Treatment scheme matching method for rehabilitation treatment of chronic obstructive pulmonary disease patient

Abstract

The invention relates to the technical field of rehabilitation treatment of chronic obstructive pulmonary patients, and discloses a therapeutic scheme matching method for rehabilitation treatment of chronic obstructive pulmonary patients, comprising the following steps: step 1, a data acquisition and integration module which is responsible for collecting multi-dimensional data of patients; according to the treatment scheme matching method for the rehabilitation treatment of the chronic obstructive pulmonary disease patient, unnecessary medical resource waste can be avoided, the patient can be timely and properly treated, the overall utilization efficiency of the medical resources is improved, detailed treatment information and self-management suggestions are provided for the patient, and the patient is prevented from suffering from chronic obstructive pulmonary disease. A patient can better know the illness state and the treatment process of the patient, a personalized treatment scheme is easier to accept and execute by the patient, the treatment compliance of the patient can be improved, a large amount of real-world patient data is accumulated by the system, rich materials are provided for medical research, and through deep mining and analysis of the data, the treatment compliance of the patient can be improved. Pathogenesis, treatment effect influence factors and the like can be further explored.

Description

Technical Field

[0001] The present invention relates to the technical field of rehabilitation treatment for patients with chronic obstructive pulmonary disease (COPD), and specifically to a treatment plan matching method for rehabilitation treatment of patients with COPD. Background Art

[0002] Chronic obstructive pulmonary disease (COPD) is a common, preventable and treatable disease characterized by persistent respiratory symptoms and airflow limitation. The airflow limitation is often progressive and caused by abnormalities in the lung airways and alveoli, just like a clogged chimney with poor ventilation. Common symptoms include chronic cough, sputum, shortness of breath or difficulty breathing, which often worsens after activity. The main causes of the disease are related to smoking, air pollution, occupational dust, chemicals, respiratory infections and other factors.

[0003] Existing treatment options include nebulization inhalation therapy, aerosol therapy, mechanical ventilation therapy, and drug therapy. Nebulization inhalation therapy involves dispersing medication or water into mist or fine particles suspended in a gas, which are then inhaled into the respiratory tract and lungs. Aerosol therapy consists of tiny liquid or solid particles suspended in the air. Liquid particle aerosols are called "mist," while solid particle aerosols are called "dust" or "smoke." Aerosol therapy involves using a device, heating, jets, or ultrasonic vibrations to impact and disperse medication liquids, powders, or saline solutions into extremely small particles suspended in a gas. These particles are then inhaled into the airways and lungs for treatment. Mechanical ventilation therapy is a non-invasive ventilation connection method that helps patients overcome airway resistance and provides external positive expiratory pressure (PEEP) to counteract intrinsic positive expiratory pressure (PEEP), thereby relieving respiratory muscle fatigue. Positive pressure ventilation therapy can promptly correct hypoxia and CO2 retention in patients, reducing damage to various tissues and organs. Drug therapy involves the administration of medications.

[0004] However, during the actual operation, the inventors found that the following problems still exist: in the prior art, doctors distinguish the early, middle and late stages of chronic obstructive pulmonary disease (COPD) based on the patient's symptoms, chest CT examination, lung function test, etc., and then perform drug treatment based on the actual diagnosis results. However, the early, middle and late stages of a patient's chronic obstructive pulmonary disease (COPD) are only diagnosed by the doctor, and then through the design of the treatment plan, the prior art mostly uses drug treatment and simple precautions to allow patients to undergo rehabilitation treatment. However, treatment plans that rely solely on drug treatment and simple precautions from the doctor may still cause the symptoms of chronic obstructive pulmonary disease (COPD) to worsen. For example, if the respiratory muscles are not exercised every day, the respiratory function is still in a poor state, and the effect of drug treatment is relatively slow or even poor.

[0005] Based on this, the present invention provides a treatment plan matching method for rehabilitation treatment of patients with COPD, which has the advantage of matching corresponding treatment plans according to the patient's specific diagnosis, and can achieve better rehabilitation effects for patients by combining multiple plans. Summary of the Invention

[0006] In response to the shortcomings of the existing technology, the present invention provides a treatment plan matching method for rehabilitation treatment of patients with chronic obstructive pulmonary disease, which has the advantage of matching corresponding treatment plans according to the patient's specific diagnosis, and solves the problems raised in the background technology.

[0007] The present invention provides the following technical solution: a method for matching treatment plans for rehabilitation treatment of patients with chronic obstructive pulmonary disease, comprising the following steps: Step 1: Data collection and integration module: responsible for collecting multi-dimensional data of patients. Patients can fill in medical record information manually, or use the QR code generator to generate QR codes for data information. Patients fill in data by scanning the QR code. The filled data includes name, gender, age, ethnicity, marital status, occupation, contact information, home address, medical insurance type, allergy history, past medical history, family medical history, symptoms, disease progression, treatment status and lifestyle habits. By integrating these data, an information database that comprehensively reflects the health status of patients is constructed. The information database is connected to the electronic medical record system server through the hospital LAN. The information is transmitted in real time or in batches at regular intervals and stored in the electronic medical record system (EMR) database. The information database uses the AES encryption algorithm to encrypt the collected data. Data is encrypted for storage and transmission to prevent data leakage, protect patient privacy, and comply with GDPR privacy protection measures. According to different access groups, it is divided into two types of access rights: patient side and doctor side. During the access process, personal authentication is performed through face recognition, and a linked list algorithm is used to record network access behavior in real time. By defining the linked list node structure and linked list structure, and realizing the functions of adding nodes and printing node data, access records can be effectively recorded. The linked list can be used to connect the access record nodes in sequence according to the order of access. Whenever a new access occurs, a new node is created and inserted into the end of the linked list. The node insertion and deletion operations of the linked list are relatively simple and efficient. When inserting a new access record, you only need to modify the pointer of the relevant node; when deleting a record, you only need to adjust the pointer. Step 2: Intelligent diagnosis and evaluation engine: (1) Utilize the artificial intelligence algorithm of the masked language model and the medical knowledge base to conduct in-depth analysis on the data input by the patient, receive the patient's basic information, vital signs, test results and other structured data from the electronic medical record system, which are stored in the form of tables and can be directly read and processed by the computer. For unstructured text data such as medical record text, perform natural language processing operations such as word segmentation, part-of-speech tagging, named entity recognition, etc., convert the text into a feature vector that can be understood by the computer, combine the diagnosis results and related features, evaluate the severity and development trend of the disease, and provide corresponding evaluation indicators and prediction results. Use the LIME model interpretation technology to explain the decision-making process and results of the model to help doctors and patients understand the basis and rationality of intelligent diagnosis and evaluation, and use association Rule mining: It can discover the correlation between different symptoms, examination results and diseases in medical data, providing a basis for diagnosis. The AI ​​algorithm based on the masked language model analyzes the patient's living habit data (such as secondhand smoke exposure). The artificial intelligence algorithm adopts a self-supervised learning training method. The AI ​​algorithm model is trained by generating pseudo labels from the input data itself. The training uses the masked language model and the causal language model to predict the masked word and the next word given the previous text, respectively, so that the model learns the relationship between the word and the context, the sequential structure of the text and the generation process of the language; (2) By combining the doctor's diagnosis results of the early, middle and late stages of chronic obstructive pulmonary disease (COPD) based on the patient's symptoms, chest CT examination, pulmonary function test, etc., the intelligent diagnosis and evaluation engine integrates the doctor's diagnosis results of pulmonary function test stages to generate a comprehensive evaluation report; Step 3: Treatment Matching and Recommendation System: Based on the combined results of intelligent diagnosis and assessment, the system selects the most suitable personalized treatment plan for the patient based on the Global Initiative for the Prevention and Control of Chronic Obstructive Pulmonary Disease and the clinical diagnosis and treatment pathway database. Treatment plans include medication, non-drug therapies, and regular follow-up appointments. Deep learning algorithms, including convolutional neural networks and recurrent neural networks, are used to formulate treatment plans. Convolutional neural networks automatically extract features from image data and other data to recommend treatment plans for image-related diseases. This is based on their strong image data processing capabilities and their ability to identify the connection between key information in images and the disease and treatment plan. Recurrent neural networks are suitable for processing sequential medical data, such as medical records. They can capture time series information in the data and recommend treatment plans that are more consistent with the progression of the disease. This is based on their ability to effectively process long-term dependencies in sequential data, reflect the dynamic changes of the disease, and thus match treatment plans that are tailored to the patient. Treatment plans are based on specific numerical values ​​such as drug blood concentration monitoring results and quantitative indicators of the patient's symptom improvement (such as the degree of reduction in dyspnea scores). Step 4: Treatment plan adjustment system: The physician can make corresponding adjustments based on the treatment plan matching and the personalized treatment plan given by the recommendation system. During the adjustment process, the physician can make adjustments based on the actual patient data. For example: (1) The treatment effect of the patient taking XXX drug is not as good as that of XXX drug. In this case, the physician can make adjustments; (2) The patient has a physical disability or serious disease that makes it impossible for the patient to recover from exercise normally. In this case, the physician can make adjustments; Step 5: Dynamic Adjustment and Tracking Module: Patients upload their daily data through the mobile app, and then use the data feedback mechanism to continuously track the patient's treatment effect and disease changes. By regularly collecting patient feedback data and sending the feedback data to the physician's mobile app, the doctor can optimize and adjust the treatment plan in a timely manner to ensure that the treatment always meets the patient's actual needs. For example: when the respiratory rate deviation in the patient's feedback data exceeds 10%, the system automatically triggers the optimization of the treatment plan.

[0008] Preferably, the intelligent diagnosis and evaluation engine uses artificial intelligence algorithms and a medical knowledge base to combine the diagnosis results of early, middle and late stages of chronic obstructive pulmonary disease (COPD) performed by doctors based on the patient's symptoms, chest CT examination, pulmonary function tests, etc., uses data and experience for comprehensive combination, and then conducts in-depth analysis and evaluation based on the combination of the two. The diagnosis results of the intelligent diagnosis and evaluation engine through artificial intelligence algorithms and a medical knowledge base are based on the "Global Initiative for the Prevention and Control of Chronic Obstructive Pulmonary Disease", and the "Global Initiative for the Prevention and Control of Chronic Obstructive Pulmonary Disease" is the knowledge content that doctors need to learn, thereby ensuring that the diagnosis results of the two are consistent.

[0009] Preferably, the drug treatment includes bronchodilators, anticholinergic drugs, theophylline drugs, and glucocorticoids. Specific drug treatment recommendations are based on AI's analysis of the patient's acute exacerbation history, adjusted according to the drug treatment effect and the patient's physical condition, and also consider drug side effects and the patient's tolerance to the drug, for example, the patient's own allergens are present in the raw materials of a certain drug: (1) Bronchodilators are the core drugs for the treatment of COPD. For example, short-acting β2 receptor agonists such as salbutamol and terbutaline have a fast onset of action and a short duration of action. They can quickly relieve dyspnea symptoms and are often used for temporary symptom relief. Long-acting β2 receptor agonists such as salmeterol and formoterol have a long duration of action that can last for more than 12 hours and are more suitable for regular use to control symptoms. (2) Anticholinergic drugs: such as ipratropium bromide, which is a short-acting anticholinergic drug; tiotropium bromide and glycopyrrolate are long-acting anticholinergic drugs that block M cholinergic receptors to dilate the bronchi and improve ventilation; (3) Theophylline drugs: aminophylline, etc., can relax bronchial smooth muscles, and also have certain anti-inflammatory and respiratory muscle strengthening effects, but they should be used with caution due to their side effects: (1) Tachycardia and arrhythmia are among the most common adverse reactions of theophylline drugs. Theophylline drugs increase the concentration of cyclic adenosine monophosphate in cells by inhibiting the activity of phosphodiesterase, thereby dilating the bronchi, but they may also stimulate the heart, leading to increased heart rate or arrhythmia; (2) Headache and anxiety are also possible adverse reactions of theophylline drugs. This is mainly because theophylline drugs can produce an excitatory effect through the central nervous system, causing excessive activity of nerve cells, thereby causing headaches and anxiety. (3) Nausea and vomiting are another common adverse reaction of theophylline drugs. This is because theophylline drugs can stimulate the gastrointestinal tract, leading to gastrointestinal dysfunction, thereby causing nausea and vomiting; (4) Glucocorticoids: For patients with moderate to severe COPD, when bronchodilators are ineffective, inhaled glucocorticoids, such as budesonide and fluticasone, can be combined with long-acting β2 receptor agonists (such as salmeterol fluticasone powder inhaler®) to more effectively control symptoms and reduce acute exacerbations. The non-drug treatments include smoking cessation, oxygen therapy, rehabilitation therapy, nutritional support and health teas. Different non-drug treatment plans can be formulated based on the patient's age, gender, physical condition, lifestyle habits and other factors: (1) Smoking cessation: For patients who smoke, smoking cessation is the most important intervention measure. Smoking cessation counseling, nicotine replacement therapy, and medication assistance (such as bupropion and varenicline) can help patients quit smoking; (2) Oxygen therapy: The duration of oxygen therapy (≥15 hours / day) is dynamically adjusted based on the patient's arterial oxygen saturation AI prediction model; (3) Rehabilitation treatment: Respiratory muscle exercises: ① Pursed lip breathing (close your mouth and inhale through your nose, then exhale slowly with pursed lips, with the exhalation time being about twice the inhalation time); ② Abdominal breathing (the abdomen bulges when inhaling and contracts when exhaling) can strengthen the respiratory muscles and improve respiratory function; ③ Exercise, such as walking and Tai Chi, can gradually increase the intensity and duration of exercise according to the patient's condition, which will help improve the patient's exercise endurance and ability to take care of themselves; (4) Nutritional support: COPD patients often suffer from malnutrition due to increased consumption caused by dyspnea and affected gastrointestinal function. They should ensure adequate intake of protein (such as lean meat, fish, eggs, beans, etc.), carbohydrates and fats, and pay attention to supplementing vitamins and minerals; (5) Health tea: Add or subtract herbs based on the patient's constitution. Recommendations can also be made based on the severity of the patient's condition and TCM constitution classification (such as Qi deficiency, Yang deficiency, phlegm-dampness, etc.) to ensure the targeted nature of health-preserving tea as an auxiliary treatment. For example, according to the prescriptions in the Complete Collection of Chinese Herbal Health Teas: ① Qi, Blood, Yin and Yang Replenishing Prescription: Composed of 10-15g of Astragalus, 3-5g of American Ginseng, 6-10g of Lycium barbarum, and 10g of Polygonatum, it can simultaneously regulate Qi and blood, replenish the spleen and kidneys, and harmonize Yin and Yang; ② Astragalus, Angelica and Red Dates Tea: Take 12 grams of Astragalus, 5 grams of Angelica and 3 dates, simmer in boiling water for 20 minutes. Depending on your constitution, you can add 10 grams of Ophiopogon japonicus or American ginseng to prevent internal heat. For those with cold deficiency, add two slices of ginger. For those with heavy phlegm and dampness, add 10 grams of dried tangerine peel. It has the effects of tonifying the middle qi, nourishing the blood and calming the mind. ③ Wolfberry and Angelica Tea: 10 grams each of longan meat, wolfberry and angelica, soak them together in water and drink as tea. It has the effect of nourishing the liver and kidneys, invigorating qi and activating blood circulation.

[0010] Preferably, the treatment plan adjustment system is a replacement of the therapeutic drug in the drug treatment in the personalized treatment plan given by the physician through treatment plan matching and recommendation system, and the non-drug treatment is adapted according to the patient's own personalization. The adaptation includes increasing the drug dosage, frequency of use, treatment cycle and other aspects, for example: (1) The patient does not smoke, but whether there is a lot of secondhand smoke in his work environment, home environment and travel environment, then the patient needs to be advised to wear a corresponding mask or avoid entering the above environment. The mask should be an N95 mask; (2) During the rehabilitation treatment, the corresponding rehabilitation plan is selected according to the patient's own condition. For example, patients with chronic obstructive pulmonary disease (COPD) in the early, middle and late stages can make corresponding choices based on whether they can perform strenuous exercise and their respiratory status. For example: ① Patients with difficulty in movement and breathing can use the method of respiratory muscle training + respiratory muscle training. This method requires a small range of motion and can exercise the respiratory muscles; ② Patients who can perform normal exercise can use the method of respiratory muscle training + respiratory muscle training + exercise training for rehabilitation treatment.

[0011] Preferably, the artificial intelligence algorithm performs patient symptom analysis from a medical knowledge base, mainly through the following steps: Data collection and preprocessing: ① Data collection: Collect patient symptom data from multiple channels such as electronic medical record systems and medical sensors, including symptom descriptions and test results; ②Data cleaning: remove duplicate, erroneous and incomplete data to ensure data quality; ③Data standardization: unify the standards of data from different sources and formats, such as standardizing symptom descriptions.

[0012] Knowledge Representation and Modeling: ① Knowledge representation: Represent the knowledge in the medical knowledge base in a computer-understandable form, such as using an ontology model to describe concepts and relationships such as diseases and symptoms; ②Build a model: Use machine learning to build an analysis model. For example, a decision tree model can be used to classify diseases based on symptom characteristics, and a neural network model can be used to automatically learn the complex mapping relationship between symptoms and diseases.

[0013] Feature extraction and selection: ① Feature extraction: Extract features related to symptom analysis from preprocessed data, such as extracting key symptom words from text symptom descriptions and extracting numerical features from examination data; ② Feature selection: Use information gain, mutual information and other methods to screen out the most representative and discriminative features, reduce data dimensions, and improve analysis efficiency and accuracy.

[0014] Model training and optimization: ① Model training: Use labeled symptom-disease data to train the model and adjust the model parameters so that the model can accurately infer the disease from the symptom data; ②Model optimization: Optimize the model through cross-validation, adjustment of hyperparameters, and other methods to improve its generalization ability and accuracy.

[0015] Symptom analysis and reasoning: ① Symptom matching: The patient's symptom data is input into the trained model. The model uses the learned knowledge and patterns to find the disease or symptom combination that best matches the patient's symptoms. ② Reasoning and explanation: Not only does it provide possible disease diagnosis results, but it also uses technologies such as knowledge graphs to explain the reasoning process to help doctors understand the basis for model judgments.

[0016] Result evaluation and feedback: ① Result evaluation: Use indicators such as accuracy and recall to evaluate the accuracy and reliability of model analysis results; ② Feedback and improvement: Based on the evaluation results and doctors’ actual application feedback, the model is adjusted and improved to continuously improve the performance of symptom analysis.

[0017] Preferably, the nutritional therapy can focus on the allergy history in the basic information provided by the patient to avoid the occurrence of other diseases caused by allergies. The drug therapy and non-drug therapy are judged on the rehabilitation effect based on weekly examinations, which include changes in 6-minute walking distance and cardiopulmonary function test indicators, and are judged based on the test results.

[0018] Compared with the prior art, the present invention has the following beneficial effects: This treatment plan matching method for rehabilitation treatment of patients with chronic obstructive pulmonary disease (COPD) can better target individual differences of patients through personalized treatment plans, improve the accuracy and effectiveness of treatment, and thus ensure the effectiveness of chronic obstructive pulmonary disease (COPD) treatment. By providing patients with reasonable treatment plans and follow-up arrangements, the system can avoid unnecessary waste of medical resources, while also ensuring that patients receive timely and appropriate treatment, improving the overall utilization efficiency of medical resources, and providing patients with detailed treatment information and self-management suggestions so that patients can better understand their condition and treatment process. Personalized treatment plans are easier for patients to accept and implement, which helps to improve patients' treatment compliance. The large amount of real-world patient data accumulated by the system provides rich material for medical research. Through in-depth mining and analysis of these data, we can further explore the pathogenesis, factors affecting treatment effects, etc., and can make significant progress in improving treatment accuracy and effectiveness. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 Schematic diagram of the process of the present invention; Figure 2 This is a schematic diagram of the basic information of the present invention; Figure 3 Schematic diagram of the treatment plan matching and recommendation system of the present invention; Figure 4 Schematic diagram of input and output of the convolutional neural network algorithm of the present invention; Figure 5 Schematic diagram of input and output of the recurrent neural network algorithm of the present invention; Figure 6 Schematic diagram of the processing steps of the artificial intelligence algorithm of the present invention. DETAILED DESCRIPTION

[0020] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only 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 ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0021] See also Figure 1 、 Figure 2 、 Figure 3 、 Figure 4 、 Figure 5 and Figure 6 A method for matching treatment plans for rehabilitation treatment of patients with chronic obstructive pulmonary disease (COPD) comprises the following steps: Step 1: Data collection and integration module: responsible for collecting multi-dimensional data of patients ( Figure 2Detailed flow chart for patient data collection), patients can manually fill in medical record information, or use a QR code generator to generate a QR code for the data information. Patients fill in the data by scanning the QR code. The data filled in includes name, gender, age, ethnicity, marital status, occupation, contact information, home address, medical insurance type, allergy history, past medical history, family medical history, symptoms, disease progression, treatment status and lifestyle habits. By integrating these data, an information database that comprehensively reflects the patient's health status is constructed. The information database is connected to the electronic medical record system server through the hospital's LAN. Information is transmitted in real time or in scheduled batches and stored in the electronic medical record system (EMR) database. The information database uses the AES encryption algorithm to encrypt the collected data Storage and transmission, prevent data leakage, protect patient privacy, and comply with GDPR's privacy protection measures. According to different access groups, there are two access rights: patient side and doctor side. During the access process, personal authentication is performed through face recognition, and a linked list algorithm is used to record network access behavior in real time. By defining the linked list node structure and linked list structure, and realizing the functions of adding nodes and printing node data, access records can be effectively recorded. The linked list can be used to connect the access record nodes in sequence according to the order of access. Whenever a new access occurs, a new node is created and inserted into the end of the linked list. The node insertion and deletion operations of the linked list are relatively simple and efficient. When inserting a new access record, you only need to modify the pointer of the relevant node; when deleting a record, you only need to adjust the pointer. Step 2: Intelligent diagnosis and evaluation engine: (1) Utilize the artificial intelligence algorithm of the masked language model and the medical knowledge base to conduct in-depth analysis on the data input by the patient, receive the patient's basic information, vital signs, test results and other structured data from the electronic medical record system, which are stored in the form of tables and can be directly read and processed by the computer. For unstructured text data such as medical record text, perform natural language processing operations such as word segmentation, part-of-speech tagging, named entity recognition, etc., convert the text into a feature vector that can be understood by the computer, combine the diagnosis results and related features, evaluate the severity and development trend of the disease, and provide corresponding evaluation indicators and prediction results. Use the LIME model interpretation technology to explain the decision-making process and results of the model to help doctors and patients understand the basis and rationality of intelligent diagnosis and evaluation, and use association Rule mining: It can discover the correlation between different symptoms, examination results and diseases in medical data, providing a basis for diagnosis. The AI ​​algorithm based on the masked language model analyzes the patient's living habit data (such as secondhand smoke exposure). The artificial intelligence algorithm adopts a self-supervised learning training method. The AI ​​algorithm model is trained by generating pseudo labels from the input data itself. The training uses the masked language model and the causal language model to predict the masked word and the next word given the previous text, respectively, so that the model learns the relationship between the word and the context, the sequential structure of the text and the generation process of the language; (2) By combining the doctor's diagnosis results of the early, middle and late stages of chronic obstructive pulmonary disease (COPD) based on the patient's symptoms, chest CT examination, pulmonary function test, etc., the intelligent diagnosis and evaluation engine integrates the doctor's diagnosis results of pulmonary function test stages to generate a comprehensive evaluation report; Step 3: Treatment Matching and Recommendation System: Based on the combined results of intelligent diagnosis and assessment, the system selects the most suitable personalized treatment plan for the patient based on the Global Initiative for the Prevention and Control of Chronic Obstructive Pulmonary Disease and the clinical diagnosis and treatment pathway database. Treatment plans include medication, non-drug therapies, and regular follow-up appointments. Deep learning algorithms, including convolutional neural networks and recurrent neural networks, are used to formulate treatment plans. Convolutional neural networks automatically extract features from image data and other data to recommend treatment plans for image-related diseases. This is based on their strong image data processing capabilities and their ability to identify the connection between key information in images and the disease and treatment plan. Recurrent neural networks are suitable for processing sequential medical data, such as medical records. They can capture time series information in the data and recommend treatment plans that are more consistent with the progression of the disease. This is based on their ability to effectively process long-term dependencies in sequential data, reflect the dynamic changes of the disease, and thus match treatment plans that are tailored to the patient. Treatment plans are based on specific numerical values ​​such as drug blood concentration monitoring results and quantitative indicators of the patient's symptom improvement (such as the degree of reduction in dyspnea scores). Step 4: Treatment plan adjustment system: The physician can make corresponding adjustments based on the treatment plan matching and the personalized treatment plan given by the recommendation system. During the adjustment process, the physician can make adjustments based on the actual patient data. For example: (1) The treatment effect of the patient taking XXX drug is not as good as that of XXX drug. In this case, the physician can make adjustments; (2) The patient has a physical disability or serious disease that makes it impossible for the patient to recover from exercise normally. In this case, the physician can make adjustments; Step 5: Dynamic Adjustment and Tracking Module: Patients upload their daily data through the mobile app, and then use the data feedback mechanism to continuously track the patient's treatment effect and disease changes. By regularly collecting patient feedback data and sending the feedback data to the physician's mobile app, the doctor can optimize and adjust the treatment plan in a timely manner to ensure that the treatment always meets the patient's actual needs. For example: when the respiratory rate deviation in the patient's feedback data exceeds 10%, the system automatically triggers the optimization of the treatment plan.

[0022] Among them, the intelligent diagnosis and evaluation engine uses artificial intelligence algorithms and medical knowledge bases to combine the diagnosis results of chronic obstructive pulmonary disease (COPD) in the early, middle and late stages with the doctor's symptoms, chest CT examinations, pulmonary function tests, etc., and uses data and experience for comprehensive combination, and then conducts in-depth analysis and evaluation based on the combination of the two. The diagnosis results of the intelligent diagnosis and evaluation engine through artificial intelligence algorithms and medical knowledge base are based on the "Global Initiative for the Prevention and Control of Chronic Obstructive Pulmonary Disease", and the "Global Initiative for the Prevention and Control of Chronic Obstructive Pulmonary Disease" is the knowledge content that doctors need to learn, thereby ensuring that the diagnostic results of the two are consistent.

[0023] Drug treatments include bronchodilators, anticholinergics, theophylline drugs, and glucocorticoids. Specific drug treatment recommendations are based on AI's analysis of the patient's history of acute exacerbations. These recommendations are adjusted based on the drug's effectiveness and the patient's physical condition. Side effects and the patient's tolerance to the drug should also be considered. For example, if the patient's own allergens are present in the raw materials of a certain drug, (1) Bronchodilators are the core drugs for the treatment of COPD. For example, short-acting β2 receptor agonists such as salbutamol and terbutaline have a fast onset of action and a short duration of action. They can quickly relieve dyspnea symptoms and are often used for temporary symptom relief. Long-acting β2 receptor agonists such as salmeterol and formoterol have a long duration of action that can last for more than 12 hours and are more suitable for regular use to control symptoms. (2) Anticholinergic drugs: such as ipratropium bromide, which is a short-acting anticholinergic drug; tiotropium bromide and glycopyrrolate are long-acting anticholinergic drugs that block M cholinergic receptors to dilate the bronchi and improve ventilation; (3) Theophylline drugs: aminophylline, etc., can relax bronchial smooth muscles, and also have certain anti-inflammatory and respiratory muscle strengthening effects, but they should be used with caution due to their side effects: (1) Tachycardia and arrhythmia are among the most common adverse reactions of theophylline drugs. Theophylline drugs increase the concentration of cyclic adenosine monophosphate in cells by inhibiting the activity of phosphodiesterase, thereby dilating the bronchi, but they may also stimulate the heart, leading to increased heart rate or arrhythmia; (2) Headache and anxiety are also possible adverse reactions of theophylline drugs. This is mainly because theophylline drugs can produce an excitatory effect through the central nervous system, causing excessive activity of nerve cells, thereby causing headaches and anxiety. (3) Nausea and vomiting are another common adverse reaction of theophylline drugs. This is because theophylline drugs can stimulate the gastrointestinal tract, leading to gastrointestinal dysfunction, thereby causing nausea and vomiting; (4) Glucocorticoids: For patients with moderate to severe COPD, when bronchodilators are ineffective, inhaled glucocorticoids, such as budesonide and fluticasone, can be combined with long-acting β2 receptor agonists (such as salmeterol fluticasone powder inhaler®) to more effectively control symptoms and reduce acute exacerbations. The non-drug treatments include smoking cessation, oxygen therapy, rehabilitation therapy, nutritional support and health teas. Different non-drug treatment plans can be formulated based on the patient's age, gender, physical condition, lifestyle habits and other factors: (1) Smoking cessation: For patients who smoke, smoking cessation is the most important intervention measure. Smoking cessation counseling, nicotine replacement therapy, and medication assistance (such as bupropion and varenicline) can help patients quit smoking; (2) Oxygen therapy: The duration of oxygen therapy (≥15 hours / day) is dynamically adjusted based on the patient's arterial oxygen saturation AI prediction model; (3) Rehabilitation treatment: Respiratory muscle exercises: ① Pursed lip breathing (close your mouth and inhale through your nose, then exhale slowly with pursed lips, with the exhalation time being about twice the inhalation time); ② Abdominal breathing (the abdomen bulges when inhaling and contracts when exhaling) can strengthen the respiratory muscles and improve respiratory function; ③ Exercise, such as walking and Tai Chi, can gradually increase the intensity and duration of exercise according to the patient's condition, which will help improve the patient's exercise endurance and ability to take care of themselves; (4) Nutritional support: COPD patients often suffer from malnutrition due to increased consumption caused by dyspnea and affected gastrointestinal function. They should ensure adequate intake of protein (such as lean meat, fish, eggs, beans, etc.), carbohydrates and fats, and pay attention to supplementing vitamins and minerals; (5) Health tea: Add or subtract herbs based on the patient's constitution. Recommendations can also be made based on the severity of the patient's condition and TCM constitution classification (such as Qi deficiency, Yang deficiency, phlegm-dampness, etc.) to ensure the targeted nature of health-preserving tea as an auxiliary treatment. For example, according to the prescriptions in the Complete Collection of Chinese Herbal Health Teas: ① Qi, Blood, Yin and Yang Replenishing Prescription: Composed of 10-15g of Astragalus, 3-5g of American Ginseng, 6-10g of Lycium barbarum, and 10g of Polygonatum, it can simultaneously regulate Qi and blood, replenish the spleen and kidneys, and harmonize Yin and Yang; ② Astragalus, Angelica and Red Dates Tea: Take 12 grams of Astragalus, 5 grams of Angelica and 3 dates, simmer in boiling water for 20 minutes. Depending on your constitution, you can add 10 grams of Ophiopogon japonicus or American ginseng to prevent internal heat. For those with cold deficiency, add two slices of ginger. For those with heavy phlegm and dampness, add 10 grams of dried tangerine peel. It has the effects of tonifying the middle qi, nourishing the blood and calming the mind. ③ Wolfberry and Angelica Tea: 10 grams each of longan meat, wolfberry and angelica, soak them together in water and drink as tea. It has the effect of nourishing the liver and kidneys, invigorating qi and activating blood circulation.

[0024] Among them, the treatment plan adjustment system is a replacement of the therapeutic drugs in the drug treatment in the personalized treatment plan given by the physician through the treatment plan matching and recommendation system. In non-drug treatment, the adaptation is based on the patient's own personalization. The adaptation includes increasing the drug dosage, frequency of use, treatment cycle, etc. For example: (1) The patient does not smoke, but whether there is a lot of secondhand smoke in his work environment, home environment and travel environment, then the patient needs to be advised to wear a corresponding mask or avoid entering the above environment. The mask should be an N95 mask; (2) During the rehabilitation treatment, the corresponding rehabilitation plan is selected according to the patient's own condition. For example, patients with chronic obstructive pulmonary disease (COPD) in the early, middle and late stages can make corresponding choices based on whether they can perform strenuous exercise and their respiratory status. For example: ① Patients with difficulty in movement and breathing can use the method of respiratory muscle training + respiratory muscle training. This method requires a small range of motion and can exercise the respiratory muscles; ② Patients who can perform normal exercise can use the method of respiratory muscle training + respiratory muscle training + exercise training for rehabilitation treatment.

[0025] Among them, the artificial intelligence algorithm analyzes patient symptoms from the medical knowledge base, mainly through the following steps: Data collection and preprocessing: ① Data collection: Collect patient symptom data from multiple channels such as electronic medical record systems and medical sensors, including symptom descriptions and test results; ②Data cleaning: remove duplicate, erroneous and incomplete data to ensure data quality; ③Data standardization: unify the standards of data from different sources and formats, such as standardizing symptom descriptions.

[0026] Knowledge Representation and Modeling: ① Knowledge representation: Represent the knowledge in the medical knowledge base in a computer-understandable form, such as using an ontology model to describe concepts and relationships such as diseases and symptoms; ②Build a model: Use machine learning to build an analysis model. For example, a decision tree model can be used to classify diseases based on symptom characteristics, and a neural network model can be used to automatically learn the complex mapping relationship between symptoms and diseases.

[0027] Feature extraction and selection: ① Feature extraction: Extract features related to symptom analysis from preprocessed data, such as extracting key symptom words from text symptom descriptions and extracting numerical features from examination data; ② Feature selection: Use information gain, mutual information and other methods to screen out the most representative and discriminative features, reduce data dimensions, and improve analysis efficiency and accuracy.

[0028] Model training and optimization: ① Model training: Use labeled symptom-disease data to train the model and adjust the model parameters so that the model can accurately infer the disease from the symptom data; ②Model optimization: Optimize the model through cross-validation, adjustment of hyperparameters, and other methods to improve its generalization ability and accuracy.

[0029] Symptom analysis and reasoning: ① Symptom matching: The patient's symptom data is input into the trained model. The model uses the learned knowledge and patterns to find the disease or symptom combination that best matches the patient's symptoms. ② Reasoning and explanation: Not only does it provide possible disease diagnosis results, but it also uses technologies such as knowledge graphs to explain the reasoning process to help doctors understand the basis for model judgments.

[0030] Result evaluation and feedback: ① Result evaluation: Use indicators such as accuracy and recall to evaluate the accuracy and reliability of model analysis results; ② Feedback and improvement: Based on the evaluation results and doctors’ actual application feedback, the model is adjusted and improved to continuously improve the performance of symptom analysis.

[0031] Among them, nutritional therapy can focus on the allergy history according to the basic information provided by the patient to avoid other diseases caused by allergies. Drug therapy and non-drug therapy are used to judge the recovery effect based on weekly examinations. The examinations include changes in 6-minute walking distance and cardiopulmonary function test indicators, and the evaluation is based on the test results. Example

[0032] Patient A: Matching and recommending through the treatment plan matching and recommendation system. According to the matching and recommendation results, the patient has early symptoms of COPD, and a treatment plan can be designed through a drug treatment + rehabilitation treatment plan. In addition, the patient is advised to take daily nutritional support and arrange regular follow-up visits every 15 days. The follow-up results show that the patient's recovery effect is excellent. Example

[0033] Patient B: Matching and recommending through the treatment plan matching and recommendation system. According to the matching and recommendation results, the patient has mid-term COPD symptoms, and a treatment plan can be designed with medication + oxygen therapy + a small amount of rehabilitation therapy. In addition, the patient is advised based on whether he has a history of smoking. Finally, the patient needs to be advised to receive daily nutritional support and arrange regular follow-up visits every 7 days. The follow-up results show that the patient's recovery effect is good. Example

[0034] Patient C Matching and recommendation are carried out through the treatment plan matching and recommendation system. According to the matching and recommendation results, the patient has symptoms of late-stage COPD, and a treatment plan can be designed through drug treatment + oxygen therapy, reducing strenuous exercise, and giving advice based on whether the patient has a history of smoking. Finally, the patient needs to be advised to obtain daily nutritional support and arrange regular follow-up visits every 3 days. The follow-up results show that the patient's recovery effect is relatively improved.

[0035] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0036] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A treatment plan matching method for rehabilitation treatment of chronic obstructive pulmonary disease patients, characterized in that: The following steps are involved: Step 1: Data collection and integration module: responsible for collecting multi-dimensional data of patients. Patients can fill in medical record information manually, or use the QR code generator to generate QR codes for data information. Patients fill in data by scanning the QR code. The filled data includes name, gender, age, ethnicity, marital status, occupation, contact information, home address, medical insurance type, allergy history, past medical history, family medical history, symptoms, disease progression, treatment status and lifestyle habits. By integrating these data, an information database that comprehensively reflects the health status of patients is constructed. The information database is connected to the electronic medical record system server through the hospital LAN. The information is transmitted in real time or in batches at regular intervals and stored in the electronic medical record system (EMR) database. The information database uses the AES encryption algorithm to encrypt the collected data. Data is encrypted for storage and transmission to prevent data leakage, protect patient privacy, and comply with GDPR privacy protection measures. According to different access groups, it is divided into two types of access rights: patient side and doctor side. During the access process, personal authentication is performed through face recognition, and a linked list algorithm is used to record network access behavior in real time. By defining the linked list node structure and linked list structure, and realizing the functions of adding nodes and printing node data, access records can be effectively recorded. The linked list can be used to connect the access record nodes in sequence according to the order of access. Whenever a new access occurs, a new node is created and inserted into the end of the linked list. The node insertion and deletion operations of the linked list are relatively simple and efficient. When inserting a new access record, you only need to modify the pointer of the relevant node; when deleting a record, you only need to adjust the pointer. Step 2: Intelligent diagnosis and evaluation engine: (1) Utilize the artificial intelligence algorithm of the masked language model and the medical knowledge base to conduct in-depth analysis on the data input by the patient, receive the patient's basic information, vital signs, test results and other structured data from the electronic medical record system, which are stored in the form of tables and can be directly read and processed by the computer. For unstructured text data such as medical record text, perform natural language processing operations such as word segmentation, part-of-speech tagging, named entity recognition, etc., convert the text into a feature vector that can be understood by the computer, combine the diagnosis results and related features, evaluate the severity and development trend of the disease, and provide corresponding evaluation indicators and prediction results. Use the LIME model interpretation technology to explain the decision-making process and results of the model to help doctors and patients understand the basis and rationality of intelligent diagnosis and evaluation, and use association Rule mining: It can discover the correlation between different symptoms, examination results and diseases in medical data, providing a basis for diagnosis. The AI ​​algorithm based on the masked language model analyzes the patient's living habit data (such as secondhand smoke exposure). The artificial intelligence algorithm adopts a self-supervised learning training method. The AI ​​algorithm model is trained by generating pseudo labels from the input data itself. The training uses the masked language model and the causal language model to predict the masked word and the next word given the previous text, respectively, so that the model learns the relationship between the word and the context, the sequential structure of the text and the generation process of the language; (2) By combining the doctor's diagnosis results of the early, middle and late stages of chronic obstructive pulmonary disease (COPD) based on the patient's symptoms, chest CT examination, pulmonary function test, etc., the intelligent diagnosis and evaluation engine integrates the doctor's diagnosis results of pulmonary function test stages to generate a comprehensive evaluation report; Step 3: Treatment Matching and Recommendation System: Based on the combined results of intelligent diagnosis and assessment, the system selects the most suitable personalized treatment plan for the patient based on the Global Initiative for the Prevention and Control of Chronic Obstructive Pulmonary Disease and the clinical diagnosis and treatment pathway database. Treatment plans include medication, non-drug therapies, and regular follow-up appointments. Deep learning algorithms, including convolutional neural networks and recurrent neural networks, are used to formulate treatment plans. Convolutional neural networks automatically extract features from image data and other data to recommend treatment plans for image-related diseases. This is based on their strong image data processing capabilities and their ability to identify the connection between key information in images and the disease and treatment plan. Recurrent neural networks are suitable for processing sequential medical data, such as medical records. They can capture time series information in the data and recommend treatment plans that are more consistent with the progression of the disease. This is based on their ability to effectively process long-term dependencies in sequential data, reflect the dynamic changes of the disease, and thus match treatment plans that are tailored to the patient. Treatment plans are based on specific numerical values ​​such as drug blood concentration monitoring results and quantitative indicators of the patient's symptom improvement (such as the degree of reduction in dyspnea scores). Step 4: Treatment plan adjustment system: The physician can make corresponding adjustments based on the treatment plan matching and the personalized treatment plan given by the recommendation system. During the adjustment process, the physician can make adjustments based on the actual patient data. For example: (1) The treatment effect of the patient taking XXX drug is not as good as that of XXX drug. In this case, the physician can make adjustments; (2) The patient has a physical disability or serious disease that makes it impossible for the patient to recover from exercise normally. In this case, the physician can make adjustments; Step 5: Dynamic Adjustment and Tracking Module: Patients upload their daily data through the mobile app, and then use the data feedback mechanism to continuously track the patient's treatment effect and disease changes. By regularly collecting patient feedback data and sending the feedback data to the physician's mobile app, the doctor can optimize and adjust the treatment plan in a timely manner to ensure that the treatment always meets the patient's actual needs. For example: when the respiratory rate deviation in the patient's feedback data exceeds 10%, the system automatically triggers the optimization of the treatment plan.

2. A treatment plan matching method for rehabilitation treatment of COPD patients according to claim 1, characterized in that: The intelligent diagnosis and evaluation engine uses artificial intelligence algorithms and a medical knowledge base to combine the physician's diagnosis results of early, middle and late stages of chronic obstructive pulmonary disease (COPD) based on the patient's symptoms, chest CT examinations, pulmonary function tests, etc., and uses data and experience for comprehensive combination, and then conducts in-depth analysis and evaluation based on the combination of the two. The diagnosis results of the intelligent diagnosis and evaluation engine through artificial intelligence algorithms and a medical knowledge base are based on the "Global Initiative for the Prevention and Control of Chronic Obstructive Pulmonary Disease", which is the knowledge content that doctors need to learn, thereby ensuring that the diagnostic results of the two are consistent.

3. The method for matching treatment plans for rehabilitation treatment of COPD patients according to claim 1, characterized in that: The drug treatment includes bronchodilators, anticholinergic drugs, theophylline drugs, and glucocorticoids. Specific drug treatment recommendations are based on AI's analysis of the patient's acute exacerbation history. These recommendations are adjusted based on the drug treatment effect and the patient's physical condition. Side effects and the patient's tolerance to the drug should also be considered. For example, if the patient's own allergens are present in the raw materials of a certain drug: (1) Bronchodilators are the core drugs for the treatment of COPD. For example, short-acting β2 receptor agonists such as salbutamol and terbutaline have a fast onset of action and a short duration of action. They can quickly relieve dyspnea symptoms and are often used for temporary symptom relief. Long-acting β2 receptor agonists such as salmeterol and formoterol have a long duration of action that can last for more than 12 hours and are more suitable for regular use to control symptoms. (2) Anticholinergic drugs: such as ipratropium bromide, which is a short-acting anticholinergic drug; tiotropium bromide and glycopyrrolate are long-acting anticholinergic drugs that block M cholinergic receptors to dilate the bronchi and improve ventilation; (3) Theophylline drugs: aminophylline, etc., can relax bronchial smooth muscles, and also have certain anti-inflammatory and respiratory muscle strengthening effects, but they should be used with caution due to their side effects: (1) Tachycardia and arrhythmia are among the most common adverse reactions of theophylline drugs. Theophylline drugs increase the concentration of cyclic adenosine monophosphate in cells by inhibiting the activity of phosphodiesterase, thereby dilating the bronchi, but they may also stimulate the heart, leading to increased heart rate or arrhythmia; (2) Headache and anxiety are also possible adverse reactions of theophylline drugs. This is mainly because theophylline drugs can produce an excitatory effect through the central nervous system, causing excessive activity of nerve cells, thereby causing headaches and anxiety. (3) Nausea and vomiting are another common adverse reaction of theophylline drugs. This is because theophylline drugs can stimulate the gastrointestinal tract, leading to gastrointestinal dysfunction, thereby causing nausea and vomiting; (4) Glucocorticoids: For patients with moderate to severe COPD, when bronchodilators are ineffective, inhaled glucocorticoids, such as budesonide and fluticasone, can be combined with long-acting β2 receptor agonists (such as salmeterol fluticasone powder inhaler®) to more effectively control symptoms and reduce acute exacerbations. The non-drug treatments include smoking cessation, oxygen therapy, rehabilitation therapy, nutritional support and health teas. Different non-drug treatment plans can be formulated based on the patient's age, gender, physical condition, lifestyle habits and other factors: (1) Smoking cessation: For patients who smoke, smoking cessation is the most important intervention measure. Smoking cessation counseling, nicotine replacement therapy, and medication assistance (such as bupropion and varenicline) can help patients quit smoking; (2) Oxygen therapy: The duration of oxygen therapy (≥15 hours / day) is dynamically adjusted based on the patient's arterial oxygen saturation AI prediction model; (3) Rehabilitation treatment: Respiratory muscle exercises: ① Pursed lip breathing (close your mouth and inhale through your nose, then exhale slowly with pursed lips, with the exhalation time being about twice the inhalation time); ② Abdominal breathing (the abdomen bulges when inhaling and contracts when exhaling) can strengthen the respiratory muscles and improve respiratory function; ③ Exercise, such as walking and Tai Chi, can gradually increase the intensity and duration of exercise according to the patient's condition, which will help improve the patient's exercise endurance and ability to take care of themselves; (4) Nutritional support: COPD patients often suffer from malnutrition due to increased consumption caused by dyspnea and affected gastrointestinal function. They should ensure adequate intake of protein (such as lean meat, fish, eggs, beans, etc.), carbohydrates and fats, and pay attention to supplementing vitamins and minerals; (5) Health tea: Add or subtract herbs based on the patient's constitution. Recommendations can also be made based on the severity of the patient's condition and TCM constitution classification (such as Qi deficiency, Yang deficiency, phlegm-dampness, etc.) to ensure the targeted nature of health-preserving tea as an auxiliary treatment. For example, according to the prescriptions in the Complete Collection of Chinese Herbal Health Teas: ① Qi, Blood, Yin and Yang Replenishing Prescription: Composed of 10-15g of Astragalus, 3-5g of American Ginseng, 6-10g of Lycium barbarum, and 10g of Polygonatum, it can simultaneously regulate Qi and blood, replenish the spleen and kidneys, and harmonize Yin and Yang; ② Astragalus, Angelica and Red Dates Tea: Take 12 grams of Astragalus, 5 grams of Angelica and 3 dates, simmer in boiling water for 20 minutes. Depending on your constitution, you can add 10 grams of Ophiopogon japonicus or American ginseng to prevent internal heat. For those with cold deficiency, add two slices of ginger. For those with heavy phlegm and dampness, add 10 grams of dried tangerine peel. It has the effects of tonifying the middle qi, nourishing the blood and calming the mind. ③ Wolfberry and Angelica Tea: 10 grams each of longan meat, wolfberry and angelica, soak them together in water and drink as tea. It has the effect of nourishing the liver and kidneys, invigorating qi and activating blood circulation.

4. The method for matching treatment plans for rehabilitation treatment of COPD patients according to claim 1, characterized in that: The treatment plan adjustment system is a system that allows doctors to replace the therapeutic drugs in the personalized treatment plan given by the treatment plan matching and recommendation system, and to adapt the non-drug treatment to the patient's own personal needs. The adaptation includes increasing the drug dosage, frequency of use, treatment cycle, etc. For example: (1) The patient does not smoke, but whether there is a lot of secondhand smoke in his work environment, home environment and travel environment, then the patient needs to be advised to wear a corresponding mask or avoid entering the above environment. The mask should be an N95 mask; (2) During the rehabilitation treatment, the corresponding rehabilitation plan is selected according to the patient's own condition. For example, patients with chronic obstructive pulmonary disease (COPD) in the early, middle and late stages can make corresponding choices based on whether they can perform strenuous exercise and their respiratory status. For example: ① Patients with difficulty in movement and breathing can use the method of respiratory muscle training + respiratory muscle training. This method requires a small range of motion and can exercise the respiratory muscles; ② Patients who can perform normal exercise can use the method of respiratory muscle training + respiratory muscle training + exercise training for rehabilitation treatment.

5. The method for matching treatment plans for rehabilitation treatment of COPD patients according to claim 1, characterized in that: The artificial intelligence algorithm analyzes patient symptoms from the medical knowledge base, mainly through the following steps: Data collection and preprocessing: ① Data collection: Collect patient symptom data from multiple channels such as electronic medical record systems and medical sensors, including symptom descriptions and test results; ②Data cleaning: remove duplicate, erroneous and incomplete data to ensure data quality; ③Data standardization: unify the standards of data from different sources and formats, such as standardizing symptom descriptions.

6. Knowledge Representation and Modeling: ① Knowledge representation: Represent the knowledge in the medical knowledge base in a computer-understandable form, such as using an ontology model to describe concepts and relationships such as diseases and symptoms; ②Build a model: Use machine learning to build an analysis model. For example, a decision tree model can be used to classify diseases based on symptom characteristics, and a neural network model can be used to automatically learn the complex mapping relationship between symptoms and diseases.

7. Feature extraction and selection: ① Feature extraction: Extract features related to symptom analysis from preprocessed data, such as extracting key symptom words from text symptom descriptions and extracting numerical features from examination data; ② Feature selection: Use information gain, mutual information and other methods to screen out the most representative and discriminative features, reduce data dimensions, and improve analysis efficiency and accuracy.

8. Model training and optimization: ① Model training: Use labeled symptom-disease data to train the model and adjust the model parameters so that the model can accurately infer the disease from the symptom data; ②Model optimization: Optimize the model through cross-validation, adjustment of hyperparameters, and other methods to improve its generalization ability and accuracy.

9. Symptom analysis and reasoning: ① Symptom matching: The patient's symptom data is input into the trained model. The model uses the learned knowledge and patterns to find the disease or symptom combination that best matches the patient's symptoms. ② Reasoning and explanation: Not only does it provide possible disease diagnosis results, but it also uses technologies such as knowledge graphs to explain the reasoning process to help doctors understand the basis for model judgment.

10. Result evaluation and feedback: ① Result evaluation: Use indicators such as accuracy and recall to evaluate the accuracy and reliability of the model analysis results; ② Feedback and improvement: Based on the evaluation results and doctors’ actual application feedback, the model is adjusted and improved to continuously improve the performance of symptom analysis.

11. A treatment plan matching method for rehabilitation treatment of COPD patients according to claim 3, characterized in that: The nutritional therapy may focus on the allergy history in the basic information provided by the patient to avoid other symptoms caused by allergies. The drug therapy and non-drug therapy may be used to judge the recovery effect based on weekly examinations. The examinations include changes in 6-minute walking distance and cardiopulmonary function test indicators, and the evaluation is based on the examination results.

Images