Construction method of hepatobiliary surgery decision support system based on case-based reasoning

Abstract

The present invention relates to the technical field of information system construction, and in particular to a method for constructing a hepatobiliary surgery decision support system based on case reasoning. The method comprises the following steps: obtaining historical case information data of hepatobiliary surgery and establishing a connection with a hepatobiliary surgery case library to generate a hepatobiliary surgery case decision management database; obtaining new case information data of hepatobiliary surgery input by a user and performing case feature vector conversion and KD tree similar case retrieval processing, while performing case neighbor reasoning decision support analysis to obtain new case neighbor reasoning decision support results of hepatobiliary surgery; performing system architecture development and design on the case neighbor reasoning mechanism process corresponding to the new case neighbor reasoning decision support results of hepatobiliary surgery and uploading it to a cloud platform for system integration and pilot testing to construct a hepatobiliary surgery case decision support system. The present invention can help hepatobiliary surgery cases improve the scientificity and accuracy of decision support.

Description

Technical Field

[0001] The present invention relates to the technical field of information system construction, and in particular to a method for constructing a hepatobiliary surgery decision support system based on case reasoning. Background Art

[0002] Hepatobiliary surgery involves multiple anatomical structures and their physiological functions. The choice of surgical decision support scheme and postoperative management directly affect the patient's prognosis and quality of life. In recent years, case-based reasoning (CBR) has gradually attracted attention as an emerging intelligent decision support method. CBR extracts successful experiences through analysis and comparison of historical cases to provide a basis for clinical decision-making. This method emphasizes learning and memory, and enhances the intelligence level and adaptability of the system by continuously accumulating and updating the case library. Compared with traditional clinical decision support systems (CDSS), CBR methods can respond to complex clinical situations more flexibly and realize personalized medical decision support services. However, traditional clinical decision support systems are often based on static medical knowledge bases and lack in-depth analysis of specific hepatobiliary surgical cases, resulting in limited effects in clinical applications, thus lacking sufficient flexibility and adaptability. Summary of the invention

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

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

[0005] Step S1: Acquire historical case information data of hepatobiliary surgery, wherein the historical case information data of hepatobiliary surgery includes basic patient information data, clinical signs data, laboratory test data, surgical decision support program data, and postoperative follow-up information data; establish a hepatobiliary surgery case library connection based on the historical case information data of hepatobiliary surgery to generate a hepatobiliary surgery case decision management database;

[0006] Step S2: obtaining the new case information data of hepatobiliary surgery input by the user, and performing case feature vector conversion on the new case information data of hepatobiliary surgery to obtain the new case feature vector of hepatobiliary surgery; performing KD tree similar case retrieval processing on the hepatobiliary surgery case decision management database based on the new case feature vector of hepatobiliary surgery to obtain a new retrieval similar case information set of the hepatobiliary surgery case library; performing case nearest neighbor reasoning decision support analysis on the case information feature vector of each retrieved similar case in the new retrieval similar case information set of the hepatobiliary surgery case library based on the new case feature vector of hepatobiliary surgery to obtain the new case nearest neighbor reasoning decision support result of hepatobiliary surgery;

[0007] Step S3: developing and designing the system architecture of the case nearest neighbor reasoning mechanism process corresponding to the decision support result of the new case of hepatobiliary surgery, so as to generate the case reasoning decision support system architecture of hepatobiliary surgery;

[0008] Step S4: Upload the hepatobiliary surgery case reasoning decision support system architecture to the cloud platform for system integration and pilot testing to build a hepatobiliary surgery case decision support system.

[0009] Beneficial effects of the present invention:

[0010] Compared with the prior art, the construction method of the hepatobiliary surgery decision support system based on case reasoning proposed by the present invention has the beneficial effect that, by acquiring the corresponding hepatobiliary surgery historical case information data from the hepatobiliary surgery information database, clinical reports, medical journals and surgical records, it can lay an important foundation for the subsequent analysis and processing process. This process involves a large amount of real clinical case data, including the patient's basic information, clinical signs, laboratory test results, surgical decision support plan and postoperative follow-up information. These data not only provide rich case support for hepatobiliary surgery decision support research, but also ensure the diversity and comprehensiveness of the data, and enhance the representativeness and scientificity of decision support research. Through the acquisition of multi-channel data, researchers can have an in-depth understanding of the clinical manifestations and effects of various hepatobiliary diseases, thereby providing an important evidence basis for subsequent decision support. By connecting and establishing the hepatobiliary surgery case library based on the hepatobiliary surgery historical case information data, a hepatobiliary surgery case decision management database is formed. The establishment of this database marks a major progress in data management and decision support in the field of hepatobiliary surgery. By integrating various types of structured data, it can not only store rich historical case data, but also continuously absorb new clinical data through a dynamic update mechanism, providing long-term and effective support for clinical decision support practice. More importantly, this decision management database can analyze the best decision support path and management strategy for new hepatobiliary surgery cases through data mining technology, and can achieve in-depth analysis of specific historical hepatobiliary surgery cases, help provide data support when formulating clinical guidelines and pathways, thereby optimizing resource allocation and promoting the innovation and development of the hepatobiliary surgery decision support process. Secondly, by obtaining new hepatobiliary surgery-related case information input by users, these data usually include basic information of patients, medical records and follow-up results, etc. The key to this step is to ensure the accuracy and completeness of new hepatobiliary surgery case information data. Through a variety of data input methods (such as questionnaires, database import, electronic medical record system integration, etc.), structured and unstructured new hepatobiliary surgery case data can be collected, which not only provides basic data support for subsequent analysis, but also ensures data diversity and representativeness, covering different patient groups and symptoms, and also promotes the scientificity and accuracy of clinical decision-making for new hepatobiliary surgery cases. By converting the case feature vectors of new case information data of hepatobiliary surgery, the feature vector quantization process converts the corresponding case features in the new case information of hepatobiliary surgery into feature vectors, which can better perform feature selection and dimensionality reduction processing, reduce noise and redundant information, and provide a clearer direction for subsequent data mining and pattern recognition.Based on the feature vectors of new hepatobiliary surgery cases, KD tree similar case retrieval processing is performed on the hepatobiliary surgery case decision management database. KD tree is an efficient retrieval data structure for multidimensional spatial data. By organizing the feature vectors corresponding to historical hepatobiliary surgery cases in the hepatobiliary surgery case decision management database into a tree structure, cases similar to the current new cases can be quickly found in a large number of cases. The efficiency of this method lies in its ability to reduce the search space, significantly speed up the retrieval speed, and provide timely support for clinical decision-making. This step uses KD tree to retrieve similar cases, and can quickly find relevant cases in massive case databases. This method improves the adaptability and flexibility of the new case decision support process, and can provide references for multiple decision support plans in a short period of time, which helps to optimize the hepatobiliary surgery decision support management process. Based on the feature vector of new hepatobiliary surgery cases, case nearest neighbor reasoning decision support analysis is performed on the case information feature vector of each retrieved similar case in the new retrieved similar case information set of the hepatobiliary surgery case library. This analysis process not only involves summarizing and generalizing similar cases, but also combines statistical models and machine learning algorithms to infer and predict the decision support effect and prognosis of specific cases. This method can provide optimal suggestions and decision support for the management of hepatobiliary surgery cases, and help formulate more scientific decision support plans. In addition, nearest neighbor reasoning can effectively integrate a variety of information, provide more personalized decision support services for new hepatobiliary surgery cases, and provide important theoretical support for future research and practice, promoting the sustainable development of the field of hepatobiliary surgery. Then, by developing and designing the system architecture of the case neighbor reasoning mechanism process corresponding to the neighbor reasoning decision support results of new cases in hepatobiliary surgery, an efficient and intelligent case reasoning decision support system for hepatobiliary surgery is constructed. By designing the system architecture, an effective decision support reasoning processing process for new case information can be realized, ensuring that all relevant data can be quickly acquired and integrated during the case analysis process. The system architecture can realize efficient processing of complex data, while ensuring the accuracy and reliability of the decision support reasoning results of new cases in hepatobiliary surgery, further enhancing its decision support capabilities.Finally, uploading the hepatobiliary surgery case reasoning decision support system architecture to the cloud platform for system integration and pilot testing is an important step to ensure that the decision support system can run smoothly in actual applications. The cloud platform can not only achieve efficient data storage and calculation, but also facilitate data sharing and collaboration between different hospitals and doctors. This centralized management method can improve the availability and maintainability of the decision support system, allowing users to access the required information at any location. The pilot test phase is an important link in verifying the functions of the decision support system. In this process, real user feedback can provide valuable information for the optimization of the decision support system. By analyzing the user experience during use, the deficiencies in the decision support system can be discovered in time and improved. It can also evaluate the effectiveness of the decision support system in actual clinical applications, such as improving decision support efficiency and reducing misdiagnosis rates. Through continuous iteration and optimization, the system will eventually be able to provide more effective support for clinical decision-making in hepatobiliary surgery. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments thereof made with reference to the following drawings:

[0012] Figure 1 It is a schematic diagram of the steps of the method for constructing a hepatobiliary surgery decision support system based on case reasoning of the present invention;

[0013] Figure 2 for Figure 1 Detailed step flow diagram of step S1;

[0014] Figure 3 for Figure 1 Detailed step flow chart of step S2 in FIG. DETAILED DESCRIPTION

[0015] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0016] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0017] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0018] To achieve this, please refer to Figures 1 to 3 The present invention provides a method for constructing a hepatobiliary surgery decision support system based on case reasoning, the method comprising the following steps:

[0019] Step S1: Acquire historical case information data of hepatobiliary surgery, wherein the historical case information data of hepatobiliary surgery includes basic patient information data, clinical signs data, laboratory test data, surgical decision support program data, and postoperative follow-up information data; establish a hepatobiliary surgery case library connection based on the historical case information data of hepatobiliary surgery to generate a hepatobiliary surgery case decision management database;

[0020] Step S2: obtaining the new case information data of hepatobiliary surgery input by the user, and performing case feature vector conversion on the new case information data of hepatobiliary surgery to obtain the new case feature vector of hepatobiliary surgery; performing KD tree similar case retrieval processing on the hepatobiliary surgery case decision management database based on the new case feature vector of hepatobiliary surgery to obtain a new retrieval similar case information set of the hepatobiliary surgery case library; performing case nearest neighbor reasoning decision support analysis on the case information feature vector of each retrieved similar case in the new retrieval similar case information set of the hepatobiliary surgery case library based on the new case feature vector of hepatobiliary surgery to obtain the new case nearest neighbor reasoning decision support result of hepatobiliary surgery;

[0021] Step S3: developing and designing the system architecture of the case nearest neighbor reasoning mechanism process corresponding to the decision support result of the new case of hepatobiliary surgery, so as to generate the case reasoning decision support system architecture of hepatobiliary surgery;

[0022] Step S4: Upload the hepatobiliary surgery case reasoning decision support system architecture to the cloud platform for system integration and pilot testing to build a hepatobiliary surgery case decision support system.

[0023] In the embodiment of the present invention, please refer to Figure 1 FIG. 1 is a flow chart of the steps of the method for constructing a hepatobiliary surgery decision support system based on case reasoning of the present invention. In this example, the method for constructing a hepatobiliary surgery decision support system based on case reasoning includes the following steps:

[0024] Step S1: Acquire historical case information data of hepatobiliary surgery, wherein the historical case information data of hepatobiliary surgery includes basic patient information data, clinical signs data, laboratory test data, surgical decision support program data, and postoperative follow-up information data; establish a hepatobiliary surgery case library connection based on the historical case information data of hepatobiliary surgery to generate a hepatobiliary surgery case decision management database;

[0025] In an embodiment of the present invention, corresponding historical case information of hepatobiliary surgery is collected from different sources (such as hospital information systems, clinical reports, medical journals and surgical records). The specific operation includes first extracting relevant data from different data sources through SQL query language. These data include basic patient information (such as gender, age, past medical history), clinical signs (such as jaundice, abdominal pain), laboratory test results (such as liver function indicators, imaging results), surgical decision support plans (such as preoperative evaluation and postoperative management recommendations), and postoperative follow-up information (such as complication records, recovery status). After data extraction, data cleaning technology is used to ensure data integrity and consistency and remove redundant information, thereby obtaining historical case information data of hepatobiliary surgery. The case type descriptions of the acquired historical case information data of hepatobiliary surgery were extracted by using natural language processing (NLP) technology. First, the clinical information in each case record was analyzed using text mining tools (such as NLTK or spaCy) to extract key terms and phrases related to hepatobiliary surgery. For example, terms such as "cholecystectomy", "hepatectomy" or "biliary obstruction" were identified from medical records. Then, the classification criteria for case descriptions were constructed, including but not limited to categories such as surgical type, etiology and complications. The vocabulary and term library were matched to extract the corresponding case type descriptors. The case types of the corresponding hepatobiliary surgery historical case information data were classified by combining the previously extracted historical case type descriptors, and machine learning algorithms (such as decision trees and support vector machines) were used to train each case to establish a classification model. The training data included the labeled case types and the corresponding clinical information features. After the model was trained, the model was used to predict and classify the new desensitized data, and each record was assigned to the corresponding case type. The historical case description data subsequence corresponding to each case type was structured and transformed by using data conversion tools (such as Apache NiFi or Talend) to organize and format the data, convert the description information of each case into a unified structured data set. The specific implementation process includes defining the data structure template, clarifying the name and data type of each field, such as patient ID, surgery type, clinical manifestations, postoperative follow-up results, etc., and using ETL (extraction, transformation, loading) technology to fill the classified case description data into the corresponding fields to form a standardized data set, and export the generated structured data set to a relational database (such as MySQL or PostgreSQL). After obtaining the historical case description structured data set corresponding to each case type, the decision management path is mined and analyzed. This step uses data mining technology, especially path analysis and association rule learning, and performs data analysis by using tools such as RapidMiner or KNIME to identify decision nodes and paths in the case decision support process.To analyze the decision associations between different case types, so as to mine and analyze and generate corresponding decision support management paths. Then, by combining the decision support management paths obtained by previous mining and analysis, the case library connection is established for the historical case description structured data set corresponding to each case type obtained after the previous structured transformation, so as to associate the structured data set with the decision support path by designing a data integration solution and using the ETL (extract, transform, load) process. In the specific implementation, data integration tools (such as Talend or Apache NiFi) are used to ensure that the descriptions of each case type are correctly connected with their corresponding decision paths. In this process, the consistency and accuracy of the data are ensured to facilitate subsequent query and management, so as to connect and generate the corresponding case decision management database, and finally generate the hepatobiliary surgery case decision management database. ,

[0026] Step S2: obtaining the new case information data of hepatobiliary surgery input by the user, and performing case feature vector conversion on the new case information data of hepatobiliary surgery to obtain the new case feature vector of hepatobiliary surgery; performing KD tree similar case retrieval processing on the hepatobiliary surgery case decision management database based on the new case feature vector of hepatobiliary surgery to obtain a new retrieval similar case information set of the hepatobiliary surgery case library; performing case nearest neighbor reasoning decision support analysis on the case information feature vector of each retrieved similar case in the new retrieval similar case information set of the hepatobiliary surgery case library based on the new case feature vector of hepatobiliary surgery to obtain the new case nearest neighbor reasoning decision support result of hepatobiliary surgery;

[0027] In an embodiment of the present invention, a user-friendly interface or application is designed to collect new hepatobiliary surgery case information input by users. The interface includes an input box, a drop-down menu, and a file upload function, allowing clinicians and researchers to input or upload new hepatobiliary surgery case data. The data content should include the patient's basic information (such as age, gender), medical history, symptoms, diagnosis results, and follow-up results, so as to obtain the corresponding new hepatobiliary surgery case information data. By extracting case features from each case information in the previously acquired new hepatobiliary surgery case information data and constructing a corresponding feature matrix, this process uses data mining technology to identify features related to hepatobiliary surgery, such as liver function indicators, imaging examination results, and biochemical indicators, and converting each case feature into a numerical form through standardization processing to construct a feature quantization matrix. The rows of the quantization matrix represent different cases, and the columns represent the extracted features. The quantization process includes feature selection and standardization (such as Z-score standardization or Min-Max standardization) to ensure the comparability of features in different dimensions, so as to obtain the corresponding new hepatobiliary surgery case feature quantization matrix. At the same time, each row (case feature) in the quantization matrix of new hepatobiliary surgery case features obtained by the previous quantization conversion is converted into a vector form. This process uses matrix operations in linear algebra to generate an initial feature vector set, and by combining each case sub-feature in each case feature vector in the initial feature vector set, a weighted algorithm or a machine learning algorithm (such as random forest, support vector machine, etc.) is used to evaluate the importance of the corresponding case information in the new hepatobiliary surgery case information data previously input by the user, so as to define an evaluation indicator (such as information gain, Gini coefficient, etc.), and calculate the contribution of each feature subset to the outcome variable (such as treatment effect) to obtain the score of the corresponding sub-feature in each feature vector. Coefficients, these scoring coefficients reflect the importance of each sub-feature in the case information. The feature vector dimensions of the corresponding case feature vectors in the initial feature vector set are optimized by combining the case importance scoring coefficients corresponding to each case sub-feature in each case feature vector obtained by previous quantitative calculations, so as to remove case sub-features with low importance scoring coefficients. This process is achieved by using techniques such as principal component analysis (PCA) or feature selection algorithms (such as LASSO regression). By analyzing the importance scores in the feature vectors, features with scores lower than the set threshold are deleted to form an optimized feature vector, which contains the sub-features that have the greatest influence on case decision support for hepatobiliary surgery, thereby obtaining a new case feature vector for hepatobiliary surgery.Secondly, by combining the feature vectors of new hepatobiliary surgery cases obtained after feature dimension optimization, the previously constructed hepatobiliary surgery case decision management database is retrieved for similar cases using the KD tree, so as to construct a KD tree structure based on the corresponding historical case feature data in the hepatobiliary surgery case decision management database, so as to efficiently store and retrieve feature vectors. When constructing the KD tree, appropriate segmentation dimensions and segmentation points are selected to ensure the balance and retrieval efficiency of the tree. The feature vectors of new hepatobiliary surgery cases to be retrieved are input, and a neighbor search is performed in the KD tree. The similarity is judged using a distance metric (such as Euclidean distance or Manhattan distance). During the retrieval process, a similarity threshold is set to screen out historical cases whose similarity with the new cases to be retrieved in the feature space meets the expected standard, thereby obtaining a new retrieval similar case information set for the hepatobiliary surgery case library. Then, by combining the feature vectors of new hepatobiliary surgery cases obtained by previous conversion, the case information feature vectors of each retrieved similar case in the new retrieved similar case information set of the hepatobiliary surgery case library obtained by KD tree similarity retrieval are analyzed by neighboring case reasoning. The feature vectors of each retrieved similar case are analyzed using a neighboring reasoning algorithm (such as KNN or weighted KNN) to obtain decision support results. During the analysis, the background information, decision support results and patient characteristics of each similar case are considered, and the decision support effects of known cases are combined to generate decision recommendations for new cases, including decision support plan selection, prognosis evaluation, etc., and finally the neighboring reasoning decision support results of new hepatobiliary surgery cases are obtained.

[0028] Step S3: developing and designing the system architecture of the case nearest neighbor reasoning mechanism process corresponding to the decision support result of the new case of hepatobiliary surgery, so as to generate the case reasoning decision support system architecture of hepatobiliary surgery;

[0029] In an embodiment of the present invention, a mechanism is developed for the case neighbor reasoning mechanism process corresponding to the neighbor reasoning decision support result of a new case in hepatobiliary surgery that has been previously recommended by decision support through a modular design concept, and the architecture of the hepatobiliary surgery case reasoning decision support system is determined. The architecture design is divided into multiple parts, including a data input module, a feature extraction module, a similar case retrieval module, and a neighbor reasoning decision support module. The data input module is responsible for receiving user case information; the feature extraction module is responsible for converting the input case information data into a feature vector that can be used for subsequent processing; the similar case retrieval module matches cases based on the feature vector of the new case and the feature vector of the historical case; and the neighbor reasoning decision support module provides decision support analysis results corresponding to the new case based on the reasoning analysis of historical similar cases, and communicates between modules by using a RESTful API to ensure the scalability and maintainability of the system, thereby realizing a complete design framework for the hepatobiliary surgery case reasoning decision support system, and finally designing and generating the architecture of the hepatobiliary surgery case reasoning decision support system.

[0030] Step S4: Upload the hepatobiliary surgery case reasoning decision support system architecture to the cloud platform for system integration and pilot testing to build a hepatobiliary surgery case decision support system.

[0031] In an embodiment of the present invention, a previously developed and designed hepatobiliary surgery case reasoning decision support system architecture is uploaded to a cloud platform (such as AWS, Azure or Google Cloud) for integrated configuration of the environment. Development tools provided by the cloud platform, such as Docker and Kubernetes, are used for containerized deployment to ensure the portability and scalability of the system. After the deployment is completed, the testing phase is entered, and a load testing tool is used to evaluate the performance of the hepatobiliary surgery case reasoning decision support system architecture, and user feedback is used to optimize the system. The accuracy and reliability of the system's decision support are verified by docking with actual clinical data, forming a complete hepatobiliary surgery case decision support system construction and verification process to ensure that the system can operate effectively in an actual environment and provide accurate support based on case reasoning. Finally, the test optimization construction is carried out to obtain the corresponding hepatobiliary surgery case decision support system.

[0032] Further, step S1 includes the following steps:

[0033] Step S11: obtaining historical case information data of hepatobiliary surgery from hepatobiliary surgery information database, clinical reports, medical journals and hepatobiliary surgery operation records, wherein the historical case information data of hepatobiliary surgery includes basic patient information data, clinical sign data, laboratory test data, surgical decision support plan data and postoperative follow-up information data;

[0034] Step S12: performing case information desensitization processing on the historical case information data of hepatobiliary surgery to obtain desensitized historical case data of hepatobiliary surgery;

[0035] Step S13: extracting case type descriptions from the desensitized data of historical cases of hepatobiliary surgery to obtain type descriptors of various historical cases of hepatobiliary surgery; classifying the desensitized data of historical cases of hepatobiliary surgery based on the type descriptors of various historical cases of hepatobiliary surgery to obtain subsequences of description data of historical cases of hepatobiliary surgery corresponding to various case types;

[0036] Step S14: performing structured conversion processing on the hepatobiliary surgery historical case description data subsequences corresponding to each case type to obtain a hepatobiliary surgery historical case description structured data set corresponding to each case type; performing decision management path mining analysis on the hepatobiliary surgery historical case description structured data set corresponding to each case type to generate a hepatobiliary surgery historical case decision support management path;

[0037] Step S15: Based on the hepatobiliary surgery historical case decision support management path, the hepatobiliary surgery historical case description structured data set corresponding to each case type is connected to the hepatobiliary surgery case library to generate a hepatobiliary surgery case decision management database.

[0038] As an embodiment of the present invention, refer to Figure 2 As shown, Figure 1 Detailed step flow diagram of step S1 in FIG. 1 , in this embodiment, step S1 includes the following steps:

[0039] Step S11: obtaining historical case information data of hepatobiliary surgery from hepatobiliary surgery information database, clinical reports, medical journals and hepatobiliary surgery operation records, wherein the historical case information data of hepatobiliary surgery includes basic patient information data, clinical sign data, laboratory test data, surgical decision support plan data and postoperative follow-up information data;

[0040] In an embodiment of the present invention, corresponding hepatobiliary surgery historical case information is collected from different sources (such as hospital information systems, clinical reports, medical journals and surgical records). The specific operations include first extracting relevant data from different data sources through SQL query language. These data include basic patient information (such as gender, age, past medical history), clinical signs (such as jaundice, abdominal pain), laboratory test results (such as liver function indicators, imaging results), surgical decision support plans (such as preoperative evaluation and postoperative management recommendations), and postoperative follow-up information (such as complication records, recovery status). After data extraction, data cleaning technology is used to ensure data integrity and consistency, remove redundant information, and finally obtain hepatobiliary surgery historical case information data.

[0041] Step S12: performing case information desensitization processing on the historical case information data of hepatobiliary surgery to obtain desensitized historical case data of hepatobiliary surgery;

[0042] In an embodiment of the present invention, the acquired historical case information data of hepatobiliary surgery is desensitized to protect patient privacy, and the patient's personal identity information (such as name, ID number, contact information, etc.) is replaced with a unique identifier or virtual data by adopting data desensitization technology, including data masking and randomization. In addition, sensitive information is encrypted and stored using data encryption technology. During this process, a data processing script is written in a programming language (such as Python or R), the data set is traversed, and each record is desensitized to ensure that each record protects the patient's privacy while maintaining the validity of the data, and finally the desensitized data of historical cases of hepatobiliary surgery are obtained.

[0043] Step S13: extracting case type descriptions from the desensitized data of historical cases of hepatobiliary surgery to obtain type descriptors of various historical cases of hepatobiliary surgery; classifying the desensitized data of historical cases of hepatobiliary surgery based on the type descriptors of various historical cases of hepatobiliary surgery to obtain subsequences of description data of historical cases of hepatobiliary surgery corresponding to various case types;

[0044] In an embodiment of the present invention, natural language processing (NLP) technology is used to extract case type descriptions from desensitized hepatobiliary surgery historical case data. First, a text mining tool (such as NLTK or spaCy) is used to analyze the clinical information in each case record to extract key terms and phrases related to hepatobiliary surgery. For example, terms such as "cholecystectomy", "hepatectomy" or "biliary obstruction" are identified from medical records. Subsequently, a classification standard for case descriptions is constructed, including but not limited to categories such as surgical type, etiology and complications. The vocabulary and term library are matched to extract the corresponding case type descriptors, thereby obtaining the type descriptors of each hepatobiliary surgery historical case. At the same time, the corresponding desensitized data of hepatobiliary surgery historical cases are classified and divided into case types by combining the previously extracted descriptors of each hepatobiliary surgery historical case type, and machine learning algorithms (such as decision trees and support vector machines) are used to train each case to establish a classification model. The training data includes the labeled case types and the corresponding clinical information features. After model training, the model is used to predict and classify the new desensitized data, and each record is assigned to the corresponding case type. After classification is completed, the hepatobiliary surgery historical case description data subsequences corresponding to each case type are finally obtained.

[0045] Step S14: performing structured conversion processing on the hepatobiliary surgery historical case description data subsequences corresponding to each case type to obtain a hepatobiliary surgery historical case description structured data set corresponding to each case type; performing decision management path mining analysis on the hepatobiliary surgery historical case description structured data set corresponding to each case type to generate a hepatobiliary surgery historical case decision support management path;

[0046] In an embodiment of the present invention, a structured conversion process is performed on a subsequence of hepatobiliary surgery historical case description data corresponding to each case type, so that the data is sorted and formatted by using a data conversion tool (such as Apache NiFi or Talend), and the description information of each case is converted into a unified structured data set. The specific implementation process includes defining a data structure template, clarifying the name and data type of each field, such as patient ID, surgery type, clinical manifestations, postoperative follow-up results, etc., and using ETL (extraction, transformation, loading) technology to fill the classified case description data into the corresponding fields to form a standardized data set, and exporting the generated structured data set to a relational database (such as MySQL or PostgreSQL), thereby obtaining a structured data set of hepatobiliary surgery historical case descriptions corresponding to each case type. At the same time, after obtaining the structured data set of historical case descriptions of hepatobiliary surgery corresponding to each case type, the decision management path is mined and analyzed. This step utilizes data mining technology, especially path analysis and association rule learning, and uses tools such as RapidMiner or KNIME to perform the data analysis process to identify the decision nodes and paths in the case decision support process. In specific implementation, frequent pattern mining algorithms (such as Apriori or FP-Growth) are used to analyze the decision associations between different case types and extract the key decision factors that affect the decision support effect, thereby mining and analyzing to generate the corresponding decision support management path, and finally obtaining the decision support management path for historical cases of hepatobiliary surgery.

[0047] Step S15: Based on the hepatobiliary surgery historical case decision support management path, the hepatobiliary surgery historical case description structured data set corresponding to each case type is connected to the hepatobiliary surgery case library to generate a hepatobiliary surgery case decision management database.

[0048] In an embodiment of the present invention, a case library connection is established for the structured data set of hepatobiliary surgery historical case descriptions corresponding to each case type obtained after structured conversion by combining the decision support management path of hepatobiliary surgery historical cases obtained by previous mining and analysis, so as to associate the structured data set with the decision support path by designing a data integration solution and using the ETL (extraction, transformation, loading) process. In the specific implementation, a data integration tool (such as Talend or Apache NiFi) is used to ensure that the description of each case type is correctly connected with its corresponding decision path. In this process, the consistency and accuracy of the data are ensured to facilitate subsequent query and management. At the same time, a user access interface is established so that clinicians can easily query related cases and their decision paths, thereby connecting to generate the corresponding case decision management database, and finally generating a hepatobiliary surgery case decision management database.

[0049] Further, the decision management path mining and analysis of the structured data set of hepatobiliary surgery historical case descriptions corresponding to each case type described in step S14 includes the following steps:

[0050] The case type association mining analysis was performed on the structured data set of historical hepatobiliary surgery case descriptions corresponding to each case type to obtain the common association relationship between case descriptions of different historical hepatobiliary surgery case types.

[0051] In an embodiment of the present invention, data preprocessing is first performed on a structured data set of historical case descriptions of hepatobiliary surgery corresponding to each case type, and a data cleaning tool is used to organize the corresponding structured data set of historical case descriptions of hepatobiliary surgery to remove redundant information, thereby ensuring the quality and accuracy of each case description. Association rule mining algorithms, such as the Apriori algorithm, are used to analyze the correlation between different case types, and by setting support and confidence thresholds, common symptoms, decision support methods, and postoperative recovery conditions in the descriptions are identified. The analysis results will generate an association rule library to clarify the common relationships between the case types, and ultimately obtain the common relationships between case descriptions of different historical case types of hepatobiliary surgery.

[0052] Preferably, based on the commonality of case description associations between different types of hepatobiliary surgery historical case studies, a decision support path exploration analysis is performed on the surgical decision support schemes in the structured data set of hepatobiliary surgery historical case descriptions corresponding to each case type, so as to generate an initial decision support path library for hepatobiliary surgery historical case types;

[0053] In an embodiment of the present invention, by combining the common relationships of case description associations between different hepatobiliary surgery historical case types obtained in previous analysis, a decision tree algorithm and analytic hierarchy process (AHP) are used to perform an in-depth analysis of the decision support paths of the corresponding surgical decision support plans in the structured data set of hepatobiliary surgery historical case descriptions corresponding to each case type, so as to systematically integrate the surgical methods, preoperative evaluation, risk management and other elements of each case type by establishing a decision support model, construct a preliminary decision support path library, and generate a set of systematic surgical decision support paths by simulating the effects and risks of different surgical plans, and finally generate an initial decision support path library for hepatobiliary surgery historical case types.

[0054] Preferably, statistical analysis is performed on the case decision success rate and complication rate of each hepatobiliary surgery case decision support path in the initial decision support path library of historical case types of hepatobiliary surgery, so as to obtain the case decision success rate and complication rate of hepatobiliary surgery corresponding to each decision support path;

[0055] In an embodiment of the present invention, a statistical analysis is performed on each hepatobiliary surgery case decision support path in the initial decision support path library of hepatobiliary surgery historical case types obtained through previous exploration and analysis, so as to extract the successful and failed surgical cases of each path through a retrospective study of the historical case data, and calculate the decision success rate. At the same time, relevant data on the occurrence of complications are collected, and detailed records and statistics are made to quantify and calculate the corresponding complication rates, and finally the hepatobiliary surgery case decision success rate and hepatobiliary surgery complication rate corresponding to each decision support path are obtained.

[0056] Preferably, based on the decision success rate of hepatobiliary surgery cases and the incidence rate of hepatobiliary surgery complications corresponding to each decision support path, the corresponding hepatobiliary surgery case decision support paths in the initial decision support path library of historical case types of hepatobiliary surgery are evaluated for path effectiveness, and a path decision effectiveness index corresponding to each decision support path is obtained;

[0057] In an embodiment of the present invention, the effectiveness of the hepatobiliary surgery case decision support paths corresponding to the initial decision support path library of hepatobiliary surgery historical case types obtained by previous exploration and analysis are evaluated and calculated by combining the decision success rate of hepatobiliary surgery cases and the incidence rate of hepatobiliary surgery complications corresponding to each decision support path obtained by previous quantitative calculations, so as to weight the success rate and the incidence rate of complications and set a reasonable weighting coefficient to form an effectiveness evaluation model, and use sensitivity analysis to test the impact of different weights on the effectiveness index to ensure the robustness of the model, thereby quantitatively generating the effectiveness index of each decision support path, and finally obtaining the path decision effectiveness index corresponding to each decision support path.

[0058] Preferably, the path decision effectiveness index corresponding to each decision support path is compared and judged according to the preset path decision effectiveness threshold. If the path decision effectiveness index corresponding to a decision support path is greater than or equal to the preset path decision effectiveness threshold, the next decision support path is compared and judged; if the path decision effectiveness index corresponding to a decision support path is less than the preset path decision effectiveness threshold, the corresponding hepatobiliary surgery case decision support path is virtualized until the path decision effectiveness index corresponding to each decision support path is compared and judged, so as to generate a hepatobiliary surgery historical case decision support management path.

[0059] In an embodiment of the present invention, a preset path decision effectiveness threshold is used to compare and judge the path decision effectiveness index corresponding to each decision support path previously obtained by quantitative calculation. If the path decision effectiveness index corresponding to one of the decision support paths is greater than or equal to the preset path decision effectiveness threshold, the next path is analyzed; if it is lower than the threshold, it is virtualized. The virtualization process constructs a virtual path model to simulate surgical plans under different circumstances to replace inefficient paths. This process uses dynamic programming and simulation technology to ensure that the virtualization results of each path are feasible in actual operations, thereby generating a complete decision support management path, and finally generating a decision support management path for historical cases of hepatobiliary surgery.

[0060] Further, step S2 includes the following steps:

[0061] Step S21: Acquire new case information data of hepatobiliary surgery input by the user;

[0062] Step S22: performing case feature extraction and matrix quantization on each case information in the new case information data of hepatobiliary surgery to obtain a new case feature quantization matrix of hepatobiliary surgery; performing feature vector quantization conversion on each case feature row in the new case feature quantization matrix of hepatobiliary surgery to obtain an initial feature vector set of the new case of hepatobiliary surgery;

[0063] Step S23: Based on each case sub-feature in each case feature vector in the initial feature vector set of new cases in hepatobiliary surgery, case importance evaluation calculation is performed on the corresponding case information in the new case information data of hepatobiliary surgery to obtain the case importance scoring coefficient corresponding to each case sub-feature in each case feature vector; based on the case importance scoring coefficient corresponding to each case sub-feature in each case feature vector, vector dimension optimization is performed on the corresponding case feature vector in the initial feature vector set of new cases in hepatobiliary surgery to obtain the new case feature vector of hepatobiliary surgery;

[0064] Step S24: performing KD tree similar case retrieval processing on the hepatobiliary surgery case decision management database based on the new case feature vector of the hepatobiliary surgery case to obtain a new retrieval similar case information set of the hepatobiliary surgery case database;

[0065] Step S25: Based on the feature vector of the new case of hepatobiliary surgery, the case information feature vector of each retrieved similar case in the new retrieved similar case information set of the hepatobiliary surgery case library is subjected to case nearest neighbor reasoning decision support analysis to obtain the decision support result of the new case of hepatobiliary surgery.

[0066] As an embodiment of the present invention, refer to Figure 3 As shown, Figure 1 Detailed step flow diagram of step S2 in the embodiment, step S2 includes the following steps:

[0067] Step S21: Acquire new case information data of hepatobiliary surgery input by the user;

[0068] In an embodiment of the present invention, a user-friendly interface or application is designed to collect new hepatobiliary surgery case information input by users. The interface includes an input box, a drop-down menu, and a file upload function, allowing clinicians and researchers to input or upload new hepatobiliary surgery case data. The data content should include the patient's basic information (such as age, gender), medical history, symptoms, diagnosis results and follow-up results, etc. To ensure the integrity and accuracy of the data, verification rules need to be implemented when the data is input, such as date format, numerical range check, etc. The input data will be stored in a database, and the corresponding new hepatobiliary surgery case information data will be obtained eventually.

[0069] Step S22: performing case feature extraction and matrix quantization on each case information in the new case information data of hepatobiliary surgery to obtain a new case feature quantization matrix of hepatobiliary surgery; performing feature vector quantization conversion on each case feature row in the new case feature quantization matrix of hepatobiliary surgery to obtain an initial feature vector set of the new case of hepatobiliary surgery;

[0070] In an embodiment of the present invention, case features are extracted from each case information in the previously acquired new case information data of hepatobiliary surgery and a corresponding feature matrix is ​​constructed. This process uses data mining technology to identify features related to hepatobiliary surgery, such as liver function indicators, imaging examination results and biochemical indicators, and each case feature is converted into a numerical form through standardization processing to construct a feature quantization matrix. The rows of the quantization matrix represent different cases, and the columns represent the extracted features. The quantization process includes feature selection and standardization (such as Z-score standardization or Min-Max standardization) to ensure the comparability of features of different dimensions, and finally obtain the corresponding hepatobiliary surgery new case feature quantization matrix. At the same time, each row (case feature) in the hepatobiliary surgery new case feature quantization matrix obtained by the previous quantization conversion is converted into a vector form. This process uses matrix operations in linear algebra to generate a feature vector set. The vector representation of each case feature contains all important information of the case, and its dimension is the same as the number of features. The feature vectorization process includes encoding the feature value, involving the use of One-Hot Encoding to process the classification features, and directly using the continuous features as the numerical components of the vector, and finally obtaining the initial feature vector set of the new hepatobiliary surgery case.

[0071] Step S23: Based on each case sub-feature in each case feature vector in the initial feature vector set of new cases in hepatobiliary surgery, case importance evaluation calculation is performed on the corresponding case information in the new case information data of hepatobiliary surgery to obtain the case importance scoring coefficient corresponding to each case sub-feature in each case feature vector; based on the case importance scoring coefficient corresponding to each case sub-feature in each case feature vector, vector dimension optimization is performed on the corresponding case feature vector in the initial feature vector set of new cases in hepatobiliary surgery to obtain the new case feature vector of hepatobiliary surgery;

[0072] In an embodiment of the present invention, a weighted algorithm or a machine learning algorithm (such as a random forest, a support vector machine, etc.) is used to evaluate the importance of the corresponding case information in the new hepatobiliary surgery case information data previously input by the user in combination with each case sub-feature in each case feature vector in the initial feature vector set of new hepatobiliary surgery cases, so as to define an evaluation index (such as information gain, Gini coefficient, etc.), calculate the contribution of each feature subset to the outcome variable (such as treatment effect), and ensure the reliability of the evaluation result through cross-validation to obtain the scoring coefficient of each feature. These scoring coefficients reflect the importance of each sub-feature in the case information, thereby obtaining the case importance scoring coefficient corresponding to each case sub-feature in each case feature vector. At the same time, by combining the case importance score coefficient corresponding to each case sub-feature in each case feature vector obtained by previous quantitative calculation, the feature vector dimension of the corresponding case feature vector in the initial feature vector set of new hepatobiliary surgery cases is optimized to remove the case sub-features with low importance score coefficients. This process is achieved by using techniques such as principal component analysis (PCA) or feature selection algorithms (such as LASSO regression). By analyzing the importance scores in the feature vectors, features with scores lower than the set threshold are deleted to form an optimized feature vector, which contains the most influential features for hepatobiliary surgery case decision support, and finally the feature vector of new hepatobiliary surgery cases is obtained.

[0073] Step S24: performing KD tree similar case retrieval processing on the hepatobiliary surgery case decision management database based on the new case feature vector of the hepatobiliary surgery case to obtain a new retrieval similar case information set of the hepatobiliary surgery case database;

[0074] In an embodiment of the present invention, a KD tree similar case retrieval process is performed on a previously constructed hepatobiliary surgery case decision management database by combining a feature vector of a new hepatobiliary surgery case obtained after previously optimizing the feature dimension, so as to construct a KD tree structure based on the corresponding historical case feature data in the hepatobiliary surgery case decision management database, so as to efficiently store and retrieve feature vectors. When constructing the KD tree, appropriate segmentation dimensions and segmentation points are selected to ensure the balance and retrieval efficiency of the tree, and by inputting the feature vector of the new hepatobiliary surgery case to be retrieved, a neighbor search is performed in the KD tree, and a distance metric (such as Euclidean distance or Manhattan distance) is used to judge the similarity. During the retrieval process, a similarity threshold is set to screen out historical cases whose degree of similarity with the new case to be retrieved in the feature space meets the expected standard, and finally a new retrieval similar case information set of the hepatobiliary surgery case library is obtained.

[0075] Step S25: Based on the feature vector of the new case of hepatobiliary surgery, the case information feature vector of each retrieved similar case in the new retrieved similar case information set of the hepatobiliary surgery case library is subjected to case nearest neighbor reasoning decision support analysis to obtain the decision support result of the new case of hepatobiliary surgery.

[0076] In an embodiment of the present invention, a decision support suggestion analysis of the case information feature vector of each retrieved similar case in the new retrieved similar case information set of the hepatobiliary surgery case library previously obtained by similarity retrieval of the KD tree is performed by combining the feature vector of the new hepatobiliary surgery case obtained by the previous conversion with the case information feature vector of the new retrieved similar case. A neighbor reasoning algorithm (such as KNN or weighted KNN) is used for analysis for the feature vector of each retrieved similar case to obtain a decision support result. During the analysis, the background information, decision support results and patient characteristics of each similar case are considered, and a weighting mechanism is used to give a higher weight to similar cases with greater influence. In combination with the decision support effect of known cases, decision suggestions for new cases are generated, including decision support plan selection, prognosis evaluation, etc., and finally the neighbor reasoning decision support result of the new hepatobiliary surgery case is obtained.

[0077] Further, the case importance evaluation calculation of the corresponding case information in the new case information data of hepatobiliary surgery based on each case sub-feature in each case feature vector in the initial feature vector set of new cases of hepatobiliary surgery described in step S23 includes the following steps:

[0078] The sub-feature parameter definition and calibration are performed for each case sub-feature in each case feature vector in the initial feature vector set of new hepatobiliary surgery cases, so as to obtain the sub-feature calibration parameters of each case corresponding to each case feature vector;

[0079] In an embodiment of the present invention, each case feature vector in the new case feature vector set of hepatobiliary surgery is split into multiple sub-features. These sub-features may include physiological indicators (such as liver function, gallbladder size), imaging features (such as lesion manifestations in CT or MRI images) and clinical information (such as patient age, gender, etc.). Each case sub-feature requires detailed parameter definition and calibration. Through a standardized data collection process, biostatistical software (such as R or Python's pandas library) is used to organize data. The calibration parameters include statistical features such as mean and standard deviation to ensure the standardization and consistency of the data. This process needs to ensure that the calibration of each sub-feature is independent, thereby forming a complete sub-feature calibration parameter set, and finally obtaining each case sub-feature calibration parameter corresponding to each case feature vector.

[0080] Preferably, the interaction between sub-features is analyzed for each case sub-feature calibration parameter corresponding to each case feature vector, so as to obtain the interaction relationship between each case sub-feature in each case feature vector;

[0081] In an embodiment of the present invention, a statistical analysis of the interaction is performed on each case sub-feature calibration parameter corresponding to each case feature vector, and a correlation analysis method (such as the Pearson correlation coefficient or the Spearman rank correlation coefficient) is used to perform a statistical analysis on each pair of sub-features to determine the interaction relationship between them. By constructing a correlation matrix, the relationship between the sub-features can be intuitively observed. This analysis process can be performed using data analysis tools (such as Python's SciPy or R's cor.test() function), which can effectively identify which sub-features are statistically significantly correlated, and ultimately obtain the interaction relationship between each case sub-feature in each case feature vector.

[0082] Preferably, based on the interaction relationship between each case sub-feature in each case feature vector, the sub-feature correlation metric calculation is performed on each case sub-feature calibration parameter corresponding to each case feature vector to obtain a feature correlation metric value between each case sub-feature in each case feature vector;

[0083] In an embodiment of the present invention, an association metric is calculated for each case sub-feature calibration parameter corresponding to each case feature vector by combining the interaction relationship between each case sub-feature in each case feature vector obtained by previous analysis, so as to quantitatively analyze the correlation of each pair of sub-features by adopting a weighted method, for example, by using a mutual information metric or an information gain method to quantify the feature correlation between each pair of sub-features. This process requires the use of a machine learning library (such as the metrics module in Scikit-learn) to implement it. By building a model and calculating the mutual information value of each pair of sub-features, the feature correlation metric between each case sub-feature in each case feature vector is finally obtained.

[0084] Preferably, based on the feature correlation measurement value between each case sub-feature in each case feature vector, the case importance evaluation calculation is performed on the corresponding case information in the new case information data of hepatobiliary surgery to obtain the case importance scoring coefficient corresponding to each case sub-feature in each case feature vector.

[0085] In an embodiment of the present invention, the case importance of the corresponding case information in the new case information data of hepatobiliary surgery is evaluated and calculated by combining the feature correlation measurement value between each case sub-feature in each case feature vector obtained by previous quantitative calculation, and the weighted summation method is adopted to combine the correlation measurement value of each sub-feature and its importance in the feature vector (such as the calibration parameters mentioned in the previous step) to quantitatively calculate the comprehensive score of the sub-feature in each case feature vector. This process can be performed using data processing tools (such as Excel or Python's NumPy library) to generate the final importance score of each case through a simple linear model or a more complex machine learning model, and finally obtain the case importance score coefficient corresponding to each case sub-feature in each case feature vector.

[0086] Further, step S24 includes the following steps:

[0087] Step S241: performing case feature vector extraction processing on each corresponding hepatobiliary surgery historical case in the hepatobiliary surgery case decision management database to obtain a case feature vector corresponding to each hepatobiliary surgery historical case;

[0088] In an embodiment of the present invention, an in-depth analysis is performed on each corresponding historical hepatobiliary surgery case in the previously constructed hepatobiliary surgery case decision management database to extract key features and generate a case feature vector. Specifically, relevant data of hepatobiliary surgery cases are first collected, including basic information of patients (such as age, gender), disease type (such as gallstones, liver cancer), diagnosis and treatment plans (such as surgical methods, medication), postoperative recovery conditions (such as complications, length of hospital stay), etc., and data preprocessing technology is used to clean and standardize this information to remove noise data and incomplete data. Subsequently, feature engineering methods are used to convert the data of each case into numerical features through data encoding, normalization and discretization techniques to form a feature vector for each historical case, and finally a case feature vector corresponding to each historical hepatobiliary surgery case is obtained.

[0089] Step S242: normalizing the feature dimension of the new hepatobiliary surgery case feature vector and the case feature vector corresponding to each hepatobiliary surgery historical case, so as to obtain the standardized feature vector of the new hepatobiliary surgery case and the standardized feature vector of each hepatobiliary surgery historical case in the same dimension;

[0090] In an embodiment of the present invention, the feature dimensions of the new hepatobiliary surgery case feature vectors obtained by the previous analysis and the case feature vectors corresponding to each historical hepatobiliary surgery case are normalized to ensure that the corresponding feature vectors can be compared under the same dimension, and the mean and standard deviation are calculated by performing statistical analysis on each eigenvalue in the new case feature vector and the historical case feature vector, and each eigenvalue is transformed by using the Z-score standardization method, that is, by subtracting the mean and dividing by the standard deviation to obtain a standardized feature vector. This process can effectively eliminate the dimensional differences between different eigenvectors, so that all eigenvectors are within the same scale range, and finally the standardized feature vectors of new hepatobiliary surgery cases and the standardized feature vectors of each historical hepatobiliary surgery case under the same dimensional dimension are obtained.

[0091] Step S243: constructing a KD tree for each hepatobiliary surgery historical case standard feature vector to generate a case feature KD tree structure corresponding to each hepatobiliary surgery historical case;

[0092] In an embodiment of the present invention, a KD tree is constructed by targeting the standard feature vector of each historical case of hepatobiliary surgery to facilitate efficient similarity retrieval. First, a suitable dimension is selected as the root node, usually the median of the feature value, and the data set is divided into two left and right subsets. Then, the process is recursively repeated for each subset to construct the left and right subtrees. During the construction of the KD tree, the feature vectors are sorted to ensure that the feature space represented by each node can be effectively divided, and finally a case feature KD tree structure corresponding to each historical case of hepatobiliary surgery is constructed.

[0093] Step S244: performing binary index tree conversion on the case feature KD tree structure corresponding to each historical case of hepatobiliary surgery, to obtain a case feature binary index tree corresponding to each historical case feature KD tree of hepatobiliary surgery;

[0094] In an embodiment of the present invention, the case feature KD tree structure corresponding to each previously constructed historical case of hepatobiliary surgery is converted into a secondary index tree, so as to access each node in the case feature KD tree structure, starting from the root node, and check its left and right child nodes. If the current node has child nodes, they are stored in a new binary index tree structure, in which each node points to its left and right child nodes. The conversion of the binary index tree needs to ensure that the feature information of each node can be effectively indexed and accessed. Through this structure, the corresponding feature data can be quickly located, and finally the case feature binary index tree corresponding to each historical case feature KD tree of hepatobiliary surgery is obtained.

[0095] Step S245: Based on the standard feature vector of new hepatobiliary surgery cases, the case feature binary index tree corresponding to each hepatobiliary surgery historical case feature KD tree is screened for similar cases to obtain new matching similar cases of hepatobiliary surgery; based on the new matching similar cases of hepatobiliary surgery, the hepatobiliary surgery case decision management database is retrieved for similar cases to obtain a new retrieval similar case information set of the hepatobiliary surgery case library.

[0096] In an embodiment of the present invention, similar cases are screened for the case feature binary index tree corresponding to each hepatobiliary surgery historical case feature KD tree by combining the previously normalized normative feature vector of the new hepatobiliary surgery case, so as to input the feature vector of the new case, and locate the feature vector most similar to the new case through the search algorithm of the KD tree. During the search process, the KD tree divides the feature space and narrows down the possible candidate nodes layer by layer, thereby realizing rapid retrieval. The similar cases found are sorted according to the distance metric (such as Euclidean distance), and several cases with the smallest distance are selected as the historical cases most similar to the new case, thereby obtaining new matching similar cases of hepatobiliary surgery. At the same time, similar case retrieval is performed on the previously constructed hepatobiliary surgery case decision management database based on the previously screened new matching similar cases of hepatobiliary surgery to extract relevant information from similar cases, including detailed descriptions of the cases, decision support plans, and postoperative results. By constructing query statements or using database retrieval tools, the hepatobiliary surgery case decision management database is accessed to obtain all similar case information related to the new cases. The retrieval results will be sorted according to similarity to generate a result set containing a new case retrieval similar case information set, and finally a new retrieval similar case information set of the hepatobiliary surgery case library is obtained.

[0097] Further, step S243 includes the following steps:

[0098] Divide each case sub-feature in the standard feature vector of each historical case of hepatobiliary surgery into sub-feature attribute dimensions to obtain a case sub-feature attribute dimension set corresponding to each historical case of hepatobiliary surgery;

[0099] In an embodiment of the present invention, detailed sub-feature attribute dimension division is performed on each case sub-feature in the standard feature vector of each hepatobiliary surgery historical case obtained after feature vector normalization. The process first uses a data mining tool to import a hepatobiliary surgery historical case data set, which contains the patient's basic information, medical history, laboratory test results, and imaging test results. The case sub-features are analyzed one by one to extract features such as liver function indicators, gallbladder size, tumor type, etc. On the basis of each feature, multiple dimensions are further defined, such as numerical dimensions (serum transaminase levels), classification dimensions (tumor stage, patient gender), etc., so as to form a dimension set containing all sub-feature attributes, and finally obtain a case sub-feature attribute dimension set corresponding to each hepatobiliary surgery historical case.

[0100] Preferably, a sub-feature dimension distribution analysis is performed on the case sub-feature attribute dimension set corresponding to each historical case of hepatobiliary surgery to generate a case sub-feature dimension distribution graph corresponding to each historical case of hepatobiliary surgery;

[0101] In an embodiment of the present invention, a visualization tool (such as Matplotlib or Seaborn) is used to perform a statistical analysis on the distribution of each case sub-feature attribute dimension in the case sub-feature attribute dimension set corresponding to each hepatobiliary surgery historical case previously divided. For example, for numerical dimensions, a histogram or box plot is used to display its distribution; for categorical dimensions, a bar chart is used to represent the frequency distribution of each category. The analysis results will reveal the frequency and distribution characteristics of each sub-feature in historical cases, and then generate clear dimensional distribution diagrams. These distribution diagrams not only intuitively display the distribution of features, but also provide important basic information for the subsequent KD tree construction. Finally, the case sub-feature dimension distribution diagram corresponding to each hepatobiliary surgery historical case is analyzed and generated.

[0102] Preferably, based on the case sub-feature dimensional distribution diagram corresponding to each hepatobiliary surgery historical case, each case sub-feature in the standard feature vector of each hepatobiliary surgery historical case is recursively divided and labeled by a KD tree node to obtain each case sub-feature KD tree labeling node corresponding to each hepatobiliary surgery historical case;

[0103] In an embodiment of the present invention, by combining the case sub-feature dimension distribution map corresponding to each hepatobiliary surgery historical case obtained by previous analysis, each case sub-feature in the standard feature vector of each hepatobiliary surgery historical case obtained after feature vector normalization is labeled by recursive division of the KD tree node. In this process, the KD tree algorithm is used to divide each case sub-feature recursively based on the feature dimension, select a main feature dimension as the root node, and divide the data into two parts according to the median of the dimension. Then, different feature dimensions are continued to be selected for each subset for division until a predetermined termination condition (such as the number of nodes or depth limit) is reached. Each division will generate a label for each node to indicate the sub-feature information represented by the node. This labeling process will ensure that the features of each hepatobiliary surgery historical case can be accurately located in the KD tree, and finally the case sub-feature KD tree labeling nodes corresponding to each hepatobiliary surgery historical case are obtained by division and labeling.

[0104] Preferably, a KD tree is constructed based on each case sub-feature KD tree annotation node corresponding to each historical case of hepatobiliary surgery to generate a case feature KD tree structure corresponding to each historical case of hepatobiliary surgery.

[0105] In an embodiment of the present invention, the KD tree is connected and constructed by annotating nodes of each case sub-feature KD tree corresponding to each historical case of hepatobiliary surgery obtained by previous annotation, so as to organize the data of the annotated nodes into a tree structure. Each node not only contains characteristic values, but also needs to store the left and right child node information of the node. In the specific operation, a recursive function is used to start from the root node, and the left subtree and the right subtree are constructed separately. By continuously dividing the data into their respective child nodes, a complete KD tree is formed. During the tree construction process, the characteristic values ​​are used to compare and determine the distribution order of the nodes to ensure the optimization of the query efficiency of each node, thereby constructing a complete KD tree, and finally generating a case feature KD tree structure corresponding to each historical case of hepatobiliary surgery.

[0106] Further, step S244 includes the following steps:

[0107] Perform node traversal hierarchical analysis on the case feature KD tree structure corresponding to each historical case of hepatobiliary surgery to obtain the case feature KD tree node hierarchical information set corresponding to each historical case of hepatobiliary surgery;

[0108] In an embodiment of the present invention, by performing a traversal hierarchical identification analysis on the KD tree nodes of the case feature KD tree structure corresponding to each previously constructed historical case of hepatobiliary surgery, all nodes of each layer can be accessed in turn by adopting a hierarchical traversal method (such as breadth-first search) to obtain the hierarchical information of the nodes. Each node will contain a corresponding feature value and a pointer to its child node, and by aggregating the hierarchical information of each node (including node depth and hierarchical index), a case feature KD tree node hierarchical information set is formed, and finally a case feature KD tree node hierarchical information set corresponding to each historical case of hepatobiliary surgery is obtained.

[0109] Preferably, based on the case feature KD tree node hierarchical information set corresponding to each historical case of hepatobiliary surgery, a binary tree transformation and reconstruction standard design is performed on the corresponding case feature KD tree structure to generate a binary tree transformation and reconstruction standard corresponding to each historical case feature KD tree of hepatobiliary surgery;

[0110] In an embodiment of the present invention, a binary tree conversion standard is designed for the corresponding case feature KD tree structure by combining the case feature KD tree node hierarchical information set corresponding to each hepatobiliary surgery historical case obtained by the previous traversal, so as to design a corresponding binary tree conversion and reconstruction standard. The standard will describe in detail how to convert the KD tree into a binary tree, which is mainly achieved by defining the structure and conversion logic of the tree node. First, the characteristics of each KD tree node and its position in the binary tree are clarified. Usually, each node in the KD tree is mapped to a binary tree node in turn, and a recursive algorithm is used for structural conversion to ensure that each layer of the tree structure is correctly reflected in the binary tree. The balance and efficiency of the binary tree need to be considered in the design process. By setting appropriate node rearrangement and subtree merging rules, it is ensured that the converted binary tree can maintain efficient retrieval performance during query. Finally, a binary tree conversion and reconstruction standard corresponding to each hepatobiliary surgery historical case feature KD tree is designed and generated.

[0111] Preferably, the corresponding case feature KD tree structure is converted into a binary index tree according to the binary tree conversion reconstruction standard corresponding to each hepatobiliary surgery historical case feature KD tree, so as to obtain a case feature binary index tree corresponding to each hepatobiliary surgery historical case feature KD tree.

[0112] In an embodiment of the present invention, the corresponding case feature KD tree structure is converted into a binary index tree by using the binary tree conversion and reconstruction standard corresponding to each hepatobiliary surgery historical case feature KD tree obtained by the previous design, so as to traverse the original KD tree and insert each node into the newly constructed binary index tree according to the node's hierarchical information. This process needs to maintain the sorting characteristics of the binary index tree, that is, for each node, all node values ​​of its left subtree are less than the node value, and all node values ​​of the right subtree are greater than the node value. During the insertion process, the next insertion position is determined by comparing the feature value with the value of the current node, thereby forming an efficient binary index tree to support efficient feature retrieval and decision support, and finally obtaining a case feature binary index tree corresponding to each hepatobiliary surgery historical case feature KD tree.

[0113] Further, the similar case screening process of the case feature binary index tree corresponding to each hepatobiliary surgery historical case feature KD tree based on the hepatobiliary surgery new case standard feature vector described in step S245 includes the following steps:

[0114] Perform node depth-first traversal processing on the case feature binary index tree corresponding to each hepatobiliary surgery historical case feature KD tree to obtain a depth-first traversal node sequence corresponding to each hepatobiliary surgery historical case feature binary tree;

[0115] In an embodiment of the present invention, a depth-first traversal algorithm is used to perform a depth-first traversal of the binary tree of the case feature binary index tree corresponding to each previously converted hepatobiliary surgery historical case feature KD tree. In this process, each node is accessed and the traversal order is recorded by recursive calls to form a node sequence. Each node represents a specific hepatobiliary surgery historical case and contains its feature information. The output node sequence will reflect the structure of the feature binary tree, and finally a depth-first traversal node sequence corresponding to each hepatobiliary surgery historical case feature binary tree is obtained.

[0116] Preferably, based on the standard feature vector of the new hepatobiliary surgery case, each traversal node in the depth-first traversal node sequence corresponding to each hepatobiliary surgery historical case feature binary tree is subjected to point-by-point feature similarity calculation to obtain the case feature vector similarity between the new hepatobiliary surgery case and each traversal node;

[0117] In an embodiment of the present invention, by combining the previously normalized normative feature vector of new hepatobiliary surgery cases, a quantitative calculation of the point-by-point feature similarity is performed on each traversal node in the depth-first traversal node sequence corresponding to each hepatobiliary surgery historical case feature binary tree, so as to evaluate the similarity between the new case vector and the feature vector of each traversal node by setting a similarity calculation formula (such as cosine similarity or Euclidean distance). During implementation, the feature vector of the traversal node is extracted and compared with the feature vector of the new case. This process can be performed by looping through the nodes, calculating the similarity scores one by one, and storing the results in a similarity list, so as to quantify the feature similarity between the new case and the historical case, and finally obtain the case feature vector similarity between the new hepatobiliary surgery case and each traversal node.

[0118] Preferably, based on the case feature vector similarity between the new hepatobiliary surgery case and each traversal node, each traversal node in the depth-first traversal node sequence corresponding to each hepatobiliary surgery historical case feature binary tree is subjected to similar node screening processing to obtain a case feature similar node set between each hepatobiliary surgery historical case and the new case;

[0119] In an embodiment of the present invention, by combining the case feature vector similarity between the new hepatobiliary surgery case and each traversal node obtained by previous quantitative calculation and a preset similarity threshold, each traversal node in the depth-first traversal node sequence corresponding to each hepatobiliary surgery historical case feature binary tree is screened for similar nodes, so as to determine whether the new case and the traversal node meet the similarity standard according to a preset similarity threshold (such as 0.7), and by traversing the similarity list, the nodes that meet the conditions are recorded as similar nodes. In implementation, a set data structure can be used to store similar nodes for efficient search and comparison in subsequent steps, thereby forming a similar node set between each historical case and the new case, and finally obtaining a case feature similar node set between each hepatobiliary surgery historical case and the new case.

[0120] Preferably, the case feature similarity node set between each historical case of hepatobiliary surgery and the new case is compared and judged according to a preset case feature similarity node number threshold; if the corresponding number of similar nodes in the case feature similarity node set is greater than or equal to the preset case feature similarity node number threshold, the corresponding historical case of hepatobiliary surgery is determined as a case similar to the new case of hepatobiliary surgery; otherwise, it is determined as a non-similar case, until the case feature similarity node set between each historical case of hepatobiliary surgery and the new case is compared and judged, so as to obtain a new matching similar case of hepatobiliary surgery.

[0121] In an embodiment of the present invention, a preset case feature similarity node number threshold is used to compare and judge the corresponding case feature similarity node numbers in the case feature similarity node set between each previously screened hepatobiliary surgery historical case and the new case. First, the number of nodes in each similarity node set is statistically calculated. If the number is greater than or equal to the preset similarity node number threshold (for example, 5 similar nodes), the historical case is marked as similar to the new case. Otherwise, it is marked as a non-similar case. The implementation process traverses each similarity node set, uses conditional statements to make judgments, and stores the results in the final case matching list. The entire comparison and judgment process ensures that all historical cases are evaluated, thereby ultimately outputting a report containing similar cases that match the new case, and ultimately obtaining a new matching similar case of hepatobiliary surgery.

[0122] Further, step S25 includes the following steps:

[0123] Step S251: Based on the feature vector of the new case of hepatobiliary surgery, a feature vector correlation analysis is performed on the case information feature vector of each retrieved similar case in the new retrieved similar case information set of the hepatobiliary surgery case library to obtain the case feature correlation coefficient between the new case of hepatobiliary surgery and each retrieved similar case;

[0124] In an embodiment of the present invention, feature vectors of new hepatobiliary surgery cases are extracted. These feature vectors can be obtained by integrating data such as medical image analysis, laboratory test results, and clinical symptoms, and the feature vectors of the new cases are compared with the case information feature vectors of each retrieved similar case in the new retrieved similar case information set of the hepatobiliary surgery case library by using methods such as the Pearson correlation coefficient or the cosine similarity, so as to quantitatively calculate the corresponding correlation coefficient, which can be implemented by programming languages ​​such as Python or R, and by using corresponding data processing libraries (such as NumPy or Pandas) to perform matrix operations and statistical analysis, and finally obtain the case feature correlation coefficient between the new hepatobiliary surgery case and each retrieved similar case.

[0125] Step S252: constructing a case neighborhood network for the corresponding new hepatobiliary surgery case and each retrieved similar case in the new retrieved similar case information set of the hepatobiliary surgery case library according to the case feature correlation coefficient between the new hepatobiliary surgery case and each retrieved similar case, so as to generate a case neighborhood network structure between the new hepatobiliary surgery case and the retrieved similar case;

[0126] In an embodiment of the present invention, the case feature correlation coefficient between the corresponding new hepatobiliary surgery case and each retrieved similar case is obtained by quantitative calculation and these coefficients are used to cluster the corresponding new hepatobiliary surgery case and each retrieved similar case in the new retrieved similar case information set of the hepatobiliary surgery case library, and a threshold is set, and similar cases with correlation coefficients higher than the threshold are regarded as neighborhood members of the new case. When constructing the neighborhood network, the network model in graph theory can be used, and the graph can be constructed by network analysis tools (such as the NetworkX library). During the construction process, each similar case will become a node, and the edges between the nodes represent the correlation relationship between them. The weights of the edges are set according to the correlation coefficients, thereby connecting to form a visual case neighborhood network structure, and finally connecting to generate a case neighborhood network structure between the new hepatobiliary surgery case and the retrieved similar cases.

[0127] Step S253: setting case nearest neighbor reasoning rules for the case neighborhood network structure between the new case of hepatobiliary surgery and the retrieved similar cases, and obtaining similar nearest neighbor reasoning rules for the retrieval of new cases of hepatobiliary surgery;

[0128] In an embodiment of the present invention, after the case neighborhood network structure between the new case of hepatobiliary surgery and the retrieved similar cases is constructed, corresponding reasoning rules are formulated for each neighborhood node in the case neighborhood network structure. These reasoning rules are set based on the characteristics of similar cases in the neighborhood. The characteristics of each similar case are comprehensively analyzed using methods such as Bayesian reasoning, decision trees or fuzzy logic, so as to derive the nearest neighbor reasoning rules for the new case. During the setting process, it is first necessary to define the input and output feature sets, and generate rules through algorithms. For example, "if the decision support X and Y of similar cases A and B are both E, then the decision support Z of the new case is also E", and finally the similar nearest neighbor reasoning rules for retrieving new cases of hepatobiliary surgery are obtained.

[0129] Step S254: Perform a neighbor reasoning comprehensive decision support analysis on the corresponding similar cases in the new retrieval similar case information set of the hepatobiliary surgery case library according to the similar neighbor reasoning rules for the retrieval of new cases in hepatobiliary surgery, so as to obtain the neighbor reasoning decision support results for the new cases in hepatobiliary surgery.

[0130] In an embodiment of the present invention, by using the previously set similar neighbor reasoning rules for retrieving new cases of hepatobiliary surgery, a comprehensive decision support analysis of neighbor reasoning is performed on the corresponding similar cases in the new retrieval similar case information set of the hepatobiliary surgery case library obtained by the previous similar retrieval, so as to generate decision support results by inputting the characteristics of the new case into the established reasoning rules and using the reasoning engine (such as Prolog or a custom reasoning system) to perform reasoning calculations. These results will be combined with the clinical pathway, treatment plan, prognosis and other information of each similar case to form a comprehensive decision support report for the new case, and the results will be displayed using data visualization tools (such as Tableau or Matplotlib) to facilitate medical staff to conduct intuitive analysis and decision support. On this basis, personalized decision support plans or intervention measures are further provided to improve the effectiveness and accuracy of hepatobiliary surgery case processing, and finally obtain the neighbor reasoning decision support results for new cases of hepatobiliary surgery.

[0131] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0132] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for constructing a hepatobiliary surgery decision support system based on case reasoning, characterized in that: The following steps are involved: Step S1: Acquire historical case information data of hepatobiliary surgery, wherein the historical case information data of hepatobiliary surgery includes basic patient information data, clinical signs data, laboratory test data, surgical decision support program data, and postoperative follow-up information data; establish a hepatobiliary surgery case library connection based on the historical case information data of hepatobiliary surgery to generate a hepatobiliary surgery case decision management database; Step S2: obtaining the new case information data of hepatobiliary surgery input by the user, and performing case feature vector conversion on the new case information data of hepatobiliary surgery to obtain the new case feature vector of hepatobiliary surgery; performing KD tree similar case retrieval processing on the hepatobiliary surgery case decision management database based on the new case feature vector of hepatobiliary surgery to obtain a new retrieval similar case information set of the hepatobiliary surgery case library; Based on the feature vector of new cases in hepatobiliary surgery, case nearest neighbor reasoning decision support analysis is performed on the case information feature vector of each similar case in the new similar case information set of the hepatobiliary surgery case library to obtain the decision support results of the new cases in hepatobiliary surgery; Step S2 includes the following steps: Step S21: Acquire new case information data of hepatobiliary surgery input by the user; Step S22: performing case feature extraction and matrix quantization on each case information in the new case information data of hepatobiliary surgery to obtain a new case feature quantization matrix of hepatobiliary surgery; performing feature vector quantization conversion on each case feature row in the new case feature quantization matrix of hepatobiliary surgery to obtain an initial feature vector set of the new case of hepatobiliary surgery; Step S23: Based on each case sub-feature in each case feature vector in the initial feature vector set of new cases in hepatobiliary surgery, case importance evaluation calculation is performed on the corresponding case information in the new case information data of hepatobiliary surgery to obtain the case importance scoring coefficient corresponding to each case sub-feature in each case feature vector; based on the case importance scoring coefficient corresponding to each case sub-feature in each case feature vector, vector dimension optimization is performed on the corresponding case feature vector in the initial feature vector set of new cases in hepatobiliary surgery to obtain the new case feature vector of hepatobiliary surgery; Step S24: performing KD tree similar case retrieval processing on the hepatobiliary surgery case decision management database based on the new case feature vector of the hepatobiliary surgery case to obtain a new retrieval similar case information set of the hepatobiliary surgery case database; Step S25: Based on the feature vector of the new case of hepatobiliary surgery, the case information feature vector of each retrieved similar case in the new retrieved similar case information set of the hepatobiliary surgery case library is subjected to case nearest neighbor reasoning decision support analysis to obtain the decision support result of the new case of hepatobiliary surgery; Step S3: developing and designing the system architecture of the case nearest neighbor reasoning mechanism process corresponding to the decision support result of the new case of hepatobiliary surgery, so as to generate the case reasoning decision support system architecture of hepatobiliary surgery; Step S4: Upload the hepatobiliary surgery case reasoning decision support system architecture to the cloud platform for system integration and pilot testing to build a hepatobiliary surgery case decision support system.

2. The method for constructing a hepatobiliary surgery decision support system based on case reasoning according to claim 1 is characterized in that: Step S1 includes the following steps: Step S11: obtaining historical case information data of hepatobiliary surgery from hepatobiliary surgery information database, clinical reports, medical journals and hepatobiliary surgery operation records, wherein the historical case information data of hepatobiliary surgery includes basic patient information data, clinical sign data, laboratory test data, surgical decision support plan data and postoperative follow-up information data; Step S12: performing case information desensitization processing on the historical case information data of hepatobiliary surgery to obtain desensitized historical case data of hepatobiliary surgery; Step S13: extracting case type descriptions from the desensitized data of historical cases of hepatobiliary surgery to obtain type descriptors of various historical cases of hepatobiliary surgery; classifying the desensitized data of historical cases of hepatobiliary surgery based on the type descriptors of various historical cases of hepatobiliary surgery to obtain subsequences of description data of historical cases of hepatobiliary surgery corresponding to various case types; Step S14: performing structured conversion processing on the hepatobiliary surgery historical case description data subsequences corresponding to each case type to obtain a hepatobiliary surgery historical case description structured data set corresponding to each case type; performing decision management path mining analysis on the hepatobiliary surgery historical case description structured data set corresponding to each case type to generate a hepatobiliary surgery historical case decision support management path; Step S15: Based on the hepatobiliary surgery historical case decision support management path, the hepatobiliary surgery historical case description structured data set corresponding to each case type is connected to the hepatobiliary surgery case library to generate a hepatobiliary surgery case decision management database.

3. The method for constructing a hepatobiliary surgery decision support system based on case reasoning according to claim 2 is characterized in that: The decision management path mining and analysis of the structured data set of hepatobiliary surgery historical case descriptions corresponding to each case type described in step S14 includes the following steps: The case type association mining analysis was performed on the structured data set of historical hepatobiliary surgery case descriptions corresponding to each case type to obtain the common association relationship between case descriptions of different historical hepatobiliary surgery case types. Based on the commonality of case description associations between different types of hepatobiliary surgery historical case studies, the decision support path exploration analysis was performed on the surgical decision support schemes in the structured data set of hepatobiliary surgery historical case descriptions corresponding to each case type, so as to generate an initial decision support path library for hepatobiliary surgery historical case types. Conduct statistical analysis on the case decision success rate and complication rate of each hepatobiliary surgery case decision support path in the initial decision support path library of historical case types of hepatobiliary surgery, so as to obtain the case decision success rate and complication rate of hepatobiliary surgery corresponding to each decision support path; Based on the decision success rate of hepatobiliary surgery cases and the incidence rate of hepatobiliary surgery complications corresponding to each decision support path, the path effectiveness evaluation and calculation of the corresponding hepatobiliary surgery case decision support paths in the initial decision support path library of historical case types of hepatobiliary surgery were performed to obtain the path decision effectiveness index corresponding to each decision support path; The path decision effectiveness index corresponding to each decision support path is compared and judged according to the preset path decision effectiveness threshold. If the path decision effectiveness index corresponding to a decision support path is greater than or equal to the preset path decision effectiveness threshold, the next decision support path will continue to be compared and judged; if the path decision effectiveness index corresponding to a decision support path is less than the preset path decision effectiveness threshold, the corresponding hepatobiliary surgery case decision support path will be virtualized until the path decision effectiveness index corresponding to each decision support path is compared and judged, so as to generate a hepatobiliary surgery historical case decision support management path.

4. The method for constructing a hepatobiliary surgery decision support system based on case reasoning according to claim 1, characterized in that: The step S23 of performing case importance evaluation calculation on the corresponding case information in the new case information data of hepatobiliary surgery based on each case sub-feature in each case feature vector in the initial feature vector set of new cases of hepatobiliary surgery comprises the following steps: The sub-feature parameter definition and calibration are performed for each case sub-feature in each case feature vector in the initial feature vector set of new hepatobiliary surgery cases, so as to obtain the sub-feature calibration parameters of each case corresponding to each case feature vector; Perform inter-feature interaction analysis on each case sub-feature calibration parameter corresponding to each case feature vector to obtain the interaction relationship between each case sub-feature in each case feature vector; Based on the interaction relationship between each case sub-feature in each case feature vector, the sub-feature correlation metric calculation is performed on each case sub-feature calibration parameter corresponding to each case feature vector to obtain the feature correlation metric value between each case sub-feature in each case feature vector; Based on the feature correlation measurement value between each case sub-feature in each case feature vector, the case importance evaluation calculation is performed on the corresponding case information in the new case information data of hepatobiliary surgery, and the case importance scoring coefficient corresponding to each case sub-feature in each case feature vector is obtained.

5. The method for constructing a hepatobiliary surgery decision support system based on case reasoning according to claim 1, characterized in that: Step S24 includes the following steps: Step S241: performing case feature vector extraction processing on each corresponding hepatobiliary surgery historical case in the hepatobiliary surgery case decision management database to obtain a case feature vector corresponding to each hepatobiliary surgery historical case; Step S242: normalizing the feature dimension of the new hepatobiliary surgery case feature vector and the case feature vector corresponding to each hepatobiliary surgery historical case, so as to obtain the standardized feature vector of the new hepatobiliary surgery case and the standardized feature vector of each hepatobiliary surgery historical case in the same dimension; Step S243: constructing a KD tree for each hepatobiliary surgery historical case standard feature vector to generate a case feature KD tree structure corresponding to each hepatobiliary surgery historical case; Step S244: performing binary index tree conversion on the case feature KD tree structure corresponding to each historical case of hepatobiliary surgery to obtain a case feature binary index tree corresponding to each historical case feature KD tree of hepatobiliary surgery; Step S245: Based on the standard feature vector of new hepatobiliary surgery cases, the case feature binary index tree corresponding to each hepatobiliary surgery historical case feature KD tree is screened for similar cases to obtain new matching similar cases of hepatobiliary surgery; based on the new matching similar cases of hepatobiliary surgery, the hepatobiliary surgery case decision management database is retrieved for similar cases to obtain a new retrieval similar case information set of the hepatobiliary surgery case library.

6. The method for constructing a hepatobiliary surgery decision support system based on case reasoning according to claim 5 is characterized in that: Step S243 includes the following steps: Divide each case sub-feature in the standard feature vector of each historical case of hepatobiliary surgery into sub-feature attribute dimensions to obtain a case sub-feature attribute dimension set corresponding to each historical case of hepatobiliary surgery; Perform sub-feature dimension distribution analysis on the case sub-feature attribute dimension set corresponding to each historical case of hepatobiliary surgery to generate a case sub-feature dimension distribution map corresponding to each historical case of hepatobiliary surgery; Based on the case sub-feature dimensional distribution map corresponding to each historical case of hepatobiliary surgery, each case sub-feature in the standard feature vector of each historical case of hepatobiliary surgery is recursively divided and labeled by the KD tree node to obtain each case sub-feature KD tree labeling node corresponding to each historical case of hepatobiliary surgery; A KD tree is constructed based on the case sub-feature KD tree annotation nodes corresponding to each historical case of hepatobiliary surgery to generate a case feature KD tree structure corresponding to each historical case of hepatobiliary surgery.

7. The method for constructing a hepatobiliary surgery decision support system based on case reasoning according to claim 5, characterized in that: Step S244 includes the following steps: Perform node traversal hierarchical analysis on the case feature KD tree structure corresponding to each historical case of hepatobiliary surgery to obtain the case feature KD tree node hierarchical information set corresponding to each historical case of hepatobiliary surgery; Based on the case feature KD tree node hierarchical information set corresponding to each historical case of hepatobiliary surgery, the corresponding case feature KD tree structure is converted into a binary tree and reconstructed to generate a binary tree conversion and reconstruction standard corresponding to each historical case feature KD tree of hepatobiliary surgery; According to the binary tree transformation and reconstruction standard corresponding to each historical case feature KD tree of hepatobiliary surgery, the corresponding case feature KD tree structure is transformed into a binary index tree to obtain a case feature binary index tree corresponding to each historical case feature KD tree of hepatobiliary surgery.

8. The method for constructing a hepatobiliary surgery decision support system based on case reasoning according to claim 5, characterized in that: The similar case screening process of the case feature binary index tree corresponding to each hepatobiliary surgery historical case feature KD tree based on the hepatobiliary surgery new case standard feature vector described in step S245 includes the following steps: Perform node depth-first traversal processing on the case feature binary index tree corresponding to each hepatobiliary surgery historical case feature KD tree to obtain a depth-first traversal node sequence corresponding to each hepatobiliary surgery historical case feature binary tree; Based on the standard feature vector of new hepatobiliary surgery cases, the feature similarity of each traversal node in the depth-first traversal node sequence corresponding to each hepatobiliary surgery historical case feature binary tree is calculated point by point, and the case feature vector similarity between the new hepatobiliary surgery case and each traversal node is obtained; Based on the case feature vector similarity between the new hepatobiliary surgery case and each traversal node, similar node screening is performed on each traversal node in the depth-first traversal node sequence corresponding to each hepatobiliary surgery historical case feature binary tree, and a case feature similar node set between each hepatobiliary surgery historical case and the new case is obtained; According to the preset case feature similarity node number threshold, the case feature similarity node set between each historical case of hepatobiliary surgery and the new case is compared and judged. If the corresponding number of similar nodes in the case feature similarity node set is greater than or equal to the preset case feature similarity node number threshold, the corresponding hepatobiliary surgery historical case is determined as a case similar to the new case of hepatobiliary surgery, otherwise it is determined as a non-similar case, until the case feature similarity node set between each historical case of hepatobiliary surgery and the new case is compared and judged, so as to obtain the new matching similar case of hepatobiliary surgery.

9. The method for constructing a hepatobiliary surgery decision support system based on case reasoning according to claim 1, characterized in that: Step S25 includes the following steps: Step S251: Based on the feature vector of the new case of hepatobiliary surgery, a feature vector correlation analysis is performed on the case information feature vector of each retrieved similar case in the new retrieved similar case information set of the hepatobiliary surgery case library to obtain the case feature correlation coefficient between the new case of hepatobiliary surgery and each retrieved similar case; Step S252: constructing a case neighborhood network for the corresponding new hepatobiliary surgery case and each retrieved similar case in the new retrieved similar case information set of the hepatobiliary surgery case library according to the case feature correlation coefficient between the new hepatobiliary surgery case and each retrieved similar case, so as to generate a case neighborhood network structure between the new hepatobiliary surgery case and the retrieved similar case; Step S253: setting case nearest neighbor reasoning rules for the case neighborhood network structure between the new case of hepatobiliary surgery and the retrieved similar cases, and obtaining similar nearest neighbor reasoning rules for the retrieval of new cases of hepatobiliary surgery; Step S254: Perform a neighbor reasoning comprehensive decision support analysis on the corresponding similar cases in the new retrieval similar case information set of the hepatobiliary surgery case library according to the similar neighbor reasoning rules for the retrieval of new cases in hepatobiliary surgery, so as to obtain the neighbor reasoning decision support results for the new cases in hepatobiliary surgery.

Images