Hypertension prevention and treatment knowledge system dynamic construction method and device based on double theory fusion

Abstract

The invention discloses a dynamic construction method and device for fusing a hypertension prevention and treatment knowledge system based on double theories, can systematically solve the core problems of knowledge organization, classification and iteration, and constructs a hypertension prevention and treatment knowledge system with theoretical support, precise classification and dynamic iteration by fusing Bauer double view and Anderson four-level classification theories. The method comprises the following steps: (1) performing knowledge double-view disassembly based on a Bauer health development model; (2) performing knowledge four-level classification mapping based on an Anderson knowledge classification theory; and (3) constructing a knowledge system dynamic update closed loop.

Description

Technical Field

[0001] This invention relates to the field of medical information processing technology, and in particular to a method for dynamically constructing a knowledge system for the prevention and treatment of hypertension based on the fusion of two theories, and a device for dynamically constructing a knowledge system for the prevention and treatment of hypertension based on the fusion of two theories. Background Technology

[0002] Hypertension is a prevalent chronic cardiovascular disease worldwide, and its prevention and treatment rely heavily on systematic knowledge support. However, current knowledge management solutions have not yet formed a mature technical path that balances theoretical rigor, accurate classification, and timely updates. The core technical gaps are concentrated in three aspects: knowledge organization logic, classification standards, and iteration mechanisms.

[0003] The existing technical solutions closest to this invention can be mainly divided into three categories: The first type is the static knowledge compilation scheme. This scheme manually compiles hypertension prevention and treatment guidelines, literature, and expert consensus into a knowledge manual based on themes such as "diagnosis, treatment, and intervention." Its core problems are: it fails to incorporate mature frameworks such as the Bauer health development model and Anderson's taxonomy of knowledge; knowledge organization remains at the level of thematic division, failing to reflect the scientific logic of "proactive health - three levels of prevention"; it fails to distinguish the differences in knowledge attributes, such as conflating "normal blood pressure range" (factual knowledge) with "pathological mechanisms of elevated blood pressure" (conceptual knowledge), making it difficult for users to quickly locate knowledge relevant to their specific scenarios; and its updates rely entirely on manual revisions every 1-2 years, failing to respond promptly to guideline developments or research breakthroughs, exhibiting significant lag.

[0004] The second category is knowledge graph management solutions. These solutions extract hypertension-related entities and relationships through natural language processing to construct a knowledge graph that supports personalized solution generation. Their core shortcomings are: the knowledge graph construction focuses on "entity-relationship" matching, failing to incorporate the dual perspectives of "health origins" and "pathogenesis" from Bauer's health development model, thus failing to distinguish between protective and risk factors in knowledge; knowledge types are only implicit in entity annotations, lacking a four-level classification standard based on Anderson's theory and failing to assign differentiated weights based on knowledge importance, affecting the priority of knowledge queries; updates rely on medical big data statistical feedback, lacking a specific trigger mechanism for "guideline revisions and theoretical breakthroughs," and lacking verification of knowledge's theoretical compliance, easily leading to a disconnect from proactive health concepts.

[0005] The third category is fragmented knowledge delivery solutions. This solution pushes "health tips" (such as "avoid strong tea before bed") via mobile applications, and some are matched with user data. Its core pain points are: lack of mature health management theory to support it; knowledge delivery is limited to isolated fragments, failing to form a systematic framework; lack of clear knowledge classification rules, resulting in a mix of factual, procedural, and other types, making it difficult for users to accurately filter according to their needs; updates rely on manual selection by the operations team, lacking authoritative assessment and update logs, easily pushing outdated knowledge, and unable to establish differentiated update cycles based on knowledge type.

[0006] In summary, none of the existing solutions systematically address the core issues of knowledge organization, classification, and iteration. In particular, they fail to integrate Bauer's dual perspective and Anderson's four-level classification theory, thus failing to construct a knowledge system for hypertension prevention and treatment that combines theoretical support, accurate classification, and dynamic iteration.

[0007] Based on the core features of the existing technical solutions, their main drawbacks can be summarized into the following four aspects, each of which directly affects the application value and management efficiency of hypertension prevention and treatment knowledge, and corresponds to the technical problems that this invention needs to address: First, the lack of theoretical support for knowledge organization leads to a chaotic logical hierarchy. None of the three existing solutions integrates mature theories from the fields of proactive health or knowledge classification: the static knowledge compilation solution only categorizes knowledge by themes such as "diagnosis" and "treatment," failing to reflect the dual-perspective logic of "health protection - disease prevention"; the knowledge graph management solution uses "entity-relationship" as its core framework, unable to distinguish between protective and risk factors in knowledge; and the fragmented knowledge delivery solution lacks any theoretical support, only providing isolated knowledge fragments. These deficiencies result in a lack of a unified logical thread in the knowledge system. Clinicians and community health managers need to spend extra time sorting out knowledge connections when selecting knowledge suitable for specific scenarios, reducing the efficiency of knowledge application.

[0008] Secondly, the lack of clear standards for knowledge classification leads to difficulties in accurate retrieval. Existing solutions lack systematic knowledge classification rules: static knowledge compilation schemes do not distinguish between factual, conceptual, and procedural knowledge types; knowledge graph management schemes only implicitly define knowledge types through entity attributes, without clearly defining the four-level classification standards and weight differences; fragmented knowledge push schemes lack classification rules, making it impossible for users to accurately filter knowledge according to their needs such as "understanding pathological mechanisms" or "learning practical skills." This shortcoming directly makes it difficult for different users to quickly locate the content they need, increasing the cost of knowledge acquisition.

[0009] Third, the knowledge update mechanism is inadequate, resulting in insufficient timeliness and authority. Static knowledge compilation schemes rely on manual revisions every 1-2 years, failing to respond promptly to revisions in medical guidelines or the release of core research findings, exhibiting significant lag. Knowledge graph management schemes rely solely on feedback from medical big data to trigger updates, lacking specific triggering mechanisms for "guideline revisions and theoretical breakthroughs," and lacking verification of theoretical conformity. Fragmented knowledge delivery schemes rely on manual screening and updates, lacking authority assessment and update logs, easily pushing outdated knowledge. These deficiencies lead to a disconnect between knowledge and clinical practice and the latest research, affecting the effectiveness of hypertension management.

[0010] Fourth, the poor adaptability of knowledge application leads to insufficient scenario coverage. Static knowledge compilation solutions output in the form of documented manuals, failing to adjust the knowledge presentation format according to user roles (such as community doctors or elderly patients); knowledge graph management solutions focus on personalized solution generation, neglecting the adaptation needs of knowledge in scenarios such as community lectures and patient education; fragmented knowledge push solutions only provide fragmented content, unable to support systematic knowledge dissemination. This shortcoming limits the application of hypertension prevention and treatment knowledge in multiple scenarios, making it difficult to meet the diverse needs of different user groups and reducing the practical application value of knowledge. Summary of the Invention

[0011] To overcome the shortcomings of existing technologies, the technical problem to be solved by this invention is to provide a dynamic construction method for a knowledge system of hypertension prevention and treatment based on the fusion of two theories. This method can systematically solve the core problems of knowledge organization, classification and iteration. It integrates Bauer's dual perspective and Anderson's four-level classification theory to construct a knowledge system of hypertension prevention and treatment that combines theoretical support, accurate classification and dynamic iteration.

[0012] The technical solution of this invention is: a dynamic construction method for a knowledge system for hypertension prevention and treatment based on the fusion of two theories, which includes the following steps: (1) Based on the Bauer health development model, the knowledge is decomposed from two perspectives, including the following steps: (1.1) Based on medical guidelines, pharmacopoeia, core journals and expert consensus, with medical guidelines and pharmacopoeia as Level I and core journals and expert consensus as Level II, relevant knowledge on the prevention and treatment of hypertension is collected through literature retrieval tools; (1.2) Establish a dual-perspective decomposition rule, extract protective factor knowledge based on Antonovsky's theory of health origin and label it with the health protection - proactive health tag, extract risk factor knowledge based on the disease pathology logic and label it with the disease prevention - pathological mechanism tag; (1.3) Encapsulate each piece of knowledge into a preliminary knowledge unit containing knowledge content, perspective tags, and original source to ensure that the knowledge organization has a dual-perspective logic with theoretical support; (2) Based on Anderson's knowledge classification theory, perform a four-level knowledge classification mapping, including the following steps: (2.1) Establish a four-level classification standard: For factual knowledge, mark the authoritative source number; for conceptual knowledge, mark the mechanism explanation; for procedural knowledge, mark the level of operational difficulty and applicable scenarios; for reflective knowledge, mark the context and the appropriate target audience. (2.2) Map the preliminary knowledge units to their corresponding types according to the above standards to generate standardized knowledge units, and assign initial weights based on the importance of the knowledge, thereby achieving standardization and differentiated weight control in knowledge classification; initial weights W k The formula, calculated using a weighted summation model, is as follows: ,in , , , These represent adjustment coefficients for theoretical importance, knowledge type, evidence level, and source level, respectively. (2.3) Standardized knowledge unit output: Standardized knowledge unit files are generated through annotation tools, including knowledge ID, classification label, annotation information, and initial weight; (3) Construct a dynamic update loop for the knowledge system, including the following steps: (3.1) Set three types of update trigger conditions: authoritative update trigger, practical feedback trigger, and theoretical supplement trigger; (3.2) Establish professional review; (3.3) After the review is approved, the weights will be adjusted according to the rules. When the authority of core knowledge types is improved, the weight will increase by 0.1, with a maximum of 0.5; the weight of outdated knowledge will be reduced to 0.05 and marked as to be eliminated. (3.4) Record update logs including updated knowledge ID, update reason, review expert, and effective time to form a dynamic update closed loop of trigger-review-adjustment-recording to ensure the timeliness and authority of knowledge.

[0013] Compared with the prior art, the present invention has the following beneficial technical effects: First, this invention addresses the technical problems of insufficient theoretical support and disorganized logical hierarchy in existing knowledge organization. It aims to construct a unified theoretical framework for hypertension prevention and treatment by integrating Bauer's health development model and Anderson's taxonomy of knowledge. This framework clarifies the knowledge organization logic from a dual perspective of "health origins" and "pathogenesis," as well as the classification criteria for four levels of knowledge: factual, conceptual, procedural, and reflective. This ensures the knowledge system has a clear logical thread, avoids difficulties in connecting and organizing knowledge due to theoretical deficiencies, and improves the efficiency of knowledge application.

[0014] Secondly, this invention addresses the technical problems of the lack of clear standards for knowledge classification and the difficulty in accurately acquiring knowledge in existing technologies. It establishes a systematic set of knowledge classification rules, formulating clear classification and labeling requirements and weighting rules for different types of hypertension prevention and treatment knowledge. This ensures that users can quickly locate the corresponding type of knowledge according to their needs, reducing the cost of knowledge acquisition and avoiding the time-consuming filtering problems caused by ambiguous classifications.

[0015] Third, this invention addresses the technical problems of inadequate knowledge update mechanisms, insufficient timeliness, and lack of authority in existing technologies. It requires designing a multi-dimensional knowledge update trigger mechanism covering scenarios such as medical guideline revisions, publication of core research findings, and optimization based on practical feedback. Furthermore, it establishes an update review process involving a "professional review team + theoretical compliance verification," while simultaneously recording update logs to trace the knowledge evolution process. This ensures that knowledge is synchronized with the latest clinical practices and research results, avoiding the risk of knowledge misrepresentation due to delayed updates or lack of review, and guaranteeing the authority and timeliness of the knowledge.

[0016] Fourth, this invention addresses the technical problems of poor knowledge application adaptability and insufficient scenario coverage in existing technologies. It uses a structured knowledge system catalog as the core output format, considering the needs of different user roles and application scenarios, to clarify the scenario adaptability attributes and presentation requirements of knowledge units. This allows the knowledge system to directly support knowledge dissemination and application in multiple scenarios, avoiding the limitation of scenario coverage caused by a single output format and enhancing the practical application value of knowledge.

[0017] A dynamic construction device for a knowledge system of hypertension prevention and treatment based on the fusion of two theories is also provided, which includes: The knowledge dual-perspective decomposition module, configured to perform knowledge dual-perspective decomposition based on the Bauer health development model, includes the following units: The data collection unit is based on medical guidelines, pharmacopoeias, core journals, and expert consensus. Medical guidelines and pharmacopoeias are classified as Level I, while core journals and expert consensus are classified as Level II. The unit collects knowledge related to the prevention and treatment of hypertension through literature retrieval tools. The decomposition unit establishes a dual-perspective decomposition rule, extracts protective factor knowledge based on Antonovsky's theory of health origin and labels it with the tags of health protection and proactive health, and extracts risk factor knowledge based on the logic of disease pathology and labels it with the tags of disease prevention and pathological mechanism. The encapsulation unit encapsulates each piece of knowledge into a preliminary knowledge unit containing knowledge content, perspective tags, and original sources, ensuring that the knowledge organization has a dual-perspective logic supported by theory. The knowledge four-level classification mapping module, configured to perform knowledge four-level classification mapping based on Anderson's knowledge classification theory, includes the following units: The unit is designed with a four-level classification standard: for factual knowledge, it is labeled with the authoritative source number; for conceptual knowledge, it is labeled with the mechanism explanation; for procedural knowledge, it is labeled with the level of operational difficulty and applicable scenarios; and for reflective knowledge, it is labeled with the appropriate target audience. The mapping unit maps preliminary knowledge units to corresponding types according to the aforementioned standards, generating standardized knowledge units, and assigns initial weights based on the importance of the knowledge, thus achieving standardization and differentiated weight control for knowledge classification; initial weights W k The formula, calculated using a weighted summation model, is as follows: ,in , , , These represent adjustment coefficients for theoretical importance, knowledge type, evidence level, and source level, respectively. The output unit performs standardized knowledge unit output, generating standardized knowledge unit files through annotation tools, which include knowledge ID, classification label, annotation information, and initial weights; The dynamic update module, configured to build a closed loop for dynamic updates of the knowledge system, includes the following units: The condition setting unit allows for setting three types of update trigger conditions: authoritative update trigger, practical feedback trigger, and theoretical supplementation trigger. The audit unit establishes a professional audit process. The weight adjustment unit adjusts the weights according to the rules after the review is approved. When the authority of core knowledge types increases, the weight increases by 0.1, with a maximum of 0.5; the weight of outdated knowledge is reduced to 0.05 and marked as pending elimination. The update log unit records update logs including the updated knowledge ID, update reason, review expert, and effective time, forming a dynamic update closed loop of trigger-review-adjustment-recording to ensure the timeliness and authority of knowledge. Attached Figure Description

[0018] Figure 1 A flowchart of the dynamic construction method for a knowledge system for hypertension prevention and treatment based on the dual-theory fusion according to the present invention is shown.

[0019] Figure 2 A flowchart of a specific embodiment of the dynamic construction method for a knowledge system for hypertension prevention and treatment based on dual-theory fusion according to the present invention is shown. Detailed Implementation

[0020] like Figure 1 As shown, this method for dynamically constructing a knowledge system for hypertension prevention and treatment based on the fusion of two theories includes the following steps: (1) Based on the Bauer health development model, the knowledge is decomposed from two perspectives, including the following steps: (1.1) Based on medical guidelines, pharmacopoeia, core journals and expert consensus, with medical guidelines and pharmacopoeia as Level I and core journals and expert consensus as Level II, relevant knowledge on the prevention and treatment of hypertension is collected through literature retrieval tools; (1.2) Establish a dual-perspective decomposition rule, extract protective factor knowledge based on Antonovsky's theory of health origin and label it with the health protection - proactive health tag, extract risk factor knowledge based on the disease pathology logic and label it with the disease prevention - pathological mechanism tag; (1.3) Encapsulate each piece of knowledge into a preliminary knowledge unit containing knowledge content, perspective tags, and original source to ensure that the knowledge organization has a dual-perspective logic with theoretical support; (2) Based on Anderson's knowledge classification theory, perform a four-level knowledge classification mapping, including the following steps: (2.1) Establish a four-level classification standard: For factual knowledge, mark the authoritative source number; for conceptual knowledge, mark the mechanism explanation; for procedural knowledge, mark the level of operational difficulty and applicable scenarios; for reflective knowledge, mark the context and the appropriate target audience. (2.2) Map the preliminary knowledge units to their corresponding types according to the above standards to generate standardized knowledge units, and assign initial weights based on the importance of the knowledge, thereby achieving standardization and differentiated weight control in knowledge classification; initial weights W k The formula, calculated using a weighted summation model, is as follows: ,in , , , These represent adjustment coefficients for theoretical importance, knowledge type, evidence level, and source level, respectively (this model ensures classification standardization while supporting differentiated weight control based on type and evidence differences). (2.3) Standardized knowledge unit output: Standardized knowledge unit files are generated through annotation tools, including knowledge ID, classification label, annotation information, and initial weight; (3) Construct a dynamic update loop for the knowledge system, including the following steps: (3.1) Set three types of update trigger conditions: authoritative update trigger, practical feedback trigger, and theoretical supplement trigger; (3.2) Establish professional review; (3.3) After the review is approved, the weights will be adjusted according to the rules. When the authority of core knowledge types is improved, the weight will increase by 0.1, with a maximum of 0.5; the weight of outdated knowledge will be reduced to 0.05 and marked as to be eliminated. (3.4) Record update logs including updated knowledge ID, update reason, review expert, and effective time to form a dynamic update closed loop of trigger-review-adjustment-recording to ensure the timeliness and authority of knowledge.

[0021] Compared with the prior art, the present invention has the following beneficial technical effects: First, this invention addresses the technical problems of insufficient theoretical support and disorganized logical hierarchy in existing knowledge organization. It aims to construct a unified theoretical framework for hypertension prevention and treatment by integrating Bauer's health development model and Anderson's taxonomy of knowledge. This framework clarifies the knowledge organization logic from a dual perspective of "health origins" and "pathogenesis," as well as the classification criteria for four levels of knowledge: factual, conceptual, procedural, and reflective. This ensures the knowledge system has a clear logical thread, avoids difficulties in connecting and organizing knowledge due to theoretical deficiencies, and improves the efficiency of knowledge application.

[0022] Secondly, this invention addresses the technical problems of the lack of clear standards for knowledge classification and the difficulty in accurately acquiring knowledge in existing technologies. It establishes a systematic set of knowledge classification rules, formulating clear classification and labeling requirements and weighting rules for different types of hypertension prevention and treatment knowledge. This ensures that users can quickly locate the corresponding type of knowledge according to their needs, reducing the cost of knowledge acquisition and avoiding the time-consuming filtering problems caused by ambiguous classifications.

[0023] Third, this invention addresses the technical problems of inadequate knowledge update mechanisms, insufficient timeliness, and lack of authority in existing technologies. It requires designing a multi-dimensional knowledge update trigger mechanism covering scenarios such as medical guideline revisions, publication of core research findings, and optimization based on practical feedback. Furthermore, it establishes an update review process involving a "professional review team + theoretical compliance verification," while simultaneously recording update logs to trace the knowledge evolution process. This ensures that knowledge is synchronized with the latest clinical practices and research results, avoiding the risk of knowledge misrepresentation due to delayed updates or lack of review, and guaranteeing the authority and timeliness of the knowledge.

[0024] Fourth, this invention addresses the technical problems of poor knowledge application adaptability and insufficient scenario coverage in existing technologies. It uses a structured knowledge system catalog as the core output format, considering the needs of different user roles and application scenarios, to clarify the scenario adaptability attributes and presentation requirements of knowledge units. This allows the knowledge system to directly support knowledge dissemination and application in multiple scenarios, avoiding the limitation of scenario coverage caused by a single output format and enhancing the practical application value of knowledge.

[0025] Preferably, step (1.2) includes the following sub-steps: (1.2.1) Classify the ability to improve blood pressure regulation as protective factors and the ability to increase blood pressure as risk factors, and simultaneously label them as health protection - proactive health and disease prevention - pathological mechanism; (1.2.2) Extract relevant content on prehypertension intervention from the collected literature, exclude duplicate and low-relevance information, and retain 30 core knowledge items to form a list to be decomposed; (1.2.3) Determine the knowledge attribution according to step (1.2.1), and note the intervention method or risk chain; (1.2.4) Cross-validation: A two-person team reviews the decomposition results, checks the label matching degree and the completeness of the remarks, and determines the discrepancies by combining the original literature and the manual examples to ensure that the accuracy rate is ≥95%.

[0026] Preferably, in step (2.1), the BMI control knowledge from the perspective of health origins is determined to be procedural knowledge, marked as medium in difficulty, applicable to the family / community, and its operational attributes are clearly defined; the overweight-blood pressure progression association knowledge from the perspective of pathogenesis is determined to be conceptual knowledge, marked as metabolic disorder mechanism, and its logical attributes are clearly defined.

[0027] Preferably, in step (2.2), the weights are: factual knowledge 0.3, conceptual knowledge 0.4, procedural knowledge 0.2, and reflective knowledge 0.1.

[0028] Preferably, in step (2.3), the annotation tool is developed based on Python and integrates basic interface functions. The annotation tool is developed based on Python, using the PyQt5 / Flask front-end interface framework, and includes the following modules: The text parsing module calls the spaCy and Jieba NLP libraries to complete knowledge text segmentation, terminology recognition, and keyword extraction; The automatic classification suggestion module classifies the input knowledge based on the Anderson classification feature lexicon, including: factual, conceptual, procedural, and reflective suggestions; The weight calculation module has a built-in Knowledge Weight Control Model (KWCM) to calculate the initial weights and display the scores for each item. The visual annotation interface provides drop-down selections, sliders, and radio buttons for manual confirmation of category labels and annotation information. The consistency verification module verifies the consistency of tags based on the dual-view keyword library of the Bauer model and marks conflicting items. The structured export module supports exporting in JSON, CSV, and Excel formats. The files contain knowledge IDs, category tags, perspective tags, annotation fields, and weights. The version management module records the person who made the changes, the time of the changes, and the content of each change, providing a basis for dynamic update closure.

[0029] The tool can run offline and supports batch import of literature or text data, improving the efficiency of knowledge standardization processing.

[0030] Preferably, in step (3.2), a review team is formed, consisting of a hypertension clinician with 5 or more years of experience and a theoretical researcher familiar with the Bauer model and Anderson classification, to conduct the review from three dimensions: theoretical conformity, authority verification, and conflict resolution.

[0031] Preferably, in step (3.2), an automated review system is constructed as follows: An automated review system is built, integrating natural language processing, authoritative database interfaces, and conflict comparison, with preset three review dimensions for judgment logic; the semantic matching degree of new knowledge is calculated using a built-in Bauer model dual-view keyword library and Anderson classification feature word library; ≥80% is judged as pass, <60% is marked as fail and the reason is explained; the system is connected to the CNKI core library and WHO guideline library to verify whether the source of new knowledge is Level I or II, and the evidence level is identified simultaneously; if the requirements are met, the knowledge is passed; the new knowledge is compared with the content already in the database, and in case of conflict, the content is retained according to the authority level > publication time > evidence level, while the old knowledge is eliminated; a report is generated, including the review conclusion and judgment basis; if the knowledge passes, the user jumps to step (3.3), and if the content fails, it is archived. The semantic matching degree uses cosine similarity based on a word vector model as the core indicator, and the matching degree calculation formula is as follows: , in Sim(x,y) Cosine similarity: If the matching degree is ≥80%, it is considered to be theoretically consistent; if it is <60%, it is considered not consistent and the reason is explained.

[0032] Preferably, in step (3.2), the review is synchronized in real time through the online collaboration tools of the cloud document platform.

[0033] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium. When executed, the program includes the steps of the methods of the above embodiments. The storage medium can be ROM / RAM, magnetic disk, optical disk, memory card, etc. Therefore, corresponding to the method of the present invention, the present invention also includes a dynamic construction device for a knowledge system for hypertension prevention and treatment based on dual-theory fusion. This device is typically represented in the form of functional modules corresponding to the steps of the method. The device includes: The knowledge dual-perspective decomposition module, configured to perform knowledge dual-perspective decomposition based on the Bauer health development model, includes the following units: The data collection unit is based on medical guidelines, pharmacopoeias, core journals, and expert consensus. Medical guidelines and pharmacopoeias are classified as Level I, while core journals and expert consensus are classified as Level II. The unit collects knowledge related to the prevention and treatment of hypertension through literature retrieval tools. The decomposition unit establishes a dual-perspective decomposition rule, extracts protective factor knowledge based on Antonovsky's theory of health origin and labels it with the tags of health protection and proactive health, and extracts risk factor knowledge based on the logic of disease pathology and labels it with the tags of disease prevention and pathological mechanism. The encapsulation unit encapsulates each piece of knowledge into a preliminary knowledge unit containing knowledge content, perspective tags, and original sources, ensuring that the knowledge organization has a dual-perspective logic supported by theory. The knowledge four-level classification mapping module, configured to perform knowledge four-level classification mapping based on Anderson's knowledge classification theory, includes the following units: The unit is designed with a four-level classification standard: for factual knowledge, it is labeled with the authoritative source number; for conceptual knowledge, it is labeled with the mechanism explanation; for procedural knowledge, it is labeled with the level of operational difficulty and applicable scenarios; and for reflective knowledge, it is labeled with the appropriate target audience. The mapping unit maps preliminary knowledge units to corresponding types according to the aforementioned standards, generating standardized knowledge units, and assigns initial weights based on the importance of the knowledge, thus achieving standardization and differentiated weight control for knowledge classification; initial weights W k The formula, calculated using a weighted summation model, is as follows: ,in , , , These represent adjustment coefficients for theoretical importance, knowledge type, evidence level, and source level, respectively. The output unit performs standardized knowledge unit output, generating standardized knowledge unit files through annotation tools, which include knowledge ID, classification label, annotation information, and initial weights; The dynamic update module, configured to build a closed loop for dynamic updates of the knowledge system, includes the following units: The condition setting unit allows for setting three types of update trigger conditions: authoritative update trigger, practical feedback trigger, and theoretical supplementation trigger. The audit unit establishes a professional audit process. The weight adjustment unit adjusts the weights according to the rules after the review is approved. When the authority of core knowledge types increases, the weight increases by 0.1, with a maximum of 0.5; the weight of outdated knowledge is reduced to 0.05 and marked as pending elimination. The update log unit records update logs including the updated knowledge ID, update reason, review expert, and effective time, forming a dynamic update closed loop of trigger-review-adjustment-recording to ensure the timeliness and authority of knowledge.

[0034] Preferably, the disassembly unit performs the following sub-steps: (1.2.1) Classify the ability to improve blood pressure regulation as protective factors and the ability to increase blood pressure as risk factors, and simultaneously label them as health protection - proactive health and disease prevention - pathological mechanism; (1.2.2) Extract relevant content on prehypertension intervention from the collected literature, exclude duplicate and low-relevance information, and retain 30 core knowledge items to form a list to be decomposed; (1.2.3) Determine the knowledge attribution according to step (1.2.1), and note the intervention method or risk chain; (1.2.4) Cross-validation: A two-person team reviews the decomposition results, checks the label matching degree and the completeness of the remarks, and determines the discrepancies by combining the original literature and the manual examples to ensure that the accuracy rate is ≥95%.

[0035] The core objective of this invention is to "construct a scientific, accurately classified, and dynamically iterative knowledge system for the prevention and treatment of hypertension." There are three alternative solutions that can achieve this objective, but all of them have obvious limitations and cannot replace the core value of this invention. Alternative Option 1: Single Theory Deepening (Bauer Model Only, without Integrating Anderson Theory) Based solely on the Bauer health development model, this approach subdivides the dual perspectives of "health origins and pathogenesis" into "primary, secondary, and tertiary prevention" sub-dimensions, assigning weights according to prevention level (e.g., primary prevention knowledge has a weight of 0.4) to form a theoretical framework. However, this approach fails to distinguish knowledge type attributes (e.g., classifying "BMI control steps" and "BMI-induced hypertension mechanisms" into the same dimension), making it difficult for users to filter based on "practical / logical needs." Furthermore, the weights are set only according to prevention level, without considering the functional positioning of knowledge, resulting in coarse granularity and low adaptability. Alternative Solution 2: Purely automated update triggering (no human practical feedback or theoretical breakthrough triggering) Relying on a medical big data platform, the system automatically monitors A-level evidence papers in core journals and updates to authoritative guidelines (such as WHO push notifications), triggering updates as soon as new data is captured, ensuring timeliness. However, this solution omits updates based on "practice feedback" (such as feedback from community doctors that certain knowledge is poorly applicable), leading to a disconnect between knowledge and real-world scenarios; it also cannot identify updates based on "theoretical breakthroughs" (requiring manual judgment of relevance), easily overlooking key content and resulting in insufficient update completeness. Alternative Option 3: Pure Scenario-Driven Classification (Not Classifying Knowledge Types According to Anderson's Theory) Knowledge is categorized by application scenario (community lectures / family management / clinical assistance) and weighted according to usage frequency (e.g., high-frequency knowledge has a weight of 0.4) to achieve scenario adaptation. However, this approach makes it easy for knowledge to be repeatedly categorized (e.g., "blood pressure measurement steps" is adapted to both family and community scenarios); it fails to reflect the logical hierarchy of knowledge (factual and procedural knowledge are mixed in different scenarios), making it difficult for users to understand the connections between knowledge and resulting in a confusing classification logic. In summary, the above-mentioned solutions can only partially achieve the purpose of the invention, and all suffer from drawbacks such as "coarse classification, incomplete updates, and logical imbalance." The "dual-theory fusion + multi-condition update + type-scene dual annotation" solution of this invention can avoid all defects and is irreplaceable in terms of scientific validity and practicality.

[0036] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A dynamic construction method for a knowledge system of hypertension prevention and treatment based on the fusion of two theories, characterized by: It includes the following steps: (1) Based on the Bauer health development model, the knowledge is decomposed from two perspectives, including the following steps: (1.1) Based on medical guidelines, pharmacopoeia, core journals and expert consensus, with medical guidelines and pharmacopoeia as Level I and core journals and expert consensus as Level II, relevant knowledge on the prevention and treatment of hypertension is collected through literature retrieval tools; (1.2) Establish a dual-perspective decomposition rule, extract protective factor knowledge based on Antonovsky's theory of health origin and label it with the health protection - proactive health tag, extract risk factor knowledge based on the disease pathology logic and label it with the disease prevention - pathological mechanism tag; (1.3) Encapsulate each piece of knowledge into a preliminary knowledge unit containing knowledge content, perspective tags, and original source to ensure that the knowledge organization has a dual-perspective logic with theoretical support; (2) Based on Anderson's knowledge classification theory, perform a four-level knowledge classification mapping, including the following steps: (2.1) Establish a four-level classification standard: For factual knowledge, mark the authoritative source number; for conceptual knowledge, mark the mechanism explanation; for procedural knowledge, mark the level of operational difficulty and applicable scenarios; for reflective knowledge, mark the context and the appropriate target audience. (2.2) Map the preliminary knowledge units to their corresponding types according to the above standards to generate standardized knowledge units, and assign initial weights based on the importance of the knowledge, thereby achieving standardization and differentiated weight control in knowledge classification; initial weights W k The formula, calculated using a weighted summation model, is as follows: ,in , , , These represent the moderating coefficients for theoretical importance, knowledge type, evidence level, and source level, respectively. (2.3) Standardized knowledge unit output: Standardized knowledge unit files are generated through annotation tools, including knowledge ID, classification label, annotation information, and initial weight; (3) Construct a dynamic update loop for the knowledge system, including the following steps: (3.1) Set three types of update trigger conditions: authoritative update trigger, practical feedback trigger, and theoretical supplement trigger; (3.2) Establish professional review; (3.3) After the review is approved, the weights will be adjusted according to the rules. When the authority of core knowledge types is improved, the weight will increase by 0.1, with a maximum of 0.5; the weight of outdated knowledge will be reduced to 0.05 and marked as to be eliminated. (3.4) Record update logs including updated knowledge ID, update reason, review expert, and effective time to form a dynamic update closed loop of trigger-review-adjustment-recording to ensure the timeliness and authority of knowledge.

2. The method for dynamically constructing a knowledge system for hypertension prevention and treatment based on the fusion of dual theories as described in claim 1, characterized in that: Step (1.2) includes the following sub-steps: (1.2.1) Classify the ability to improve blood pressure regulation as a protective factor and the ability to increase blood pressure as a risk factor, and simultaneously label it with the tags of health protection - proactive health and disease prevention - pathological mechanism; (1.2.2) Extract relevant content on prehypertension intervention from the collected literature, exclude duplicate and low-relevance information, and retain 30 core knowledge items to form a list to be decomposed; (1.2.3) Determine the knowledge attribution according to step (1.2.1), and note the intervention method or risk chain; (1.2.4) Cross-validation: A two-person team reviews the decomposition results, checks the label matching degree and the completeness of the remarks, and determines the discrepancies by combining the original literature and the manual examples to ensure that the accuracy rate is ≥95%.

3. The method for dynamically constructing a knowledge system for hypertension prevention and treatment based on the fusion of two theories as described in claim 2, characterized in that: In step (2.1), the BMI control knowledge from the perspective of health origins is determined to be procedural knowledge, marked as medium in difficulty, and applicable to the family / community, thus clarifying its operational attributes; the overweight-blood pressure progression association knowledge from the perspective of pathogenesis is determined to be conceptual knowledge, and the marked mechanism is explained as a metabolic disorder mechanism, thus clarifying its logical attributes.

4. The method for dynamically constructing a knowledge system for hypertension prevention and treatment based on the fusion of two theories as described in claim 3, characterized in that: In step (2.2), the weights are: factual knowledge 0.3, conceptual knowledge 0.4, procedural knowledge 0.2, and reflective knowledge 0.

1.

5. The method for dynamically constructing a knowledge system for hypertension prevention and treatment based on the fusion of two theories as described in claim 4, characterized in that: In step (2.3), the annotation tool is developed based on Python and uses the PyQt5 / Flask front-end interface framework. The tool includes the following modules: The text parsing module calls the spaCy and Jieba NLP libraries to complete knowledge text segmentation, terminology recognition, and keyword extraction; The automatic classification suggestion module classifies the input knowledge based on the Anderson classification feature lexicon, including: factual, conceptual, procedural, and reflective suggestions; The weight calculation module has a built-in Knowledge Weight Control Model (KWCM) to calculate the initial weights and display the scores for each item. The visual annotation interface provides drop-down selections, sliders, and radio buttons for manual confirmation of category labels and annotation information. The consistency verification module verifies the consistency of tags based on the dual-view keyword library of the Bauer model and marks conflicting items. The structured export module supports exporting in JSON, CSV, and Excel formats. The files contain knowledge IDs, category tags, perspective tags, annotation fields, and weights. The version management module records the person who made the changes, the time of the changes, and the content of each change, providing a basis for dynamic update closure.

6. The method for dynamically constructing a knowledge system for hypertension prevention and treatment based on the fusion of two theories as described in claim 5, characterized in that: In step (3.2), an audit team is formed, consisting of a hypertension clinician with 5 or more years of experience and a theoretical researcher familiar with the Bauer model and Anderson classification, to conduct the audit from three dimensions: theoretical conformity, authority verification, and conflict resolution.

7. The method for dynamically constructing a knowledge system for hypertension prevention and treatment based on the fusion of two theories as described in claim 5, characterized in that: In step (3.2), an automated review system is constructed, specifically as follows: An automated review system is built, integrating natural language processing, authoritative database interfaces, and conflict comparison, with preset three review dimensions for judgment logic; the semantic matching degree of new knowledge is calculated using a built-in Bauer model dual-view keyword library and Anderson classification feature word library; the system connects to the CNKI core library and WHO guideline library to verify whether the source of new knowledge is Level I or II, and simultaneously identifies the evidence level; if the requirements are met, the knowledge is approved; the new knowledge is compared with the content already in the database, and in case of conflict, the content is retained and the old knowledge is discarded based on the authority level > publication time > evidence level; a report is generated, including the review conclusion and judgment basis; if approved, the user proceeds to step (3.3), and if not approved, the content is archived; among which, the semantic matching degree uses cosine similarity based on word vector model as the core indicator, and the matching degree calculation formula is as follows: , in Sim(x,y) Cosine similarity: If the matching degree is ≥80%, it is considered to be theoretically consistent; if it is <60%, it is considered not consistent and the reason is explained.

8. The method for dynamically constructing a knowledge system for hypertension prevention and treatment based on the fusion of two theories as described in claim 6 or 7, characterized in that: In step (3.2), the review is synchronized in real time through the online collaboration tools of the cloud document platform.

9. A dynamic construction device for a knowledge system of hypertension prevention and treatment based on the fusion of two theories, characterized in that: It includes: The knowledge dual-perspective decomposition module, configured to perform knowledge dual-perspective decomposition based on the Bauer health development model, includes the following units: The data collection unit is based on medical guidelines, pharmacopoeias, core journals, and expert consensus. Medical guidelines and pharmacopoeias are classified as Level I, while core journals and expert consensus are classified as Level II. The unit collects knowledge related to the prevention and treatment of hypertension through literature retrieval tools. The decomposition unit establishes a dual-perspective decomposition rule, extracts protective factor knowledge based on Antonovsky's theory of health origin and labels it with the tags of health protection and proactive health, and extracts risk factor knowledge based on the logic of disease pathology and labels it with the tags of disease prevention and pathological mechanism. The encapsulation unit encapsulates each piece of knowledge into a preliminary knowledge unit containing knowledge content, perspective tags, and original sources, ensuring that the knowledge organization has a dual-perspective logic supported by theory. The knowledge four-level classification mapping module, configured to perform knowledge four-level classification mapping based on Anderson's knowledge classification theory, includes the following units: The unit is designed with a four-level classification standard: for factual knowledge, it is labeled with the authoritative source number; for conceptual knowledge, it is labeled with the mechanism explanation; for procedural knowledge, it is labeled with the level of operational difficulty and applicable scenarios; and for reflective knowledge, it is labeled with the appropriate target audience. The mapping unit maps preliminary knowledge units to corresponding types according to the aforementioned standards, generating standardized knowledge units, and assigns initial weights based on the importance of the knowledge, thus achieving standardization and differentiated weight control for knowledge classification; initial weights W k The formula, calculated using a weighted summation model, is as follows: ,in , , , These represent the moderating coefficients for theoretical importance, knowledge type, evidence level, and source level, respectively. The output unit performs standardized knowledge unit output, generating standardized knowledge unit files through annotation tools, which include knowledge ID, classification label, annotation information, and initial weights; The dynamic update module, configured to build a closed loop for dynamic updates of the knowledge system, includes the following units: The condition setting unit allows for setting three types of update trigger conditions: authoritative update trigger, practical feedback trigger, and theoretical supplementation trigger. The audit unit establishes a professional audit process. The weight adjustment unit adjusts the weights according to the rules after the review is approved. When the authority of core knowledge types increases, the weight increases by 0.1, with a maximum of 0.5; the weight of outdated knowledge is reduced to 0.05 and marked as pending elimination. The update log unit records update logs including the updated knowledge ID, update reason, review expert, and effective time, forming a dynamic update closed loop of trigger-review-adjustment-recording to ensure the timeliness and authority of knowledge.

10. The dynamic construction device for a knowledge system for hypertension prevention and treatment based on the fusion of dual theories as described in claim 9, characterized in that: The disassembly unit performs the following steps: (1.2.1) Classify the ability to improve blood pressure regulation as a protective factor and the ability to increase blood pressure as a risk factor, and simultaneously label it with the tags of health protection - proactive health and disease prevention - pathological mechanism; (1.2.2) Extract relevant content on prehypertension intervention from the collected literature, exclude duplicate and low-relevance information, and retain 30 core knowledge items to form a list to be decomposed; (1.2.3) Determine the knowledge attribution according to step (1.2.1), and note the intervention method or risk chain; (1.2.4) Cross-validation: A two-person team reviews the decomposition results, checks the label matching degree and the completeness of the remarks, and determines any discrepancies by combining the original literature and the manual examples to ensure that the accuracy rate is ≥95%.

Images