A diabetic foot risk assessment method and system based on knowledge graph

Abstract

The present invention relates to the field of diabetic foot, and discloses a diabetic foot risk assessment method and system based on a knowledge graph, including: collecting patient data of diabetic foot, identifying patient group characteristics, risk factors and risk characteristics of diabetic foot, building a risk prediction model for diabetic foot, extracting keywords from the risk prediction data of the risk prediction model to query the domain knowledge of diabetic foot, and building a diabetic foot ontology structure, identifying the entity objects of the ontology structure for disambiguation processing, and extracting the entity data of the entity objects from the domain knowledge, according to the disambiguation result and the entity data, integrating the stored knowledge of diabetic foot, combining the inference rules and the inference type to obtain the inference conclusion of diabetic foot, so as to identify the risk factors and risk levels of patient data, and generate a risk assessment report for diabetic foot. The present invention can improve the accuracy of diabetic foot risk assessment.

Description

Technical Field

[0001] The present invention relates to the field of diabetic foot, and in particular to a diabetic foot risk assessment method and system based on a knowledge graph. Background Art

[0002] The risk of diabetic foot refers to the possibility that diabetic patients may suffer from ischemia, loss of sensation, infection, ulcers, and even deep tissue damage in the lower limbs due to vascular and nerve lesions. With the changes in the lifestyle of modern human society, diabetic foot has become the most common complication among diabetic patients. Its high risk seriously threatens the health and quality of life of patients. In order to improve the risk prevention and control of diabetic foot, people have put forward higher requirements for the risk assessment of diabetic foot.

[0003] In the existing technology, the ankle-brachial index method is mainly used to assess the risk of diabetic foot. The blood flow in the lower limbs is assessed by measuring the ratio of ankle and upper arm blood pressure. Although this method can detect lower limb arterial blood flow disorders at an early stage, it is more suitable for assessing large blood vessel lesions and is difficult to assess microvascular lesions, which affects the accuracy of diabetic foot risk assessment. Summary of the invention

[0004] In order to solve the above problems, the present invention provides a diabetic foot risk assessment method and system based on knowledge graph, which can improve the accuracy of diabetic foot risk assessment.

[0005] In a first aspect, the present invention provides a diabetic foot risk assessment method based on a knowledge graph, comprising:

[0006] Obtaining a diabetic foot to be evaluated, and collecting patient data of the diabetic foot, identifying classification data of the patient data, analyzing patient population characteristics and risk factors of the diabetic foot based on the classification data, identifying risk characteristics of the diabetic foot according to the patient population characteristics and the risk factors, and constructing a risk prediction model for the diabetic foot in combination with the patient population characteristics, the risk factors and the risk characteristics;

[0007] Scheduling risk prediction data of the risk prediction model, extracting keywords of the diabetic foot from the risk prediction data, querying domain knowledge of the diabetic foot based on the keywords, identifying core concepts of the diabetic foot based on the domain knowledge and the keywords, and constructing an ontology structure of the diabetic foot based on the core concepts;

[0008] Identify the entity objects of the ontology structure, perform disambiguation processing on the entity objects to obtain disambiguation results, extract entity data corresponding to the entity objects from the domain knowledge, fuse the disambiguation results with the entity data to obtain fused knowledge, and store the fused knowledge to obtain stored knowledge;

[0009] According to the ontology structure and the stored knowledge, the inference rule and the inference type of the diabetic foot are set, and the diabetic foot is inferred by combining the inference rule, the inference type and the stored knowledge to obtain an inference conclusion;

[0010] Based on the reasoning conclusion, the patient data is comprehensively evaluated to obtain comprehensive evaluation data, the risk factors of diabetic foot are extracted from the comprehensive evaluation data, and the risk levels of the risk factors are marked, and the risk assessment report of the diabetic foot is generated by combining the risk factors and the risk level.

[0011] In a possible implementation of the first aspect, analyzing the patient population characteristics and risk factors of diabetic foot based on the classified data includes:

[0012] identifying data items of the categorical data, and analyzing probability distribution of the data items in the categorical data;

[0013] Extracting a maximum probability value from the probability distribution, and extracting key data items of the data items based on the maximum probability value;

[0014] According to the key data items, identifying the patient group characteristics of diabetic foot;

[0015] Based on the characteristics of the patient population, the risk factors for diabetic foot are analyzed.

[0016] In a possible implementation of the first aspect, the risk prediction model for diabetic foot is constructed by combining the patient population characteristics, the risk factors and the risk characteristics, including:

[0017] According to the characteristics of the patient group, setting the patient sample of diabetic foot;

[0018] Combining the risk factors and the risk characteristics, identifying the risk factors for diabetic foot;

[0019] analyzing the patient sample for risk variables based on the risk factors;

[0020] Calculating the indicator weight of the risk variable in the diabetic foot;

[0021] Analyzing the correlation strength between the risk variable and the risk factor according to the indicator weight;

[0022] Based on the indicator weights, determining the model type of the diabetic foot;

[0023] A risk prediction model for diabetic foot is constructed based on the risk factors and the model type.

[0024] In a possible implementation of the first aspect, identifying the core concept of diabetic foot according to the domain knowledge and the keywords includes:

[0025] Analyze the concept definition of the keyword based on the domain knowledge;

[0026] Based on the concept definition, analyzing the role of the keyword in the domain knowledge;

[0027] identifying correlations between the keywords;

[0028] Based on the correlation and the performance, the core concept of diabetic foot is identified.

[0029] In a possible implementation of the first aspect, constructing the ontology structure of diabetic foot based on the core concept includes:

[0030] Based on the core concepts, identify the scope of the diabetic foot disease;

[0031] Analyzing the hierarchical relationship between the core concepts, and setting the relationship type of the core concepts according to the hierarchical relationship;

[0032] Based on the relationship type, setting attribute restrictions for the core concept;

[0033] According to the attribute restriction, setting constraint conditions between the core concept and the relationship type;

[0034] Extracting the concept object of the core concept and analyzing the attribute items of the concept object;

[0035] According to the attribute items, the constraint conditions and the hierarchical relationship, setting a filling instance of the core concept;

[0036] The ontology structure of the diabetic foot is constructed by combining the domain scope, the hierarchical relationship and the filling instance.

[0037] In a possible implementation manner of the first aspect, performing disambiguation processing on the entity object to obtain a disambiguation result includes:

[0038] Identify an entity referent of the entity object and generate candidate entities of the entity referent;

[0039] Extracting association features between the entity referent and the candidate entity;

[0040] Based on the association features, the similarity between the entity reference and the candidate entity is calculated using the following formula:

[0041] ;

[0042] Among them, b represents the similarity between the entity referent and the candidate entity, m represents the vector dimension of the vector entity referent and the vector candidate entity, and e represents the associated feature index between the entity referent and the candidate entity. represents the e-th associated feature of the vector entity referent, Represents the e-th associated feature of the vector candidate entity;

[0043] Based on the similarity and the candidate entity, setting a link object of the entity referent;

[0044] Based on the candidate entity and the link object, the entity object is disambiguated to obtain a disambiguation result.

[0045] In a possible implementation of the first aspect, setting the inference rule and inference type of the diabetic foot according to the ontology structure and the stored knowledge includes:

[0046] Extracting the core concept of diabetic foot according to the ontology structure and the stored knowledge;

[0047] Identify the performance characteristics of the ontology structure, and identify the reasoning target of the diabetic foot based on the core concept and the performance characteristics;

[0048] In combination with the performance characteristics and the reasoning target, setting the reasoning type of the diabetic foot;

[0049] Based on the ontology structure and the reasoning target, the reasoning rules for diabetic foot are set.

[0050] In a possible implementation of the first aspect, combining the inference rule, the inference type, and the stored knowledge to reason about the diabetic foot and obtain a reasoning conclusion includes:

[0051] According to the inference rule and the inference type, configuring the inference engine of diabetic foot;

[0052] In combination with the inference rule, the inference type and the inference engine, data deduction is performed on the stored knowledge to obtain deduced data;

[0053] The derived data is verified using the following formula to obtain a verification result:

[0054] ;

[0055] in, It represents the probability of the verification result of the derived data reaching conclusion d under the known evidence f, It represents the probability of the verification result of conclusion d under the known condition c and evidence f. It represents the probability that condition c is true under the known evidence f of the derived data, P represents the derived data, and C represents the set where condition c is located;

[0056] Scheduling the reasoning explanation of the verification results in the derivation process;

[0057] According to the verification result and the reasoning explanation, the reasoning conclusion of the diabetic foot is generated.

[0058] In a possible implementation of the first aspect, combining the risk factor and the risk level to generate the risk assessment report of diabetic foot includes:

[0059] In combination with the risk factor and the risk level, scheduling the patient risk data of diabetic foot;

[0060] Analyzing the risk status of the patient with diabetic foot according to the patient risk data;

[0061] Calculating the weight level of the risk factor;

[0062] According to the weight level, the risk status of the patient is scored using the following formula to obtain a risk score result:

[0063] ;

[0064] in, represents the risk score result of the patient's risk status at time t, represents the baseline risk score of the patient's risk status at time t, The risk split representing the patient's risk status, , The risk factor coefficient representing the patient's risk status, , Risk factor values ​​representing the patient's risk status;

[0065] Setting the result explanation of the risk scoring result;

[0066] Based on the result interpretation and the risk scoring result, a risk assessment report for diabetic foot is generated.

[0067] In a second aspect, the present invention provides a diabetic foot risk assessment system based on a knowledge graph, the system comprising:

[0068] A model prediction module is used to obtain diabetic foot to be evaluated, collect patient data of the diabetic foot, identify classification data of the patient data, analyze patient population characteristics and risk factors of the diabetic foot based on the classification data, identify risk characteristics of the diabetic foot according to the patient population characteristics and the risk factors, and construct a risk prediction model for the diabetic foot in combination with the patient population characteristics, the risk factors and the risk characteristics;

[0069] an ontology construction module, for scheduling the risk prediction data of the risk prediction model, extracting keywords of the diabetic foot from the risk prediction data, querying the domain knowledge of the diabetic foot based on the keywords, identifying the core concepts of the diabetic foot based on the domain knowledge and the keywords, and constructing the ontology structure of the diabetic foot based on the core concepts;

[0070] A knowledge storage module, used for identifying entity objects of the ontology structure, performing disambiguation processing on the entity objects to obtain disambiguation results, extracting entity data corresponding to the entity objects from the domain knowledge, fusing the disambiguation results with the entity data to obtain fused knowledge, and storing the fused knowledge to obtain stored knowledge;

[0071] A knowledge reasoning module, used to set the reasoning rules and reasoning types of the diabetic foot according to the ontology structure and the stored knowledge, and to reason about the diabetic foot in combination with the reasoning rules, the reasoning type and the stored knowledge to obtain a reasoning conclusion;

[0072] The risk assessment module is used to perform a comprehensive assessment on the patient data based on the reasoning conclusion to obtain comprehensive assessment data, extract the risk factors of diabetic foot from the comprehensive assessment data, and mark the risk level of the risk factors, and generate a risk assessment report for diabetic foot by combining the risk factors and the risk level.

[0073] Compared with the prior art, the technical principle and beneficial effects of this solution are:

[0074] The embodiment of the present invention obtains the diabetic foot to be evaluated, collects the patient data of the diabetic foot, and identifies the classification data of the patient data, so as to identify the risk factors related to the occurrence and development of diabetic foot, help understand the basic characteristics of diabetic foot patients, so as to accurately assess the risk of patients developing diabetic foot, thereby constructing the risk prediction model of diabetic foot, and improving the accuracy of the model in predicting the risk of diabetic foot. At the same time, the construction results of the risk prediction model can be converted into rules and paths in the knowledge graph to enhance the interpretability of the model; further, the embodiment of the present invention schedules the risk prediction data of the risk prediction model to construct the ontology structure of diabetic foot, so as to better identify and understand the core concepts of diabetic foot, thereby constructing a comprehensive and accurate ontology structure of diabetic foot, and helping medical personnel to quickly identify the key symptoms, risk factors and treatment points of diabetic foot to reduce the risk of diabetic foot; the embodiment of the present invention obtains fused knowledge by fusing the disambiguation result with the entity data, and stores the fused knowledge to obtain stored knowledge, which can avoid the occurrence of diabetic foot due to homonymy or heteronymy. The confusion caused by synonyms can be eliminated, thereby improving the quality of the knowledge graph, realizing the interconnection of data, and ensuring that the data in the knowledge graph is consistent and accurate, thereby improving the accuracy of diabetic foot risk assessment; further, according to the ontology structure and the stored knowledge, setting the inference rules and inference types of the diabetic foot can help the system automatically identify and infer the knowledge implicit in the ontology structure, thereby discovering new facts and relationships, so that the knowledge graph can handle complex queries, improve the accuracy and comprehensiveness of diabetic foot risk assessment, and combine the stored knowledge to reason about the diabetic foot to obtain inference conclusions, which can automatically identify high-risk factors for diabetic foot from a large amount of patient data, thereby improving the accuracy of risk assessment; the embodiment of the present invention generates the risk assessment report of diabetic foot by combining the risk factors and the risk level, which can improve the accuracy of risk identification, and provide personalized risk assessment by integrating the individual data of patients, and at the same time can help diabetic patients and their families understand the prevention and care knowledge of diabetic foot, improve self-management ability, and thus prevent the occurrence of diabetic foot. Therefore, the diabetic foot risk assessment method based on knowledge graph proposed in the present invention can improve the accuracy of diabetic foot risk assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0076] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0077] Figure 1 A schematic diagram of a process of a diabetic foot risk assessment method based on a knowledge graph provided in one embodiment of the present invention;

[0078] Figure 2 A module schematic diagram of a knowledge graph-based diabetic foot risk assessment system provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0079] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0080] The embodiment of the present invention provides a diabetic foot risk assessment method based on a knowledge graph, and the execution subject of the diabetic foot risk assessment method based on the knowledge graph includes but is not limited to at least one of the electronic devices such as a server and a terminal that can be configured to execute the method provided by the embodiment of the present invention. In other words, the diabetic foot risk assessment method based on the knowledge graph can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc. The server can be an independent server, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (Content Delivery Network, CDN), and big data and artificial intelligence platforms.

[0081] See also Figure 1 FIG. 1 is a flow chart of a method for risk assessment of diabetic foot based on a knowledge graph provided by an embodiment of the present invention. Figure 1 The knowledge graph-based diabetic foot risk assessment method described in includes:

[0082] S1. Obtain the diabetic foot to be evaluated, and collect the patient data of the diabetic foot, identify the classification data of the patient data, analyze the patient group characteristics and risk factors of the diabetic foot based on the classification data, identify the risk characteristics of the diabetic foot according to the patient group characteristics and the risk factors, and construct a risk prediction model for the diabetic foot by combining the patient group characteristics, the risk factors and the risk characteristics.

[0083] The embodiment of the present invention obtains the diabetic foot to be evaluated, collects the patient data of the diabetic foot, and identifies the classification data of the patient data, so as to identify the risk factors related to the occurrence and development of diabetic foot, help understand the basic characteristics of diabetic foot patients, and thus accurately evaluate the risk of patients developing diabetic foot. The diabetic foot refers to foot infection, ulcer or deep tissue damage caused by peripheral neuropathy, peripheral arterial disease or both caused by long-term hyperglycemia in diabetic patients. The patient data refers to all information related to the diabetic foot patient, such as diabetes type, course of disease, blood sugar control level, etc. The classification data refers to the classification data obtained by systematically classifying the collected patient data according to certain standards or characteristics.

[0084] Optionally, the diabetic foot patient data collection can be obtained through an electronic health record system (EHR), and the classified data identification of the patient data can be implemented using a relational database, such as PostgreSQL, to classify and retrieve patient data through SQL query statements.

[0085] Furthermore, the embodiments of the present invention can more accurately assess the patient's diabetic foot risk and identify key pathological features to predict the risk of diabetic foot by analyzing the patient group characteristics and risk factors of diabetic foot based on the classified data. The patient group characteristics refer to the common attributes or characteristics of group members suffering from diabetic foot, such as hypertension, and the risk factors refer to the direct causes or promoting factors of diabetic foot.

[0086] As an embodiment of the present invention, the analysis of the patient group characteristics and risk factors of diabetic foot based on the classified data includes: identifying data items of the classified data, and analyzing the probability distribution of the data items in the classified data; extracting the maximum probability value in the probability distribution, and based on the maximum probability value, extracting the key data items of the data items; identifying the patient group characteristics of diabetic foot according to the key data items; and analyzing the risk factors of diabetic foot based on the patient group characteristics.

[0087] Among them, the data item refers to a single variable or attribute in the categorical data, the probability distribution refers to the probability of a random variable taking each possible value, such as the probability of different blood sugar levels in a group of diabetic patients, the maximum probability value refers to the value with the highest probability that the random variable takes a certain value, and the key data item refers to the data item with the largest probability value in the probability distribution, including age range, gender ratio, common complications, general level of blood sugar control, prevalence of foot problems, etc.

[0088] Optionally, the probability distribution analysis of the data items in the classified data can be obtained by calculating the number of data in different intervals, the maximum probability value extraction in the probability distribution can be achieved by utilizing the peak value of the probability distribution graph, and the risk factor analysis of diabetic foot based on the patient group characteristics can be obtained by Spearman correlation analysis.

[0089] The embodiment of the present invention can more accurately predict which patients have a higher risk of diabetic foot by identifying the risk characteristics of diabetic foot according to the patient group characteristics and the risk factors, and use it as an important variable in model construction to help researchers select the most influential factors to incorporate into the model, thereby improving the discrimination and calibration of the model. The risk characteristics refer to characteristic factors that play a decisive or important role in the development of diabetic foot, such as increased glycated hemoglobin levels.

[0090] Optionally, the risk feature identification of diabetic foot can be obtained through a decision tree model based on the patient group characteristics and the risk factors.

[0091] Furthermore, the embodiment of the present invention constructs the risk prediction model of diabetic foot by combining the patient group characteristics, the risk factors and the risk characteristics, which can improve the accuracy of the model in predicting the risk of diabetic foot. At the same time, the construction results of the risk prediction model can be converted into rules and paths in the knowledge graph to enhance the interpretability of the model. The risk prediction model refers to a statistical model for predicting the risk of diabetic foot.

[0092] As an embodiment of the present invention, the risk prediction model of diabetic foot is constructed by combining the patient group characteristics, the risk factors and the risk characteristics, including: setting the patient samples of diabetic foot according to the patient group characteristics; identifying the risk factors of diabetic foot in combination with the risk factors and the risk characteristics; analyzing the risk variables of the patient samples based on the risk factors; calculating the indicator weights of the risk variables in the diabetic foot; analyzing the correlation strength between the risk variables and the risk factors based on the indicator weights; determining the model type of diabetic foot based on the indicator weights; and constructing the risk prediction model of diabetic foot based on the risk factors and the model type.

[0093] Among them, the patient sample refers to a part of individuals selected from the diabetic patient group, the risk factor refers to the disease risk caused by diabetic foot, such as ulcers, the risk variable refers to the factors that can cause or increase the chance of risk accidents or expand the degree of loss, such as age, blood sugar level, blood pressure, etc., the indicator weight refers to the influence of risk variables on diabetic foot, the association strength refers to the correlation strength between risk variables and risk factors, and the model type refers to the statistical method or model framework used to construct a risk prediction model, such as the Logistic regression model and the Cox proportional risk model.

[0094] Optionally, the analysis of the strength of association between the risk variable and the risk factor based on the indicator weight can be obtained through Logistic regression analysis, the determination of the model type of diabetic foot based on the indicator weight can be achieved by interpreting and utilizing the indicator weight, and the calculation of the indicator weight of the risk variable in the diabetic foot can be obtained through the entropy method.

[0095] S2. Schedule the risk prediction data of the risk prediction model, and extract the keywords of the diabetic foot from the risk prediction data; based on the keywords, query the domain knowledge of the diabetic foot; according to the domain knowledge and the keywords, identify the core concepts of the diabetic foot; based on the core concepts, construct the ontology structure of the diabetic foot.

[0096] The embodiment of the present invention can better identify and understand the core concepts of diabetic foot by scheduling the risk prediction data of the risk prediction model and extracting the keywords of diabetic foot from the risk prediction data, thereby constructing a comprehensive and accurate diabetic foot ontology structure. The risk prediction data refers to a data set related to the risk of diabetic foot output by the risk prediction model, and the keywords refer to features or concepts that are highly correlated with diabetic foot extracted from the risk prediction data, such as glycated hemoglobin, history of foot ulcers, history of foot amputation, etc.

[0097] Furthermore, the embodiment of the present invention can identify the core concepts in the field of diabetic foot and help construct the ontology structure of diabetic foot by querying the domain knowledge of diabetic foot based on the keywords. The domain knowledge refers to professional information such as medical information, clinical practice guidelines, treatment methods, risk factors, pathological mechanisms, etc. related to diabetic foot.

[0098] Optionally, the domain knowledge query of diabetic foot based on the keywords can be obtained through a search engine, such as PubMed database.

[0099] The embodiment of the present invention identifies the core concept of diabetic foot based on the domain knowledge and the keywords, thereby clarifying the key terms related to diabetic foot and their interrelationships, and effectively integrating knowledge from different sources to form a unified knowledge framework, thereby providing a basis for constructing the ontology structure of diabetic foot. The core concept refers to the most critical and basic concept in the field or subject of diabetic foot, such as diabetes, neuropathy, infection and the like.

[0100] As an embodiment of the present invention, identifying the core concept of diabetic foot based on the domain knowledge and the keywords includes: analyzing the concept definition of the keywords based on the domain knowledge; analyzing the role performance of the keywords in the domain knowledge based on the concept definition; identifying the correlation between the keywords; and identifying the core concept of diabetic foot based on the correlation and the role performance.

[0101] Among them, the concept definition refers to the basic meaning of the keywords in the field of diabetic foot, including the characteristics, attributes, categories and examples covered by the terms; the role performance refers to the specific functions or impacts of the keywords in the field of diabetic foot, that is, the role of these keywords in the actual pathological process, diagnosis, treatment and management; the correlation refers to the mutual connection and interaction between keywords.

[0102] Optionally, the conceptual definition analysis of the keywords based on the domain knowledge can be implemented using a professional medical database, such as a PubMed database, the performance analysis of the keywords in the domain knowledge based on the conceptual definition can be obtained through correlation analysis, and the correlation identification between the keywords can be implemented using an Apriori algorithm.

[0103] Furthermore, the embodiment of the present invention can help medical personnel quickly identify the key symptoms, risk factors and treatment points of diabetic foot to reduce the risk of diabetic foot by constructing the ontology structure of diabetic foot based on the core concepts. The ontology structure refers to the formal expression of concepts and their relationships in the field of diabetic foot.

[0104] As an embodiment of the present invention, the ontology structure of diabetic foot is constructed based on the core concept, including: identifying the domain scope of diabetic foot based on the core concept; analyzing the hierarchical relationship between the core concepts, and setting the relationship type of the core concept according to the hierarchical relationship; setting the attribute restriction of the core concept based on the relationship type; setting the constraint conditions between the core concept and the relationship type according to the attribute restriction; extracting the concept object of the core concept, and analyzing the attribute items of the concept object; setting the filling instance of the core concept according to the attribute items, the constraint conditions and the hierarchical relationship; and constructing the ontology structure of diabetic foot in combination with the domain scope, the hierarchical relationship and the filling instance.

[0105] Among them, the domain scope refers to the specific knowledge domain covered by the diabetic foot ontology structure, including all concepts, entities, processes and relationships related to diabetic foot. The hierarchical relationship refers to the classification and inheritance relationship between core concepts in the ontology structure. For example, diabetic foot, as a general concept, may contain sub-concepts of neurogenic diabetic foot and ischemic diabetic foot. The relationship type refers to various semantic connections between core concepts, such as cause, accompany, treat, prevent, etc. The attribute restriction refers to the constraints on the attributes related to the core concepts. The constraints refer to the logical rules and restrictions imposed on the core concepts and their relationships in the ontology structure. The concept object refers to a specific entity in the ontology structure, such as diabetic patients, foot ulcers, vascular lesions, etc. The attribute item refers to the characteristics or states of the concept object, such as age, blood sugar level, ulcer area, etc. The filled instance refers to a specific object instance created based on the core concept, attribute item and constraint conditions, such as a specific diabetic foot patient record, including his personal information, medical history, clinical manifestations, etc.

[0106] Optionally, the hierarchical relationship analysis between the core concepts can be implemented using natural language processing methods, the relationship type setting of the core concepts based on the hierarchical relationship can be obtained through a statistical template extraction method, and the attribute restriction setting of the core concepts based on the relationship type can be implemented using a rule definition method. For example, it can be stipulated that a certain attribute can only be a specific data type, such as a date. The filling instance setting of the core concepts based on the attribute items, the constraints and the hierarchical relationship can be obtained through an ontology editor, such as the Protégé ontology editor.

[0107] S3. Identify the entity objects of the ontology structure, and perform disambiguation processing on the entity objects to obtain disambiguation results, extract entity data corresponding to the entity objects from the domain knowledge, fuse the disambiguation results with the entity data to obtain fused knowledge, and store the fused knowledge to obtain stored knowledge.

[0108] The embodiment of the present invention obtains a disambiguation result by identifying the entity objects of the ontology structure and performing disambiguation processing on the entity objects, thereby avoiding confusion caused by homonymy or synonymy, thereby improving the quality of the knowledge graph. At the same time, through the entity disambiguation technology, the search engine can more accurately understand the user's query intention and return more accurate search results. The entity object refers to a specific entity in the ontology structure, such as diabetic patients, foot ulcers, vascular lesions, etc. The disambiguation result refers to the result of the process of determining whether entity objects with the same name or similar features refer to the same entity when identifying and processing them.

[0109] As an embodiment of the present invention, the entity object is disambiguated to obtain a disambiguation result, including: identifying an entity referent of the entity object and generating a candidate entity of the entity referent; extracting association features between the entity referent and the candidate entity; based on the association features, calculating the similarity between the entity referent and the candidate entity; based on the similarity and the candidate entity, setting a link object of the entity referent; based on the candidate entity and the link object, disambiguating the entity object to obtain a disambiguation result.

[0110] Among them, the entity reference term refers to the specific name or title of the entity object, the candidate entity refers to a group of entities retrieved from the knowledge base and may match the entity reference term, the association feature refers to various association features between the entity reference term and the candidate entity, including entity attributes, contextual information, semantic features, co-occurrence frequency, dependency relationships, etc., the similarity refers to the degree of match or similarity between the entity reference term and the candidate entity, and the link object refers to the final result obtained by entity linking after determining the candidate entity that best matches the entity reference term based on the similarity calculation result.

[0111] Optionally, the generation of candidate entities for the entity referent term can be implemented using a knowledge base, the extraction of associated features between the entity referent term and the candidate entity can be obtained through a DeepDive feature extraction tool, and the setting of link objects for the entity referent term based on the similarity and the candidate entity can be implemented using a Learning to Rank Methods method.

[0112] In an optional embodiment of the present invention, based on the association feature, the similarity between the entity referent and the candidate entity is calculated using the following formula:

[0113] ;

[0114] Among them, b represents the similarity between the entity referent and the candidate entity, m represents the vector dimension of the vector entity referent and the vector candidate entity, and e represents the associated feature index between the entity referent and the candidate entity. represents the e-th associated feature of the vector entity referent, Represents the e-th associated feature of the vector candidate entity.

[0115] Furthermore, the embodiments of the present invention can extract entity data corresponding to the entity object from the domain knowledge and perform data fusion with the disambiguation result of the entity object, thereby improving the generalization ability of the knowledge graph and realizing data interconnection and interoperability. The entity data refers to data that describes and defines specific characteristics and behaviors of an entity, such as age data of diabetic patients and etiology data of diabetic foot.

[0116] Optionally, the extraction of entity data corresponding to the entity object may be achieved by using data mining and analysis tools.

[0117] The embodiment of the present invention obtains fused knowledge by fusing the disambiguation result with the entity data, and stores the fused knowledge to obtain stored knowledge, thereby ensuring that the data in the knowledge graph is consistent and accurate, thereby improving the accuracy of diabetic foot risk assessment. At the same time, the fused knowledge and stored knowledge can provide more comprehensive support for medical decision-making, help doctors and researchers better understand the risk factors of diabetic foot, and formulate more effective prevention and treatment strategies. The fused knowledge refers to a knowledge system formed by integrating and fusing the disambiguation result data of diabetic foot with the data, relationships, functions and other information of the entity data. The stored knowledge refers to the result data generated by storing the fused knowledge system in a certain form of database or knowledge base.

[0118] Optionally, the fusion processing of the disambiguation result and the entity data can be obtained through a SAP Data Services data fusion tool, and the storage of the fusion knowledge can be implemented using a graph database, such as a Neo4j property graph database.

[0119] S4. According to the ontology structure and the stored knowledge, the inference rules and inference types of the diabetic foot are set, and the diabetic foot is inferred by combining the inference rules, the inference type and the stored knowledge to obtain an inference conclusion.

[0120] The embodiment of the present invention can help the system automatically identify and infer the knowledge implicit in the ontology structure by setting the inference rules and inference types of diabetic foot according to the ontology structure and the stored knowledge, thereby discovering new facts and relationships, so that the knowledge graph can handle complex queries and improve the accuracy and comprehensiveness of diabetic foot risk assessment. The inference rules refer to a set of statements or formulas used to define the logical relationship between knowledge elements, and the inference type refers to a specific algorithm or process for applying inference rules to derive knowledge.

[0121] As an embodiment of the present invention, the embodiment of the present invention sets the reasoning rules and reasoning types of diabetic foot according to the ontology structure and the stored knowledge, including: extracting the core concept of diabetic foot according to the ontology structure and the stored knowledge; identifying the performance characteristics of the ontology structure, and identifying the reasoning target of diabetic foot based on the core concept and the performance characteristics; setting the reasoning type of diabetic foot in combination with the performance characteristics and the reasoning target; setting the reasoning rules of diabetic foot based on the ontology structure and the reasoning target.

[0122] Among them, the performance characteristics refer to the manifestation forms of the core concepts of diabetic foot in actual situations, such as neuropathy, infection, vascular lesions, etc. The reasoning goals refer to the specific goals expected to be achieved by applying reasoning rules, such as predicting the risk of diabetic foot. The reasoning type refers to the logical reasoning method or strategy used to achieve the reasoning goal, such as deductive reasoning, inductive reasoning, statistical reasoning, causal reasoning, analogical reasoning, etc.

[0123] Optionally, based on the core concept and the performance characteristics, the reasoning target identification of diabetic foot can be determined by the clinical needs and medical practices of diabetic foot, and in combination with the performance characteristics and the reasoning target, the reasoning type setting of diabetic foot can be achieved by statistical reasoning.

[0124] Furthermore, the embodiment of the present invention combines the inference rules, the inference type and the stored knowledge to reason about the diabetic foot and obtains an inference conclusion, which can automatically identify high-risk factors for diabetic foot from a large amount of patient data, thereby improving the accuracy of risk assessment. The inference conclusion refers to the result obtained by logically deducing the data and knowledge related to diabetic foot by applying inference rules and inference types.

[0125] As an embodiment of the present invention, the combination of the inference rule, the inference type and the stored knowledge is used to reason about the diabetic foot to obtain an inference conclusion, including: configuring the inference engine of the diabetic foot according to the inference rule and the inference type; performing data deduction on the stored knowledge in combination with the inference rule, the inference type and the inference engine to obtain deduced data; performing verification processing on the deduced data to obtain a verification result; scheduling the reasoning interpretation of the verification result in the derivation process; and generating the reasoning conclusion of the diabetic foot according to the verification result and the reasoning interpretation.

[0126] Among them, the inference engine refers to a component that can apply predefined inference rules and inference types to process input data to derive new conclusions or knowledge, the derived data refers to data generated by the inference engine based on stored knowledge and inference rules, the inference explanation refers to a detailed description of the logic, rules and steps applied by the inference engine in the deduction process, and the verification result refers to the derived data result that has been verified.

[0127] Optionally, according to the inference rule and the inference type, the inference engine configuration of diabetic foot can be implemented using Rule Engine, and the scheduling of the inference interpretation of the verification data in the derivation process can be obtained through the metadata recorded in the inference process.

[0128] In an optional embodiment of the present invention, the derived data is verified using the following formula to obtain a verification result:

[0129] ;

[0130] in, It represents the probability of the verification result of the derived data reaching conclusion d under the known evidence f, It represents the probability of the verification result of conclusion d under the known condition c and evidence f. It represents the probability that condition c is true under the known evidence f of the derived data, P represents the derived data, and C represents the set to which condition c belongs.

[0131] S5. Based on the reasoning conclusion, the patient data is comprehensively evaluated to obtain comprehensive evaluation data, the risk factors of diabetic foot are extracted from the comprehensive evaluation data, and the risk levels of the risk factors are marked, and a risk assessment report of the diabetic foot is generated by combining the risk factors and the risk level.

[0132] The embodiment of the present invention obtains comprehensive evaluation data by comprehensively evaluating the patient data based on the reasoning conclusion, and can conduct in-depth analysis of the patient data to discover potential connections and patterns between the data, thereby identifying various risk factors related to diabetic foot and providing personalized risk assessment for each patient. The comprehensive evaluation data refers to data that can comprehensively reflect the current health status of diabetic foot patients.

[0133] Optionally, the comprehensive evaluation data of the patient data based on the inference conclusion can be implemented using the knowledge base of the knowledge graph.

[0134] Furthermore, the embodiment of the present invention can accurately identify the patient's diabetic foot risk by extracting the risk factors of diabetic foot from the comprehensive assessment data and marking the risk levels of the risk factors, and propose personalized intervention measures based on the risk levels to improve the efficiency and accuracy of diabetic foot risk assessment. The risk factors refer to conditions that can prompt or cause risk events to occur, as well as conditions that cause increased and expanded losses when risk events occur. The risk level refers to dividing risks into different levels based on the characteristics and degree of risks.

[0135] Optionally, the extraction of the risk factors of diabetic foot from the comprehensive assessment data may be achieved by using text mining technology, such as co-word analysis technology, and the risk level markers of the risk factors may be obtained through a risk grading system.

[0136] The embodiment of the present invention generates the risk assessment report of diabetic foot by combining the risk factors and the risk levels, thereby improving the accuracy of risk identification, and providing personalized risk assessment by integrating the patient's individual data. At the same time, it can help diabetic patients and their families understand the prevention and care knowledge of diabetic foot and improve self-management capabilities, thereby preventing the occurrence of diabetic foot. The risk assessment report refers to an assessment conclusion drawn by comprehensively evaluating the data of diabetic foot patients based on the reasoning conclusion of the knowledge graph.

[0137] As an embodiment of the present invention, the risk assessment report of diabetic foot is generated by combining the risk factors and the risk level, including: scheduling the patient risk data of diabetic foot in combination with the risk factors and the risk level; analyzing the patient risk status of diabetic foot based on the patient risk data; calculating the weight level of the risk factor; performing risk scoring on the patient risk status based on the weight level to obtain a risk scoring result; setting a result interpretation of the risk scoring result; and generating the risk assessment report of diabetic foot based on the result interpretation and the risk scoring result.

[0138] Among them, the patient risk data refers to all data related to patients with diabetic foot, the patient risk status refers to the patient's current health status and the potential risk of developing diabetic foot in the future, the weight level refers to a quantitative indicator of the relative importance or influence assigned to each risk factor, the risk score result refers to the patient's overall risk score calculated after comprehensively considering all risk factors and their weight levels, and the result interpretation refers to a detailed description and explanation of the risk score results, including the specific meaning of the score, the reasoning behind the score, and the specific impact of the score on the patient's health.

[0139] Optionally, the result interpretation setting of the risk scoring result can be implemented by using a risk assessment report generator, and the weight level calculation of the risk factor can be obtained by a LASSO regression function.

[0140] In an optional embodiment of the present invention, the risk status of the patient is scored according to the weight level using the following formula to obtain a risk scoring result:

[0141] ;

[0142] in, represents the risk score result of the patient's risk status at time t, represents the baseline risk score of the patient's risk status at time t, The risk split representing the patient's risk status, , The risk factor coefficient representing the patient's risk status, , The risk factor value that represents the patient's risk status.

[0143] It can be seen that the embodiment of the present invention can identify the risk factors related to the occurrence and development of diabetic foot by acquiring the diabetic foot to be evaluated, collecting the patient data of the diabetic foot, and identifying the classification data of the patient data, so as to help understand the basic characteristics of diabetic foot patients, so as to accurately assess the risk of patients developing diabetic foot, thereby constructing the risk prediction model of diabetic foot, and improving the accuracy of the model in predicting the risk of diabetic foot. At the same time, the construction results of the risk prediction model can be converted into rules and paths in the knowledge graph to enhance the interpretability of the model; further, the embodiment of the present invention can better identify and understand the core concepts of diabetic foot by scheduling the risk prediction data of the risk prediction model to construct the ontology structure of diabetic foot, thereby constructing a comprehensive and accurate ontology structure of diabetic foot, helping medical personnel to quickly identify the key symptoms, risk factors and treatment points of diabetic foot to reduce the risk of diabetic foot; the embodiment of the present invention obtains fused knowledge by fusing the disambiguation result with the entity data, and stores the fused knowledge to obtain stored knowledge, which can avoid the occurrence of diabetic foot due to homonymy or heteronymy. The confusion caused by synonyms can be eliminated, thereby improving the quality of the knowledge graph, realizing the interconnection of data, and ensuring that the data in the knowledge graph is consistent and accurate, thereby improving the accuracy of diabetic foot risk assessment; further, the embodiment can help the system automatically identify and infer the knowledge implicit in the ontology structure by setting the inference rules and inference types of the diabetic foot according to the ontology structure and the stored knowledge, thereby discovering new facts and relationships, so that the knowledge graph can handle complex queries, improve the accuracy and comprehensiveness of diabetic foot risk assessment, and combine the stored knowledge to reason about the diabetic foot to obtain inference conclusions, and automatically identify high-risk factors for diabetic foot from a large amount of patient data, thereby improving the accuracy of risk assessment; the embodiment of the present invention can improve the accuracy of risk identification by combining the risk factors and the risk level to generate the risk assessment report of diabetic foot, and provide personalized risk assessment by integrating the individual data of patients, and at the same time can help diabetic patients and their families understand the prevention and care knowledge of diabetic foot, improve self-management ability, and thus prevent the occurrence of diabetic foot. Therefore, the diabetic foot risk assessment method based on knowledge graph proposed in the present invention can improve the accuracy of diabetic foot risk assessment.

[0144] like Figure 2 Shown is a system functional module diagram of diabetic foot risk assessment based on knowledge graph of the present invention.

[0145] The diabetic foot risk assessment system 200 based on knowledge graph of the present invention can be installed in an electronic device. According to the functions implemented, the diabetic foot risk assessment system based on knowledge graph can include a model prediction module 201, an ontology construction module 202, a knowledge storage module 203, a knowledge reasoning module 204 and a risk assessment module 205. The module of the present invention can also be called a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0146] In the embodiment of the present invention, the functions of each module / unit are as follows:

[0147] The model prediction module 201 is used to obtain the diabetic foot to be evaluated, collect the patient data of the diabetic foot, identify the classification data of the patient data, analyze the patient group characteristics and risk factors of the diabetic foot based on the classification data, identify the risk characteristics of the diabetic foot according to the patient group characteristics and the risk factors, and construct the risk prediction model of the diabetic foot in combination with the patient group characteristics, the risk factors and the risk characteristics;

[0148] An ontology construction module 202 is used to schedule the risk prediction data of the risk prediction model, extract the keywords of the diabetic foot from the risk prediction data, query the domain knowledge of the diabetic foot based on the keywords, identify the core concepts of the diabetic foot based on the domain knowledge and the keywords, and construct the ontology structure of the diabetic foot based on the core concepts;

[0149] The knowledge storage module 203 is used to identify the entity objects of the ontology structure, perform disambiguation processing on the entity objects to obtain disambiguation results, extract entity data corresponding to the entity objects from the domain knowledge, fuse the disambiguation results with the entity data to obtain fused knowledge, and store the fused knowledge to obtain stored knowledge;

[0150] The knowledge reasoning module 204 is used to set the reasoning rules and reasoning types of the diabetic foot according to the ontology structure and the stored knowledge, and to reason about the diabetic foot in combination with the reasoning rules, the reasoning types and the stored knowledge to obtain a reasoning conclusion;

[0151] The risk assessment module 205 is used to perform a comprehensive assessment on the patient data based on the reasoning conclusion to obtain comprehensive assessment data, extract the risk factors of diabetic foot from the comprehensive assessment data, and mark the risk level of the risk factors, and generate a risk assessment report for diabetic foot by combining the risk factors and the risk level.

[0152] In detail, each module in the diabetic foot risk assessment system 200 based on the knowledge graph in the embodiment of the present invention is used in the same manner as described above. Figure 1 The method for diabetic foot risk assessment based on knowledge graph described in the present invention has the same technical means and can produce the same technical effects, so it will not be repeated here.

[0153] 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 diabetic foot risk assessment method based on knowledge graph, characterized in that: The method comprises: Obtaining a diabetic foot to be evaluated, and collecting patient data of the diabetic foot, identifying classification data of the patient data, analyzing patient population characteristics and risk factors of the diabetic foot based on the classification data, identifying risk characteristics of the diabetic foot according to the patient population characteristics and the risk factors, and constructing a risk prediction model for the diabetic foot in combination with the patient population characteristics, the risk factors and the risk characteristics; Scheduling risk prediction data of the risk prediction model, extracting keywords of the diabetic foot from the risk prediction data, querying domain knowledge of the diabetic foot based on the keywords, identifying core concepts of the diabetic foot based on the domain knowledge and the keywords, and constructing an ontology structure of the diabetic foot based on the core concepts; Identify the entity objects of the ontology structure, perform disambiguation processing on the entity objects to obtain disambiguation results, extract entity data corresponding to the entity objects from the domain knowledge, fuse the disambiguation results with the entity data to obtain fused knowledge, and store the fused knowledge to obtain stored knowledge; According to the ontology structure and the stored knowledge, the inference rule and the inference type of the diabetic foot are set, and the diabetic foot is inferred by combining the inference rule, the inference type and the stored knowledge to obtain an inference conclusion; Based on the inference conclusion, the patient data is comprehensively evaluated to obtain comprehensive evaluation data, the risk factors of diabetic foot are extracted from the comprehensive evaluation data, and the risk levels of the risk factors are marked, and a risk assessment report of diabetic foot is generated by combining the risk factors and the risk levels; The analyzing the patient population characteristics and risk factors of diabetic foot based on the classified data includes: identifying data items of the categorical data, and analyzing probability distribution of the data items in the categorical data; Extracting a maximum probability value from the probability distribution, and extracting key data items of the data items based on the maximum probability value; According to the key data items, identifying the patient group characteristics of diabetic foot; Analyze the risk factors of diabetic foot based on the characteristics of the patient population; The risk prediction model for diabetic foot is constructed by combining the patient population characteristics, the risk factors and the risk characteristics, including: According to the characteristics of the patient group, setting the patient sample of diabetic foot; Combining the risk factors and the risk characteristics, identifying the risk factors for diabetic foot; analyzing the patient sample for risk variables based on the risk factors; Calculating the indicator weight of the risk variable in the diabetic foot; Analyzing the correlation strength between the risk variable and the risk factor according to the indicator weight; Based on the indicator weights, determining the model type of the diabetic foot; Constructing the risk prediction model of diabetic foot according to the risk factors and the model type; The identifying the core concepts of diabetic foot according to the domain knowledge and the keywords includes: Analyze the concept definition of the keyword based on the domain knowledge; Based on the concept definition, analyzing the role of the keyword in the domain knowledge; identifying correlations between the keywords; According to the association and the performance, the core concept of diabetic foot is identified; The risk variables refer to factors that can cause or increase the chance of risk accidents or expand the extent of losses. The risk factors refer to the direct causes or promoting factors of diabetic foot. The risk characteristics refer to characteristic factors that play a decisive or important role in the development of diabetic foot. The risk factors refer to conditions that can promote or cause risk events to occur, as well as conditions that cause losses to increase and expand when risk events occur.

2. The method according to claim 1, characterized in that Based on the core concept, the ontology structure of diabetic foot is constructed, including: Based on the core concepts, identify the scope of the diabetic foot disease; Analyzing the hierarchical relationship between the core concepts, and setting the relationship type of the core concepts according to the hierarchical relationship; Based on the relationship type, setting attribute restrictions for the core concept; According to the attribute restriction, setting constraint conditions between the core concept and the relationship type; Extracting the concept object of the core concept and analyzing the attribute items of the concept object; According to the attribute items, the constraint conditions and the hierarchical relationship, setting a filling instance of the core concept; The ontology structure of the diabetic foot is constructed by combining the domain scope, the hierarchical relationship and the filling instance.

3. The method according to claim 1, characterized in that: The performing disambiguation processing on the entity object to obtain a disambiguation result includes: Identify an entity referent of the entity object and generate candidate entities of the entity referent; Extracting association features between the entity referent and the candidate entity; Based on the association features, the similarity between the entity reference and the candidate entity is calculated using the following formula: ； Among them, b represents the similarity between the entity referent and the candidate entity, m represents the vector dimension of the vector entity referent and the vector candidate entity, j represents the associated feature index between the entity referent and the candidate entity, represents the jth associated feature of the vector entity referent, Represents the jth associated feature of the vector candidate entity; Based on the similarity and the candidate entity, setting a link object of the entity referent; Based on the candidate entity and the link object, the entity object is disambiguated to obtain a disambiguation result.

4. The method according to claim 1, characterized in that: According to the ontology structure and the stored knowledge, the inference rules and inference types of the diabetic foot are set, including: Extracting the core concept of diabetic foot according to the ontology structure and the stored knowledge; Identify the performance characteristics of the ontology structure, and identify the reasoning target of the diabetic foot based on the core concept and the performance characteristics; In combination with the performance characteristics and the reasoning target, setting the reasoning type of the diabetic foot; Based on the ontology structure and the reasoning target, the reasoning rules for diabetic foot are set.

5. The method according to claim 1, characterized in that The combining of the inference rule, the inference type and the stored knowledge to infer the diabetic foot to obtain an inference conclusion includes: According to the inference rule and the inference type, configuring the inference engine of diabetic foot; In combination with the inference rule, the inference type and the inference engine, data deduction is performed on the stored knowledge to obtain deduced data; The derived data is verified using the following formula to obtain a verification result: ； in, It represents the probability of the verification result of the derived data reaching conclusion d under the known evidence f, It represents the probability of the verification result of conclusion d under the known condition c and evidence f. represents the probability that condition c is true given the known evidence f of the derived data, Indicates the set where condition c is located; Scheduling the reasoning explanation of the verification results in the derivation process; According to the verification result and the reasoning explanation, the reasoning conclusion of the diabetic foot is generated.

6. The method according to claim 1, characterized in that The step of combining the risk factors and the risk levels to generate a risk assessment report for diabetic foot includes: In combination with the risk factor and the risk level, scheduling the patient risk data of diabetic foot; Analyzing the risk status of the patient with diabetic foot according to the patient risk data; Calculating the weight level of the risk factor; According to the weight level, the risk status of the patient is scored using the following formula to obtain a risk score result: ； in, represents the risk score result of the patient's risk status at time t, represents the baseline risk score of the patient's risk status at time t, a risk score representing the patient's risk status, , The risk factor coefficient representing the patient's risk status, , Risk factor values ​​representing the patient's risk status; Setting a result explanation for the risk scoring result; Based on the result interpretation and the risk scoring result, a risk assessment report for diabetic foot is generated.

7. A system for diabetic foot risk assessment based on knowledge graph, the system implementing the method according to claim 1, characterized in that: The system comprises: A model prediction module is used to obtain diabetic foot to be evaluated, collect patient data of the diabetic foot, identify classification data of the patient data, analyze patient population characteristics and risk factors of the diabetic foot based on the classification data, identify risk characteristics of the diabetic foot according to the patient population characteristics and the risk factors, and construct a risk prediction model for the diabetic foot in combination with the patient population characteristics, the risk factors and the risk characteristics; an ontology construction module, for scheduling the risk prediction data of the risk prediction model, extracting keywords of the diabetic foot from the risk prediction data, querying the domain knowledge of the diabetic foot based on the keywords, identifying the core concepts of the diabetic foot based on the domain knowledge and the keywords, and constructing the ontology structure of the diabetic foot based on the core concepts; A knowledge storage module, used for identifying entity objects of the ontology structure, performing disambiguation processing on the entity objects to obtain disambiguation results, extracting entity data corresponding to the entity objects from the domain knowledge, fusing the disambiguation results with the entity data to obtain fused knowledge, and storing the fused knowledge to obtain stored knowledge; A knowledge reasoning module, used to set the reasoning rules and reasoning types of the diabetic foot according to the ontology structure and the stored knowledge, and to reason about the diabetic foot in combination with the reasoning rules, the reasoning type and the stored knowledge to obtain a reasoning conclusion; The risk assessment module is used to perform a comprehensive assessment on the patient data based on the reasoning conclusion to obtain comprehensive assessment data, extract the risk factors of diabetic foot from the comprehensive assessment data, and mark the risk level of the risk factors, and generate a risk assessment report for diabetic foot by combining the risk factors and the risk level.

Images