Traditional Chinese medicine informatization processing system and method for acute lung injury

Abstract

The invention relates to the technical field of information processing, in particular to a traditional Chinese medicine informatization processing system and method for acute lung injury. The system comprises a traditional Chinese medicine information slice storage module, a traditional Chinese medicine knowledge graph construction module, a dialectical auxiliary reasoning module and a prescription sorting recommendation module. An acute lung injury traditional Chinese medicine information database can be generated by collecting and integrating ancient traditional Chinese medicine poeia information corresponding to acute lung injury, domestic and overseas clinical research information and acute lung injury patient record data extracted from electronic medical record systems corresponding to various hospitals and performing slice coding storage; performing entity and relation extraction and knowledge graph construction according to the acute lung injury traditional Chinese medicine information database, and generating an acute lung injury traditional Chinese medicine knowledge graph; dialectical auxiliary reasoning and traditional Chinese medicine prescription sorting recommendation are carried out based on the acute lung injury traditional Chinese medicine knowledge map, and a traditional Chinese medicine prescription suggestion scheme sequence corresponding to the acute lung injury syndrome type is generated. The information processing efficiency of acute lung injury traditional Chinese medicine prescription recommendation can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of information processing, and particularly to an information processing system and method for traditional Chinese medicine for acute lung injury. Background Art

[0002] Acute lung injury (ALI) refers to an acute pulmonary inflammatory response caused by various factors (such as infection, trauma, drugs, poisons, etc.), leading to a series of clinical manifestations such as alveolar epithelial cell injury, pulmonary edema, and gas exchange disorders. In severe cases, it can develop into acute respiratory distress syndrome (ARDS). In recent years, with the development of information technology, the intellectualization and informatization in the medical field have gradually become research hotspots. The introduction of information processing methods provides new possibilities for the information processing of traditional Chinese medicine. Especially in the treatment of traditional Chinese medicine for acute lung injury, through information processing technology, the specific condition of patients can be monitored in real time. Combining with the theoretical guidance of traditional Chinese medicine, factors such as the cause, symptoms, and constitution of patients can be accurately analyzed, and then a personalized traditional Chinese medicine prescription combination plan can be formulated. However, there are still some challenges in the current information processing methods of traditional Chinese medicine. The knowledge system of traditional Chinese medicine is huge and complex, including a large amount of traditional Chinese medicine prescriptions, medicinal material information, clinical experience, etc. The traditional manual search and analysis methods are inefficient and difficult to quickly and accurately obtain and integrate relevant information. In addition, there are differences in the syndrome differentiation and treatment of acute lung injury among different doctors, resulting in low processing efficiency of traditional Chinese medicine prescription recommendations. Summary of the Invention

[0003] Based on this, it is necessary for the present invention to provide an information processing system and method for traditional Chinese medicine for acute lung injury to solve at least one of the above technical problems.

[0004] To achieve the above object, an information processing system for traditional Chinese medicine for acute lung injury includes the following modules: A traditional Chinese medicine information slicing storage module, which is used to collect and integrate the information of ancient traditional Chinese medicine classics corresponding to acute lung injury, domestic and foreign clinical research information, and the recorded data of acute lung injury patients extracted from the electronic medical record systems of major hospitals, and perform slicing coding storage on the information of ancient traditional Chinese medicine classics, domestic and foreign modern clinical research information, and the recorded data of acute lung injury patients to generate an acute lung injury traditional Chinese medicine information database; A traditional Chinese medicine knowledge graph construction module, which is used to extract entities and relationships based on the information of ancient traditional Chinese medicine classics corresponding to acute lung injury and domestic and foreign clinical research information in the acute lung injury traditional Chinese medicine information database to generate the entity of the traditional Chinese medicine knowledge structure corresponding to acute lung injury and the relationship of the traditional Chinese medicine knowledge chain; construct a knowledge graph based on the entity of the traditional Chinese medicine knowledge structure corresponding to acute lung injury and the relationship of the traditional Chinese medicine knowledge chain to generate an acute lung injury traditional Chinese medicine knowledge graph; A dialectical auxiliary reasoning module, which is used to perform dialectical auxiliary reasoning based on the traditional Chinese medicine knowledge graph of acute lung injury and in combination with the recorded data of acute lung injury patients in the traditional Chinese medicine information database of acute lung injury, so as to generate the dialectical suspected syndrome types corresponding to acute lung injury; A prescription sorting and recommendation module, which is used to sort and recommend traditional Chinese medicine prescriptions for the acute lung injury traditional Chinese medicine knowledge graph based on the dialectical suspected syndrome types corresponding to acute lung injury, so as to generate a sequence of traditional Chinese medicine prescription suggestion plans corresponding to the syndrome types of acute lung injury.

[0005] Furthermore, the traditional Chinese medicine information slicing storage module includes the following functions: By collecting the information records in ancient traditional Chinese medicine classics regarding acute lung injury, including symptom descriptions, prescription compositions, and medication experiences, to obtain the ancient traditional Chinese medicine classic information corresponding to acute lung injury; By collecting and integrating the clinical research information on the traditional Chinese medicine treatment of acute lung injury in domestic and foreign modern medical journals and research reports, to obtain the domestic and foreign clinical research information corresponding to acute lung injury; By extracting the clinical diagnosis information, medication records, and clinical effects corresponding to acute lung injury patients from the electronic medical record systems of major hospitals, to obtain the recorded data of acute lung injury patients; Removing duplicate, incorrect, and invalid data from the ancient traditional Chinese medicine classic information, domestic and foreign modern clinical research information, and the recorded data of acute lung injury patients, to obtain the cleaned traditional Chinese medicine classic information, domestic and foreign clinical research information, and acute lung injury recorded data; Performing slicing coding storage on the cleaned traditional Chinese medicine classic information, domestic and foreign clinical research information, and acute lung injury recorded data, to generate a traditional Chinese medicine information database of acute lung injury.

[0006] Furthermore, the slicing coding storage of the cleaned traditional Chinese medicine classic information, domestic and foreign clinical research information, and acute lung injury recorded data includes: Extracting symptom attributes of the acute lung injury conditions corresponding to the cleaned traditional Chinese medicine classic information, domestic and foreign clinical research information, and acute lung injury recorded data, to obtain the acute lung injury symptom attribute elements corresponding to the pharmacopoeia, research, and patient records; Classifying the acute lung injury symptom attribute elements corresponding to the pharmacopoeia, research, and patient records into symptom type clusters, to generate different acute lung injury symptom attribute type cluster groups; Based on different acute lung injury symptom attribute type cluster groups, dividing the traditional Chinese medicine classic information, domestic and foreign clinical research information, and acute lung injury recorded data under the corresponding type clusters into symptom type slices, to generate traditional Chinese medicine information fragments, domestic and foreign research information fragments, and patient record data fragments under different symptom type clusters; Index and encode the traditional Chinese medicine information segments, domestic and foreign research information segments, and patient record data segments under different symptom type clusters for storage, so as to generate an acute lung injury traditional Chinese medicine information database.

[0007] Further, the classification of the acute lung injury symptom attribute elements corresponding to the pharmacopoeia, research, and patient records into symptom type clusters includes: Perform word vector embedding on the acute lung injury symptom attribute elements corresponding to the pharmacopoeia, research, and patient records to generate acute lung injury symptom word embedding vectors corresponding to the pharmacopoeia, research, and patient records; Perform semantic feature analysis on the acute lung injury symptom word embedding vectors corresponding to the pharmacopoeia, research, and patient records to obtain the acute lung injury symptom semantic features corresponding to the pharmacopoeia, research, and patient records; Evaluate the symptom scores according to the acute lung injury symptom semantic features corresponding to the pharmacopoeia, research, and patient records to obtain the acute lung injury symptom scores corresponding to the pharmacopoeia, research, and patient records; Classify the acute lung injury symptom attribute elements corresponding to the pharmacopoeia, research, and patient records into symptom type clusters based on the acute lung injury symptom scores corresponding to them, so as to generate different acute lung injury symptom attribute type cluster groups.

[0008] Further, the traditional Chinese medicine knowledge graph construction module includes the following functions: Identify and extract knowledge entity according to the ancient traditional Chinese medicine pharmacopoeia information and domestic and foreign clinical research information corresponding to different types of acute lung injury in the acute lung injury traditional Chinese medicine information database, so as to generate traditional Chinese medicine knowledge structure entities corresponding to acute lung injury; Measure the knowledge association between the traditional Chinese medicine knowledge structure entities corresponding to each acute lung injury to obtain the knowledge association coefficient between each traditional Chinese medicine knowledge entity; Connect the traditional Chinese medicine knowledge structure entities corresponding to each acute lung injury based on the knowledge association coefficient between each traditional Chinese medicine knowledge entity to generate an acute lung injury action relationship chain between each traditional Chinese medicine knowledge entity; Extract the knowledge chain relationship of the acute lung injury action relationship chain between each traditional Chinese medicine knowledge entity based on the traditional Chinese medicine knowledge structure entities corresponding to each acute lung injury to generate the traditional Chinese medicine knowledge chain relationship corresponding to acute lung injury; Construct a knowledge graph according to the traditional Chinese medicine knowledge structure entities corresponding to acute lung injury and the traditional Chinese medicine knowledge chain relationship, using each traditional Chinese medicine knowledge structure entity as a node and the corresponding traditional Chinese medicine knowledge chain relationship as a connecting edge, so as to generate an acute lung injury traditional Chinese medicine knowledge graph.

[0009] Further, the extraction of the knowledge chain relationship of the acute lung injury effect relationship between each traditional Chinese medicine knowledge entity based on the traditional Chinese medicine knowledge structure entity corresponding to each acute lung injury includes: Perform a pharmacodynamic synergy evaluation between the traditional Chinese medicine knowledge structure entities corresponding to each acute lung injury to obtain the acute lung injury pharmacodynamic synergy coefficient between each traditional Chinese medicine knowledge entity; Obtain the traditional Chinese medicine pharmacological action components corresponding to each traditional Chinese medicine knowledge entity through the traditional Chinese medicine knowledge structure entities corresponding to each acute lung injury, and perform a pharmacological conflict probability analysis based on the traditional Chinese medicine pharmacological action components corresponding to each traditional Chinese medicine knowledge entity to obtain the pharmacological action conflict probability between each traditional Chinese medicine knowledge entity; Quantify the compatibility combination degree between the traditional Chinese medicine knowledge structure entities corresponding to each acute lung injury based on the acute lung injury pharmacodynamic synergy coefficient and the pharmacological action conflict probability between each traditional Chinese medicine knowledge entity to obtain the pharmacological action compatibility combination degree between each traditional Chinese medicine knowledge entity; Based on the pharmacological action compatibility combination degree between each traditional Chinese medicine knowledge entity, screen and extract the compatibility combination chain of the acute lung injury effect relationship between each traditional Chinese medicine knowledge entity, so as to extract the acute lung injury effect relationship chain corresponding to the traditional Chinese medicine knowledge entities with a pharmacological action compatibility combination degree greater than the preset threshold as the corresponding knowledge chain relationship, so as to generate the traditional Chinese medicine knowledge chain relationship corresponding to acute lung injury.

[0010] Further, the dialectical auxiliary reasoning module includes the following functions: Perform a deep analysis of the dialectical knowledge of the acute lung injury traditional Chinese medicine knowledge graph to extract the dialectical knowledge related to acute lung injury, including the etiology, pathogenesis and pathological characteristics corresponding to acute lung injury, and obtain the acute lung injury traditional Chinese medicine dialectical knowledge; Extract the dialectical relationship and rules of the acute lung injury traditional Chinese medicine knowledge graph based on the acute lung injury traditional Chinese medicine dialectical knowledge, so as to generate the dialectical relationship and dialectical rules corresponding to each acute lung injury dialectical knowledge in the graph; Perform dialectical auxiliary reasoning based on the dialectical relationship and dialectical rules corresponding to each acute lung injury dialectical knowledge in the graph and combine the acute lung injury patient record data in the acute lung injury traditional Chinese medicine information database to generate the dialectical suspected syndrome types corresponding to acute lung injury.

[0011] Further, the dialectical auxiliary reasoning based on the dialectical relationship and dialectical rules corresponding to each acute lung injury dialectical knowledge in the graph and combining the acute lung injury patient record data in the acute lung injury traditional Chinese medicine information database includes: Perform a clinical symptom analysis of the acute lung injury patient record data in the acute lung injury traditional Chinese medicine information database to obtain the clinical symptom characteristics corresponding to the acute lung injury patients; According to the corresponding dialectical relationships and dialectical rules of each acute lung injury dialectical knowledge in the atlas, and the corresponding acute lung injury syndrome types in the traditional Chinese medicine knowledge atlas of acute lung injury, a dialectical correlation analysis is carried out to obtain the internal connection between each acute lung injury dialectical effect and the acute lung injury syndrome type; Based on the internal connection between each acute lung injury dialectical effect and the acute lung injury syndrome type, a dialectical auxiliary reasoning is carried out on the corresponding clinical symptom characteristics of acute lung injury patients to generate a dialectical suspected syndrome type corresponding to acute lung injury.

[0012] Furthermore, the prescription sorting and recommendation module includes the following functions: Based on the dialectical suspected syndrome type corresponding to acute lung injury, a similarity measurement calculation is carried out between the corresponding symptoms of acute lung injury patients in the traditional Chinese medicine knowledge atlas of acute lung injury to obtain the similarity between the dialectical suspected syndrome type and each symptom in the knowledge atlas; Based on the similarity between the dialectical suspected syndrome type and each symptom in the knowledge atlas, a similarity sorting process is carried out on the corresponding symptoms of acute lung injury patients in the traditional Chinese medicine knowledge atlas of acute lung injury to generate a similar symptom sequence of patients corresponding to the acute lung injury syndrome type; Based on the traditional Chinese medicine knowledge atlas of acute lung injury, a sorting and recommendation of traditional Chinese medicine prescriptions is carried out for the corresponding symptoms of acute lung injury patients in the similar symptom sequence of patients corresponding to the acute lung injury syndrome type, so as to find out the corresponding classic traditional Chinese medicine prescription combination suggestion plan in the knowledge atlas in sequence, and generate a traditional Chinese medicine prescription suggestion plan sequence corresponding to the acute lung injury syndrome type.

[0013] Furthermore, the present invention also provides a method for information processing of traditional Chinese medicine for acute lung injury. The method is implemented based on the above-mentioned information processing system of traditional Chinese medicine for acute lung injury. The method for information processing of traditional Chinese medicine for acute lung injury includes: By collecting and integrating the information of ancient traditional Chinese medicine classics corresponding to acute lung injury, domestic and foreign clinical research information, and the record data of acute lung injury patients extracted from the electronic medical record systems of major hospitals, and performing slice coding storage on the information of ancient traditional Chinese medicine classics, domestic and foreign modern clinical research information, and the record data of acute lung injury patients, an information database of traditional Chinese medicine for acute lung injury is generated; According to the information of ancient traditional Chinese medicine classics corresponding to acute lung injury and domestic and foreign clinical research information in the information database of traditional Chinese medicine for acute lung injury, entity and relationship extraction are carried out to generate the entity of traditional Chinese medicine knowledge structure corresponding to acute lung injury and the chain relationship of traditional Chinese medicine knowledge; according to the entity of traditional Chinese medicine knowledge structure corresponding to acute lung injury and the chain relationship of traditional Chinese medicine knowledge, a knowledge atlas is constructed to generate a traditional Chinese medicine knowledge atlas of acute lung injury; Based on the traditional Chinese medicine knowledge atlas of acute lung injury and combined with the record data of acute lung injury patients in the information database of traditional Chinese medicine for acute lung injury, a dialectical auxiliary reasoning is carried out to generate a dialectical suspected syndrome type corresponding to acute lung injury; Sort and recommend traditional Chinese medicine prescriptions for the traditional Chinese medicine knowledge graph of acute lung injury based on the dialectical suspected syndrome types corresponding to acute lung injury, so as to generate a sequence of recommended traditional Chinese medicine prescription solutions corresponding to the syndrome types of acute lung injury.

[0014] The beneficial effects of the present invention: The acute lung injury traditional Chinese medicine information processing system proposed by the present invention is generally composed of a traditional Chinese medicine information slicing storage module, a traditional Chinese medicine knowledge graph construction module, a dialectical auxiliary reasoning module, and a prescription sorting and recommendation module. Compared with the prior art, the beneficial effects of this application are as follows: by collecting and integrating the information of ancient traditional Chinese medicine classics corresponding to acute lung injury (ALI), domestic and foreign clinical research information, and the record data of acute lung injury patients in hospitals, and slicing and encoding these data for storage, a comprehensive multi-dimensional knowledge framework can be assembled, including traditional ancient book materials, modern clinical research results, and actual clinical case data. The information of ancient traditional Chinese medicine classics can provide rich theories and medicinal material application experiences, the domestic and foreign clinical research information shows the latest research results of modern medicine on acute lung injury, and the patient record data can further reflect the disease manifestations and their changing trends in the actual clinical environment. By effectively combining these information and storing them in a standardized manner, the richness and practicability of the database are ensured. This database not only provides a solid foundation for the subsequent construction of the knowledge graph, but also makes future knowledge retrieval more efficient and information more accurate, and can meet different clinical practice needs, improving the clinical auxiliary decision-making ability of traditional Chinese medicine. Secondly, by extracting entities and relationships from the information of ancient traditional Chinese medicine classics, modern clinical research results, and patient record data in the acute lung injury traditional Chinese medicine information database, a traditional Chinese medicine knowledge structure entity and a knowledge chain relationship corresponding to acute lung injury can be constructed. The construction of the knowledge graph has far-reaching significance. Through the structured processing of a large amount of data, traditional Chinese medicine knowledge can be effectively inherited and expanded through modern information technology. Through entity extraction, information such as traditional Chinese medicines, prescriptions, symptoms, and etiologies related to acute lung injury is refined to form a knowledge structure. Relationship extraction helps to understand the action mechanisms of different traditional Chinese medicines and prescriptions in the treatment of acute lung injury and their synergistic effects by establishing the connections between these entities. The traditional Chinese medicine knowledge graph can not only display traditional Chinese medicine solutions in a more intuitive way, but also provide cross-disciplinary and cross-field knowledge integration and discovery, so as to quickly and accurately obtain and integrate corresponding traditional Chinese medicine prescriptions, medicinal material information, clinical experience and other information. Then, through dialectical auxiliary reasoning based on the acute lung injury traditional Chinese medicine knowledge graph and the acute lung injury patient record data, it combines traditional Chinese medicine dialectical theory with modern technology, realizing the intelligence and personalization of dialectical treatment. Dialectical treatment is one of the core ideas of traditional Chinese medicine, emphasizing individualized analysis according to the specific manifestations and overall conditions of patients. It can infer the suspected syndrome types of patients through reasoning analysis, and then provide an auxiliary decision-making basis for the subsequent prescription recommendation plan. The automation and intelligence of this reasoning method can reduce the differences in dialectical treatment among different doctors.Finally, by sorting and recommending prescriptions based on the dialectical suspected syndrome types corresponding to acute lung injury in the traditional Chinese medicine knowledge graph, it can provide doctors with accurate solutions based on traditional Chinese medicine theory and big data analysis. In actual operation, the selection of prescriptions is often affected by the diversity and complexity of patients' symptoms. Traditional manual selection methods are restricted by subjective factors and experience. Through automated prescription recommendation, different syndrome types of acute lung injury can be analyzed more scientifically, providing personalized traditional Chinese medicine prescription combinations for patients. In this way, the most suitable prescriptions can be pushed to doctors, reducing drug conflicts and side effects, ensuring the optimal effect of traditional Chinese medicine prescriptions, and thus improving the information processing efficiency of traditional Chinese medicine prescription recommendation. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments read in conjunction with the accompanying drawings: Figure 1 It is a schematic diagram of the modules of the traditional Chinese medicine information processing system for acute lung injury of the present invention; Figure 2 For Figure 1 It is a schematic diagram of the functional flow of the traditional Chinese medicine information slicing storage module in Figure 3 For Figure 1 It is a schematic diagram of the functional flow of the traditional Chinese medicine knowledge graph construction module in Figure 4 It is a schematic diagram of the semantic similarity matrix of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0016] The following clearly and completely describes the technical system of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are some but not all of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0017] In addition, the accompanying drawings are only schematic diagrams 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 thus their repeated description will be omitted. Some of the block diagrams shown in the 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 in one or more hardware modules or integrated circuits, or in different networks and / or processor systems and / or microcontroller systems.

[0018] It should be understood that although terms such as "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are only used to distinguish one unit from another. 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.

[0019] To achieve the above object, please refer to Figures 1 to 3 , the present invention provides an information processing system for traditional Chinese medicine for acute lung injury, and the system includes the following modules: A traditional Chinese medicine information slicing storage module, which is used to collect and integrate the information of ancient traditional Chinese medicine classics corresponding to acute lung injury, domestic and foreign clinical research information, and the recorded data of acute lung injury patients extracted from the corresponding electronic medical record systems of major hospitals, and slice and encode the information of ancient traditional Chinese medicine classics, domestic and foreign modern clinical research information, and the recorded data of acute lung injury patients for storage, so as to generate a traditional Chinese medicine information database for acute lung injury; A traditional Chinese medicine knowledge graph construction module, which is used to extract entities and relationships according to the information of ancient traditional Chinese medicine classics corresponding to acute lung injury and domestic and foreign clinical research information in the traditional Chinese medicine information database for acute lung injury, so as to generate the entity of the traditional Chinese medicine knowledge structure corresponding to acute lung injury and the relationship of the traditional Chinese medicine knowledge chain; construct a knowledge graph according to the entity of the traditional Chinese medicine knowledge structure corresponding to acute lung injury and the relationship of the traditional Chinese medicine knowledge chain, so as to generate a traditional Chinese medicine knowledge graph for acute lung injury; A dialectical auxiliary reasoning module, which is used to perform dialectical auxiliary reasoning based on the traditional Chinese medicine knowledge graph for acute lung injury and in combination with the recorded data of acute lung injury patients in the traditional Chinese medicine information database for acute lung injury, so as to generate the dialectical suspected syndrome types corresponding to acute lung injury; A prescription sorting and recommendation module, which is used to sort and recommend traditional Chinese medicine prescriptions for the traditional Chinese medicine knowledge graph for acute lung injury based on the dialectical suspected syndrome types corresponding to acute lung injury, so as to generate a sequence of recommended traditional Chinese medicine prescription schemes corresponding to the syndrome types of acute lung injury.

[0020] In the embodiment of the present invention, please refer to Figure 1 shown, which is a schematic diagram of the modules of the information processing system for traditional Chinese medicine for acute lung injury of the present invention. In this example, the information processing system for traditional Chinese medicine for acute lung injury includes the following modules: S1: Traditional Chinese medicine information slicing storage module, which is used to collect and integrate ancient Chinese medicine pharmacopoeia information corresponding to acute lung injury, domestic and foreign clinical research information, and acute lung injury patient record data extracted from the corresponding electronic medical record systems of major hospitals, and slice and encode and store the ancient Chinese medicine pharmacopoeia information, domestic and foreign modern clinical research information, and acute lung injury patient record data to generate an acute lung injury traditional Chinese medicine information database; In the embodiment of the present invention, by collecting and integrating data related to acute lung injury and performing slice coding and storage, a professional team is formed to carry out the work. For the information of ancient Chinese medicine pharmacopoeia, high-definition scans of more than 200 ancient books such as "Huangdi Neijing" and "Shanghan Zabing Lun" are obtained. The Tesseract OCR optical character recognition technology is used to convert the content of the ancient books into text format. The text is proofread and annotated by experts in Chinese medicine literature, and the description of the symptoms of acute lung injury, the composition of the prescription and the experience of using the medicine are extracted. In terms of collecting clinical research information at home and abroad, the Chinese medicine information database is used to "acute Lung injury, "acute lung injury", "traditional Chinese medicine treatment" and other search terms were used to retrieve literature data from the past 20 years. The literature was deduplicated using EndNote software, and medical professionals were organized to screen out valid literature related to traditional Chinese medicine treatment of acute lung injury. Through the data interface of the hospital information system (HIS), a stored procedure was written using the SQL Server database management system to extract the clinical diagnosis information, medication records and clinical effect data of patients with acute lung injury. After cleaning the above three types of data, the Hadoop distributed file system (HDFS) was used to store them. According to the data type, they were divided into three directories: "ancient book information", "clinical research information" and "patient record information". Each directory was sliced by year and quarter. A program was written in Java, and each slice data was encoded using the Base64 encoding algorithm. The encoded data was stored in the HBase database, and an index was established at the same time to finally generate a traditional Chinese medicine information database for acute lung injury.

[0021] S2: A Chinese medicine knowledge graph construction module is used to extract entities and relationships based on the ancient Chinese medicine pharmacopoeia information and domestic and foreign clinical research information corresponding to acute lung injury in the Chinese medicine information database to generate Chinese medicine knowledge structure entities and Chinese medicine knowledge chain relationships corresponding to acute lung injury; and to construct a knowledge graph based on the Chinese medicine knowledge structure entities and Chinese medicine knowledge chain relationships corresponding to acute lung injury to generate a Chinese medicine knowledge graph for acute lung injury; In the embodiment of the present invention, when extracting entities and relationships from the traditional Chinese medicine information database for acute lung injury and constructing a knowledge graph, for the information of ancient traditional Chinese medicine classics and domestic and foreign clinical research information, the spaCy library of Python is used for natural language processing to identify entities such as traditional Chinese medicine names, symptom descriptions, formula names, etiologies and pathogenesis. For example, from the sentence "Mahuang Decoction can treat the cough and asthma symptoms of patients with acute lung injury", entities such as "Mahuang Decoction", "acute lung injury", "cough", and "asthma" are extracted. A method combining rules and machine learning is used to extract the relationships between entities. Relationship types such as "treatment", "composition", and "causing" are predefined, and relationships are identified through regular expression matching and dependency syntax analysis. For example, for "Mahuang Decoction is composed of Ephedra, Cinnamon Twig, Apricot Kernel, and Licorice Root", the "composition" relationship between "Mahuang Decoction" and "Ephedra", "Cinnamon Twig", "Apricot Kernel", and "Licorice Root" can be identified. The extracted traditional Chinese medicine knowledge structure entities and traditional Chinese medicine knowledge chain relationships are stored in a MySQL database, and an acute lung injury traditional Chinese medicine knowledge graph is constructed using a Neo4j graph database. The traditional Chinese medicine knowledge structure entities are used as nodes, and different labels are assigned according to the entity type, such as "traditional Chinese medicine", "symptom", "formula", etc.; the traditional Chinese medicine knowledge chain relationships are used as connecting edges, and the types and attributes of the edges are set. For example, the relationship type is "treatment", and the attributes include literature source, evidence strength, etc. Finally, a visual acute lung injury traditional Chinese medicine knowledge graph is generated.

[0022] S3: A dialectical auxiliary reasoning module, which is used to perform dialectical auxiliary reasoning based on the acute lung injury traditional Chinese medicine knowledge graph and in combination with the record data of acute lung injury patients in the acute lung injury traditional Chinese medicine information database to generate the dialectical suspected syndrome types corresponding to acute lung injury; In the embodiment of the present invention, when performing dialectical auxiliary reasoning based on the acute lung injury traditional Chinese medicine knowledge graph and patient record data, the pandas library of Python is used to read the patient record data in the acute lung injury traditional Chinese medicine information database, and information such as the symptoms and examination results of the patients is preprocessed and feature extracted. For example, symptom features such as "fever", "cough", and "decrease in blood oxygen saturation" are extracted from the patient records. In the Neo4j database, the Cypher query language is used to traverse the acute lung injury traditional Chinese medicine knowledge graph to find the syndrome type nodes and relationship paths related to the patient's symptoms, and by combining the pre-set traditional Chinese medicine dialectical rules, these rules are stored in the "dialectical rules" table of the MySQL database. The forward reasoning algorithm is adopted to start from the patient's symptoms and reason based on the relationships and dialectical rules in the knowledge graph. For example, if a patient has symptoms such as "fever" and "yellow and thick sputum", and there is an associated relationship between the "syndrome of phlegm-heat accumulating in the lung" in the knowledge graph and these symptoms and it conforms to the corresponding dialectical rules, then the "syndrome of phlegm-heat accumulating in the lung" is used as the dialectical suspected syndrome type for this patient. Reasoning is performed on all patient record data to generate a list of dialectical suspected syndrome types corresponding to acute lung injury and stored in the database.

[0023] S4: A formula sorting and recommendation module, which is used to sort and recommend traditional Chinese medicine formulas in the traditional Chinese medicine knowledge graph for acute lung injury based on the dialectical suspected syndrome types corresponding to acute lung injury, so as to generate a sequence of traditional Chinese medicine formula suggestion plans corresponding to the syndrome types of acute lung injury.

[0024] In the embodiment of the present invention, when sorting and recommending traditional Chinese medicine formulas based on dialectical suspected syndrome types, the cosine similarity algorithm is used to calculate the similarity between the dialectical suspected syndrome types and each symptom in the traditional Chinese medicine knowledge graph for acute lung injury, and the algorithm is implemented by using the numpy library of Python. The dialectical suspected syndrome types and symptoms are respectively represented in vector form. For example, the typical symptoms associated with the syndrome of phlegm-heat accumulating in the lung are "fever", "coughing up thick yellow phlegm", etc., which are encoded as feature vectors, and the symptom of "coughing and wheezing" in the knowledge graph is also encoded as a vector. By calculating the cosine similarity between the vectors, the similarity degree between the two is obtained, and the calculated similarities are sorted by using the pandas library of Python. For each dialectical suspected syndrome type, the related symptoms are arranged in descending order of similarity to generate a sequence of patient symptom similarities. At the same time, based on the sequence of patient symptom similarities, the Cypher query language of Neo4j is used to search for the related traditional Chinese medicine formula nodes in the traditional Chinese medicine knowledge graph for acute lung injury to count the number of times each traditional Chinese medicine formula is associated. In the case of the same number of times, a secondary sorting is performed with reference to the "recommended priority" attribute of the traditional Chinese medicine formula nodes in the knowledge graph (this attribute is set according to factors such as the frequency of ancient medical records and the clinical application effect). According to the sorted order, the detailed information of the traditional Chinese medicine formulas is extracted, including the formula name, medicinal material composition, dosage ratio, etc., to generate a sequence of traditional Chinese medicine formula suggestion plans corresponding to the syndrome types of acute lung injury, which is stored in the "traditional Chinese medicine formula suggestion plan" table in the MySQL database to provide accurate medication recommendations for clinical treatment.

[0025] Furthermore, the traditional Chinese medicine information slicing storage module includes the following functions: Collect information records about acute lung injury in ancient traditional Chinese medicine classics, including symptom descriptions, formula compositions, and medication experiences, to obtain ancient traditional Chinese medicine classic information corresponding to acute lung injury; Collect and integrate clinical research information on the traditional Chinese medicine treatment of acute lung injury in domestic and foreign modern medical journals and research reports to obtain domestic and foreign clinical research information corresponding to acute lung injury; Extract the corresponding clinical diagnosis information, medication records, and clinical effects of acute lung injury patients from the electronic medical record systems of major hospitals to obtain the record data of acute lung injury patients; Remove duplicate, incorrect, and invalid data from the information in ancient Chinese medical classics, modern clinical research information at home and abroad, and the recorded data of patients with acute lung injury to obtain the cleaned Chinese medical classic information, domestic and foreign clinical research information, and acute lung injury recorded data; Slice and encode the stored cleaned Chinese medical classic information, domestic and foreign clinical research information, and acute lung injury recorded data to generate a Chinese medicine information database for acute lung injury.

[0026] As an embodiment of the present invention, refer to Figure 2 As shown in Figure 1 The functional flowchart of the Chinese medicine information slicing storage module in S11: By collecting the information records in ancient Chinese medical classics corresponding to acute lung injury, including symptom descriptions, prescription compositions, and medication experiences, to obtain the ancient Chinese medical classic information corresponding to acute lung injury; In the embodiment of the present invention, when collecting the information records in ancient Chinese medical classics corresponding to acute lung injury, a professional ancient book research team is formed, whose members include experts in Chinese medical literature and Chinese medicine clinicians. By visiting professional ancient book collection institutions, more than 200 Chinese medical ancient books such as "Huangdi Neijing", "Treatise on Febrile and Miscellaneous Diseases", "Qianjin Fang", and "Jingyue Quanshu" in the collection are page-by-page browsed. Then, using a high-precision scanner, the pages related to acute lung injury are scanned at a resolution of 600 dpi to generate high-definition PDF format files. Using the intelligent annotation method, information such as symptom descriptions, prescription compositions, and medication experiences are extracted according to the content of the ancient books. For example, the symptom description and prescription information of "For those with panting and coughing upwards and a floating pulse, treat with Houpo Mahuang Decoction" are extracted from "Treatise on Febrile and Miscellaneous Diseases" and sorted into a structured data table, including fields such as the name of the ancient book, era, chapter, original text content, symptom keywords (such as panting and coughing, upwards), prescription name, medicinal material composition (Houpo, Mahuang, etc.), and dosage of medication. Finally, the ancient Chinese medical classic information corresponding to acute lung injury is obtained.

[0027] S12: By collecting and integrating the clinical research information on the treatment of acute lung injury with Chinese medicine in modern medical journals and research reports at home and abroad, to obtain the domestic and foreign clinical research information corresponding to acute lung injury; In the embodiments of the present invention, when collecting and integrating clinical research information on the treatment of acute lung injury with traditional Chinese medicine in domestic and foreign modern medical journals and research reports, a traditional Chinese medicine professional database is utilized. With "acute lung injury" and "traditional Chinese medicine" as search terms, the time range is set to the past 20 years, and more than 3,000 English literature are retrieved. With "acute lung injury" and "treatment with traditional Chinese medicine" as search terms, more than 5,000 Chinese literature are retrieved. Then, the retrieved literature is processed to remove duplicates through the literature management software EndNote X9. By comparing information such as the literature title, author, and publication year, 1,200 duplicate literature are removed. Medical professionals are organized to screen the remaining literature. According to the inclusion criteria (the research object is patients with acute lung injury, involving traditional Chinese medicine treatment intervention, and including clinical efficacy evaluation indicators), 2,500 literature that do not meet the requirements are excluded. For the 4,300 screened literature, key information such as research purpose, method, result, and conclusion is extracted by means of natural language processing technology abstract extraction, and organized into structured data, including literature title, author, publication journal, publication year, research type, intervention measures (composition, dosage, administration method of traditional Chinese medicine formula), main efficacy indicators (such as changes in oxygenation index, inflammatory factor levels), etc. Finally, domestic and foreign clinical research information corresponding to acute lung injury is obtained.

[0028] S13: Extract the corresponding clinical diagnosis information, medication records, and clinical effects of patients with acute lung injury from the electronic medical record systems of major hospitals to obtain the recorded data of patients with acute lung injury; In an embodiment of the present invention, when extracting the corresponding clinical diagnosis information, medication records, and clinical effects of patients with acute lung injury from the electronic medical record systems of major hospitals, a data extraction script is written through the data interface of the hospital information system (HIS) using the SQL Server database management system. For example, a SELECT statement "SELECT patient ID, admission date, discharge date, diagnosis name, main symptoms, blood routine examination results, chest CT imaging diagnosis, traditional Chinese medicine prescription name, medicinal material dosage, medication frequency, post-treatment oxygenation index, post-treatment inflammatory factor level FROM the electronic medical record table WHERE diagnosis name LIKE '%acute lung injury%'" is written to extract relevant data from the electronic medical record system. The extracted data is exported in a CSV format file, with each file containing 1000 patient records. The exported data is subjected to a preliminary format conversion, using the text-to-columns function in Excel to split the traditional Chinese medicine prescription name according to the medicinal material name and dosage. For example, "Mahuang Decoction (Ephedra 6g, Cinnamon Twig 4g, Apricot Kernel 9g, Licorice Root 3g)" is split into fields such as "Mahuang Decoction", "Ephedra", "6g", "Cinnamon Twig", "4g", "Apricot Kernel", "9g", "Licorice Root", "3g", etc., and finally, the patient record data of patients with acute lung injury is obtained.

[0029] S14: Remove duplicate, incorrect, and invalid data from the ancient traditional Chinese medicine classic information, domestic and foreign modern clinical research information, and the patient record data of patients with acute lung injury to obtain the cleaned traditional Chinese medicine classic information, domestic and foreign clinical research information, and acute lung injury record data; In the embodiments of the present invention, when cleaning the information of ancient Chinese medical classics, modern clinical research information at home and abroad, and the recorded data of patients with acute lung injury, for the information of ancient Chinese medical classics, regular expressions are used to uniformly standardize the names of medicinal materials. For example, "Astragalus membranaceus var. mongholicus" is replaced with "Astragalus membranaceus", and "Magnolia officinalis var. biloba" is replaced with "Magnolia officinalis". By means of computer verification, check whether there are logical errors in the dosages of medicinal materials in the composition of prescriptions, such as the situation of excessive or too small dosages. For the clinical research information at home and abroad, the SPSS statistical software is used for detecting data outliers. The efficacy index data beyond the normal range is identified through box plots and compared with the original literature to correct the wrong data. The pandas library of Python is used for detecting duplicate literatures. By comparing information such as the literature title, author, and research content, duplicate researches are removed. For the recorded data of patients with acute lung injury, the data validation function of Excel is used to check whether the date format is correct, such as the wrong data where the admission date is earlier than the discharge date. The records lacking key information are screened out through the SQL statement "SELECT * FROM the patient record table WHERE the traditional Chinese medicine prescription name IS NULL OR the post-treatment oxygenation index IS NULL", and the hospital is contacted to supplement them. After the above operations, the information of Chinese medical classics, modern clinical research information at home and abroad, and the recorded data of acute lung injury after cleaning are finally obtained.

[0030] S15: Slice and encode and store the information of Chinese medical classics, modern clinical research information at home and abroad, and the recorded data of acute lung injury after cleaning to generate a Chinese medicine information database for acute lung injury.

[0031] In the embodiments of the present invention, when slicing and encoding the stored traditional Chinese medicine (TCM) pharmacopoeia information, domestic and foreign clinical research information, and acute lung injury record data after cleaning, a distributed database management system MongoDB is used. The data is divided into an ancient TCM pharmacopoeia information set, a domestic and foreign clinical research information set, and an acute lung injury patient record information set according to the information type. For the corresponding ancient TCM pharmacopoeia information set, domestic and foreign clinical research information set, and acute lung injury patient record information set, slicing is performed according to the analyzed corresponding symptom attribute type clusters. Using the pandas library of Python and data screening techniques, taking the "respiratory dysfunction cluster" as an example, in the TCM pharmacopoeia information, the classical paragraphs containing symptom keywords such as "dyspnea", "difficulty in breathing", and "shortness of breath" are screened out, and these paragraphs and their related prescriptions and medication information are extracted to form TCM information fragments. In the domestic and foreign clinical research information, the literature involving the assessment and treatment of respiratory dysfunction symptoms is retrieved, and relevant research methods, experimental data, conclusions, etc. are extracted to generate domestic and foreign research information fragments. For the acute lung injury patient record data, the patient records with the main symptom fields containing respiratory dysfunction-related symptoms are screened out, and their diagnosis information, medication records, treatment effects, etc. are extracted to obtain patient record data fragments. At the same time, by performing the above operations on each symptom attribute type cluster, each fragment corresponds to a specific symptom type cluster. At the same time, by using the distributed file system HDFS and Lucene indexing technology, each symptom type cluster is used as an independent storage unit, and a corresponding folder is created in HDFS. For example, under the "respiratory dysfunction cluster" folder, the corresponding TCM information fragments, domestic and foreign research information fragments, and patient record data fragments are stored respectively, and the file format is JSON to facilitate data storage and reading. And by using Lucene to index and encode each fragment, an inverted index is established for the key information in each fragment (such as symptom names, prescription names, research conclusion keywords, patient IDs, etc.). For example, for the TCM information fragment containing "Mahuang Decoction", Lucene will use "Mahuang Decoction" as the index term and record the storage location of this fragment in HDFS. The index file and the data fragment are stored together in HDFS. By establishing an index mapping table, the correspondence between each symptom type cluster and the index file is recorded, and finally, an acute lung injury TCM information database is stored and generated.

[0032] Further, the slicing and encoding storage of the cleaned traditional Chinese medicine pharmacopoeia information, domestic and foreign clinical research information, and acute lung injury record data includes: Extracting symptom attributes of the corresponding acute lung injury conditions in the cleaned traditional Chinese medicine pharmacopoeia information, domestic and foreign clinical research information, and acute lung injury record data to obtain the acute lung injury symptom attribute elements corresponding to the pharmacopoeia, research, and patient records; In the embodiments of the present invention, when extracting symptom attributes of the corresponding acute lung injury conditions from the information of traditional Chinese medicine classics, domestic and foreign clinical research information, and acute lung injury record data after cleaning, a combination of natural language processing (NLP) technology and manual verification is adopted. For the information of traditional Chinese medicine classics, the NLTK (Natural Language Toolkit) and spaCy libraries of Python are used to perform word segmentation, part-of-speech tagging, and named entity recognition on the classic texts. For example, from the record in "Qianjin Fang" that "lung injury with dyspnea, hemoptysis, shortness of breath, and weakness of qi", symptom attributes such as "dyspnea", "hemoptysis", "shortness of breath", and "weakness of qi" are identified, and they are classified into symptom categories such as respiration, bleeding, and qi movement through matching and verification with a pre-constructed traditional Chinese medicine symptom dictionary. For domestic and foreign clinical research information, regular expressions and semantic analysis techniques are used to extract symptom attributes from research results and patient descriptions. For example, from the text "the patient has persistent cough accompanied by a decrease in blood oxygen saturation", attributes such as "persistent cough" and "decrease in blood oxygen saturation" are extracted. In the acute lung injury patient record data, it is directly extracted through structured data fields. For example, information such as "fever" and "dyspnea" is obtained from the "main symptoms" field. By organizing all the extracted symptom attributes into a structured table, including fields such as information source (pharmacopoeia, research, patient record), specific symptoms, and symptom categories, the acute lung injury symptom attribute elements corresponding to the pharmacopoeia, research, and patient record are obtained.

[0033] Preferably, the acute lung injury symptom attribute elements corresponding to the pharmacopoeia, research, and patient record are classified into symptom type clusters to generate different acute lung injury symptom attribute type cluster groups; In the embodiments of the present invention, when classifying the symptom type clusters for the acute lung injury symptom attributes corresponding to the pharmacopoeia, research, and patient records, the hierarchical clustering algorithm (HAC) is used. First, a symptom attribute similarity matrix is constructed by calculating the similarity degree between different symptom attributes using cosine similarity. For example, "dyspnea" and "difficult breathing" have a high similarity in semantics and medical manifestations, and their cosine similarity can reach 0.85. The distance threshold is set to 0.8. For example, taking a dataset containing 5 acute lung injury symptom attributes such as "difficult breathing", "dyspnea", "fever", "elevated white blood cell count", and "decreased blood oxygen saturation" as an example, first, the spaCy library of Python is used to vectorize each symptom attribute and convert it into a word vector with a fixed dimension (assumed to be 300 dimensions). For example, the word vector of "difficult breathing" is [0.12, 0.34, 0.21,..., 0.08], and the word vector of "dyspnea" is [0.10, 0.32, 0.23,..., 0.10]. Then, the cosine_similarity function in the scikit-learn library of Python is used to calculate the similarity between any two symptom attribute word vectors. Taking the calculation of the similarity between "difficult breathing" and "dyspnea" as an example, their word vectors are passed into the function as parameters: from sklearn.metrics.pairwise import cosine_similarity symptom1_vector = [0.12, 0.34, 0.21,..., 0.08] symptom2_vector = [0.10, 0.32, 0.23,..., 0.10] similarity = cosine_similarity([symptom1_vector], [symptom2_vector])[0][0] After calculation, the cosine similarity between the two is 0.85. Similarly, the similarities between other symptom attributes are calculated. For example, the similarity between "difficult breathing" and "fever" is 0.12, and the similarity between "dyspnea" and "elevated white blood cell count" is 0.08, etc. Finally, a 5×5 semantic similarity matrix is constructed. The rows and columns of the matrix respectively correspond to these 5 symptom attributes. The element M[i][j] in the matrix represents the similarity between the i-th symptom attribute and the j-th symptom attribute. The specific matrix is as follows Figure 4As shown, this matrix comprehensively demonstrates the semantic similarity degree among these symptom attribute elements. The closer the value is to 1, the stronger the semantic association between two symptom attributes; the closer the value is to 0, the weaker the semantic association. Subsequently, based on this matrix, a distance threshold of 0.8 is set. When the similarity between two symptom attribute elements is greater than this threshold, they can be merged into one cluster. That is, "dyspnea" and "difficulty in breathing" can be combined into one cluster. Through iterative merging, all symptom attribute elements are gradually clustered. For example, respiratory-related symptom attributes such as "dyspnea", "difficulty in breathing", and "shortness of breath" are clustered into the "respiratory dysfunction" cluster; "fever", "shivering", etc. are classified into the "systemic symptoms" cluster. During the clustering process, medical experts evaluate and adjust the clustering results to ensure that the classification conforms to medical logic. Finally, different clusters of acute lung injury symptom attribute types are generated. Each cluster contains pharmacopoeias, research, and patient records corresponding to multiple related symptom attributes, and a clear medical category name is given, such as "respiratory dysfunction cluster", "inflammatory response cluster", "circulatory system symptom cluster", etc.

[0034] Preferably, based on different clusters of acute lung injury symptom attribute types, slice division of traditional Chinese medicine pharmacopoeia information, domestic and foreign clinical research information, and acute lung injury record data corresponding to the corresponding type clusters is performed to generate traditional Chinese medicine information segments, domestic and foreign research information segments, and patient record data segments under different symptom type clusters. In the embodiment of the present invention, when performing slice division of traditional Chinese medicine pharmacopoeia information, domestic and foreign clinical research information, and acute lung injury record data corresponding to the corresponding type clusters based on different clusters of acute lung injury symptom attribute types, the pandas library of Python and data screening techniques are used. Taking the "respiratory dysfunction cluster" as an example, in the traditional Chinese medicine pharmacopoeia information, the classical paragraphs containing symptom keywords such as "dyspnea", "difficulty in breathing", and "shortness of breath" are screened out, and these paragraphs and their related prescriptions and medication information are extracted to form traditional Chinese medicine information segments. In the domestic and foreign clinical research information, the literature whose research content involves the assessment and treatment of respiratory dysfunction symptoms is retrieved, and relevant research methods, experimental data, conclusions, etc. are extracted to generate domestic and foreign research information segments. For the acute lung injury patient record data, the patient records whose main symptom fields contain respiratory dysfunction-related symptoms are screened out, and their diagnostic information, medication records, treatment effects, etc. are extracted to obtain patient record data segments. At the same time, by performing the above operations on each symptom attribute type cluster, traditional Chinese medicine information segments, domestic and foreign research information segments, and patient record data segments under different symptom type clusters are finally generated. Each segment corresponds to a specific symptom type cluster, facilitating subsequent retrieval and analysis.

[0035] Preferably, the traditional Chinese medicine information fragments, domestic and foreign research information fragments, and patient record data fragments under different symptom type clusters are indexed and encoded for storage to generate an acute lung injury traditional Chinese medicine information database.

[0036] In the embodiment of the present invention, when indexing and encoding the storage of traditional Chinese medicine information fragments, domestic and foreign research information fragments, and patient record data fragments under different symptom type clusters, the distributed file system HDFS and Lucene indexing technology are adopted. Each symptom type cluster is used as an independent storage unit, and a corresponding folder is created in HDFS. For example, under the "respiratory dysfunction cluster" folder, the corresponding traditional Chinese medicine information fragments, domestic and foreign research information fragments, and patient record data fragments are stored respectively, and the file format is JSON to facilitate data storage and reading. And each fragment is indexed and encoded by using Lucene to establish an inverted index for the key information in each fragment (such as symptom name, prescription name, research conclusion keywords, patient ID, etc.). For example, for the traditional Chinese medicine information fragment containing "Mahuang Decoction", Lucene will take "Mahuang Decoction" as the index term and record the storage location of this fragment in HDFS. The index file and the data fragment are stored together in HDFS. By establishing an index mapping table, the corresponding relationship between each symptom type cluster and the index file is recorded. Finally, all storage units are integrated to generate an acute lung injury traditional Chinese medicine information database, realizing efficient data retrieval and query based on symptom types. For example, by inputting "dyspnea", all the traditional Chinese medicine classic information, clinical research information, and patient record data related to this symptom can be quickly retrieved.

[0037] Furthermore, the classification of the acute lung injury symptom attribute elements corresponding to the pharmacopoeia, research, and patient records into symptom type clusters includes: Performing word vector embedding on the acute lung injury symptom attribute elements corresponding to the pharmacopoeia, research, and patient records to generate acute lung injury symptom word embedding vectors corresponding to the pharmacopoeia, research, and patient records; In an embodiment of the present invention, when performing word vector embedding on the acute lung injury symptom attribute elements corresponding to the pharmacopoeia, research, and patient records, a pre-trained BERT (Bidirectional Encoder Representations from Transformers) model is used. Specifically, the Chinese BERT-base version is used. The symptom attribute elements extracted from the traditional Chinese medicine pharmacopoeia information, domestic and foreign clinical research information, and acute lung injury patient record data, such as "rapid breathing", "hemoptysis", "decrease in blood oxygen saturation", etc., are sorted into text sequences in a fixed format. When the length of each text sequence is less than 128 tokens, special padding symbols are used for completion; when it exceeds 128 tokens, truncation processing is performed. The BERT model and tokenizer are loaded through the Hugging Face Transformers library of Python, and the text sequence is input into the model to obtain a 768-dimensional word embedding vector corresponding to each symptom attribute element. For example, for the symptom attribute element "rapid breathing", after being processed by the BERT model, a vector representation containing 768 floating-point numbers is obtained, and this vector contains the characteristic information of "rapid breathing" in the semantic space. The word embedding vectors of all symptom attribute elements are stored in a NumPy array, and finally, an acute lung injury symptom word embedding vector dataset corresponding to the pharmacopoeia, research, and patient records is formed.

[0038] Preferably, semantic feature analysis is performed on the acute lung injury symptom word embedding vectors corresponding to the pharmacopoeia, research, and patient records to obtain the acute lung injury symptom semantic features corresponding to the pharmacopoeia, research, and patient records; In an embodiment of the present invention, when performing semantic feature analysis on the acute lung injury symptom word embedding vectors corresponding to the pharmacopoeia, research, and patient records, the principal component analysis (PCA) algorithm is used to reduce the dimension of the 768-dimensional word embedding vectors to 50 dimensions to reduce data redundancy and highlight the main semantic features. The PCA algorithm is implemented through the Scikit-learn library of Python, and the number of principal components is set to 50. The word embedding vectors of each symptom attribute element are transformed. After dimensionality reduction, the values of each dimension are analyzed to extract representative semantic features. For example, in the vector after dimensionality reduction, the magnitude of a certain dimension value reflects the semantic feature related to the breathing frequency of the symptom, and another dimension is related to the severity of the symptom. At the same time, the cosine similarity between the word embedding vectors of different symptom attribute elements is calculated to construct a semantic similarity matrix to further analyze the semantic association between symptoms. For example, the cosine similarity of the word embedding vectors of "dyspnea" and "rapid breathing" is relatively high, indicating that they are closely related semantically. Finally, the information such as the vector values after dimensionality reduction and the semantic similarity matrix is integrated to finally obtain the acute lung injury symptom semantic features corresponding to the pharmacopoeia, research, and patient records.

[0039] Preferably, the symptom scores corresponding to the acute lung injury symptoms in the pharmacopoeia, research, and patient records are evaluated based on the semantic features of the acute lung injury symptom words, and the symptom scores corresponding to the acute lung injury symptoms in the pharmacopoeia, research, and patient records are obtained; In the embodiment of the present invention, when evaluating the symptom scores according to the semantic features of the acute lung injury symptom words corresponding to the pharmacopoeia, research, and patient records, first, an evaluation group composed of medical experts and data analysts assigns weights to each semantic feature according to the medical knowledge and clinical experience of acute lung injury. For example, the weight of the semantic feature directly related to respiratory function is set to 0.3, the weight of the semantic feature reflecting the severity of the symptom is set to 0.4, and the weight of other relevant semantic features is set to 0.3. For each symptom attribute element, the value of its word semantic feature is multiplied by the corresponding weight and then summed to obtain the symptom score of the symptom attribute element. Taking "decrease in blood oxygen saturation" as an example, the values of its word semantic features in each dimension are [0.2, 0.3, 0.1,...], and the corresponding weights are [0.3, 0.4, 0.2,...]. The symptom score is calculated by weighted summation as 0.2×0.3 + 0.3×0.4 + 0.1×0.2 +... = 0.22 +..., and this operation is performed on all acute lung injury symptom attribute elements corresponding to the pharmacopoeia, research, and patient records. Finally, a complete list of acute lung injury symptom scores is obtained, and each score corresponds to a specific symptom attribute element.

[0040] Preferably, based on the acute lung injury symptom scores corresponding to the pharmacopoeia, research, and patient records, the corresponding acute lung injury symptom attribute elements are classified into symptom type clusters to generate different acute lung injury symptom attribute type clusters.

[0041] In an embodiment of the present invention, when classifying symptom type clusters for the corresponding acute lung injury symptom attribute elements based on the acute lung injury symptom scores corresponding to the pharmacopoeia, research, and patient records, the K-Means clustering algorithm is used to determine the optimal number of clusters K through the Elbow Method. The specific operation is to run the K-Means algorithm respectively within the range of K from 2 to 10, calculate the sum of squared errors within the clusters (SSE) for each K value, draw the relationship curve between K and SSE, and select the K value corresponding to the inflection point of the curve as the optimal number of clusters. Assume that K = 5 is determined through the Elbow Method. To implement the K-Means algorithm using the Scikit-learn library in Python, the symptom attribute elements are clustered with their symptom scores as feature vectors. For example, the symptom score vectors of symptom attribute elements such as "dyspnea", "difficulty in breathing", and "shortness of breath" are input into the K-Means algorithm, and the algorithm divides these symptom attribute elements into different clusters according to the similarity of the scores. After clustering, medical experts evaluate and name each cluster. For example, the cluster containing symptoms related to respiratory rate is named "abnormal respiratory rate cluster", and the cluster where symptoms reflecting a higher severity level are located is named "severe symptom-related cluster", etc. Finally, different acute lung injury symptom attribute type clusters are generated, providing structured data support for subsequent traditional Chinese medicine treatment research and clinical applications.

[0042] Furthermore, the traditional Chinese medicine knowledge graph construction module includes the following functions: Identify and extract knowledge entity according to the ancient traditional Chinese medicine pharmacopoeia information and domestic and foreign clinical research information corresponding to different types of acute lung injury in the acute lung injury traditional Chinese medicine information database, so as to generate the traditional Chinese medicine knowledge structure entity corresponding to acute lung injury; Measure the knowledge association between the traditional Chinese medicine knowledge structure entities corresponding to each acute lung injury to obtain the knowledge association coefficient between each traditional Chinese medicine knowledge entity; Connect the traditional Chinese medicine knowledge structure entities corresponding to each acute lung injury based on the knowledge association coefficient between each traditional Chinese medicine knowledge entity to generate the acute lung injury action relationship chain between each traditional Chinese medicine knowledge entity; Extract the knowledge chain relationship of the acute lung injury action relationship chain between each traditional Chinese medicine knowledge entity based on the traditional Chinese medicine knowledge structure entities corresponding to each acute lung injury to generate the traditional Chinese medicine knowledge chain relationship corresponding to acute lung injury; Construct a knowledge graph according to the traditional Chinese medicine knowledge structure entity corresponding to acute lung injury and the traditional Chinese medicine knowledge chain relationship, using each traditional Chinese medicine knowledge structure entity as a node and the corresponding traditional Chinese medicine knowledge chain relationship as a connecting edge to generate the acute lung injury traditional Chinese medicine knowledge graph.

[0043] As an embodiment of the present invention, refer to Figure 2As shown, it is Figure 1 a schematic diagram of the functional process of the traditional Chinese medicine knowledge graph construction module. In this embodiment, the traditional Chinese medicine knowledge graph construction module includes the following functions: S21: Identify and extract knowledge entity according to the ancient traditional Chinese medicine classic information and domestic and foreign clinical research information corresponding to different types of acute lung injury in the acute lung injury traditional Chinese medicine information database, so as to generate a traditional Chinese medicine knowledge structure entity corresponding to acute lung injury; In the embodiment of the present invention, when identifying and extracting knowledge entity according to the ancient traditional Chinese medicine classic information and domestic and foreign clinical research information corresponding to different types of acute lung injury in the acute lung injury traditional Chinese medicine information database, a method combining natural language processing (NLP) technology and manual verification is adopted. For the ancient traditional Chinese medicine classic information, using the spaCy library of Python to load the pre-trained Chinese medical model, tokenize the classic text, perform part-of-speech tagging and named entity recognition, and entities such as traditional Chinese medicine names (such as ephedra, cassia twig), symptom descriptions (such as dyspnea, coughing and expectorating purulent blood), prescription names (such as mahuang decoction, shegan mahuang decoction) can be identified. For domestic and foreign clinical research information, use regular expressions and rule matching methods to extract entities such as traditional Chinese medicine ingredients (such as baicalin, berberine), efficacy indicators (such as oxygenation index, white blood cell count), and experimental methods from the research text. For example, from the sentence "Mahuang decoction can improve the oxygenation index of patients with acute lung injury", extract the three entities of "Mahuang decoction", "acute lung injury", and "oxygenation index", process the extracted entities in a unified format, add entity type tags (such as "traditional Chinese medicine", "symptom", "prescription"), and manually review and correct the misidentified entities. Finally, generate a structured traditional Chinese medicine knowledge structure entity corresponding to acute lung injury and store it in the "knowledge_entities" table of the MySQL database.

[0044] S22: Measure the knowledge association between the traditional Chinese medicine knowledge structure entities corresponding to each acute lung injury to obtain the knowledge association coefficient between each traditional Chinese medicine knowledge entity; In the embodiments of the present invention, when measuring the knowledge association among the traditional Chinese medicine knowledge structure entities corresponding to each acute lung injury, a calculation method based on co-occurrence frequency and semantic similarity is adopted. First, all text records in the "knowledge_entities" table are traversed through the pandas library of Python to count the number of times any two entities appear in the same paragraph or sentence. For example, among 1000 records, "Ephedra" and "Ramulus Cinnamomi" co-occur 300 times, and "Ephedra" and "oxygenation index" co-occur 100 times. Then, the pre-trained BERT model is used to calculate the semantic similarity between entities. The entity text is input into the model to obtain word vectors, and the similarity between vectors is calculated through the cosine similarity formula. For example, the semantic similarity between "Ephedra" and "Ramulus Cinnamomi" is 0.8, and the semantic similarity between "Ephedra" and "Coptis chinensis" is 0.3. Finally, the co-occurrence frequency and semantic similarity are weighted and fused. The co-occurrence frequency weight is set to 0.6, and the semantic similarity weight is set to 0.4. The knowledge association effect coefficient is calculated through the formula: knowledge association effect coefficient = co-occurrence frequency × 0.6 + semantic similarity × 0.4. The knowledge association effect coefficient between "Ephedra" and "Ramulus Cinnamomi" is calculated as (300 / 1000) × 0.6 + 0.8 × 0.4 = 0.5. Calculations are performed among all traditional Chinese medicine knowledge entities, and finally, the knowledge association effect coefficients among each traditional Chinese medicine knowledge entity are obtained.

[0045] S23: Based on the knowledge association effect coefficients among each traditional Chinese medicine knowledge entity, perform association connection among the traditional Chinese medicine knowledge structure entities corresponding to each acute lung injury to generate an acute lung injury effect relationship chain among each traditional Chinese medicine knowledge entity; In the embodiments of the present invention, when performing association connection among the traditional Chinese medicine knowledge structure entities based on the knowledge association effect coefficients among each traditional Chinese medicine knowledge entity, the association threshold is set to 0.3, and the corresponding association effect coefficients are read through the pandas library of Python. Entity pairs with coefficients greater than 0.3 are screened out through boolean indexing. For example, entity pairs such as "Ephedra - Ramulus Cinnamomi", "Belamcanda chinensis - Ephedra", and "Mahuang Decoction - oxygenation index" are screened out. For each pair of qualified entities, nodes and relationships are created in the Neo4j graph database. Taking "Ephedra - Ramulus Cinnamomi" as an example, two nodes, namely "Ephedra" and "Ramulus Cinnamomi", are created respectively. The node label is "traditional Chinese medicine", and the detailed information of the entity is stored in the node. Then, a relationship of type "association" is created to connect the two nodes, and the knowledge association effect coefficient is stored as an attribute of the relationship. For example, the association coefficient = 0.5. All the screened entity pairs are traversed, and the above operations are repeated. Finally, an acute lung injury effect relationship chain among each traditional Chinese medicine knowledge entity is generated in the Neo4j database, forming a preliminary knowledge network structure.

[0046] S24: Extract the knowledge chain relationship from the acute lung injury-related Chinese medicine knowledge entity pairs for the acute lung injury action relationship chain among various Chinese medicine knowledge entities, so as to generate the Chinese medicine knowledge chain relationship corresponding to acute lung injury; In the embodiment of the present invention, when extracting the knowledge chain relationship from the acute lung injury action relationship chain based on the Chinese medicine knowledge structure entities corresponding to each acute lung injury, the relationship chain is screened in combination with the clinical application scenario and medical logic, and the Cypher query language of Neo4j is used to write a query statement to screen out the relationship chain with practical clinical significance. For example, query the relationship chain in which the "Chinese medicine" nodes are connected by the "association" relationship, and the correlation coefficient is greater than 0.4, and at the same time involves the treatment of symptoms related to acute lung injury: MATCH (a: Chinese medicine)-[r: association {coefficient: >0.4}]-(b: Chinese medicine) WHERE (a)-[: used for treating]->(: symptom {name:'symptoms related to acute lung injury'}) OR (b)-[: used for treating]->(: symptom {name:'symptoms related to acute lung injury'}) RETURN a, r, b Extract the relationship chains that meet the conditions, supplement relevant information such as literature sources and experimental data to form a complete knowledge chain relationship. For example, for the "Ephedra-Cassia Twig" relationship chain, supplement the original text recorded in "Treatise on Febrile and Miscellaneous Diseases" and the treatment effect data of the combination of the two on acute lung injury in modern clinical research, etc., and store all the screened knowledge chain relationships in a new database table, including fields such as entity 1, relationship type, correlation coefficient, entity 2, literature source, and clinical evidence, and finally generate the Chinese medicine knowledge chain relationship corresponding to acute lung injury.

[0047] S25: Construct a knowledge graph according to the Chinese medicine knowledge structure entities corresponding to acute lung injury and the Chinese medicine knowledge chain relationship, using each Chinese medicine knowledge structure entity as a node and the corresponding Chinese medicine knowledge chain relationship as a connecting edge to generate an acute lung injury Chinese medicine knowledge graph.

[0048] In the embodiments of the present invention, when constructing a knowledge graph based on the traditional Chinese medicine knowledge structure entities corresponding to acute lung injury and the traditional Chinese medicine knowledge chain relationship, the Neo4j graph database is used as the basis. The corresponding traditional Chinese medicine knowledge structure entities are imported into Neo4j as nodes, and different node labels are assigned according to the entity types (traditional Chinese medicine, symptoms, prescriptions, etc.), and the detailed attributes of the entities (such as name, efficacy, ingredients) are stored as node attributes. To read the corresponding knowledge chain relationship, corresponding relationships are created in Neo4j to connect the nodes, and the relationship types and attributes are consistent with the knowledge chain relationship. For example, for the knowledge chain relationship of "Ephedra - associated (coefficient = 0.5) - Cinnamon Twig", nodes and relationships are accurately created in Neo4j, and by using the visualization interface of Neo4j, the layout of the knowledge graph is adjusted and the style is optimized to display the distribution of nodes and relationships. At the same time, a simple query interface is developed through the py2neo library of Python, supporting users to input keywords such as traditional Chinese medicine names and symptoms to query relevant sub-graphs of the knowledge graph, and finally a complete traditional Chinese medicine knowledge graph of acute lung injury is generated, providing an intuitive knowledge display and query tool for traditional Chinese medicine clinical research, prescription analysis and teaching.

[0049] Further, the extraction of the knowledge chain relationship of the acute lung injury action relationship chain between each traditional Chinese medicine knowledge entity based on each traditional Chinese medicine knowledge structure entity corresponding to acute lung injury includes: Performing a pharmacodynamic synergy evaluation on the traditional Chinese medicine knowledge structure entities corresponding to each acute lung injury to obtain the acute lung injury pharmacodynamic synergy coefficient between each traditional Chinese medicine knowledge entity; In the embodiments of the present invention, when evaluating the efficacy synergy among the traditional Chinese medicine knowledge structure entities corresponding to various acute lung injuries, the prescription composition and clinical efficacy data are extracted from the traditional Chinese medicine information database for acute lung injuries, and the data is sorted using the pandas library of Python. Taking "Mahuang Decoction" and "Shegan Mahuang Decoction" as examples, the efficacy index data such as the improvement rate of oxygenation index and the degree of reduction of inflammatory factors for treating patients with acute lung injuries using these two prescriptions are obtained from the database, and the grey relational analysis method is adopted to calculate the data sequence correlation degrees of different traditional Chinese medicine knowledge entities (such as Ephedra and Cinnamon Twig, Shegan and Ephedra) under the same efficacy index. The higher the correlation degree, the stronger the efficacy synergy. The specific operation is to compare and calculate the efficacy data sequences when traditional Chinese medicines are used in combination with the efficacy data sequences when individual traditional Chinese medicines are used through the greyatom library of Python. If the data sequence of the improvement rate of oxygenation index for patients when Ephedra is used alone is [0.1, 0.15, 0.2], and the corresponding sequence when Ephedra and Cinnamon Twig are used in combination is [0.2, 0.25, 0.3], and the grey relational degree calculated between the two is 0.8, this value is used as the acute lung injury efficacy synergy coefficient between the two. This operation is repeated for all combinations of traditional Chinese medicine knowledge entities in the database, and finally the acute lung injury efficacy synergy coefficients among all traditional Chinese medicine knowledge entities are obtained.

[0050] Preferably, the traditional Chinese medicine pharmacological action components corresponding to each traditional Chinese medicine knowledge entity are obtained through the traditional Chinese medicine knowledge structure entities corresponding to various acute lung injuries, and the pharmacological conflict probability analysis is carried out according to the traditional Chinese medicine pharmacological action components corresponding to each traditional Chinese medicine knowledge entity to obtain the pharmacological action conflict probability among each traditional Chinese medicine knowledge entity; In the embodiments of the present invention, when obtaining the pharmacological active components of traditional Chinese medicines through the traditional Chinese medicine knowledge structure entities corresponding to various acute lung injuries and conducting a pharmacological conflict probability analysis, relying on the traditional Chinese medicine ingredient databases TCMSP and TCMID, using the requests library and the BeautifulSoup library of Python, with the traditional Chinese medicine name as the search term, the main active components of traditional Chinese medicines such as Coptis chinensis and Scutellaria baicalensis are extracted, such as berberine, baicalin, etc. The obtained ingredient data is matched with the chemical property data in the PubChem database to obtain information such as the chemical structure and target protein of the ingredients. A Bayesian network model is used to conduct a pharmacological conflict probability analysis. The network structure is constructed through the pgmpy library of Python. Taking berberine of Coptis chinensis and glycyrrhizic acid of Glycyrrhiza uralensis as examples, it is known that the two compete for the same metabolic enzyme CYP3A4. This information is used as the conditional probability of the network nodes and edges. By analyzing a large number of reported drug interaction cases in the literature, the probability relationship between the nodes is determined. If in 100 existing relevant studies, it is found that there are 20 cases of metabolic abnormalities when Coptis chinensis and Glycyrrhiza uralensis are used simultaneously, then the pharmacological action conflict probability between the two is 0.2. The ingredient combinations between all traditional Chinese medicine knowledge entities are calculated, and finally the pharmacological action conflict probabilities between each traditional Chinese medicine knowledge entity are obtained.

[0051] Preferably, based on the acute lung injury pharmacodynamic synergistic coefficient and the pharmacological action conflict probability between each traditional Chinese medicine knowledge entity, the compatibility combination degree between the traditional Chinese medicine knowledge structure entities corresponding to each acute lung injury is quantified to obtain the pharmacological action compatibility combination degree between each traditional Chinese medicine knowledge entity; In the embodiments of the present invention, when quantifying the compatibility combination degree between traditional Chinese medicine knowledge structure entities based on the pharmacodynamic synergistic coefficient and the pharmacological conflict probability, a weighted comprehensive evaluation model is constructed. The weight of the pharmacodynamic synergistic coefficient in this model is set to 0.7, and the weight of the pharmacological action conflict probability is set to 0.3. The compatibility combination degree is calculated through the following formula: Compatibility combination degree = Pharmacodynamic synergistic coefficient × 0.7 - Pharmacological action conflict probability × 0.3. Taking Ephedra sinica and Cinnamomum cassia as examples, it is known that their pharmacodynamic synergistic coefficient is 0.8 and the pharmacological action conflict probability is 0.1. Then the compatibility combination degree between the two is 0.8 × 0.7 - 0.1 × 0.3 = 0.53. And through the use of the numpy library of Python, the calculation process between all traditional Chinese medicine knowledge entities is batch processed, traversing the pharmacodynamic synergistic coefficient and the pharmacological action conflict probability between each entity, substituting the corresponding elements into the model formula for calculation, and finally obtaining the pharmacological action compatibility combination degree of each traditional Chinese medicine knowledge entity combination, forming a new matrix and storing it in the database to provide a quantitative basis for subsequent screening.

[0052] Preferably, based on the compatibility combination degree of pharmacological effects among various traditional Chinese medicine knowledge entities, the compatibility combination chain screening and extraction of the acute lung injury effect relationship chain among various traditional Chinese medicine knowledge entities is carried out, so as to extract the acute lung injury effect relationship chain corresponding to the traditional Chinese medicine knowledge entities with the pharmacological effect compatibility combination degree greater than the preset threshold as the corresponding knowledge chain relationship, and generate the traditional Chinese medicine knowledge chain relationship corresponding to acute lung injury.

[0053] In the embodiment of the present invention, when carrying out the compatibility combination chain screening and extraction of the acute lung injury effect relationship chain based on the pharmacological effect compatibility combination degree, the preset threshold is set to 0.5. By using the pandas library of Python to read the pharmacological effect compatibility combination degree matrix data, the combination of traditional Chinese medicine knowledge entities greater than the threshold is screened out through boolean indexing. For example, combinations such as Ephedra and Cinnamon Twig (combination degree 0.53), Belamcanda Root and Ephedra (combination degree 0.6) are found in the matrix. For each qualified combination, relevant symptom descriptions, formula compositions, clinical curative effects and other information are extracted from the acute lung injury traditional Chinese medicine information database to construct the knowledge chain relationship. Taking Ephedra and Cinnamon Twig as an example, the original text record in Treatise on Febrile and Miscellaneous Diseases, the symptom improvement situation during the treatment of acute lung injury, the specific formula composition and other information are extracted to form a knowledge chain record including fields such as traditional Chinese medicine name, compatibility relationship, pharmacodynamic description, literature source, etc. All the knowledge chain relationships that meet the conditions are summarized and stored in a new database table, and finally the traditional Chinese medicine knowledge chain relationship corresponding to acute lung injury is generated, providing a reference basis for clinical medication and formula optimization.

[0054] Furthermore, the dialectical auxiliary reasoning module includes the following functions: Conduct in-depth dialectical knowledge analysis on the acute lung injury traditional Chinese medicine knowledge graph to extract the dialectical knowledge related to acute lung injury, including the etiology, pathogenesis and pathological characteristics corresponding to acute lung injury, and obtain the acute lung injury traditional Chinese medicine dialectical knowledge; In the embodiments of the present invention, when conducting in-depth analysis of the dialectical knowledge of the traditional Chinese medicine knowledge graph for acute lung injury, relying on the knowledge graph stored in the Neo4j graph database, the Cypher query language is used for data extraction. For the nodes related to acute lung injury, the query statement MATCH (a: Acute Lung Injury Associated Node)-[r]-(b) WHERE a.type = "Acute Lung Injury" RETURN a, r, b is executed to obtain all the node and relationship data directly or indirectly associated with acute lung injury. Then, the networkx library of Python is used to construct the query results into a graph structure, and potential knowledge is mined through the topological analysis of the graph. For example, in the knowledge graph, the "External Sensation of Pathogenic Factors" node is connected to the "Stagnation of Lung Qi" node through the "causing" relationship, and "Stagnation of Lung Qi" is associated with "Acute Lung Injury". Thus, it is determined that "External Sensation of Pathogenic Factors" is one of the causes of acute lung injury. For the pathogenesis and pathological features, the relationship paths between nodes are analyzed. For example, "Internal Generation of Phlegm Turbidity" affects the "Qi and Blood Circulation in the Lungs" through the "obstructing qi movement" relationship, ultimately leading to acute lung injury, and this process is identified as the relevant pathogenesis. The extracted knowledge of causes, pathogenesis, and pathological features is sorted in a unified format. Each knowledge entry includes the knowledge type (cause / pathogenesis / pathological feature), knowledge content, associated nodes, and relationship description. For example, for the cause "External Sensation of Pathogenic Factors", the associated nodes are "External Sensation of Pathogenic Factors", "Stagnation of Lung Qi", and "Acute Lung Injury", and the relationship is "External Sensation of Pathogenic Factors → causing → Stagnation of Lung Qi → triggering → Acute Lung Injury". Finally, the traditional Chinese medicine dialectical knowledge for acute lung injury is formed and stored in the "Dialectical Knowledge" table of the MySQL database.

[0055] Preferably, based on the traditional Chinese medicine dialectical knowledge for acute lung injury, the dialectical relationships and rules of the traditional Chinese medicine knowledge graph for acute lung injury are extracted to generate the corresponding dialectical relationships and dialectical rules of each acute lung injury dialectical knowledge in the knowledge graph; In the embodiments of the present invention, when extracting dialectical relationships and rules from the traditional Chinese medicine knowledge graph for acute lung injury based on the dialectical knowledge of traditional Chinese medicine for acute lung injury, data is read from the "dialectical knowledge" table and analyzed in combination with the node and relationship structure of the knowledge graph. The dialectical knowledge is classified and sorted using the pandas library in Python. For knowledge related to the cause of the disease, relationship patterns such as "cause of disease → influencing factor" and "cause of disease → resulting outcome" are extracted. For example, from "exogenous pathogenic factors causing qi stagnation in the lungs", the relationship is extracted as follows: the subject is "exogenous pathogenic factors", the relationship type is "causing", and the object is "qi stagnation in the lungs". For the extraction of dialectical rules, the IF-THEN rule template is used to traverse the dialectical knowledge. If a stable correlation is found between "phlegm-heat accumulating in the lungs" and symptoms such as "coughing, expectorating thick yellow phlegm", and "fever", then the rule "IF the patient shows symptoms of coughing, expectorating thick yellow phlegm, and fever, THEN it is inclined to the syndrome of phlegm-heat accumulating in the lungs" is generated. Regular expressions and logical judgments are used to verify the rules to ensure the accuracy and integrity of the rules. The extracted dialectical relationships and rules are stored in the "dialectical relationship" table and the "dialectical rule" table in the MySQL database respectively. The "dialectical relationship" table contains fields such as the subject node ID, relationship type, and object node ID; the "dialectical rule" table contains fields such as rule ID, condition description, and conclusion description, completing the generation of the corresponding dialectical relationships and dialectical rules for each acute lung injury dialectical knowledge in the graph.

[0056] Preferably, based on the corresponding dialectical relationships and dialectical rules for each acute lung injury dialectical knowledge in the graph and in combination with the acute lung injury patient record data in the acute lung injury traditional Chinese medicine information database, dialectical auxiliary reasoning is performed to generate the dialectical suspected syndromes corresponding to acute lung injury.

[0057] In the embodiments of the present invention, when performing dialectical auxiliary reasoning based on dialectical relationships, rules, and the recorded data of patients with acute lung injury, the pandas library of Python is used to read the clinical data of the "patient records" table in the MySQL database, including information such as symptoms and examination results. For each patient record, the spaCy library is used for natural language processing to extract symptom keywords and match them with the conditions in the dialectical rules. For example, a patient record shows "fever, yellow and thick sputum, shortness of breath". By matching, it is found that it meets the condition part of the rule "IF the patient has symptoms of cough, yellow and thick sputum, and fever, THEN it tends to be the syndrome of phlegm-heat accumulating in the lung". The forward reasoning algorithm is adopted. When the patient's symptoms meet the conditions of a certain rule, the syndrome type conclusion corresponding to the rule is obtained. To improve the accuracy of reasoning, the confidence level is calculated for each possible syndrome type, and the confidence level value is determined by combining factors such as the association strength and occurrence frequency between the syndrome type and other relevant nodes in the knowledge graph. The confidence level threshold is set to 0.7, and the syndrome types with a confidence level higher than this threshold are used as the dialectical suspected syndrome types corresponding to acute lung injury. At the same time, information such as traditional Chinese medicines, prescriptions, and ancient book records related to the suspected syndrome types is extracted from the knowledge graph and the database to generate a detailed dialectical auxiliary report, providing a comprehensive reference basis for clinical diagnosis.

[0058] Further, the dialectical auxiliary reasoning based on the dialectical relationships and dialectical rules corresponding to each piece of acute lung injury dialectical knowledge in the graph and combined with the recorded data of acute lung injury patients in the acute lung injury traditional Chinese medicine information database includes: Analyze the clinical symptoms of patients with acute lung injury in the recorded data of acute lung injury patients in the acute lung injury traditional Chinese medicine information database to obtain the clinical symptom characteristics corresponding to patients with acute lung injury; In the embodiments of the present invention, when analyzing the clinical symptoms of patients recorded in the traditional Chinese medicine information database for acute lung injury, the pandas library of Python is used to read all the data in the "patient record" table in the database. First, text cleaning is performed on fields such as "main symptoms" and "examination results" to remove special characters and invalid spaces, and the expression of symptom terms is unified. For example, "shortness of breath and rapid breathing" is replaced with "shortness of breath" to utilize natural language processing technology. The cleaned text is tokenized and part-of-speech tagged through the SpaCy library to identify symptom keywords. For example, from the patient record "fever, cough, expectoration, chest CT shows patchy shadows in both lungs, oxygen saturation 88%", symptom features such as "fever", "cough", "expectoration", "patchy shadows in both lungs", and "oxygen saturation 88%" are extracted. To quantify the symptom features, a symptom feature vector is constructed. The common acute lung injury symptoms are sorted into a standard symptom list containing 200 elements. If a certain symptom appears in the patient record, the corresponding vector position is assigned 1, otherwise 0. For example, if a patient has "fever" and "cough" symptoms, and "fever" is the 30th element and "cough" is the 50th element in the standard list, then the 30th and 50th positions of the symptom feature vector are 1, and the rest are 0. Finally, the symptom feature vectors of all patients are stored in a new "symptom feature" table, and the clinical symptom features corresponding to acute lung injury patients are finally obtained.

[0059] Preferably, according to the dialectical relationships and dialectical rules corresponding to each acute lung injury dialectical knowledge in the atlas and the acute lung injury syndrome types corresponding in the traditional Chinese medicine knowledge atlas for acute lung injury, a dialectical correlation analysis is carried out to obtain the internal connection between each acute lung injury dialectical effect and the acute lung injury syndrome type; In the embodiments of the present invention, when performing dialectical correlation analysis on the corresponding dialectical relationships and dialectical rules of each acute lung injury dialectical knowledge in the atlas and the corresponding acute lung injury syndrome types in the traditional Chinese medicine knowledge atlas of acute lung injury, relying on the acute lung injury traditional Chinese medicine knowledge atlas stored in the Neo4j graph database, and by using the Cypher query language, extract the relationships between the "syndrome type" nodes (such as "syndrome of phlegm-heat congesting the lung" and "syndrome of deficiency of lung qi") and the "symptoms", "traditional Chinese medicines", and "prescriptions" nodes related to the syndrome type in the atlas. For example, the query statement MATCH (z:syndrome type)-[r]->(s: symptom) RETURN z.name, s.name, r can obtain the association relationships between each syndrome type and symptoms. A rule library containing 50 traditional Chinese medicine dialectical rules is pre-established and stored in the "dialectical rules" table of the MySQL database. The rules are expressed in the form of IF-THEN, such as "IF the symptoms include 'fever' and 'yellow and thick sputum', THEN it tends to be the'syndrome of phlegm-heat congesting the lung'". By using the pandas library of Python to read the data of the rule library, and by looping through each rule and combining the relationship data in the knowledge atlas for matching. For example, for a certain rule, if there is a corresponding relationship between the patient's symptoms in the knowledge atlas and the symptoms in the rule, then calculate the association strength between the rule and the syndrome type. The association strength calculation uses the Boolean weighting method. If all the symptoms in the rule exist in the atlas, the association strength is 1. If some exist, it is calculated according to the existing proportion. Finally, obtain the internal connection between each acute lung injury dialectical effect and the acute lung injury syndrome type, and store it in the database in matrix form.

[0060] Preferably, based on the internal connection between each acute lung injury dialectical effect and the acute lung injury syndrome type, perform dialectical auxiliary reasoning on the clinical symptom characteristics corresponding to the acute lung injury patients to generate the dialectical suspected syndrome types corresponding to the acute lung injury.

[0061] In the embodiments of the present invention, when performing dialectical auxiliary reasoning on the clinical symptom characteristics corresponding to acute lung injury patients based on the internal relationship between the dialectical effects of each acute lung injury and the acute lung injury syndrome types, data processing is carried out by using the numpy library and pandas library of Python. First, the patient clinical symptom feature vectors in the "symptom characteristics" table and the "dialectical association" matrix are read. For each patient, the dot product operation is performed between the symptom feature vector and the association vectors of each syndrome type in the dialectical association matrix to obtain the preliminary scores of each syndrome type. For example, if a patient's symptom feature vector is [0, 1, 0...1] and the association vector of the "syndrome of phlegm-heat congesting the lung" is [0, 1, 1...0], the preliminary score of the "syndrome of phlegm-heat congesting the lung" for this patient is 1 after the dot product. To exclude interference factors, a score threshold of 0.6 is set, and the syndrome types with scores lower than this threshold are directly excluded. For the remaining syndrome types, they are sorted according to the scores, and the top 3 syndrome types with the highest scores are selected as the suspected syndrome types. At the same time, in combination with the traditional Chinese medicine and prescription information related to each syndrome type in the knowledge graph, an auxiliary description document is generated for each suspected syndrome type. For example, for the suspected syndrome type of "syndrome of phlegm-heat congesting the lung", the relevant traditional Chinese medicines (such as Scutellaria baicalensis and Morus alba root bark), prescriptions (such as Qingqi Huatan Decoction), and the original text records of ancient medical books corresponding to this syndrome type are listed in the document. Finally, the dialectical suspected syndrome types corresponding to acute lung injury are generated, providing a reference basis for clinical diagnosis.

[0062] Furthermore, the prescription sorting and recommendation module includes the following functions: Performing a similarity measurement calculation between the symptoms of the acute lung injury patients corresponding to the dialectical suspected syndrome types of acute lung injury in the acute lung injury traditional Chinese medicine knowledge graph to obtain the similarity between the dialectical suspected syndrome types and each symptom in the knowledge graph; In the embodiments of the present invention, when calculating the similarity measure between the symptoms of acute lung injury patients corresponding to the traditional Chinese medicine knowledge graph of acute lung injury based on the dialectical suspected syndromes corresponding to acute lung injury, the cosine similarity algorithm is used to quantify the similarity degree between symptoms. The dialectical suspected syndrome information corresponding to acute lung injury is read from the relevant tables of the MySQL database, and the symptom node data in the traditional Chinese medicine knowledge graph of acute lung injury stored in the Neo4j graph database is read. Each dialectical suspected syndrome and the symptoms in the knowledge graph are respectively represented in vector form. For the dialectical suspected syndromes, the typical symptom keywords associated with them are extracted to construct feature vectors; for the symptom nodes in the knowledge graph, their relevant attributes and associated relationship information are also extracted to construct vectors. For example, the typical symptom keywords associated with a certain dialectical suspected syndrome "syndrome of phlegm-heat accumulating in the lung" are "cough", "yellow and thick sputum", "fever", which are encoded as a vector [1, 1, 1, 0, 0,...] (assuming there are a total of 100 common symptom dimensions); the "cough and asthma" symptom node in the knowledge graph is encoded as a vector [1, 0, 0, 1, 0,...] according to its attributes and associated relationships, and the cosine similarity between the vectors is calculated using the numpy library of Python. The formula is: , where and are the vectors of the dialectical suspected syndrome and the symptom respectively, is the dot product of the vectors, and are the norms of the vectors. By traversing all symptom nodes, the similarity between each dialectical suspected syndrome and the symptoms in the knowledge graph is calculated, and the results are stored in the "symptom similarity" table of the MySQL database. The table contains fields such as dialectical suspected syndrome ID, symptom ID, similarity value, etc.

[0063] Preferably, based on the similarity between the dialectical suspected syndromes and the symptoms in the knowledge graph, the symptoms of acute lung injury patients corresponding to the traditional Chinese medicine knowledge graph of acute lung injury are sorted in terms of similarity to generate a similar sequence of symptoms corresponding to the acute lung injury syndromes; In the embodiments of the present invention, when sorting the symptoms of acute lung injury patients corresponding to the traditional Chinese medicine knowledge graph of acute lung injury in terms of similarity based on the similarity between the dialectical suspected syndromes and the symptoms in the knowledge graph, the data in the "symptom similarity" table is read using the pandas library of Python. For each dialectical suspected syndrome, the relevant symptoms are sorted in descending order according to the similarity value. For example, for the "syndrome of phlegm-heat accumulating in the lung", all the related symptoms and their corresponding similarity values are queried in the "symptom similarity" table, and then the sort_values() function is used to sort the data in descending order: import pandas as pd data = pd.read_csv('Symptom Similarity.csv') sorted_data = data[data['Dialectical Suspected Syndrome Type ID'] == 'Phlegm-Heat Accumulating in the Lung Syndrome_ID'].sort_values(by='Similarity Value', ascending=False) Store the sorted symptom IDs in sequence to generate the patient symptom similarity sequence corresponding to each acute lung injury syndrome type, and store it in the new "Symptom Similarity Sequence" table. The table structure includes fields such as Dialectical Suspected Syndrome Type ID, Symptom Similarity Sequence (stored in the form of a comma-separated string of symptom IDs, such as "Cough_ID, Yellow and Thick Sputum_ID, Fever_ID"), etc., and finally generate the patient symptom similarity sequence corresponding to the acute lung injury syndrome type.

[0064] Preferably, based on the traditional Chinese medicine knowledge graph of acute lung injury, perform sorting and recommendation of traditional Chinese medicine prescriptions for the symptoms of acute lung injury patients corresponding to the patient symptom similarity sequence of the acute lung injury syndrome type, so as to find the corresponding classic traditional Chinese medicine prescription combination suggestion plan in the knowledge graph in order, and generate the traditional Chinese medicine prescription suggestion plan sequence corresponding to the acute lung injury syndrome type.

[0065] In the embodiments of the present invention, when ranking and recommending traditional Chinese medicine prescriptions for the symptoms of acute lung injury patients corresponding to the syndrome types of acute lung injury based on the traditional Chinese medicine knowledge graph of acute lung injury, it is implemented by using the Cypher query language of Neo4j in combination with Python scripts. The symptom similarity sequences corresponding to each dialectical suspected syndrome type are read from the "symptom similarity sequence" table. For each symptom ID in the sequence, the traditional Chinese medicine prescription nodes associated with it are searched in the knowledge graph. For example, for the symptom similarity sequence "cough_ID, yellow and thick sputum_ID, fever_ID" of the "syndrome of phlegm-heat accumulating in the lung", the Cypher query statement MATCH (s: symptom {id: "cough_ID"})-[:associated]->(p: traditional Chinese medicine prescription) RETURN p is executed in sequence to search for the traditional Chinese medicine prescription nodes associated with the "cough" symptom, and all associated traditional Chinese medicine prescription IDs are recorded. Such queries are performed for each symptom in the sequence, and the number of times each traditional Chinese medicine prescription is associated is counted. At the same time, the traditional Chinese medicine prescriptions are sorted from high to low according to the number of times they are associated. In the case of the same number of times, a secondary sorting is performed with reference to the "recommended priority" attribute of the traditional Chinese medicine prescription nodes in the knowledge graph (this attribute can be preset according to factors such as the frequency of ancient medical records and the wide range of clinical applications). According to the sorted order, the corresponding classical traditional Chinese medicine prescription combination information, including the prescription name, medicinal material composition, dosage ratio, etc., is extracted from the knowledge graph to generate a sequence of traditional Chinese medicine prescription suggestion schemes corresponding to the syndrome types of acute lung injury, which is stored in the "traditional Chinese medicine prescription suggestion scheme" table of the MySQL database. The table contains fields such as the dialectical suspected syndrome type ID and the sequence of traditional Chinese medicine prescription suggestion schemes (stored in the form of a comma-separated string of prescription information), providing a detailed reference scheme for clinical medication.

[0066] Furthermore, the present invention also provides a method for information processing of traditional Chinese medicine for acute lung injury. The method is implemented based on the above-mentioned information processing system of traditional Chinese medicine for acute lung injury. The method for information processing of traditional Chinese medicine for acute lung injury includes: By collecting and integrating the information of ancient traditional Chinese medicine classics corresponding to acute lung injury, domestic and foreign clinical research information, and the data of acute lung injury patients extracted from the electronic medical record systems of major hospitals, and performing slice coding storage on the information of ancient traditional Chinese medicine classics, domestic and foreign modern clinical research information, and the data of acute lung injury patients to generate an information database of traditional Chinese medicine for acute lung injury; Performing entity and relationship extraction based on the information of ancient traditional Chinese medicine classics corresponding to acute lung injury and domestic and foreign clinical research information in the information database of traditional Chinese medicine for acute lung injury to generate the entity of traditional Chinese medicine knowledge structure corresponding to acute lung injury and the relationship of traditional Chinese medicine knowledge chain; constructing a knowledge graph based on the entity of traditional Chinese medicine knowledge structure corresponding to acute lung injury and the relationship of traditional Chinese medicine knowledge chain to generate a traditional Chinese medicine knowledge graph of acute lung injury; Based on the traditional Chinese medicine knowledge graph for acute lung injury and combined with the recorded data of acute lung injury patients in the traditional Chinese medicine information database for acute lung injury, dialectical auxiliary reasoning is carried out to generate the dialectical suspected syndromes corresponding to acute lung injury; Based on the dialectical suspected syndromes corresponding to acute lung injury, the traditional Chinese medicine prescriptions in the traditional Chinese medicine knowledge graph for acute lung injury are sorted and recommended to generate a sequence of traditional Chinese medicine prescription suggestion schemes corresponding to the syndromes of acute lung injury.

[0067] Therefore, from any point of view, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the application documents are intended to be included in the present invention.

[0068] The above are only the specific embodiments of the present invention, enabling those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can 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 these embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. An information processing system for traditional Chinese medicine for acute lung injury, characterized in that, It includes the following modules: A traditional Chinese medicine information slicing and storage module, which is used to collect and integrate the information of ancient traditional Chinese medicine classics corresponding to acute lung injury, domestic and foreign clinical research information, and the recorded data of acute lung injury patients extracted from the electronic medical record systems of major hospitals, and slice, encode and store the information of ancient traditional Chinese medicine classics, domestic and foreign modern clinical research information, and the recorded data of acute lung injury patients to generate an acute lung injury traditional Chinese medicine information database; A traditional Chinese medicine knowledge graph construction module, which is used to extract entities and relationships based on the information of ancient traditional Chinese medicine classics and domestic and foreign clinical research information corresponding to acute lung injury in the acute lung injury traditional Chinese medicine information database to generate the entity of the traditional Chinese medicine knowledge structure corresponding to acute lung injury and the relationship of the traditional Chinese medicine knowledge chain; Construct a knowledge graph based on the entity of the traditional Chinese medicine knowledge structure corresponding to acute lung injury and the relationship of the traditional Chinese medicine knowledge chain to generate an acute lung injury traditional Chinese medicine knowledge graph; A dialectical auxiliary reasoning module, which is used to perform dialectical auxiliary reasoning based on the acute lung injury traditional Chinese medicine knowledge graph and combined with the recorded data of acute lung injury patients in the acute lung injury traditional Chinese medicine information database to generate the dialectical suspected syndrome types corresponding to acute lung injury; A prescription sorting and recommendation module, which is used to sort and recommend traditional Chinese medicine prescriptions for the acute lung injury traditional Chinese medicine knowledge graph based on the dialectical suspected syndrome types corresponding to acute lung injury to generate a sequence of recommended traditional Chinese medicine prescription schemes corresponding to the acute lung injury syndrome types.

2. The traditional Chinese medicine information processing system for acute lung injury according to claim 1, wherein The traditional Chinese medicine information slicing and storage module includes the following functions: Collect the information records about acute lung injury in ancient traditional Chinese medicine classics, including symptom descriptions, prescription compositions, and medication experiences, to obtain the information of ancient traditional Chinese medicine classics corresponding to acute lung injury; Collect and integrate the clinical research information about the traditional Chinese medicine treatment of acute lung injury in domestic and foreign modern medical journals and research reports to obtain the domestic and foreign clinical research information corresponding to acute lung injury; Extract the clinical diagnosis information, medication records, and clinical effects of acute lung injury patients from the electronic medical record systems of major hospitals to obtain the recorded data of acute lung injury patients; Remove duplicate, incorrect, and invalid data from the information of ancient traditional Chinese medicine classics, domestic and foreign modern clinical research information, and the recorded data of acute lung injury patients to obtain the cleaned traditional Chinese medicine classic information, domestic and foreign clinical research information, and acute lung injury recorded data; Slice, encode, and store the cleaned traditional Chinese medicine classic information, domestic and foreign clinical research information, and acute lung injury recorded data to generate an acute lung injury traditional Chinese medicine information database.

3. The traditional Chinese medicine information processing system for acute lung injury according to claim 2, characterized in that, The slicing, encoding, and storage of the cleaned traditional Chinese medicine classic information, domestic and foreign clinical research information, and acute lung injury recorded data include: Extract the symptom attributes of the acute lung injury conditions corresponding to the cleaned traditional Chinese medicine classic information, domestic and foreign clinical research information, and the recorded data of acute lung injury patients to obtain the acute lung injury symptom attribute elements corresponding to the pharmacopoeia, research, and patient records; Classify the acute lung injury symptom attribute elements corresponding to the pharmacopoeia, research, and patient records into symptom type clusters to generate different acute lung injury symptom attribute type clusters; Slice and divide the traditional Chinese medicine (TCM) pharmacopoeia information, domestic and international clinical research information, and acute lung injury record data corresponding to the corresponding type clusters based on different acute lung injury symptom attribute type clusters to generate TCM information fragments, domestic and international research information fragments, and patient record data fragments under different symptom type clusters; Index, encode, and store the TCM information fragments, domestic and international research information fragments, and patient record data fragments under different symptom type clusters to generate an acute lung injury TCM information database.

4. The traditional Chinese medicine information processing system for acute lung injury according to claim 3, wherein, The classification of acute lung injury symptom attribute elements corresponding to the pharmacopoeia, research, and patient records into symptom type clusters includes: Perform word vector embedding on the acute lung injury symptom attribute elements corresponding to the pharmacopoeia, research, and patient records to generate acute lung injury symptom word embedding vectors corresponding to the pharmacopoeia, research, and patient records; Conduct semantic feature analysis on the acute lung injury symptom word embedding vectors corresponding to the pharmacopoeia, research, and patient records to obtain the acute lung injury symptom semantic features corresponding to the pharmacopoeia, research, and patient records; Evaluate the symptom scores based on the acute lung injury symptom semantic features corresponding to the pharmacopoeia, research, and patient records to obtain the acute lung injury symptom scores corresponding to the pharmacopoeia, research, and patient records; Classify the acute lung injury symptom attribute elements corresponding to the pharmacopoeia, research, and patient records into symptom type clusters based on the acute lung injury symptom scores corresponding to the pharmacopoeia, research, and patient records to generate different acute lung injury symptom attribute type clusters.

5. The traditional Chinese medicine information processing system for acute lung injury according to claim 1, wherein The TCM knowledge graph construction module includes the following functions: Identify and extract knowledge entity structures from the ancient TCM pharmacopoeia information and domestic and international clinical research information corresponding to different types of acute lung injury in the acute lung injury TCM information database to generate TCM knowledge structure entities corresponding to acute lung injury; Measure the knowledge associations between the TCM knowledge structure entities corresponding to each acute lung injury to obtain the knowledge association coefficient between each TCM knowledge entity; Connect the TCM knowledge structure entities corresponding to each acute lung injury based on the knowledge association coefficient between each TCM knowledge entity to generate the acute lung injury action relationship chain between each TCM knowledge entity; Extract the knowledge chain relationship from the acute lung injury action relationship chain between each TCM knowledge entity based on the TCM knowledge structure entities corresponding to each acute lung injury to generate the TCM knowledge chain relationship corresponding to acute lung injury; Construct a knowledge graph based on the TCM knowledge structure entities corresponding to acute lung injury and the TCM knowledge chain relationship, using each TCM knowledge structure entity as a node and the corresponding TCM knowledge chain relationship as a connection edge to generate an acute lung injury TCM knowledge graph.

6. The information processing system for traditional Chinese medicine for acute lung injury according to claim 5, wherein The extraction of the knowledge chain relationship from the acute lung injury action relationship chain between each TCM knowledge entity based on the TCM knowledge structure entities corresponding to each acute lung injury includes: Evaluate the pharmacodynamic synergy between the TCM knowledge structure entities corresponding to each acute lung injury to obtain the acute lung injury pharmacodynamic synergy coefficient between each TCM knowledge entity; Obtain the traditional Chinese medicine pharmacological action components corresponding to each traditional Chinese medicine knowledge entity through the traditional Chinese medicine knowledge structure entities corresponding to each acute lung injury, and conduct a pharmacological conflict probability analysis based on the traditional Chinese medicine pharmacological action components corresponding to each traditional Chinese medicine knowledge entity to obtain the pharmacological action conflict probability between each traditional Chinese medicine knowledge entity; Quantify the compatibility combination degree between the traditional Chinese medicine knowledge structure entities corresponding to each acute lung injury based on the acute lung injury pharmacodynamic synergistic coefficient and the pharmacological action conflict probability between each traditional Chinese medicine knowledge entity, and obtain the pharmacological action compatibility combination degree between each traditional Chinese medicine knowledge entity; Based on the pharmacological action compatibility combination degree between each traditional Chinese medicine knowledge entity, screen and extract the compatibility combination chains of the acute lung injury action relationship chains between each traditional Chinese medicine knowledge entity, and extract the acute lung injury action relationship chains corresponding to the traditional Chinese medicine knowledge entities with a pharmacological action compatibility combination degree greater than the preset threshold as the corresponding knowledge chain relationships to generate the traditional Chinese medicine knowledge chain relationships corresponding to acute lung injury.

7. The traditional Chinese medicine information processing system for acute lung injury according to claim 1, wherein The dialectical auxiliary reasoning module includes the following functions: Conduct a deep analysis of the dialectical knowledge of the acute lung injury traditional Chinese medicine knowledge graph to extract the dialectical knowledge related to acute lung injury, including the etiology, pathogenesis, and pathological characteristics corresponding to acute lung injury, and obtain the acute lung injury traditional Chinese medicine dialectical knowledge; Extract the dialectical relationships and rules of the acute lung injury traditional Chinese medicine knowledge graph based on the acute lung injury traditional Chinese medicine dialectical knowledge to generate the dialectical relationships and dialectical rules corresponding to each acute lung injury dialectical knowledge in the graph; Conduct dialectical auxiliary reasoning based on the dialectical relationships and rules corresponding to each acute lung injury dialectical knowledge in the graph and combined with the acute lung injury patient record data in the acute lung injury traditional Chinese medicine information database to generate the dialectical suspected syndromes corresponding to acute lung injury.

8. The traditional Chinese medicine information processing system for acute lung injury according to claim 7, wherein The conducting dialectical auxiliary reasoning based on the dialectical relationships and rules corresponding to each acute lung injury dialectical knowledge in the graph and combined with the acute lung injury patient record data in the acute lung injury traditional Chinese medicine information database includes: Conduct a clinical symptom analysis of the acute lung injury patient record data in the acute lung injury traditional Chinese medicine information database to obtain the clinical symptom characteristics corresponding to the acute lung injury patients; Conduct a dialectical correlation analysis based on the dialectical relationships and rules corresponding to each acute lung injury dialectical knowledge in the graph and the acute lung injury syndromes corresponding to the acute lung injury traditional Chinese medicine knowledge graph to obtain the internal connections between each acute lung injury dialectical action and the acute lung injury syndromes; Conduct dialectical auxiliary reasoning on the clinical symptom characteristics corresponding to the acute lung injury patients based on the internal connections between each acute lung injury dialectical action and the acute lung injury syndromes to generate the dialectical suspected syndromes corresponding to acute lung injury.

9. The traditional Chinese medicine information processing system for acute lung injury according to claim 1, wherein, The prescription sorting and recommendation module includes the following functions: Calculate the similarity measure between the symptoms of the acute lung injury patients corresponding to the acute lung injury traditional Chinese medicine knowledge graph based on the dialectical suspected syndromes corresponding to acute lung injury to obtain the similarity between the dialectical suspected syndromes and each symptom in the knowledge graph; Performing similarity ranking processing on the symptoms of acute lung injury patients corresponding to the acute lung injury traditional Chinese medicine knowledge graph based on the similarity between the dialectical suspected syndrome types and each symptom in the knowledge graph, so as to generate a patient symptom similarity sequence corresponding to the acute lung injury syndrome type; Performing traditional Chinese medicine prescription ranking recommendation on the symptoms of acute lung injury patients corresponding to the patient symptom similarity sequence corresponding to the acute lung injury syndrome type based on the acute lung injury traditional Chinese medicine knowledge graph, so as to search for the corresponding classic traditional Chinese medicine prescription combination suggestion plan in the knowledge graph in sequence, so as to generate a traditional Chinese medicine prescription suggestion plan sequence corresponding to the acute lung injury syndrome type.

10. A Chinese medicine information processing method for acute lung injury, characterized in that, The method is implemented based on the acute lung injury traditional Chinese medicine information processing system described in claim 1, and the acute lung injury traditional Chinese medicine information processing method includes: By collecting and integrating the information of ancient traditional Chinese medicine classics corresponding to acute lung injury, domestic and foreign clinical research information, and the record data of acute lung injury patients extracted from the electronic medical record systems of major hospitals, and performing slice coding storage on the information of ancient traditional Chinese medicine classics, domestic and foreign modern clinical research information, and the record data of acute lung injury patients, so as to generate an acute lung injury traditional Chinese medicine information database; Performing entity and relationship extraction according to the information of ancient traditional Chinese medicine classics corresponding to acute lung injury and domestic and foreign clinical research information in the acute lung injury traditional Chinese medicine information database, so as to generate the traditional Chinese medicine knowledge structure entities corresponding to acute lung injury and the traditional Chinese medicine knowledge chain relationship; constructing a knowledge graph according to the traditional Chinese medicine knowledge structure entities corresponding to acute lung injury and the traditional Chinese medicine knowledge chain relationship, so as to generate an acute lung injury traditional Chinese medicine knowledge graph; Performing dialectical auxiliary reasoning based on the acute lung injury traditional Chinese medicine knowledge graph and combining the record data of acute lung injury patients in the acute lung injury traditional Chinese medicine information database, so as to generate a dialectical suspected syndrome type corresponding to acute lung injury; Performing traditional Chinese medicine prescription ranking recommendation on the acute lung injury traditional Chinese medicine knowledge graph based on the dialectical suspected syndrome type corresponding to acute lung injury, so as to generate a traditional Chinese medicine prescription suggestion plan sequence corresponding to the acute lung injury syndrome type.

Images