Construction method of myocardial rehabilitation decision support system based on case-based reasoning

Abstract

The present invention relates to the field of data mining technology, and in particular to a method for constructing a myocardial rehabilitation decision support system based on case reasoning. The method comprises the following steps: collecting and integrating historical myocardial case data, desensitizing sensitive features, and obtaining desensitized data; dividing myocardial rehabilitation cases and integrating abnormal electrocardiogram waveform features to obtain abnormal waveform data of rehabilitation cases; constructing an abnormal waveform detection model, and using it to detect abnormal waveforms to obtain abnormal electrocardiogram waveform myocardial case data; performing trend analysis on abnormal waveform data and calculating myocardial rehabilitation influencing factors; evaluating the doctor's treatment priority based on the influencing factors, and constructing a treatment doctor recommendation model; combining the recommendation model and the detection model, constructing and optimizing the myocardial rehabilitation decision system, thereby obtaining a myocardial rehabilitation decision support system. Based on data mining, the present invention realizes the accuracy and effectiveness of myocardial rehabilitation research on historical myocardial cases.

Description

Technical Field

[0001] The present invention relates to the technical field of data mining, and in particular to a method for constructing a myocardial rehabilitation decision support system based on case reasoning. Background Art

[0002] Myocardial rehabilitation is a rehabilitation treatment process for patients with cardiovascular diseases, especially those after myocardial infarction. Its goal is to improve the quality of life and survival rate of patients through systematic medical intervention. The myocardial rehabilitation process involves the design of individualized treatment plans, which usually requires a combination of information such as the patient's medical history, symptoms, and rehabilitation effects. In traditional myocardial rehabilitation practice, the formulation of treatment plans often relies on the doctor's experience and intuition. The lack of systematic and personalized support will cause the effect of the treatment plan to vary depending on the individual differences of the doctor. At the same time, in traditional myocardial rehabilitation practice, it is difficult to effectively integrate data from different sources (such as medical records, imaging data, and laboratory test results), resulting in information islands and incomplete decision-making basis. Summary of the invention

[0003] Based on this, it is necessary for the present invention to provide a method for constructing a myocardial rehabilitation decision support system based on case reasoning to solve at least one of the above technical problems.

[0004] To achieve the above purpose, a method for constructing a myocardial rehabilitation decision support system based on case reasoning includes the following steps:

[0005] Step S1: Acquire historical myocardial case data, and fuse the historical myocardial case data with patient case data to obtain patient myocardial case fusion data, and desensitize the patient's sensitive features on the patient's myocardial case fusion data to obtain historical myocardial case desensitization data;

[0006] Step S2: performing myocardial rehabilitation case intersection division on the historical myocardial rehabilitation case desensitization data, thereby obtaining historical myocardial rehabilitation case data, and integrating abnormal electrocardiogram waveform features based on the historical myocardial rehabilitation case data, thereby obtaining abnormal electrocardiogram waveform data of rehabilitation cases;

[0007] Step S3: constructing an abnormal ECG waveform detection model based on the abnormal ECG waveform data of the rehabilitation cases, and performing abnormal waveform detection on the historical myocardial rehabilitation case data through the abnormal ECG waveform detection model, thereby obtaining abnormal ECG waveform myocardial case data;

[0008] Step S4: performing an electrocardiogram waveform trend analysis based on the abnormal electrocardiogram waveform myocardial case data to obtain electrocardiogram waveform trend data, and calculating a myocardial rehabilitation impact factor on the electrocardiogram waveform trend data based on historical myocardial rehabilitation case data to obtain a myocardial rehabilitation impact factor;

[0009] Step S5: evaluating the treatment priority of myocardiologists on historical myocardial case data according to the myocardial rehabilitation influencing factors, thereby obtaining the treatment priority data of myocardiologists, and constructing a myocardial treatment physician recommendation model according to the treatment priority data of myocardiologists and the myocardial rehabilitation influencing factors;

[0010] Step S6: Construct a myocardial rehabilitation decision system based on the myocardial physician recommendation model and the electrocardiogram abnormal waveform detection model, and iteratively optimize and adjust the parameters of the myocardial rehabilitation decision system based on historical myocardial case data to obtain a myocardial rehabilitation decision support system.

[0011] The present invention can integrate medical records, imaging materials and laboratory test results from different sources by acquiring and integrating historical myocardial case data, thereby forming a comprehensive patient data archive. This integration helps overcome information islands and enables data to be better utilized in the treatment decision-making process. After desensitizing the patient case data, the patient's privacy can be effectively protected and the security of the data can be ensured. This is crucial for the research and analysis of sensitive medical data. The fused data is more consistent and complete, which is helpful for subsequent analysis and model training, thereby improving the reliability of the research results. By dividing the desensitized data into myocardial rehabilitation case intersections, different cases related to myocardial rehabilitation can be more accurately screened out, data irrelevant to the research purpose can be excluded, and the pertinence of the data can be enhanced. Integrating abnormal electrocardiogram waveform features helps to identify and analyze abnormal patterns appearing in electrocardiograms and improve the diagnostic ability of abnormal electrocardiograms. The abnormal electrocardiogram waveform data of rehabilitation cases obtained can provide a basis for subsequent model construction and trend analysis, ensuring the depth and breadth of the analysis. By constructing an abnormal electrocardiogram waveform detection model, abnormal waveforms can be automatically detected and identified, improving detection efficiency and accuracy, and reducing the burden of manual analysis. The model can be trained with a large amount of historical data to improve the accuracy of abnormal waveform detection and help doctors identify potential ECG abnormalities earlier. By analyzing the trend of ECG waveforms, the long-term pattern of ECG changes in patients can be found, which helps evaluate the effect and progress of myocardial rehabilitation. Calculating the influencing factors of myocardial rehabilitation helps to identify factors that have a significant impact on the effect of myocardial rehabilitation and provide a basis for formulating personalized treatment plans. Based on the analysis results of ECG waveform trends and influencing factors, rehabilitation treatment plans can be optimized to improve treatment effects. Evaluating the treatment priority of historical cases can help the system prioritize doctors to treat cases that need intervention most according to the specific conditions of patients, thereby improving the efficiency and effectiveness of treatment. By building a recommendation model for myocardial treatment doctors, the most suitable doctors and treatment plans can be recommended according to the specific conditions of patients, enhancing the ability of personalized medical services. The data-driven decision support provided by the recommendation model helps doctors make more scientific decisions on the selection of treatment plans. The myocardial rehabilitation decision system integrates various data and models, which can provide comprehensive decision support for doctors and improve the scientificity and rationality of treatment plans. Through iterative optimization and parameter adjustment of the decision-making system, the accuracy and reliability of the system can be continuously improved to ensure its effectiveness in practical applications. The decision-making system can speed up the process of formulating treatment plans, reduce human errors, and improve the efficiency and quality of patient rehabilitation. Overall, these steps can significantly improve the personalization and scientific nature of myocardial rehabilitation research through systematic data processing and model construction.

[0012] Optionally, step S1 specifically includes:

[0013] Step S11: Acquire historical myocardial case data, and perform structured data classification on the historical myocardial case data, thereby obtaining structured myocardial case data and unstructured myocardial case data;

[0014] Step S12: performing unstructured patient myocardial feature fusion on the unstructured myocardial case data, thereby obtaining unstructured patient myocardial fusion data;

[0015] Step S13: fusing the patient's treatment course physiological characteristics according to the structured myocardial case data, thereby obtaining structured patient physiological fusion data;

[0016] Step S14: merging the unstructured patient myocardial fusion data and the structured patient physiological fusion data into patient information cases, thereby obtaining the patient myocardial case fusion data;

[0017] Step S15: Desensitize the patient's sensitive features on the patient's myocardial case fusion data, so as to obtain historical myocardial case desensitized data.

[0018] The present invention acquires and classifies historical myocardial case data, which helps to divide the data into two categories: structured and unstructured, ensuring that subsequent processing is more efficient and accurate. Structured data can be directly used for analysis and modeling, while the classification of unstructured data lays the foundation for further processing. Feature fusion of unstructured myocardial case data can extract myocardial feature information, improve the comprehensive utilization efficiency of data, and thus improve the accuracy and comprehensiveness of myocardial research. According to the structured myocardial case data, the physiological characteristics of the patient's treatment course are fused, which can integrate the patient's physiological data, form a more comprehensive description of physiological characteristics, and enhance the effectiveness of the prediction model. Merging unstructured patient myocardial fusion data with structured patient physiological fusion data can fully integrate the patient's myocardial and physiological information, provide more comprehensive patient case data, and improve the formulation of personalized treatment plan research. The sensitive features of the fused patient myocardial case data are desensitized to ensure patient privacy and data security, thereby maintaining patient privacy trust and promoting the safe sharing and use of data.

[0019] Optionally, step S12 is specifically:

[0020] Step S121: extracting medical record summary records, patient electrocardiogram records and medical image records from the unstructured myocardial case data, thereby obtaining myocardial medical record summary data, myocardial patient electrocardiogram sets and myocardial patient medical image sets;

[0021] Step S122: integrating myocardial description features of the cardiology medical record summary data to obtain myocardial description data;

[0022] Step S123: extracting dynamic structural features of the myocardium scanned by CT according to the medical image set of the patients in the myocardium department, thereby obtaining dynamic structural data of the myocardium scanned by CT, and constructing a three-dimensional dynamic structural model of the myocardium of the patient according to the dynamic structural data of the myocardium scanned by CT;

[0023] Step S124: filling the patient's myocardial three-dimensional dynamic structure model with myocardial structure details according to the myocardial description data, thereby obtaining the patient's myocardial dynamic structure model;

[0024] Step S125: integrating the dynamic fluctuation characteristics of the electrocardiogram according to the electrocardiogram set of the patients in the myocardium department, thereby obtaining the dynamic fluctuation data of the electrocardiogram, and performing myocardial dynamic motion simulation on the dynamic fluctuation data of the electrocardiogram through the dynamic structural model of the myocardium of the patients, thereby obtaining the dynamic simulation data of the myocardium of the patients;

[0025] Step S126: performing patient myocardial dynamic feature fusion according to the patient myocardial dynamic simulation data, thereby obtaining unstructured patient myocardial fusion data.

[0026] The present invention extracts the medical record summary, electrocardiogram record and medical imaging record of the cardiology department, which is helpful to systematically collect and integrate all relevant data of the patient and provide basic data support for subsequent analysis. Integrating the myocardial description features can summarize the myocardial features in the medical record data to form unified myocardial description data, which is convenient for subsequent analysis and comparison. Extracting the dynamic structural features of the myocardium scanned by CT and constructing a three-dimensional dynamic structural model can provide detailed structural information of the myocardium in space, which is helpful for accurate three-dimensional analysis and simulation. Filling the details of the three-dimensional dynamic structural model according to the myocardial description data makes the model more realistic and accurate, thereby enhancing the reliability of simulation and analysis. Integrating the dynamic fluctuation features of the electrocardiogram and performing motion simulation through the myocardial dynamic structure model can dynamically associate the electrocardiogram data with the myocardial structure, which is helpful for studying the real-time changes of myocardial function. Feature fusion based on the myocardial dynamic simulation data can integrate the information of various data sources into a comprehensive view, improve the comprehensive understanding of the patient's myocardial state and the accuracy of myocardial research.

[0027] Optionally, step S13 is specifically:

[0028] Step S131: extracting drug dosage records, disease course records and physiological coefficient records from the structured myocardial case data, thereby obtaining drug dosage data, disease course data and physiological coefficient data of the patient;

[0029] Step S132: dividing the patient's treatment stage according to the patient's disease course data, thereby obtaining the patient's treatment stage data;

[0030] Step S133: performing a drug characteristic analysis on the drug dosage data, thereby obtaining drug dosage characteristic data;

[0031] Step S134: performing physiological coefficient fluctuation correlation according to the drug dosage characteristic data and the patient's physiological coefficient data, thereby obtaining drug dosage correlated physiological fluctuation data;

[0032] Step S135: fusing the patient treatment stage data with the treatment stage physiological characteristics according to the drug dosage-associated physiological fluctuation data, thereby obtaining structured patient physiological fusion data.

[0033] The present invention extracts drug dosage records, course records and physiological coefficient records, which helps to disassemble structured myocardial case data into specific and operational sub-datasets. This extraction provides basic data for subsequent detailed analysis, including drug use, disease progression and physiological state, and lays a data foundation for accurate treatment analysis. Dividing the treatment stage according to the patient's course data helps to systematize the patient's treatment process. This process can identify different stages of treatment (such as acute phase, recovery phase, etc.), thereby providing phased data support for analyzing the relationship between drug dosage and physiological response, and helping to study the treatment plan in the case. Drug characteristic analysis of drug dosage data can reveal the dosage characteristics of different drugs, including medication frequency, dosage changes, etc. This analysis helps to understand the actual application of drugs in different patients, and then optimize the research of drug treatment plans and dosage adjustment strategies. According to the drug dosage characteristic data and the patient's physiological coefficient data, the physiological coefficient fluctuation association is performed, and the relationship between drug dosage changes and physiological parameters can be identified. This association analysis helps to understand the impact of drugs on physiological fluctuations, evaluate drug effects and side effects, and thus support the study of individualized treatment. By integrating the physiological characteristics of the patient's treatment stage data based on the physiological fluctuation data associated with the dosage, the relationship between the dosage and the physiological response at different treatment stages can be integrated. This fusion provides a comprehensive view of physiological data, supports more in-depth evaluation and research of treatment effects, and improves the research effect of the overall treatment plan.

[0034] Optionally, step S2 specifically includes:

[0035] Step S21: extracting treatment and rehabilitation stage features from the historical myocardial case desensitization data, thereby obtaining historical myocardial case treatment stage data, and classifying the historical myocardial case treatment stage data by treatment stage length, thereby obtaining long treatment stage case data and short treatment stage case data;

[0036] Step S22: classifying the historical myocardial case treatment stage data by treatment stage review, thereby obtaining case treatment stage review data;

[0037] Step S23: performing myocardial variable statistics review according to the case treatment stage review data, thereby obtaining the myocardial variable data review, and classifying the myocardial variable fluctuation amplitude of the myocardial variable data review, thereby obtaining high-amplitude fluctuation myocardial variable review data and low-amplitude fluctuation myocardial variable review data;

[0038] Step S24: classifying the high-amplitude fluctuation myocardial variable review data and the low-amplitude fluctuation myocardial variable review data by the high-proportion myocardial variable fluctuation amplitude ratio according to the historical myocardial case treatment stage data, thereby obtaining high-proportion case data of the high-amplitude fluctuation treatment stage and high-proportion case data of the low-amplitude fluctuation treatment stage;

[0039] Step S25: performing a treatment case intersection operation on the long treatment stage case data and the low amplitude fluctuation treatment stage case data with a high proportion, thereby obtaining slow-acting treatment and rehabilitation case data; performing a treatment case intersection operation on the short treatment stage case data and the high amplitude fluctuation treatment stage case data with a high proportion, thereby obtaining fast-acting treatment and rehabilitation case data;

[0040] Step S26: integrating the slow-acting treatment rehabilitation case data and the fast-acting treatment rehabilitation case data, thereby obtaining historical myocardial rehabilitation case data;

[0041] Step S27: Integrate abnormal electrocardiogram waveform features based on historical myocardial rehabilitation case data to obtain abnormal electrocardiogram waveform data of rehabilitation cases.

[0042] The present invention can structure the complex treatment process data in the case by extracting the characteristics of the treatment and rehabilitation stage, and provide a clear data basis for subsequent analysis. Dividing the treatment stage data into a long treatment stage and a short treatment stage can help understand the impact of different stages on the rehabilitation effect and provide a reference for personalized treatment. Further classifying the treatment stage data into review data helps to focus on the review results after treatment in the case, which is crucial for evaluating the treatment effect and further improving the treatment strategy. By classifying the review data, the changes and trends in the review process can be clearly identified, and the tracking and management capabilities of the patient's recovery in the case can be improved. Statistics and fluctuation range classification of the reviewed myocardial variables can provide an in-depth understanding of the changes in myocardial variables in the case, so as to more accurately evaluate the treatment effect. By distinguishing between myocardial variables with high amplitude fluctuations and low amplitude fluctuations, key indicators that have a significant impact on patient recovery can be identified, and the treatment plan can be adjusted in a targeted manner. Classifying myocardial variables with high amplitude fluctuations and low amplitude fluctuations by proportion helps to understand the impact of different fluctuation ranges on the treatment stage in the case. By analyzing the proportion of myocardial variable fluctuations, the parts that may need to be adjusted in the treatment can be identified, and the effectiveness and accuracy of the treatment plan research can be improved. Separating the slow-acting and fast-acting treatment case data through intersection operations helps to clarify the effects of different treatment methods and then study and optimize treatment plans. Integrating the rehabilitation case data of slow-acting and fast-acting treatments can summarize the comprehensive rehabilitation information in the case set, providing a basis for further analysis and research. By integrating various types of rehabilitation case data, it is possible to comprehensively evaluate the historical myocardial rehabilitation situation and discover common and individual problems in treatment. By integrating abnormal electrocardiogram waveform features, potential abnormalities or complications in the case can be discovered, providing a basis for timely intervention.

[0043] Optionally, step S27 is specifically:

[0044] Step S271: extracting electrocardiogram records from historical myocardial rehabilitation case data, thereby obtaining an electrocardiogram set of slow-acting treatment rehabilitation cases and an electrocardiogram set of fast-acting treatment rehabilitation cases;

[0045] Step S272: performing Fourier transform on the slow-acting treatment rehabilitation case electrocardiogram set and the fast-acting treatment rehabilitation case electrocardiogram set respectively, so as to obtain the slow-acting rehabilitation case electrocardiogram spectrum and the fast-acting rehabilitation case electrocardiogram spectrum;

[0046] Step S273: performing spectrum amplitude threshold statistics on the ECG spectrum of the slow-acting rehabilitation case and the ECG spectrum of the fast-acting rehabilitation case, respectively, so as to obtain the slow-acting spectrum amplitude threshold and the fast-acting spectrum amplitude threshold;

[0047] Step S274: classifying the abnormal spectrum amplitude of the slow-acting rehabilitation case ECG spectrum according to the slow-acting spectrum amplitude threshold, thereby obtaining the abnormal slow-acting rehabilitation case ECG spectrum; classifying the abnormal spectrum amplitude of the fast-acting rehabilitation case ECG spectrum according to the fast-acting spectrum amplitude threshold, thereby obtaining the abnormal fast-acting rehabilitation case ECG spectrum;

[0048] Step S275: performing time domain conversion on the ECG spectrum of the abnormal slow-acting recovery case and the ECG spectrum of the abnormal fast-acting recovery case, respectively, so as to obtain ECG fluctuation data of the abnormal slow-acting recovery case and ECG fluctuation data of the abnormal fast-acting recovery case;

[0049] Step S276: Integrate the abnormal electrocardiogram waveform features of the abnormal slow-acting rehabilitation case electrocardiogram fluctuation data and the abnormal fast-acting rehabilitation case electrocardiogram fluctuation data, so as to obtain the abnormal electrocardiogram waveform data of the rehabilitation case.

[0050] By extracting electrocardiogram records, the present invention can provide a detailed understanding of the electrocardiogram characteristics of slow-acting and fast-acting treatments, and provide basic data for subsequent analysis. Fourier transform helps convert time domain signals into frequency domain signals, making spectrum analysis intuitive and revealing the frequency characteristics of the electrocardiogram under different treatment effects in the case. Spectral amplitude threshold statistics can quantify abnormal fluctuations in the electrocardiogram, provide standardized indicators for identifying abnormalities, and help discover differences in treatment effects. Abnormal spectrum amplitude classification can accurately identify abnormal characteristics of the electrocardiogram after treatment, helping to distinguish abnormal situations under different treatment effects. Converting abnormal spectrum data into time domain data makes the abnormal fluctuation characteristics clearer, which is convenient for further analysis and understanding of the time domain changes of the electrocardiogram. Integrating abnormal waveform features can summarize specific abnormal patterns, help evaluate treatment effects, and provide data support for formulating more effective rehabilitation strategies.

[0051] Optionally, step S4 is specifically:

[0052] Step S41: segmenting the electrocardiogram waveform signal according to the abnormal electrocardiogram waveform myocardial case data, thereby obtaining abnormal electrocardiogram waveform signal data;

[0053] Step S42: performing time series trend line fitting on the abnormal electrocardiogram waveform signal data, thereby obtaining electrocardiogram waveform signal trend data;

[0054] Step S43: performing fluctuation measurement classification on the electrocardiogram waveform signal trend data, thereby obtaining excessive fluctuation waveform signal trend data and normal fluctuation waveform signal trend data;

[0055] Step S44: performing waveform signal trend correlation on the abnormal electrocardiogram waveform signal data according to the excessive fluctuation waveform signal trend data, thereby obtaining electrocardiogram waveform degradation trend data; performing waveform signal trend correlation on the abnormal electrocardiogram waveform signal data according to the normal fluctuation waveform signal trend data, thereby obtaining electrocardiogram waveform improvement trend data;

[0056] Step S45: merging the electrocardiogram waveform degradation trend data and the electrocardiogram waveform improvement trend data to obtain the electrocardiogram waveform trend data;

[0057] Step S46: Calculate the myocardial rehabilitation impact factor on the electrocardiogram waveform trend data according to the historical myocardial rehabilitation case data, so as to obtain the myocardial rehabilitation impact factor.

[0058] By segmenting the electrocardiogram waveform signal, the present invention can accurately extract abnormal signal data and provide a clear abnormal signal basis for subsequent analysis. Time series trend line fitting helps reveal the long-term trend of the electrocardiogram waveform signal, making the waveform change pattern more intuitive and easy to understand. The fluctuation metric classification divides the signal into excessive fluctuations and normal fluctuations, which helps to distinguish the severity of abnormal waveforms and normal variations, and improves the accuracy of the analysis. Correlating the deterioration and improvement trends of abnormal signals helps to understand the impact of different fluctuation patterns on the electrocardiogram waveform, thereby identifying the deterioration or improvement of the condition. Data merging provides comprehensive electrocardiogram waveform trend data, comprehensively considering the deterioration and improvement trends, and facilitating the overall assessment of the progress of myocardial rehabilitation. Calculating the myocardial rehabilitation influencing factor can evaluate the effects of different treatment options based on historical case data, providing a scientific basis for optimizing rehabilitation strategies.

[0059] Optionally, step S46 is specifically:

[0060] Step S461: extracting left ventricular ejection fraction features and electrocardiogram waveform features from historical myocardial rehabilitation case data, thereby obtaining rehabilitation case left ventricular ejection fraction data and rehabilitation case electrocardiogram waveform data;

[0061] Step S462: constructing a myocardial rehabilitation index evaluation system based on the left ventricular ejection fraction data of the rehabilitation case and the electrocardiogram waveform data of the rehabilitation case, and performing myocardial rehabilitation index evaluation according to the myocardial rehabilitation index evaluation system, thereby obtaining myocardial rehabilitation index evaluation data;

[0062] Step S463: extracting waveform amplitude features from the electrocardiogram waveform trend data, thereby obtaining trend waveform amplitude data;

[0063] Step S464: constructing a myocardial rehabilitation linear regression model according to the myocardial rehabilitation index evaluation data and the trend waveform amplitude data;

[0064] Step S465: Calculate the myocardial rehabilitation influencing factor according to the myocardial rehabilitation linear regression model, so as to obtain the myocardial rehabilitation influencing factor.

[0065] By extracting the left ventricular ejection fraction and electrocardiogram waveform characteristics, the present invention can fully understand the key physiological and electrophysiological indicators of myocardial rehabilitation and provide basic data for analysis and evaluation. By constructing a myocardial rehabilitation indicator evaluation system and conducting an evaluation, the rehabilitation effect in the case can be systematically evaluated to help determine the key physiological and electrophysiological changes in the rehabilitation process. Extracting the waveform amplitude characteristics makes the changing characteristics of the electrocardiogram trend data clearer, which helps to evaluate the specific changes in the electrocardiogram waveform. By constructing a myocardial rehabilitation linear regression model and combining the evaluation data with the trend waveform amplitude data, the specific mechanism of action of the myocardial rehabilitation influencing factors can be revealed, and quantitative analysis results can be provided. By calculating the myocardial rehabilitation influencing factors based on the linear regression model, the actual effects of different treatments on myocardial rehabilitation in the case can be quantified, providing data support for optimizing the treatment plan.

[0066] Optionally, step S5 specifically includes:

[0067] Step S51: extracting electrocardiogram waveform features according to the myocardial rehabilitation influencing factor, thereby obtaining the myocardial rehabilitation electrocardiogram waveform influencing factor;

[0068] Step S52: extracting the characteristics of the myocardial doctor and the electrocardiogram of the cases from the historical myocardial case data, thereby obtaining the myocardial doctor data and the electrocardiogram data of the cases;

[0069] Step S53: performing electrocardiogram waveform similarity calculation on the case electrocardiogram data and the myocardial rehabilitation electrocardiogram waveform influencing factor, thereby obtaining case electrocardiogram similarity data, and performing high-similarity case statistics on the case electrocardiogram similarity data, thereby obtaining high-similarity case electrocardiogram data;

[0070] Step S54: Associating treatment cases according to the ECG data of cases with high similarity and the data of myocardial treating doctors, thereby obtaining treating doctor-ECG associated data;

[0071] Step S55: evaluating the cardiotocologist's treatment priority according to the treating physician-ECG association data, thereby obtaining the cardiotocologist's treatment priority data;

[0072] Step S56: Construct a myocardial physician recommendation model based on the cardiologist treatment priority data and myocardial rehabilitation influencing factors.

[0073] The present invention can obtain specific waveform influencing factors for rehabilitation effects by extracting electrocardiogram waveform features according to myocardial rehabilitation influencing factors. This process provides individualized electrocardiogram feature data, which helps to more accurately evaluate and monitor electrocardiogram changes in the myocardial rehabilitation process in the case. The characteristics of different treating doctors and the specific conditions of the case electrocardiogram are fully understood by extracting the characteristics of historical myocardial case data. This process helps to build a detailed archive between doctors and patients' electrocardiogram data, and supports subsequent association analysis. The similarity between the case electrocardiogram data and the myocardial rehabilitation electrocardiogram waveform influencing factors is calculated, and the high-similarity cases are counted, so that the electrocardiogram cases closest to the target waveform influencing factors can be identified. This process can help find similar cases, thereby providing targeted treatment and rehabilitation plans. According to the high-similarity case electrocardiogram data and the myocardial treatment doctor data, the treatment cases are associated, which can clarify the treatment plans and effects of which treating doctors when processing similar electrocardiogram data. This process helps to reveal the actual operations and effects of different doctors when processing similar cases, thereby providing a basis for personalized treatment. By evaluating the treatment priority of myocardiologists through the treating physician-ECG correlation data, doctors can be ranked according to their treatment effects and the similarity of cases. This process can help hospitals or medical institutions optimize resource allocation and assign efficient doctors to priority cases. The treating physician recommendation model constructed based on the treatment priority data of myocardiologists and myocardial rehabilitation influencing factors can recommend the most suitable myocardiologist for patients. This process can ensure that patients receive the best medical services, improve treatment effects, and improve the efficiency and quality of overall myocardial rehabilitation.

[0074] Optionally, step S55 is specifically:

[0075] Step S551: performing doctor treatment frequency statistics and doctor-associated electrocardiogram statistics according to the treating doctor-electrocardiogram association data, thereby obtaining doctor treatment frequency data and doctor-associated electrocardiogram data;

[0076] Step S552: classifying the doctor treatment frequency data by treatment frequency, thereby obtaining the doctor data with high treatment frequency and the doctor data with low treatment frequency;

[0077] Step S553: ​​classifying the doctor-associated ECG data by the number of associated ECGs, thereby obtaining doctor data with a high number of associated ECGs and doctor data with a low number of associated ECGs;

[0078] Step S554: performing doctor intersection calculation on the doctor data with high treatment times and the doctor data with high electrocardiogram association quantity, so as to obtain the cardiomyologist data with high treatment priority; performing doctor intersection calculation on the doctor data with low treatment times and the doctor data with low electrocardiogram association quantity, so as to obtain the cardiomyologist data with low treatment priority;

[0079] Step S555: Merge data based on the high treatment priority cardiotocologist data and the low treatment priority cardiotocologist data to obtain cardiotocologist treatment priority data.

[0080] The present invention can clarify the treatment frequency of each doctor and the number of associated electrocardiograms by statistically correlating the treating doctors with the electrocardiogram data. This step lays the foundation for subsequent data classification and analysis, and helps to identify the workload and professional practice of different doctors. Classifying the number of times doctors treat helps to distinguish between doctors with high treatment frequency and doctors with low treatment frequency. In this way, doctors who are more active in treatment can be identified, thereby providing a basis for resource allocation and priority adjustment. Classifying the electrocardiogram-related data of doctors helps to distinguish between doctors with high electrocardiogram correlation numbers and doctors with low electrocardiogram correlation numbers. This step reveals the degree of activity of doctors in electrocardiogram cases and helps to evaluate their professional ability in electrocardiograms. Through intersection calculation, doctors with high treatment times and high electrocardiogram correlation numbers are combined, which helps to identify doctors with strong comprehensive treatment capabilities and mark doctors with high treatment priority. Intersection calculation of doctors with low treatment times and low electrocardiogram correlation numbers helps to mark doctors with low priority. This classification can optimize resource allocation and improve treatment efficiency. The data of high-priority and low-priority doctors are combined to obtain comprehensive treatment priority data. This will enable better physician management and scheduling, ensuring that high-priority physicians receive more resources and support, thereby improving overall care quality and efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments thereof made with reference to the following drawings:

[0082] Figure 1 It is a schematic diagram of the steps of the method for constructing a myocardial rehabilitation decision support system based on case reasoning of the present invention;

[0083] Figure 2 Detailed step flow diagram of step S1 in the present invention;

[0084] Figure 3 Detailed step flow diagram of step S12 in the present invention;

[0085] 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

[0086] 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.

[0087] 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.

[0088] 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.

[0089] To achieve this, please refer to Figures 1 to 3 The present invention provides a method for constructing a myocardial rehabilitation decision support system based on case reasoning, the method comprising the following steps:

[0090] Step S1: Acquire historical myocardial case data, and fuse the historical myocardial case data with patient case data to obtain patient myocardial case fusion data, and desensitize the patient's sensitive features on the patient's myocardial case fusion data to obtain historical myocardial case desensitization data;

[0091] In this embodiment, historical patient case data of the hospital's myocardium department are collected. These data include the patient's basic information, diagnosis records, treatment plans and efficacy, etc. Then, these data are integrated together through data fusion technology to form patient myocardium case fusion data. This process includes unifying the format of data from different sources (such as outpatient records, hospitalization records, etc.) and removing duplicate data. Next, these fused data are desensitized for sensitive information, for example, by replacing sensitive fields such as name, ID number, contact information with anonymous identifiers to ensure patient privacy protection. The final generated historical myocardium case desensitized data can be used for further analysis and model training.

[0092] Step S2: performing myocardial rehabilitation case intersection division on the historical myocardial rehabilitation case desensitization data, thereby obtaining historical myocardial rehabilitation case data, and integrating abnormal electrocardiogram waveform features based on the historical myocardial rehabilitation case data, thereby obtaining abnormal electrocardiogram waveform data of rehabilitation cases;

[0093] In this embodiment, the desensitized historical myocardial case data is used to perform the intersection division of myocardial rehabilitation cases. Specifically, by screening out the patient cases that have undergone the myocardial rehabilitation process, extracting their data and dividing them, historical myocardial rehabilitation case data is formed. Next, the electrocardiogram waveform data is extracted from these data, and the waveform features are integrated with expert knowledge and machine learning algorithms to identify abnormal electrocardiogram waveforms. This includes analyzing the peak, band, and timing characteristics of the electrocardiogram waveform, combining known abnormal patterns (such as ST segment elevation and T wave inversion) to mark abnormal waveforms, and generating abnormal electrocardiogram waveform data for rehabilitation cases.

[0094] Step S3: constructing an abnormal ECG waveform detection model based on the abnormal ECG waveform data of the rehabilitation cases, and performing abnormal waveform detection on the historical myocardial rehabilitation case data through the abnormal ECG waveform detection model, thereby obtaining abnormal ECG waveform myocardial case data;

[0095] In this embodiment, an abnormal ECG waveform detection model is constructed based on the abnormal ECG waveform data of rehabilitation cases. Deep learning methods, such as convolutional neural networks (CNN) or recurrent neural networks (RNN), can be used to extract features and classify ECG data. By training the model, abnormal waveforms in the ECG are detected and labeled. The training data includes ECG samples marked with abnormal and normal waveforms. After the model is trained, abnormal waveform detection can be performed on historical myocardial rehabilitation case data, potential abnormal ECG waveforms can be identified, and abnormal ECG waveform myocardial case data can be generated.

[0096] Step S4: performing an electrocardiogram waveform trend analysis based on the abnormal electrocardiogram waveform myocardial case data to obtain electrocardiogram waveform trend data, and calculating a myocardial rehabilitation impact factor on the electrocardiogram waveform trend data based on historical myocardial rehabilitation case data to obtain a myocardial rehabilitation impact factor;

[0097] In this embodiment, an ECG waveform trend analysis is performed on abnormal ECG waveform myocardial case data. Time series analysis methods (such as moving average method and exponential smoothing method) are used to evaluate the changing trend of the ECG waveform. For example, the changing trend of the ST segment in the ECG is analyzed to identify potential signs of myocardial ischemia. Then, combined with historical myocardial rehabilitation case data, the influencing factors of the ECG waveform trend on myocardial rehabilitation are calculated. This can be done through statistical analysis methods, such as regression analysis, to determine the degree of influence of different waveform characteristics on the rehabilitation effect and generate myocardial rehabilitation influencing factors.

[0098] Step S5: evaluating the treatment priority of myocardiologists on historical myocardial case data according to the myocardial rehabilitation influencing factors, thereby obtaining the treatment priority data of myocardiologists, and constructing a myocardial treatment physician recommendation model according to the treatment priority data of myocardiologists and the myocardial rehabilitation influencing factors;

[0099] In this embodiment, the doctor treatment priority of historical myocardial case data is evaluated based on the calculated myocardial rehabilitation impact factor. The specific method includes combining the impact factor with the doctor's treatment history and other data, and using a weighted scoring system to determine the treatment priority of each doctor. Then, based on these priority data and myocardial rehabilitation impact factors, a myocardial treatment doctor recommendation model is constructed. Algorithms such as decision trees and random forests can be used to predict and recommend the most suitable doctors to improve treatment outcomes.

[0100] Step S6: Construct a myocardial rehabilitation decision system based on the myocardial physician recommendation model and the electrocardiogram abnormal waveform detection model, and iteratively optimize and adjust the parameters of the myocardial rehabilitation decision system based on historical myocardial case data to obtain a myocardial rehabilitation decision support system.

[0101] In this embodiment, a comprehensive myocardial rehabilitation decision system is constructed based on the myocardial physician recommendation model and the electrocardiogram abnormal waveform detection model. The system may include user interface, data input module, model interface and other parts. First, it is necessary to determine the system functional requirements, which include user interface, data input module, model interface and the like. Then the system architecture is designed to clarify the functions and interfaces of each functional module. The myocardial physician recommendation model and the electrocardiogram abnormal waveform detection model are integrated into the system architecture, and then the system is adjusted and tested to optimize, so as to obtain a myocardial rehabilitation decision system. In the actual application process, the system is iteratively optimized according to the new historical myocardial case data, and the accuracy and practicality of the model are improved by continuously adjusting the model parameters (such as learning rate, regularization coefficient, etc.) and introducing new features, and finally an efficient myocardial rehabilitation decision support system is formed.

[0102] Optionally, step S1 specifically includes:

[0103] Step S11: Acquire historical myocardial case data, and perform structured data classification on the historical myocardial case data, thereby obtaining structured myocardial case data and unstructured myocardial case data;

[0104] In this embodiment, obtaining historical myocardial case data includes extracting electronic health records (EHR) and electrocardiogram (ECG) data from a hospital database. The data is classified according to the format, for example, structured data (such as medical records, laboratory test results) and unstructured data (such as doctor's handwritten notes, image files) are stored separately. In a specific implementation, a data annotation tool is used to convert ECG images and text annotations into structured fields and store them in a SQL database, while images and text files are saved in a file system to form structured and unstructured myocardial case data.

[0105] Step S12: performing unstructured patient myocardial feature fusion on the unstructured myocardial case data, thereby obtaining unstructured patient myocardial fusion data;

[0106] In this embodiment, unstructured myocardial case data (such as electrocardiogram images and doctor's notes) are processed. Image processing technology (such as convolutional neural network, CNN) is used to extract features from electrocardiogram images, and natural language processing (NLP) technology is used to perform semantic analysis on doctor's notes. Through feature fusion, the extracted feature data is integrated with the text analysis results to form unstructured patient myocardial fusion data, for example, the abnormal waveform features in the image and the condition description in the notes are integrated into a unified data set.

[0107] Step S13: fusing the patient's treatment course physiological characteristics according to the structured myocardial case data, thereby obtaining structured patient physiological fusion data;

[0108] In this embodiment, the treatment course physiological feature fusion is performed based on the information in the structured myocardial case data (such as the patient's physiological parameters and treatment records). Statistical analysis and machine learning methods (such as principal component analysis, PCA) are used to reduce the dimension and select features of physiological parameters to form structured patient physiological fusion data. For example, the main features are extracted from the heart rate and blood pressure data of multiple examinations, and a comprehensive physiological indicator data set is synthesized.

[0109] Step S14: merging the unstructured patient myocardial fusion data and the structured patient physiological fusion data into patient information cases, thereby obtaining the patient myocardial case fusion data;

[0110] In this embodiment, the unstructured patient myocardial fusion data is merged with the structured patient physiological fusion data. Data fusion technology, such as multimodal data fusion method, is used to integrate data from different sources (ECG features and physiological features). By building a joint database or data table, the two types of data are associated by patient ID to form comprehensive patient myocardial case fusion data, such as combining the fused ECG features and physiological features into a unified data set for further analysis.

[0111] Step S15: Desensitize the patient's sensitive features on the patient's myocardial case fusion data, so as to obtain historical myocardial case desensitized data.

[0112] In this embodiment, the sensitive features of the patient's myocardial case fusion data are desensitized. Data desensitization techniques, such as data masking or data encryption, are used to remove or replace the patient's personal identification information (such as name, ID number) from the data set. For example, anonymization tools are used to convert personal identification information into a universal identifier to generate historical myocardial case desensitized data to protect patient privacy.

[0113] Optionally, step S12 is specifically:

[0114] Step S121: extracting medical record summary records, patient electrocardiogram records and medical image records from the unstructured myocardial case data, thereby obtaining myocardial medical record summary data, myocardial patient electrocardiogram sets and myocardial patient medical image sets;

[0115] In this embodiment, the medical record summary data of the myocardium patients, including the chief complaint, diagnosis and treatment records, are extracted from the unstructured myocardium case data. Secondly, the electrocardiogram records of the patients stored in DICOM format are extracted. Finally, the medical imaging records of the patients, such as cardiac CT scan images, are extracted and format converted for easy processing.

[0116] Step S122: integrating myocardial description features of the cardiology medical record summary data to obtain myocardial description data;

[0117] In this embodiment, the main descriptive features of the myocardium (such as myocardial thickness, myocardial function, etc.) are extracted from the cardiology medical record summary data, and these features are integrated and summarized using text mining technology to form structured myocardial description data. For example, the myocardium-related descriptions in the medical records are identified by natural language processing technology and converted into a standardized feature list.

[0118] Step S123: extracting dynamic structural features of the myocardium scanned by CT according to the medical image set of the patients in the myocardium department, thereby obtaining dynamic structural data of the myocardium scanned by CT, and constructing a three-dimensional dynamic structural model of the myocardium of the patient according to the dynamic structural data of the myocardium scanned by CT;

[0119] In this embodiment, the CT scan images of patients in the cardiology department are processed to extract the dynamic structural characteristics of the myocardium, such as the contraction and relaxation of the myocardium. The myocardial tissue is separated by image segmentation technology, and a three-dimensional dynamic structural model of the myocardium is generated by applying three-dimensional reconstruction technology, which can accurately reflect the structural changes of the myocardium under different cardiac cycles.

[0120] Step S124: filling the patient's myocardial three-dimensional dynamic structure model with myocardial structure details according to the myocardial description data, thereby obtaining the patient's myocardial dynamic structure model;

[0121] In this embodiment, the myocardial description data is used to fill in the patient's myocardial three-dimensional dynamic structure model in detail. By comparing the myocardial function description in the medical record with the existing data of the model, the myocardial thickness, morphology and other details in the model are adjusted to obtain a more accurate patient's myocardial dynamic structure model.

[0122] Step S125: integrating the dynamic fluctuation characteristics of the electrocardiogram according to the electrocardiogram set of the patients in the myocardium department, thereby obtaining the dynamic fluctuation data of the electrocardiogram, and performing myocardial dynamic motion simulation on the dynamic fluctuation data of the electrocardiogram through the dynamic structural model of the myocardium of the patients, thereby obtaining the dynamic simulation data of the myocardium of the patients;

[0123] In this embodiment, the extracted ECG data is subjected to dynamic fluctuation feature analysis, such as the amplitude, frequency and rhythm changes of the ECG. By combining with the patient's myocardial dynamic structural model, the influence of ECG fluctuations on the dynamic movement of the myocardium is simulated, and myocardial dynamic simulation data is generated to help evaluate the actual functional state of the myocardium.

[0124] Step S126: performing patient myocardial dynamic feature fusion according to the patient myocardial dynamic simulation data, thereby obtaining unstructured patient myocardial fusion data.

[0125] In this embodiment, the patient's myocardial dynamic simulation data and myocardial description data are combined to perform feature fusion. Through data fusion technology, dynamic features from different sources are comprehensively analyzed to form a comprehensive unstructured myocardial fusion data set for further clinical analysis and personalized treatment plan formulation.

[0126] Optionally, step S13 is specifically:

[0127] Step S131: extracting drug dosage records, disease course records and physiological coefficient records from the structured myocardial case data, thereby obtaining drug dosage data, disease course data and physiological coefficient data of the patient;

[0128] In this embodiment, drug dosage records, medical history records, and physiological coefficient records are extracted from structured myocardial case data. Specifically, drug dosage records are extracted as the dosage and time of each medication, medical history records are extracted as the disease progression and symptom changes of each examination, and physiological coefficient records are extracted as the patient's physiological data during the examination, such as blood pressure, heart rate, etc. For example, by parsing the data stored in the electronic health record system, the drug dosage, disease description, and physiological parameter data of a patient at a certain point in time are obtained to form a structured patient data set.

[0129] Step S132: dividing the patient's treatment stage according to the patient's disease course data, thereby obtaining the patient's treatment stage data;

[0130] In this embodiment, the treatment stages are divided according to the patient's medical history data. The specific method is to divide the entire treatment process into different stages, such as early, middle and late stages, according to the timestamp and condition description in the medical history record. For example, if the medical history record shows that the patient's condition is stable in the first three months, deteriorates in the middle stage, and improves in the late stage, these three periods can be marked as early, middle and late stages, and the specific time range and condition characteristics of each stage can be recorded.

[0131] Step S133: performing a drug characteristic analysis on the drug dosage data, thereby obtaining drug dosage characteristic data;

[0132] In this embodiment, the drug dosage data is analyzed for drug characteristics. The specific analysis includes calculating the average dosage, dosage frequency, dosage range, etc. of the drug. For example, by statistically analyzing the usage data of anti-heart failure drugs, it is determined that the average daily dosage of a certain drug is 50 mg, the dosage range is 30-70 mg, and the usage frequency is once a day. These data help understand the characteristics and variations of drugs in actual applications.

[0133] Step S134: performing physiological coefficient fluctuation correlation according to the drug dosage characteristic data and the patient's physiological coefficient data, thereby obtaining drug dosage correlated physiological fluctuation data;

[0134] In this embodiment, a physiological coefficient fluctuation correlation analysis is performed based on the drug dosage characteristic data and the patient's physiological coefficient data. The specific method is to statistically analyze the relationship between the change in drug dosage and the fluctuation of the physiological coefficient, for example, using regression analysis or correlation analysis to determine whether the increase in drug dosage is related to a change in a physiological indicator such as blood pressure. For example, if it is found that the fluctuation amplitude of the patient's blood pressure increases after increasing the dosage of a certain drug, the correlation data between the drug dosage and the blood pressure fluctuation is recorded.

[0135] Step S135: fusing the patient treatment stage data with the treatment stage physiological characteristics according to the drug dosage-associated physiological fluctuation data, thereby obtaining structured patient physiological fusion data.

[0136] In this embodiment, the data of physiological fluctuations associated with the dosage of the patient's treatment stage are used to fuse the physiological characteristics of the treatment stage. Specifically, the data of association between the dosage of the dosage and the physiological fluctuations are combined with the treatment stage data to extract the physiological characteristics of each treatment stage. For example, by integrating the dosage of the dosage and the physiological fluctuation data in the mid-term stage, the average physiological index characteristics of the stage are obtained. Finally, a structured patient physiological fusion data set is formed, including the comprehensive physiological characteristics of each stage, for further analysis and decision-making.

[0137] Optionally, step S2 specifically includes:

[0138] Step S21: extracting treatment and rehabilitation stage features from the historical myocardial case desensitization data, thereby obtaining historical myocardial case treatment stage data, and classifying the historical myocardial case treatment stage data by treatment stage length, thereby obtaining long treatment stage case data and short treatment stage case data;

[0139] In this embodiment, the original data of historical myocardial cases are obtained, including the patient's electrocardiogram, medical records and treatment plan. By applying data preprocessing algorithms (such as denoising and standardization), the relevant features of the treatment stage, such as treatment duration, electrocardiogram parameters, drug dosage, etc., are extracted. The length of the treatment stage is classified using a machine learning algorithm (such as a support vector machine SVM or a random forest), and the treatment stage is divided into a long stage (such as more than 6 months) and a short stage (such as less than 6 months). The classification results obtained will be long treatment stage case data and short treatment stage case data.

[0140] Step S22: classifying the historical myocardial case treatment stage data by treatment stage review, thereby obtaining case treatment stage review data;

[0141] In this embodiment, the treatment stage data is reviewed and classified. The review can be performed by checking the review date, electrocardiogram waveform and doctor's review suggestion in the patient record. The review data is marked by a classification algorithm (such as K-nearest neighbor algorithm KNN or decision tree) and divided into two categories: "reviewed" and "not reviewed", so as to obtain the review data of the case treatment stage.

[0142] Step S23: performing myocardial variable statistics review according to the case treatment stage review data, thereby obtaining the myocardial variable data review, and classifying the myocardial variable fluctuation amplitude of the myocardial variable data review, thereby obtaining high-amplitude fluctuation myocardial variable review data and low-amplitude fluctuation myocardial variable review data;

[0143] In this embodiment, the review data in the historical myocardial case is used to perform statistical analysis on myocardial variables (such as heart rate, ST segment changes, etc.). The fluctuation amplitude (such as standard deviation or range) of the myocardial variables during the review period is calculated. According to the statistical results, the reviewed myocardial variables are divided into two categories: high-amplitude fluctuation (such as standard deviation greater than 0.5mV) and low-amplitude fluctuation (such as standard deviation less than 0.5mV). The generated classification results will include high-amplitude fluctuation myocardial variable review data and low-amplitude fluctuation myocardial variable review data.

[0144] Step S24: classifying the high-amplitude fluctuation myocardial variable review data and the low-amplitude fluctuation myocardial variable review data by the high-proportion myocardial variable fluctuation amplitude ratio according to the historical myocardial case treatment stage data, thereby obtaining high-proportion case data of the high-amplitude fluctuation treatment stage and high-proportion case data of the low-amplitude fluctuation treatment stage;

[0145] In this embodiment, the high-amplitude fluctuation and low-amplitude fluctuation myocardial variable review data are further classified according to the treatment stage data in the historical myocardial department cases. The proportion of each type of fluctuation amplitude in the treatment stage is calculated (for example, the proportion of high-amplitude fluctuation in the amplitude fluctuation of the entire treatment stage). The cases are classified using a proportion threshold (such as 70%) to obtain case data with a high proportion of high-amplitude fluctuation treatment stages and case data with a high proportion of low-amplitude fluctuation treatment stages.

[0146] Step S25: performing a treatment case intersection operation on the long treatment stage case data and the low amplitude fluctuation treatment stage case data with a high proportion, thereby obtaining slow-acting treatment and rehabilitation case data; performing a treatment case intersection operation on the short treatment stage case data and the high amplitude fluctuation treatment stage case data with a high proportion, thereby obtaining fast-acting treatment and rehabilitation case data;

[0147] In this embodiment, the case data of the long treatment stage and the case data of the low amplitude fluctuation treatment stage with a high proportion are intersected to obtain the case data of the slow-acting treatment and rehabilitation. The case data of the short treatment stage and the case data of the high amplitude fluctuation treatment stage with a high proportion are intersected to obtain the case data of the fast-acting treatment and rehabilitation. The common features of these cases are determined using a set operation method, such as a set intersection operation.

[0148] Step S26: integrating the slow-acting treatment rehabilitation case data and the fast-acting treatment rehabilitation case data, thereby obtaining historical myocardial rehabilitation case data;

[0149] In this embodiment, the slow-acting treatment rehabilitation case data and the fast-acting treatment rehabilitation case data are integrated. Specifically, the integration process includes data cleaning, standardization, and feature matching. For example, if the slow-acting treatment rehabilitation case data includes electrocardiogram features A, B, C and treatment time X, and the fast-acting treatment rehabilitation case data includes electrocardiogram features D, E, F and treatment time Y, then during the integration process, it is necessary to unify the representation of these features and combine them into a unified data set to generate historical myocardial rehabilitation case data. These integrated data will be used for further analysis and modeling.

[0150] Step S27: Integrate abnormal electrocardiogram waveform features based on historical myocardial rehabilitation case data to obtain abnormal electrocardiogram waveform data of rehabilitation cases.

[0151] In this embodiment, abnormal ECG waveform features are integrated based on historical myocardial rehabilitation case data. Specifically, it is first necessary to extract all ECG waveform data from the integrated data set, and then apply an anomaly detection algorithm (e.g., a classification model based on machine learning or a statistical analysis method) to identify and mark abnormal waveforms. For example, by comparing the ECG waveform in the rehabilitation case data with the normal waveform model, abnormal waveform patterns can be detected. Finally, these abnormal ECG waveform data are organized into a data set and marked as abnormal ECG waveform data for rehabilitation cases. These data will be used for further abnormal pattern analysis and clinical research.

[0152] Optionally, step S27 is specifically:

[0153] Step S271: extracting electrocardiogram records from historical myocardial rehabilitation case data, thereby obtaining an electrocardiogram set of slow-acting treatment rehabilitation cases and an electrocardiogram set of fast-acting treatment rehabilitation cases;

[0154] In this embodiment, the electrocardiogram records of the slow-acting treatment rehabilitation cases and the fast-acting treatment rehabilitation cases are extracted from the historical myocardial rehabilitation case data to ensure that the data contains multiple electrocardiogram cycles. The records are classified to form the electrocardiogram set of the slow-acting treatment rehabilitation cases and the electrocardiogram set of the fast-acting treatment rehabilitation cases.

[0155] Step S272: performing Fourier transform on the slow-acting treatment rehabilitation case electrocardiogram set and the fast-acting treatment rehabilitation case electrocardiogram set respectively, so as to obtain the slow-acting rehabilitation case electrocardiogram spectrum and the fast-acting rehabilitation case electrocardiogram spectrum;

[0156] In this embodiment, Fourier transform is applied to the two sets of ECG data. Fourier transform converts time domain signals into frequency domain signals so as to analyze their spectrum characteristics. For example, the scipy.fft library in Python is used to perform Fourier transform on each ECG data to generate the ECG spectrum of the slow-acting rehabilitation case and the ECG spectrum of the fast-acting rehabilitation case.

[0157] Step S273: performing spectrum amplitude threshold statistics on the ECG spectrum of the slow-acting rehabilitation case and the ECG spectrum of the fast-acting rehabilitation case, respectively, so as to obtain the slow-acting spectrum amplitude threshold and the fast-acting spectrum amplitude threshold;

[0158] In this embodiment, the amplitude threshold of each group of spectrum is counted. The ECG spectrum of the slow-acting rehabilitation case and the ECG spectrum of the fast-acting rehabilitation case are statistically analyzed respectively, the amplitude distribution of each spectrum is calculated, and a threshold is set. For example, the 95% quantile of the spectrum amplitude is used as the threshold. This will obtain the amplitude threshold of the slow-acting spectrum and the amplitude threshold of the fast-acting spectrum respectively.

[0159] Step S274: classifying the abnormal spectrum amplitude of the slow-acting rehabilitation case ECG spectrum according to the slow-acting spectrum amplitude threshold, thereby obtaining the abnormal slow-acting rehabilitation case ECG spectrum; classifying the abnormal spectrum amplitude of the fast-acting rehabilitation case ECG spectrum according to the fast-acting spectrum amplitude threshold, thereby obtaining the abnormal fast-acting rehabilitation case ECG spectrum;

[0160] In this embodiment, abnormal spectrum amplitudes are classified according to the slow-acting spectrum amplitude threshold and the fast-acting spectrum amplitude threshold. For the slow-acting recovery case, the spectrum with an amplitude exceeding the threshold is marked as abnormal. Similarly, the fast-acting recovery case is classified in the same way. These abnormal spectra are recorded for subsequent analysis, and abnormal slow-acting recovery case ECG spectra and abnormal fast-acting recovery case ECG spectra are generated respectively.

[0161] Step S275: performing time domain conversion on the ECG spectrum of the abnormal slow-acting recovery case and the ECG spectrum of the abnormal fast-acting recovery case, respectively, so as to obtain ECG fluctuation data of the abnormal slow-acting recovery case and ECG fluctuation data of the abnormal fast-acting recovery case;

[0162] In this embodiment, the ECG spectrum of the abnormal slow-acting recovery case and the ECG spectrum of the abnormal fast-acting recovery case are respectively converted into the time domain. The abnormal spectrum is converted back into time domain data by applying inverse Fourier transform. For example, scipy.fft.ifft is used for inverse transformation to obtain ECG fluctuation data of the abnormal slow-acting recovery case and ECG fluctuation data of the abnormal fast-acting recovery case.

[0163] Step S276: Integrate the abnormal electrocardiogram waveform features of the abnormal slow-acting rehabilitation case electrocardiogram fluctuation data and the abnormal fast-acting rehabilitation case electrocardiogram fluctuation data, so as to obtain the abnormal electrocardiogram waveform data of the rehabilitation case.

[0164] In this embodiment, the abnormal electrocardiogram waveform features are integrated. Waveform features are extracted from the acquired abnormal time domain data, for example, by detecting the R wave peak value, interval and other features of the waveform, and integrating to form abnormal electrocardiogram waveform data. These features are extracted and classified using image processing tools, and finally abnormal electrocardiogram waveform data of rehabilitation cases are generated.

[0165] Optionally, step S4 is specifically:

[0166] Step S41: segmenting the electrocardiogram waveform signal according to the abnormal electrocardiogram waveform myocardial case data, thereby obtaining abnormal electrocardiogram waveform signal data;

[0167] In this embodiment, during the electrocardiogram monitoring process, an electrocardiogram instrument is first used to collect electrocardiogram waveform signals. For the continuous electrocardiogram signals in the abnormal electrocardiogram waveform myocardial case data, an adaptive threshold algorithm is used to segment the signals. These thresholds are determined by a series of preset standards (such as R wave peak, QRS complex wave). By detecting and locating the R wave in the electrocardiogram waveform, the electrocardiogram signal is divided into several cardiac cycle signal segments. Each signal segment represents a complete heartbeat cycle, thereby extracting electrocardiogram waveform signal data containing abnormalities. These abnormal signals can be features such as ST segment elevation and T wave inversion for subsequent processing.

[0168] Step S42: performing time series trend line fitting on the abnormal electrocardiogram waveform signal data, thereby obtaining electrocardiogram waveform signal trend data;

[0169] In this embodiment, the least square method is used to fit the time series trend line for the obtained abnormal electrocardiogram waveform signal data. By plotting the signal points in each cycle on the time axis and fitting a trend line with a linear regression algorithm, the overall trend of the electrocardiogram signal within a certain time range can be described. This trend line can be linear or nonlinear, and the specific selection is based on the fluctuation characteristics of the signal. The trend data after fitting can help identify the basic change pattern and long-term trend of the electrocardiogram signal.

[0170] Step S43: performing fluctuation measurement classification on the electrocardiogram waveform signal trend data, thereby obtaining excessive fluctuation waveform signal trend data and normal fluctuation waveform signal trend data;

[0171] In this embodiment, the obtained ECG waveform trend data is subjected to fluctuation measurement. Statistical indicators such as standard deviation and variance can be used to measure the volatility of the signal. According to the calculation results, the trend data is classified into excessive fluctuation and normal fluctuation. For example, if the standard deviation exceeds a preset threshold, it is classified as "excessive fluctuation waveform signal trend data", otherwise it is classified as "normal fluctuation waveform signal trend data". This classification helps to further analyze the stability and abnormality of the signal.

[0172] Step S44: performing waveform signal trend correlation on the abnormal electrocardiogram waveform signal data according to the excessive fluctuation waveform signal trend data, thereby obtaining electrocardiogram waveform degradation trend data; performing waveform signal trend correlation on the abnormal electrocardiogram waveform signal data according to the normal fluctuation waveform signal trend data, thereby obtaining electrocardiogram waveform improvement trend data;

[0173] In this embodiment, waveform signal trend correlation analysis is performed on abnormal electrocardiogram waveform signal data based on excessive fluctuation waveform signal trend data and normal fluctuation waveform signal trend data. A correlation analysis algorithm (such as Pearson correlation coefficient) is applied to each trend data to determine the relationship between the abnormal signal and excessive fluctuation or normal fluctuation. Through this analysis, degradation trend data (such as aggravation of waveform distortion) and improvement trend data (such as gradual return of the waveform to normal) of the electrocardiogram waveform can be obtained.

[0174] Step S45: merging the electrocardiogram waveform degradation trend data and the electrocardiogram waveform improvement trend data to obtain the electrocardiogram waveform trend data;

[0175] In this embodiment, the obtained ECG waveform degradation trend data and improvement trend data are merged. The merging operation can be performed by time series merging or data aggregation. The specific method includes splicing or reorganizing the degradation trend and improvement trend data according to time or waveform characteristics to form a comprehensive ECG waveform trend data set. This data set provides a comprehensive perspective, reflecting the overall changes in the ECG waveform.

[0176] Step S46: Calculate the myocardial rehabilitation impact factor on the electrocardiogram waveform trend data according to the historical myocardial rehabilitation case data, so as to obtain the myocardial rehabilitation impact factor.

[0177] In this embodiment, historical myocardial rehabilitation case data is used to calculate the influencing factors of the obtained ECG waveform trend data. By comparing and analyzing the ECG waveform trend data with the rehabilitation data (such as recovery speed and rehabilitation effect) in historical rehabilitation cases, regression analysis or other statistical methods are used to calculate the influencing factors of myocardial rehabilitation. Specifically, a regression model can be constructed, with the waveform trend data as the independent variable and the rehabilitation effect as the dependent variable, so as to determine the weight and effect of each influencing factor. This will help predict and evaluate the effect of myocardial rehabilitation and provide personalized rehabilitation advice.

[0178] Optionally, step S46 is specifically:

[0179] Step S461: extracting left ventricular ejection fraction features and electrocardiogram waveform features from historical myocardial rehabilitation case data, thereby obtaining rehabilitation case left ventricular ejection fraction data and rehabilitation case electrocardiogram waveform data;

[0180] In this embodiment, when processing historical myocardial rehabilitation case data, it is first necessary to extract cardiac ultrasound data from the historical myocardial rehabilitation case data to calculate the left ventricular ejection fraction (LVEF). Next, waveform features such as R wave amplitude, PR interval, and QT interval are extracted from the electrocardiogram (ECG) data. Specialized cardiac ultrasound analysis software and ECG feature extraction algorithms can be used. The extracted LVEF data should include detailed statistical data for each rehabilitation case, and the ECG waveform data should be converted into digital features for subsequent analysis.

[0181] Step S462: constructing a myocardial rehabilitation index evaluation system based on the left ventricular ejection fraction data of the rehabilitation case and the electrocardiogram waveform data of the rehabilitation case, and performing myocardial rehabilitation index evaluation according to the myocardial rehabilitation index evaluation system, thereby obtaining myocardial rehabilitation index evaluation data;

[0182] In this embodiment, a myocardial rehabilitation index evaluation system is constructed using the extracted LVEF data and ECG waveform features. This can be done by determining the weights of each indicator and standardizing these indicators. For example, a weighted scoring system is used to combine LVEF and ECG features to evaluate the effect of myocardial rehabilitation. The evaluation system can include indicators such as recovery progress and electrocardiogram stability. Through statistical analysis, the generated evaluation data should include a comprehensive score and sub-item scores for each case.

[0183] Step S463: extracting waveform amplitude features from the electrocardiogram waveform trend data, thereby obtaining trend waveform amplitude data;

[0184] In this embodiment, when processing the ECG waveform trend data, it is first necessary to arrange the ECG waveform data in time series, and then use Fourier transform and other methods to extract waveform amplitude features. Signal processing tools can be used to extract features such as the maximum amplitude, minimum amplitude, and average amplitude of the waveform to generate trend waveform amplitude data for further analysis.

[0185] Step S464: constructing a myocardial rehabilitation linear regression model according to the myocardial rehabilitation index evaluation data and the trend waveform amplitude data;

[0186] In this embodiment, the myocardial rehabilitation linear regression model is constructed using the obtained myocardial rehabilitation index evaluation data and the trend waveform amplitude data obtained in step S463. In the model construction, LVEF and waveform amplitude characteristics are used as independent variables, and myocardial rehabilitation effect is used as a dependent variable to perform linear regression analysis. Specifically, it includes selecting a regression algorithm, verifying the accuracy of the model, and adjusting the model parameters to optimize the prediction results.

[0187] Step S465: Calculate the myocardial rehabilitation influencing factor according to the myocardial rehabilitation linear regression model, so as to obtain the myocardial rehabilitation influencing factor.

[0188] In this embodiment, a linear regression model is used to calculate the influencing factors of myocardial rehabilitation. The historical myocardial rehabilitation case data is input into the model to obtain the value of each influencing factor. The calculation of the influencing factor should include the comparison of the model prediction results with the actual data, and the actual contribution of each factor to the rehabilitation effect is determined by statistical methods to generate the final influencing factor report.

[0189] Optionally, step S5 specifically includes:

[0190] Step S51: extracting electrocardiogram waveform features according to the myocardial rehabilitation influencing factor, thereby obtaining the myocardial rehabilitation electrocardiogram waveform influencing factor;

[0191] In this embodiment, the ECG waveform influencing factor data extracted from the myocardial rehabilitation influencing factors are obtained, and these data may include various features of the waveform such as amplitude, waveform interval, shape, etc. The ECG waveform is feature extracted using a signal processing algorithm (such as fast Fourier transform FFT, wavelet transform, etc.). The extracted features include the main frequency components, the amplitude of the waveform, the time interval between waveforms, and the influencing factors affecting the ECG waveform, etc. The extracted features are formatted into a data structure that can be used for subsequent analysis, such as a table or database record, for further processing.

[0192] Step S52: extracting the characteristics of the myocardial doctor and the electrocardiogram of the cases from the historical myocardial case data, thereby obtaining the myocardial doctor data and the electrocardiogram data of the cases;

[0193] In this embodiment, feature extraction is performed on the electrocardiogram of each case in the historical myocardial case data, including the use of Fourier transform to analyze the frequency domain features and time domain features of the electrocardiogram. In addition, the characteristics of the doctor are extracted, including the years of clinical experience, the number of successfully treated cases, etc. These data are organized into a database, in which the electrocardiogram data includes P wave duration, QRS wave amplitude, etc., and the doctor data includes personal information, qualifications, etc.

[0194] Step S53: performing electrocardiogram waveform similarity calculation on the case electrocardiogram data and the myocardial rehabilitation electrocardiogram waveform influencing factor, thereby obtaining case electrocardiogram similarity data, and performing high-similarity case statistics on the case electrocardiogram similarity data, thereby obtaining high-similarity case electrocardiogram data;

[0195] In this embodiment, the dynamic time warping (DTW) algorithm is used to calculate the similarity between the case ECG and the myocardial rehabilitation ECG waveform influencing factors. Specifically, the ECG of each case is compared with the standard waveform features to calculate the similarity score. Then, the similarity scores are sorted and the top 10 cases with high similarity are counted. For these high-similarity cases, detailed statistical analysis is further performed to extract the ECG data of these cases for subsequent analysis and association.

[0196] Step S54: Associating treatment cases according to the ECG data of cases with high similarity and the data of myocardial treating doctors, thereby obtaining treating doctor-ECG associated data;

[0197] In this embodiment, the ECG data of high-similarity cases are matched with the treating doctor data. The specific operation is to associate the ECG data of each high-similarity case with the processed doctor feature data to construct a treating doctor-ECG association data table. For example, the SQL database query statement "SELECT * FROM case_data JOINdoctor_data ON case_data.doctor_id = doctor_data.id WHERE case_data.similarity_score>0.9" is used to extract relevant data. In this way, the treatment records and effect data of each doctor for high-similarity ECG cases can be obtained.

[0198] Step S55: evaluating the cardiotocologist's treatment priority according to the treating physician-ECG association data, thereby obtaining the cardiotocologist's treatment priority data;

[0199] In this embodiment, the treatment priority of each doctor is calculated based on the treating doctor-ECG association data. This can be achieved through a weighted scoring system, for example, using the success rate of each doctor in high similarity cases as a weighting factor. The specific operation is to calculate the treatment success rate and number of cases of each doctor in these cases, and then standardize them to obtain the priority score. For example, the formula "Priority Score = (Success Rate * 0.7) + (CaseCount * 0.3)" is used for weighted scoring to obtain the doctor's treatment priority data.

[0200] Step S56: Construct a myocardial physician recommendation model based on the cardiologist treatment priority data and myocardial rehabilitation influencing factors.

[0201] In this embodiment, a treatment physician recommendation model is constructed using the treatment priority data of cardiotocologists and myocardial rehabilitation influencing factors. A weighted linear regression model is used, and the doctor priority score is used as one of the input features, while the myocardial rehabilitation influencing factor is used as another input feature. The model training process uses historical data for cross-validation to optimize the model parameters. After training, the model can predict the most suitable treating physician based on the electrocardiogram waveform characteristics and rehabilitation influencing factors of new cases. For example, the linear regression model can be implemented using the scikit-learn library in Python. It is also possible to use the characteristics of priority data and rehabilitation influencing factors to build a recommendation model (such as a weighted score-based recommendation system, a collaborative filtering model).

[0202] Optionally, step S55 is specifically:

[0203] Step S551: performing doctor treatment frequency statistics and doctor-associated electrocardiogram statistics according to the treating doctor-electrocardiogram association data, thereby obtaining doctor treatment frequency data and doctor-associated electrocardiogram data;

[0204] In this embodiment, the number of treatments by each doctor (for example, the total number of ECG cases handled by each doctor in the past year) is extracted from the treating doctor-ECG association data. At the same time, the number of ECGs associated with each doctor (i.e., the number of ECGs handled by each doctor) is counted.

[0205] Step S552: classifying the doctor treatment frequency data by treatment frequency, thereby obtaining the doctor data with high treatment frequency and the doctor data with low treatment frequency;

[0206] In this embodiment, the doctor's treatment times data is statistically analyzed, and the doctors are classified according to the treatment times. Two thresholds are set, for example, those with treatment times greater than or equal to 50 times are "high treatment times doctors", and those with treatment times less than 50 times are "low treatment times doctors". Data classification tools (such as the pandas library in Python) can be used to divide doctors into two groups: high treatment times doctor data including doctors with treatment times greater than or equal to 50 times; low treatment times doctor data including doctors with treatment times less than 50 times.

[0207] Step S553: ​​classifying the doctor-associated ECG data by the number of associated ECGs, thereby obtaining doctor data with a high number of associated ECGs and doctor data with a low number of associated ECGs;

[0208] In this embodiment, the doctor's associated ECG number data is classified. Two thresholds are set. For example, doctors with an associated ECG number greater than or equal to 100 are "doctors with a high number of associated ECGs", and doctors with an associated ECG number less than 100 are "doctors with a low number of associated ECGs". The doctors are divided into two groups using a data processing tool: doctors with a high number of associated ECGs, including doctors with an associated ECG number greater than or equal to 100; and doctors with a low number of associated ECGs, including doctors with an associated ECG number less than 100.

[0209] Step S554: performing doctor intersection calculation on the doctor data with high treatment times and the doctor data with high electrocardiogram association quantity, so as to obtain the cardiomyologist data with high treatment priority; performing doctor intersection calculation on the doctor data with low treatment times and the doctor data with low electrocardiogram association quantity, so as to obtain the cardiomyologist data with low treatment priority;

[0210] In this embodiment, the "doctor data set with a high number of treatments" and the "doctor data set with a high number of electrocardiogram associations" are intersected to find doctors who meet both conditions. These doctors are considered "cardiomyologists with high treatment priority." Similarly, the "doctor data set with a low number of treatments" and the "doctor data set with a low number of electrocardiogram associations" are intersected to obtain "cardiomyologists with low treatment priority." Using set operation tools (such as Python's set operations) for calculations, doctors with a high number of treatments and a high number of electrocardiogram associations are classified as cardiomyologists with high treatment priority, and doctors with a low number of treatments and a low number of electrocardiogram associations are classified as cardiomyologists with low treatment priority.

[0211] Step S555: Merge data based on the high treatment priority cardiotocologist data and the low treatment priority cardiotocologist data to obtain cardiotocologist treatment priority data.

[0212] In this embodiment, the obtained "high treatment priority cardiomyologist dataset" and "low treatment priority cardiomyologist dataset" are merged by doctor ID to form a complete cardiomyologist treatment priority dataset. After merging, they are sorted by priority (for example, sorted by high and low treatment priority). The final output dataset contains all cardiomyologists and their treatment priorities.

[0213] 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.

[0214] 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 method for constructing a myocardial rehabilitation decision support system based on case-based reasoning, characterized in that: The following steps are involved: Step S1: Acquire historical myocardial case data, and fuse the historical myocardial case data with patient case data to obtain patient myocardial case fusion data, and desensitize the patient's sensitive features on the patient's myocardial case fusion data to obtain historical myocardial case desensitization data; Step S2: performing myocardial rehabilitation case intersection division on the historical myocardial rehabilitation case desensitization data, thereby obtaining historical myocardial rehabilitation case data, and integrating abnormal electrocardiogram waveform features based on the historical myocardial rehabilitation case data, thereby obtaining abnormal electrocardiogram waveform data of rehabilitation cases; Step S3: constructing an abnormal ECG waveform detection model based on the abnormal ECG waveform data of the rehabilitation cases, and performing abnormal waveform detection on the historical myocardial rehabilitation case data through the abnormal ECG waveform detection model, thereby obtaining abnormal ECG waveform myocardial case data; Step S4: Performing an electrocardiogram waveform trend analysis based on abnormal electrocardiogram waveform myocardial case data to obtain electrocardiogram waveform trend data, and calculating myocardial rehabilitation impact factors on the electrocardiogram waveform trend data based on historical myocardial rehabilitation case data to obtain myocardial rehabilitation impact factors; Step S4 is specifically: Step S41: segmenting the electrocardiogram waveform signal according to the abnormal electrocardiogram waveform myocardial case data, thereby obtaining abnormal electrocardiogram waveform signal data; Step S42: performing time series trend line fitting on the abnormal electrocardiogram waveform signal data, thereby obtaining electrocardiogram waveform signal trend data; Step S43: performing fluctuation measurement classification on the electrocardiogram waveform signal trend data, thereby obtaining excessive fluctuation waveform signal trend data and normal fluctuation waveform signal trend data; Step S44: performing waveform signal trend correlation on the abnormal electrocardiogram waveform signal data according to the excessive fluctuation waveform signal trend data, thereby obtaining electrocardiogram waveform degradation trend data; According to the normal fluctuation waveform signal trend data, the abnormal electrocardiogram waveform signal data is correlated with the waveform signal trend, so as to obtain the electrocardiogram waveform improvement trend data; Step S45: merging the electrocardiogram waveform degradation trend data and the electrocardiogram waveform improvement trend data to obtain the electrocardiogram waveform trend data; Step S46: Calculate the myocardial rehabilitation impact factor on the electrocardiogram waveform trend data according to the historical myocardial rehabilitation case data, thereby obtaining the myocardial rehabilitation impact factor; Step S5: Evaluate the treatment priority of myocardiologists on historical myocardial case data according to the myocardial rehabilitation influencing factors, thereby obtaining the treatment priority data of myocardial physicians, and construct a myocardial physician recommendation model based on the treatment priority data of myocardial physicians and the myocardial rehabilitation influencing factors; Step S5 is specifically as follows: Step S51: extracting electrocardiogram waveform features according to the myocardial rehabilitation influencing factor, thereby obtaining the myocardial rehabilitation electrocardiogram waveform influencing factor; Step S52: extracting the characteristics of the myocardial doctor and the electrocardiogram of the cases from the historical myocardial case data, thereby obtaining the myocardial doctor data and the electrocardiogram data of the cases; Step S53: performing electrocardiogram waveform similarity calculation on the case electrocardiogram data and the myocardial rehabilitation electrocardiogram waveform influencing factor, thereby obtaining case electrocardiogram similarity data, and performing high-similarity case statistics on the case electrocardiogram similarity data, thereby obtaining high-similarity case electrocardiogram data; Step S54: Associating treatment cases according to the ECG data of cases with high similarity and the data of myocardial treating doctors, thereby obtaining treating doctor-ECG associated data; Step S55: evaluating the cardiotocologist's treatment priority according to the treating physician-ECG association data, thereby obtaining the cardiotocologist's treatment priority data; Step S56: constructing a myocardial doctor recommendation model based on the myocardial doctor treatment priority data and myocardial rehabilitation influencing factors; Step S6: Construct a myocardial rehabilitation decision system based on the myocardial physician recommendation model and the electrocardiogram abnormal waveform detection model, and iteratively optimize and adjust the parameters of the myocardial rehabilitation decision system based on historical myocardial case data to obtain a myocardial rehabilitation decision support system.

2. The method for constructing a myocardial rehabilitation decision support system based on case reasoning according to claim 1, characterized in that: Step S1 is specifically as follows: Step S11: Acquire historical myocardial case data, and perform structured data classification on the historical myocardial case data, thereby obtaining structured myocardial case data and unstructured myocardial case data; Step S12: performing unstructured patient myocardial feature fusion on the unstructured myocardial case data, thereby obtaining unstructured patient myocardial fusion data; Step S13: fusing the patient's treatment course physiological characteristics according to the structured myocardial case data, thereby obtaining structured patient physiological fusion data; Step S14: merging the unstructured patient myocardial fusion data and the structured patient physiological fusion data into patient information cases, thereby obtaining the patient myocardial case fusion data; Step S15: Desensitize the patient's sensitive features on the patient's myocardial case fusion data, so as to obtain historical myocardial case desensitized data.

3. The method for constructing a myocardial rehabilitation decision support system based on case reasoning according to claim 2, characterized in that: Step S12 is specifically as follows: Step S121: extracting medical record summary records, patient electrocardiogram records and medical image records from the unstructured myocardial case data, thereby obtaining myocardial medical record summary data, myocardial patient electrocardiogram sets and myocardial patient medical image sets; Step S122: integrating myocardial description features of the cardiology medical record summary data to obtain myocardial description data; Step S123: extracting dynamic structural features of the myocardium scanned by CT according to the medical image set of the patients in the myocardium department, thereby obtaining dynamic structural data of the myocardium scanned by CT, and constructing a three-dimensional dynamic structural model of the myocardium of the patient according to the dynamic structural data of the myocardium scanned by CT; Step S124: filling the patient's myocardial three-dimensional dynamic structure model with myocardial structure details according to the myocardial description data, thereby obtaining the patient's myocardial dynamic structure model; Step S125: integrating the dynamic fluctuation characteristics of the electrocardiogram according to the electrocardiogram set of the patients in the myocardium department, thereby obtaining the dynamic fluctuation data of the electrocardiogram, and performing myocardial dynamic motion simulation on the dynamic fluctuation data of the electrocardiogram through the dynamic structural model of the myocardium of the patients, thereby obtaining the dynamic simulation data of the myocardium of the patients; Step S126: performing patient myocardial dynamic feature fusion according to the patient myocardial dynamic simulation data, thereby obtaining unstructured patient myocardial fusion data.

4. The method for constructing a myocardial rehabilitation decision support system based on case reasoning according to claim 2, characterized in that: Step S13 is specifically as follows: Step S131: extracting drug dosage records, disease course records and physiological coefficient records from the structured myocardial case data, thereby obtaining drug dosage data, disease course data and physiological coefficient data of the patient; Step S132: dividing the patient's treatment stage according to the patient's disease course data, thereby obtaining the patient's treatment stage data; Step S133: performing a drug characteristic analysis on the drug dosage data, thereby obtaining drug dosage characteristic data; Step S134: performing physiological coefficient fluctuation correlation according to the drug dosage characteristic data and the patient's physiological coefficient data, thereby obtaining drug dosage correlated physiological fluctuation data; Step S135: fusing the patient treatment stage data with the treatment stage physiological characteristics according to the drug dosage-associated physiological fluctuation data, thereby obtaining structured patient physiological fusion data.

5. The method for constructing a myocardial rehabilitation decision support system based on case reasoning according to claim 1, characterized in that: Step S2 is specifically as follows: Step S21: extracting treatment and rehabilitation stage features from the historical myocardial case desensitization data, thereby obtaining historical myocardial case treatment stage data, and classifying the historical myocardial case treatment stage data by treatment stage length, thereby obtaining long treatment stage case data and short treatment stage case data; Step S22: classifying the historical myocardial case treatment stage data by treatment stage review, thereby obtaining case treatment stage review data; Step S23: performing myocardial variable statistics review according to the case treatment stage review data, thereby obtaining the myocardial variable data review, and classifying the myocardial variable fluctuation amplitude of the myocardial variable data review, thereby obtaining high-amplitude fluctuation myocardial variable review data and low-amplitude fluctuation myocardial variable review data; Step S24: classifying the high-amplitude fluctuation myocardial variable review data and the low-amplitude fluctuation myocardial variable review data by the high-proportion myocardial variable fluctuation amplitude ratio according to the historical myocardial case treatment stage data, thereby obtaining high-proportion case data of the high-amplitude fluctuation treatment stage and high-proportion case data of the low-amplitude fluctuation treatment stage; Step S25: performing a treatment case intersection operation on the long treatment stage case data and the low amplitude fluctuation treatment stage case data with a high proportion, thereby obtaining slow-acting treatment and rehabilitation case data; performing a treatment case intersection operation on the short treatment stage case data and the high amplitude fluctuation treatment stage case data with a high proportion, thereby obtaining fast-acting treatment and rehabilitation case data; Step S26: integrating the slow-acting treatment rehabilitation case data and the fast-acting treatment rehabilitation case data, thereby obtaining historical myocardial rehabilitation case data; Step S27: Integrate abnormal electrocardiogram waveform features based on historical myocardial rehabilitation case data to obtain abnormal electrocardiogram waveform data of rehabilitation cases.

6. The method for constructing a myocardial rehabilitation decision support system based on case reasoning according to claim 5, characterized in that: Step S27 is specifically as follows: Step S271: extracting electrocardiogram records from historical myocardial rehabilitation case data, thereby obtaining an electrocardiogram set of slow-acting treatment rehabilitation cases and an electrocardiogram set of fast-acting treatment rehabilitation cases; Step S272: performing Fourier transform on the slow-acting treatment rehabilitation case electrocardiogram set and the fast-acting treatment rehabilitation case electrocardiogram set respectively, so as to obtain the slow-acting rehabilitation case electrocardiogram spectrum and the fast-acting rehabilitation case electrocardiogram spectrum; Step S273: performing spectrum amplitude threshold statistics on the ECG spectrum of the slow-acting rehabilitation case and the ECG spectrum of the fast-acting rehabilitation case, respectively, so as to obtain the slow-acting spectrum amplitude threshold and the fast-acting spectrum amplitude threshold; Step S274: classifying the abnormal spectrum amplitude of the slow-acting rehabilitation case ECG spectrum according to the slow-acting spectrum amplitude threshold, thereby obtaining the abnormal slow-acting rehabilitation case ECG spectrum; classifying the abnormal spectrum amplitude of the fast-acting rehabilitation case ECG spectrum according to the fast-acting spectrum amplitude threshold, thereby obtaining the abnormal fast-acting rehabilitation case ECG spectrum; Step S275: performing time domain conversion on the ECG spectrum of the abnormal slow-acting recovery case and the ECG spectrum of the abnormal fast-acting recovery case, respectively, so as to obtain ECG fluctuation data of the abnormal slow-acting recovery case and ECG fluctuation data of the abnormal fast-acting recovery case; Step S276: Integrate the abnormal electrocardiogram waveform features of the abnormal slow-acting rehabilitation case electrocardiogram fluctuation data and the abnormal fast-acting rehabilitation case electrocardiogram fluctuation data, so as to obtain the abnormal electrocardiogram waveform data of the rehabilitation case.

7. The method for constructing a myocardial rehabilitation decision support system based on case reasoning according to claim 1, characterized in that: Step S46 is specifically as follows: Step S461: extracting left ventricular ejection fraction features and electrocardiogram waveform features from historical myocardial rehabilitation case data, thereby obtaining rehabilitation case left ventricular ejection fraction data and rehabilitation case electrocardiogram waveform data; Step S462: constructing a myocardial rehabilitation index evaluation system based on the left ventricular ejection fraction data of the rehabilitation case and the electrocardiogram waveform data of the rehabilitation case, and performing myocardial rehabilitation index evaluation according to the myocardial rehabilitation index evaluation system, thereby obtaining myocardial rehabilitation index evaluation data; Step S463: extracting waveform amplitude features from the electrocardiogram waveform trend data, thereby obtaining trend waveform amplitude data; Step S464: constructing a myocardial rehabilitation linear regression model according to the myocardial rehabilitation index evaluation data and the trend waveform amplitude data; Step S465: Calculate the myocardial rehabilitation influencing factor according to the myocardial rehabilitation linear regression model, so as to obtain the myocardial rehabilitation influencing factor.

8. The method for constructing a myocardial rehabilitation decision support system based on case reasoning according to claim 1, characterized in that: Step S55 is specifically as follows: Step S551: performing doctor treatment frequency statistics and doctor-associated electrocardiogram statistics according to the treating doctor-electrocardiogram association data, thereby obtaining doctor treatment frequency data and doctor-associated electrocardiogram data; Step S552: classifying the doctor treatment frequency data by treatment frequency, thereby obtaining the doctor data with high treatment frequency and the doctor data with low treatment frequency; Step S553: ​​classifying the doctor-associated ECG data by the number of associated ECGs, thereby obtaining doctor data with a high number of associated ECGs and doctor data with a low number of associated ECGs; Step S554: performing doctor intersection calculation on the doctor data with high treatment times and the doctor data with high electrocardiogram association quantity, so as to obtain the myocardial doctor data with high treatment priority; Doctor intersection calculation is performed on the data of doctors with low treatment times and doctors with low ECG association numbers to obtain the data of myocardial doctors with low treatment priority; Step S555: Merge data based on the high treatment priority cardiotocologist data and the low treatment priority cardiotocologist data to obtain cardiotocologist treatment priority data.

Images