattern. The contribution of each independent care parameter to improving the patient's condition is quantified by regression analysis, correlation analysis, or multivariate statistical methods. The contribution of each nursing parameter is calculated based on its effect on improving the patient's pathological state. For the "acute myocardial infarction" pathological pattern, the contribution of different independent nursing parameters is calculated using regression analysis. For example, the contribution analysis of ECG monitoring shows that its contribution to the early detection of arrhythmias in patients is 0.4; the contribution analysis of antithrombotic drugs shows that its contribution to reducing thrombosis and improving blood flow is 0.35; and the contribution of pain management is 0.25. In this way, the contribution data of cardiac nursing parameters of the pathological pattern are generated to clarify the contribution of each nursing parameter in the treatment process. Combined with the analysis results of the first two steps, the contribution weight analysis of cardiac nursing intervention is carried out based on the contribution data of independent cardiac nursing parameters of the pathological pattern. This analysis assigns the weight of each nursing intervention measure according to the contribution of different nursing parameters to the pathological pattern. The weight of each nursing intervention is obtained by normalizing the contribution data to ensure that the sum is 1, so that resources can be reasonably allocated during actual nursing intervention. Using the contribution data of independent nursing parameters of the pathological pattern, the weight analysis of the intervention contribution of the "acute myocardial infarction" pattern is carried out. Assume that the contribution of ECG monitoring is 0.4, antithrombotic drugs is 0.35, and pain management is 0.25. Normalize these contribution values ​​to ensure that their sum is 1. The normalized weights are: ECG monitoring 0.4, antithrombotic drugs 0.35, pain management 0.25. These weights are the contribution weight parameters of pathological model cardiac nursing intervention. The weight parameters are used for subsequent nursing decisions and the design of intervention plans to ensure that the priority of intervention measures and the allocation of resources are reasonable.

[0119] Furthermore, the optimization prediction model design for deriving cardiac pathology patterns and nursing information from a preset deep learning network model using the common characteristic data of cardiac nursing of pathology patterns and the differential intervention characteristic data of cardiac nursing of pathology patterns includes:

[0120] Use the preset deep learning network model to design the model architecture for cardiac care derivation prediction and generate a preliminary cardiac care derivation prediction model;

[0121] Using the common characteristic data of cardiac nursing of pathological patterns as the low-level abstract training set for cardiac nursing derivation, and using the differential intervention characteristic data of cardiac nursing of pathological patterns as the high-level abstract training set for cardiac nursing derivation;

[0122] transferring the cardiac care derivation low-level abstract training set to the preliminary cardiac care derivation prediction model for network-level cardiac care derivation training processing to generate the cardiac care derivation prediction model;

[0123] The cardiac care derivation high-order abstract training set is transferred to the cardiac care derivation prediction model for optimized learning processing of the cardiac care derivation difference characteristics of the network-level pathological pattern to generate an optimized cardiac care derivation prediction model.

[0124] In an embodiment of the present invention, based on the task requirements and the needs of cardiac nursing derivation prediction, a deep learning network model architecture is designed. The design of the model needs to comprehensively consider the input data type (such as pathological pattern cardiac nursing common feature data and pathological pattern cardiac nursing differential intervention feature data), the prediction task (i.e. cardiac nursing intervention derivation) and the level of the network structure. The preset deep learning network model should have a multi-layer neural network architecture, which includes an input layer, multiple hidden layers and an output layer. The activation function and loss function of each layer will be selected according to the characteristics of the training data to achieve efficient prediction of cardiac nursing derivation. For example, a deep learning network architecture is designed, and a model architecture suitable for cardiac nursing derivation is selected, such as using a convolutional neural network (CNN) or a recurrent neural network (RNN) as the basic architecture. In the network architecture, the input layer processes the pathological pattern cardiac nursing common feature data and the pathological pattern cardiac nursing differential intervention feature data, and the hidden layer is used for feature extraction and feature fusion. Finally, the output layer predicts the recommendations for the nursing intervention plan. The preliminary design includes three hidden layers, each layer contains 128 neurons, the activation function selects ReLU, and the output layer uses a softmax activation function to output the probabilities of various nursing plans. According to the different properties of the data, the common feature data of cardiac nursing in pathological patterns and the differential intervention feature data of cardiac nursing in pathological patterns are used as low-order abstract and high-order abstract training sets, respectively. The low-order abstract training set mainly consists of relatively simple nursing features that are generally applicable to various pathological patterns, such as cardiac function monitoring and basic vital signs management; while the high-order abstract training set contains relatively complex nursing features that are closely related to specific pathological patterns, such as drug intervention and personalized nursing plans. By inputting these data into the model as low-order and high-order training sets, the learning ability of the model can be better improved. The common feature data of cardiac nursing in pathological patterns (such as conventional cardiac monitoring data and basic nursing indicators) are used as low-order training sets, and these data are transmitted to the input layer of the deep learning network model for processing. For the differential intervention feature data of cardiac nursing in pathological patterns (such as drug use and personalized intervention plans for specific pathological patterns), they are input into the high-level neurons of the model as high-order training sets. This process uses the hierarchical relationship between network layers to abstract different features, so as to realize feature learning at different levels in the subsequent training process. The low-order abstract training set is input into the preliminarily designed deep learning network model for model training. This process includes forward propagation, backpropagation and gradient update of input data to train an effective cardiac care derivation prediction model. Through multiple iterative optimizations, the weights and biases in the model are gradually adjusted to improve the prediction accuracy of the model. The network-level training process ensures that low-order features are fully learned throughout the network architecture and effectively converted into nursing intervention decisions. The data of the low-order training set (such as cardiac monitoring results, vital signs) is passed into the preliminarily designed deep learning network model through forward propagation.During the training process, the back propagation algorithm is used to adjust the network weights according to the error of each prediction result. Assuming that the low-level training set contains 1000 samples, after 200 rounds of training, the preliminary model will gradually learn how to derive corresponding nursing recommendations based on these common nursing features. In this process, the training error continues to decrease, and the generated model can more accurately predict the nursing intervention plan under the common pathological pattern. The high-level abstract training set is input into the preliminary model that has been trained at the low level and further optimized. The features contained in the high-level abstract training set have higher complexity and can better reflect the nursing needs under specific pathological patterns. By passing this data to the upper layer of the network, through multiple hidden layers in the deep learning network, the model further refines and optimizes the nursing difference features related to the pathological pattern. Through this optimization process, the model can generate a more accurate and personalized nursing derivation prediction model. The high-level training set (for example, drug intervention for acute myocardial infarction, use of pacemakers, etc.) is transferred to the low-level trained model. At this stage, the network layer will focus on optimizing the learning of these high-level features, for example, by strengthening the association between nursing measures and patient status under specific pathological patterns, optimizing the nursing derivation difference features of pathological patterns. After 50 rounds of training, the model was able to identify the subtle differences in nursing features in each pathological pattern, thereby improving the model's prediction accuracy for complex nursing interventions. Ultimately, the cardiac care derivation prediction results generated by the optimized model can provide more accurate and personalized care plans.

[0125] Furthermore, the step of transmitting the cardiac care derivation high-order abstract training set to the cardiac care derivation prediction model for performing the cardiac care derivation difference feature optimization learning process of the network-level pathological pattern includes:

[0126] Using the cardiac care derivation high-level abstract training set, the cardiac care derivation prediction model is analyzed for the cardiac care bias characteristics at the network level to generate the cardiac care bias characteristic data at the network level;

[0127] Performing network-level cardiac care feature reconstruction processing according to the network-level cardiac care bias feature data to generate network-level cardiac care feature reconstruction data;

[0128] Through network-level cardiac care feature reconstruction data, the cardiac care derivation prediction model is optimized and learned based on the cardiac care derivation difference features of each pathological pattern to generate an optimized cardiac care derivation prediction model.

[0129] In an embodiment of the present invention, a high-level abstract training set (including complex nursing features under a specific pathological mode) is used to train a cardiac nursing derivation prediction model, and the bias features that may exist in the model at each level are analyzed. These bias features refer to the phenomenon that the model responds excessively or insufficiently to certain input features (such as drug intervention, surgical measures, etc.) during the training process. By analyzing the weights and activation values ​​of the network level, it is found which features have not been fully learned during the training process. These bias features are identified and adjusted to optimize the learning process of the model. For the nursing derivation of heart disease patients, a high-level abstract training set (such as specific drug usage data for acute myocardial infarction, vasodilation therapy, etc.) is used to analyze the bias features of the middle layer (such as the third layer) of the cardiac nursing derivation prediction model. Assuming that the analysis results show that the model's response to drug intervention is weak, it may be due to the imbalance of input data, which causes the model to fail to fully learn the effect of drug treatment during training, and generates network-level cardiac nursing bias feature data to identify which bias features need to be optimized in subsequent processing. Based on the cardiac nursing bias feature data, the model is feature reconstructed. Feature reconstruction refers to adjusting the weights or input data in the network layer to eliminate biased features and ensure that the model can fully and accurately learn the impact of each nursing feature. The reconstruction process can be done by adjusting the learning rate, reweighting the training data set, or adding specific regularization terms to the model to ensure that the biased features are effectively corrected. The reconstructed network layer will be able to better process high-order feature data, thereby improving the accuracy and comprehensiveness of the derivation results. Based on the biased feature data of cardiac care, assuming that the model has a weak response when dealing with drug interventions for acute myocardial infarction, feature reconstruction can be performed in the following ways: by increasing the weight of drug intervention-related data, adjusting the sample distribution of different pathological patterns in the training data, or using a gradient penalty method to reduce the bias of specific features. At this point, assuming that the training process of the adjusted model can pay more attention to the relationship between drug intervention and patient recovery, the generated network-level cardiac care feature reconstruction data will correct the model's deficiencies in drug intervention derivation. Using the feature-reconstructed data (i.e., network-level cardiac care feature reconstruction data), the cardiac care derivation prediction model is further optimized and learned. This process involves optimizing the differential features of cardiac care derivation under different pathological modes to ensure that the model can derive accurate and clinically valuable nursing intervention plans for different pathological modes (such as acute myocardial infarction, chronic heart failure, etc.). The optimization learning process usually adopts a fine-grained adjustment strategy to optimize the parameters of the model layer by layer, thereby improving the model's ability to recognize differential features. In the optimization learning process, reconstructed data (such as drug intervention data of patients with acute myocardial infarction) is first used for training, and the parameters of the model are adjusted layer by layer to strengthen the model's learning of the nursing derivation features under the pathological mode. For example, the model's weight on drug intervention is optimized to ensure that it has a strong correlation with the recovery effect under the pathological mode.At the same time, for patients with chronic heart failure, features related to long-term care measures (such as fluid management, nutritional support, etc.) are added to gradually improve the model's derivation accuracy under multiple pathological modes. Ultimately, through this process, an optimized cardiac care derivation prediction model is generated, which can effectively identify and optimize the differential features of care derivation under each pathological mode.

[0130] This specification provides a method for constructing a cardiology nursing knowledge base, which is implemented based on the construction system for the cardiology nursing knowledge base as described above. The method for constructing a cardiology nursing knowledge base includes:

[0131] Step S1: Acquire cardiac pathology data and corresponding cardiac care data from a medical database; perform cardiac pathology data analysis on the cardiac pathology data to generate cardiac pathology analysis data, wherein the cardiac pathology analysis data includes cardiac pathology diagnosis data, cardiac pathology physiological index data, and cardiac pathology imaging data; perform cardiac pathology diagnosis feature extraction on the cardiac pathology diagnosis data to generate cardiac pathology diagnosis feature data;

[0132] Step S2: performing fuzzy logic analysis and processing of potential cardiac pathological abnormality characteristics based on cardiac pathological and physiological index data and cardiac pathological image data to generate fuzzy logic data of potential cardiac pathological abnormality characteristics;

[0133] Step S3: using cardiac pathology diagnosis feature data and potential cardiac pathology abnormality feature fuzzy logic data as cardiac state entity nodes and cardiac nursing data as cardiac nursing entity nodes, performing a mapping relationship analysis between cardiac state and nursing knowledge according to the cardiac state entity nodes and cardiac nursing entity nodes, and generating a cardiac state nursing relationship feature matrix; performing pathological pattern identification processing of cardiac state nursing relationship features on the cardiac state nursing relationship feature matrix, and generating a pathological pattern cardiac state nursing relationship feature matrix;

[0134] Step S4: Design an optimized prediction model for deriving cardiac pathology patterns and nursing information based on the pathology pattern heart state nursing relationship feature matrix, and generate an optimized cardiac nursing derivation prediction model; design and process an intelligent feedback engine for the cardiology nursing knowledge base based on the optimized cardiac nursing derivation prediction model to build an intelligent cardiology nursing knowledge base engine.

[0135] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0136] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A system for constructing a cardiology nursing knowledge base, characterized in that: Includes the following modules: A cardiac pathology diagnosis feature analysis module is used to obtain cardiac pathology data and corresponding cardiac care data from a medical database; perform cardiac pathology data analysis on the cardiac pathology data to generate cardiac pathology analysis data, wherein the cardiac pathology analysis data includes cardiac pathology diagnosis data, cardiac pathology physiological index data, and cardiac pathology imaging data; perform cardiac pathology diagnosis feature extraction based on the cardiac pathology diagnosis data to generate cardiac pathology diagnosis feature data; A potential cardiac pathological abnormality feature fuzzy logic analysis module is used to perform fuzzy logic analysis and processing of potential cardiac pathological abnormality features based on cardiac pathological physiological index data and cardiac pathological imaging data to generate potential cardiac pathological abnormality feature fuzzy logic data; A cardiac state nursing relationship feature analysis module is used to use cardiac pathology diagnosis feature data and potential cardiac pathology abnormality feature fuzzy logic data as cardiac state entity nodes and cardiac nursing data as cardiac nursing entity nodes, and to perform a mapping relationship analysis between cardiac state and nursing knowledge based on the cardiac state entity nodes and cardiac nursing entity nodes to generate a cardiac state nursing relationship feature matrix; Performing pathological pattern identification processing of the heart state nursing relationship characteristics on the heart state nursing relationship characteristic matrix to generate a pathological pattern heart state nursing relationship characteristic matrix; Cardiology nursing knowledge base construction module, used to design an optimized prediction model for deriving cardiac pathology patterns and nursing information based on the pathology pattern cardiac state nursing relationship feature matrix, and generate an optimized cardiac nursing derivation prediction model; Based on the optimized cardiac care derivation prediction model, the intelligent feedback engine design and processing of the cardiology nursing knowledge base is carried out to build an intelligent cardiology nursing knowledge base engine.

2. The system for constructing a cardiology nursing knowledge base according to claim 1, characterized in that: The cardiac pathology diagnosis feature analysis module includes the following functions: Obtain cardiac pathology data and corresponding cardiac care data from medical databases; Perform data modality type analysis based on cardiac pathology data to generate cardiac pathology modality data; Performing cardiac pathology data analysis and processing on the cardiac pathology data according to the cardiac pathology modality data to generate cardiac pathology analysis data; Performing text coding processing on the diagnostic data according to the cardiac pathology diagnostic data to generate cardiac pathology diagnostic text coding data; The cardiac pathology diagnosis feature extraction process is performed based on the cardiac pathology diagnosis text coding data to generate cardiac pathology diagnosis feature data.

3. The system for constructing a cardiology nursing knowledge base according to claim 1, characterized in that: The potential cardiac pathological abnormality feature fuzzy logic analysis module includes the following functions: Analyze and process abnormal physiological characteristics of cardiac pathological state according to cardiac pathological and physiological index data, and generate abnormal physiological index characteristic data of cardiac pathological state; Analyze abnormal cardiac pathological physiological characteristics according to cardiac pathological and physiological index data to generate abnormal cardiac pathological physiological characteristic data; Performing pathological image segmentation processing on cardiac pathological image data to divide cardiac functional structural regions, generating segmented cardiac pathological image data; performing image abnormality feature analysis on cardiac functional structure based on the segmented cardiac pathological image data, generating cardiac functional structure image abnormality feature data; Fuzzy logic conversion processing of potential cardiac pathological abnormality characteristics is performed based on cardiac pathological abnormality physiological characteristic data and cardiac functional structural imaging abnormality characteristic data to generate potential cardiac pathological abnormality characteristic fuzzy logic data.

4. The system for constructing a cardiology nursing knowledge base according to claim 1, characterized in that: The cardiac state nursing relationship feature analysis module includes the following functions: The cardiac pathology diagnosis feature data and the potential cardiac pathology abnormality feature fuzzy logic data are used as cardiac state entity nodes and the cardiac care data are used as cardiac care entity nodes. The mapping relationship between cardiac state and nursing knowledge is analyzed based on the cardiac state entity nodes and cardiac care entity nodes to generate a cardiac state nursing relationship feature matrix. Performing cardiac pathology diagnosis cluster analysis on cardiac pathology diagnosis feature data to generate cardiac pathology diagnosis cluster data; Conducting cardiac pathology diagnosis correlation analysis based on cardiac pathology diagnosis clustering data to generate cardiac pathology diagnosis correlation data; Perform cardiac pathology pattern analysis based on cardiac pathology diagnosis correlation data to generate cardiac pathology pattern data; Based on the cardiac pathological pattern data, the cardiac state nursing relationship feature matrix is ​​processed with the pathological pattern identification of the cardiac state nursing relationship features to generate a pathological pattern cardiac state nursing relationship feature matrix.

5. The system for constructing a cardiology nursing knowledge base according to claim 1, characterized in that: The cardiology nursing knowledge base building module includes the following functions: Performing pathological pattern cardiac nursing associated data analysis on the pathological pattern cardiac state nursing relationship characteristic matrix to generate pathological pattern cardiac nursing associated data; Performing cardiac care-specific analysis of the pathological pattern based on the cardiac care-related data of the pathological pattern to generate cardiac care-specific data of the pathological pattern; Extracting the common nursing features and differential nursing features of the pathological pattern from the pathological pattern heart nursing specific data, and generating the common nursing feature data of the pathological pattern heart nursing and differential nursing feature data of the pathological pattern heart nursing respectively; The intervention contribution of nursing difference characteristics is weighted on the pathological pattern heart nursing difference characteristic data to generate pathological pattern heart nursing difference weighted characteristic data; Based on the fuzzy logic data of potential cardiac pathological abnormalities, the influencing factors of cardiac nursing differences in pathological patterns are analyzed to generate data on influencing factors of cardiac nursing differences in pathological patterns; Using the data of influencing factors of cardiac nursing differences in pathological patterns, the empowerment characteristic data of cardiac nursing differences in pathological patterns are analyzed to generate intervention characteristic data of cardiac nursing differences in pathological patterns; The preset deep learning network model is used to design an optimized prediction model for deriving cardiac pathology patterns and nursing information through the common characteristic data of cardiac nursing of pathology patterns and the differential intervention characteristic data of cardiac nursing of pathology patterns, and an optimized cardiac nursing derivation prediction model is generated; Based on the optimized cardiac care derivation prediction model, the intelligent feedback engine design and processing of the cardiology nursing knowledge base is carried out to build an intelligent cardiology nursing knowledge base engine.

6. The system for constructing a cardiology nursing knowledge base according to claim 5, characterized in that: The intervention contribution weighting processing of the nursing difference characteristics of the pathological pattern heart nursing difference characteristics data includes: Based on the pathological pattern cardiac nursing associated data, the cardiac nursing intervention contribution weight analysis of each pathological pattern is performed to generate the cardiac nursing intervention contribution weight parameters of the pathological pattern; According to the pathological pattern heart nursing intervention contribution weight parameter, the pathological pattern heart nursing difference characteristic data is weighted by the intervention contribution of the nursing difference characteristics to generate the pathological pattern heart nursing difference weighted characteristic data.

7. The system for constructing a cardiology nursing knowledge base according to claim 6, characterized in that: The cardiac nursing intervention contribution weight analysis of each pathological pattern based on the pathological pattern cardiac nursing associated data includes: Perform independent cardiac care parameter analysis of each pathological pattern according to the cardiac care associated data of the pathological pattern, and generate independent cardiac care parameters of the pathological pattern; Performing pathology pattern independent heart care parameter contribution analysis according to the pathology pattern independent heart care parameters to generate pathology pattern independent heart care parameter contribution data; According to the independent cardiac nursing parameter contribution data of the pathological pattern, the cardiac nursing intervention contribution weight analysis of each pathological pattern is performed to generate the cardiac nursing intervention contribution weight parameters of the pathological pattern.

8. The system for constructing a cardiology nursing knowledge base according to claim 5, characterized in that: The design of an optimized prediction model for deriving cardiac pathology patterns and nursing information from a preset deep learning network model using the common characteristic data of cardiac nursing of pathology patterns and the differential intervention characteristic data of cardiac nursing of pathology patterns includes: Use the preset deep learning network model to design the model architecture for cardiac care derivation prediction and generate a preliminary cardiac care derivation prediction model; Using the common characteristic data of cardiac nursing of pathological patterns as a training set for deriving low-level abstractions of cardiac nursing, and using the differential intervention characteristic data of cardiac nursing of pathological patterns as a training set for deriving high-level abstractions of cardiac nursing; transferring the cardiac care derivation low-level abstract training set to the preliminary cardiac care derivation prediction model for network-level cardiac care derivation training processing to generate the cardiac care derivation prediction model; The cardiac care derivation high-order abstract training set is transferred to the cardiac care derivation prediction model for optimized learning processing of the cardiac care derivation difference features of the network-level pathological pattern to generate an optimized cardiac care derivation prediction model.

9. The system for constructing a cardiology nursing knowledge base according to claim 8, characterized in that: The transferring of the cardiac care derivation high-order abstract training set to the cardiac care derivation prediction model for the cardiac care derivation difference feature optimization learning process of the network-level pathological pattern includes: Using the cardiac care derivation high-level abstract training set, the cardiac care derivation prediction model is analyzed for the cardiac care bias characteristics at the network level to generate the cardiac care bias characteristic data at the network level; Performing network-level cardiac care feature reconstruction processing according to the network-level cardiac care bias feature data to generate network-level cardiac care feature reconstruction data; Through network-level cardiac care feature reconstruction data, the cardiac care derivation prediction model is optimized and learned based on the cardiac care derivation difference features of each pathological pattern to generate an optimized cardiac care derivation prediction model.

10. A method for constructing a cardiology nursing knowledge base, characterized in that: The method is implemented based on the construction system for cardiology nursing knowledge base according to claim 1, and the construction method for cardiology nursing knowledge base includes: Step S1: for acquiring cardiac pathology data and corresponding cardiac care data from a medical database; performing cardiac pathology data parsing processing on the cardiac pathology data to generate cardiac pathology analysis data, wherein the cardiac pathology analysis data includes cardiac pathology diagnosis data, cardiac pathology physiological index data and cardiac pathology imaging data; performing cardiac pathology diagnosis feature extraction processing on the cardiac pathology diagnosis data to generate cardiac pathology diagnosis feature data; Step S2: performing fuzzy logic analysis and processing of potential cardiac pathological abnormality characteristics based on cardiac pathological and physiological index data and cardiac pathological image data to generate fuzzy logic data of potential cardiac pathological abnormality characteristics; Step S3: using cardiac pathology diagnosis feature data and potential cardiac pathology abnormality feature fuzzy logic data as cardiac state entity nodes and cardiac nursing data as cardiac nursing entity nodes, performing a mapping relationship analysis between cardiac state and nursing knowledge according to the cardiac state entity nodes and cardiac nursing entity nodes, and generating a cardiac state nursing relationship feature matrix; performing pathological pattern identification processing of cardiac state nursing relationship features on the cardiac state nursing relationship feature matrix, and generating a pathological pattern cardiac state nursing relationship feature matrix; Step S4: Design an optimized prediction model for deriving cardiac pathology patterns and nursing information based on the pathology pattern heart state nursing relationship feature matrix, and generate an optimized cardiac nursing derivation prediction model; design and process an intelligent feedback engine for the cardiology nursing knowledge base based on the optimized cardiac nursing derivation prediction model to build an intelligent cardiology nursing knowledge base engine.

Images