Construction method of health and rehabilitation robot knowledge graph based on ai technology

Abstract

The application relates to the technical field of knowledge graph construction, in particular to a health and rehabilitation robot knowledge graph construction method based on AI technology, which comprises the following steps: based on health and rehabilitation related data, AI analysis data is utilized to extract entity data items, key entities in the data set are identified, corresponding attributes are extracted, repeated entity information is combined, data redundancy is reduced, and entity identification information is obtained. Through the utilization of the AI technology, the recognition accuracy of the entity data items and the dynamic updating of the relationships between the entities are strengthened, the health and rehabilitation related data is analyzed and applied in a deeper level, continuous entity identification, entity relationship strengthening, dynamic data tracking and graph adaptive updating are adopted in the processing logic, the utilization efficiency of the data and the accuracy of the information are enhanced, through the continuous evaluation and optimization of the entity attributes and the relationships, the knowledge graph can reflect the latest health and rehabilitation data, and the knowledge graph can maintain the latest state in the changing health and rehabilitation environment.

Description

Technical Field

[0001] This invention relates to the field of knowledge graph construction technology, and in particular to a method for constructing a knowledge graph for a health and wellness robot based on AI technology. Background Technology

[0002] The field of knowledge graph construction technology involves extracting information from various data sources and transforming it into structured knowledge for further querying and analysis. A knowledge graph is essentially a complex data structure, typically represented as a graph, containing entities, entity attributes, and relationships between entities. Entities can be specific objects, concepts, or events, while the relationships between them define their interrelationships, such as belonging, influencing, or interacting. The construction of a knowledge graph typically involves steps such as data acquisition, information extraction, knowledge abstraction, and knowledge fusion, with the aim of generating a rich and dynamic knowledge base that can support complex queries and decision support systems.

[0003] The knowledge graph construction method for elderly care robots involves designing and implementing methods for generating and managing structured knowledge graphs related to rehabilitation and elderly care. At the heart of this topic is creating a knowledge base that elderly care robots can utilize to enhance their service quality and efficiency, such as in providing daily care, health monitoring, and personalized medical advice. By integrating patients' health records, treatment histories, and relevant medical knowledge, the knowledge graph enables robots to better understand users' specific needs and provide more precise services. Furthermore, the construction of this knowledge graph also supports robots in exhibiting greater autonomy and adaptability when handling complex situations and making decisions.

[0004] Traditional knowledge graph construction methods lack sufficient dynamic update mechanisms, making it difficult for the constructed knowledge graph to reflect the latest developments in a timely manner once the environment changes. Traditional methods rely on static datasets for entity recognition and the definition of relationships between entities, which reveals their limitations when facing constantly changing healthcare scenarios. This static approach restricts the application scope and effectiveness of the knowledge graph, failing to fully utilize newly collected data for real-time updates, thus affecting the robot's responsiveness to user needs and service quality. For example, without a real-time update mechanism, healthcare robots may be unable to identify the relationship between newly emerging symptoms and treatment methods, resulting in inaccurate medical recommendations and an inability to effectively support the personalized care needs of the elderly. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing a method for constructing a knowledge graph for health and wellness robots based on AI technology.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a method for constructing a knowledge graph for a health and wellness robot based on AI technology, comprising the following steps: S1: Based on health and wellness related data, AI is used to analyze the data, extract entity data items, identify key entities in the dataset, extract corresponding attributes, merge duplicate entity information, reduce data redundancy, and obtain entity identification information. S2: Based on the entity identification information, analyze the interaction data between entities, evaluate the connections between entities, including the connection between treatment and symptoms, assess the strength of entity relationships, and obtain basic entity relationship information; S3: Based on the basic entity relationship information, analyze and remove entity pairs that do not fit the health and wellness scenario, track the changing trend of entity relationships, update the relationship strength between entities, match the current actual situation, optimize the accuracy and practicality of the data, and generate optimized entity relationship information; S4: Based on the optimized entity relationship information, import the relationships between entities into the health and wellness knowledge graph, continuously track new data and user feedback in the health and wellness environment, and adjust the addition interval of new entities according to the frequency of receiving new data and the frequency of user feedback, automatically introduce new entities, and update the attributes and relationship strength of existing entities to match the changes in the health and wellness environment, thereby obtaining a dynamically updated health and wellness knowledge graph.

[0007] The present invention improves upon the method for identifying key entities in the dataset as follows: S111: Based on health and wellness-related data, AI is used to analyze the data and extract entity data items, including user information, symptom descriptions, and treatment history; S112: Based on the entity data item, using the formula: ; Calculate word frequency parameters ,in, It is the frequency of occurrence of terms in the dataset. It is the total number of words in the dataset. It is the number of documents in which the term appears in the differentiated documents. Indicates the term frequency parameter; S113: Based on the word frequency parameters The system compares the frequency parameters with a preset word frequency threshold, marks entities whose word frequency parameters exceed the preset word frequency threshold as key entities, and uses AI to analyze and extract the attributes corresponding to the key entities to obtain the key entity recognition results.

[0008] The present invention is improved in that the step of obtaining the entity recognition information is as follows: S121: Based on the key entity identification results, for each entity, the following formula is used: ; Calculate each entity identifier ,in, It is a hash function. It is an abbreviation of entity name. For timestamps and source code; S122: Based on each entity identifier Through the formula: ; Calculate entity frequency, where, express The number of times it appears, It is the total number of entities in the dataset; S123: Based on the entity frequency, compare it with a preset frequency threshold, merge entity information whose entity frequency exceeds the preset frequency threshold, reduce data redundancy, and obtain entity recognition information.

[0009] The present invention is improved in that the method for evaluating the relationship between entities is as follows: S211: Based on the entity recognition information, traverse the already recognized entity data using the formula: ; Calculate the co-occurrence frequency of entity pairs, where, It is an entity pair and The number of documents that appear together This represents the total number of documents in the dataset. For entity pairs and The co-occurrence frequency; S212: Based on the co-occurrence frequency of the entity pairs, compare it with a preset co-occurrence threshold, mark entity pairs with a co-occurrence frequency higher than the preset co-occurrence threshold as associated, and extract the relationship type between entities through AI to obtain the relationship information between entities.

[0010] The present invention is improved in that the step of obtaining the basic entity relationship information is as follows: S221: Based on the relationship information between the entities, extract the co-occurrence frequency of the entity pairs and calculate the product of the number of documents in which the entity pairs appear in the dataset to obtain the relationship strength association information; S222: Based on the aforementioned relationship strength association information, using the formula: ; Evaluate the strength of the relationship between entity pairs to obtain basic entity relationship information, where... For entity pairs and The strength of the relationship It is an entity pair and The product of the number of documents appearing in the dataset. This is the magnification factor.

[0011] The present invention is improved in that the method for tracking the changing trend of entity relationships is as follows: S311: Based on the basic entity relationship information, AI is used to analyze the relationships between current entities, and entity pairs that do not conform to the health and wellness scenario are eliminated to obtain the filtered entity pairs; S312: Based on the filtered entity pairs, extract the entity pair relationship strength data over a period of time, using the formula: ; Calculate the degree of change in the strength of entity-to-entity relationships to obtain information on the trend of changes in entity-to-entity relationships. Represents entity pairs and The degree of change in the strength of the relationship at the two points in time. Indicates the current entity pair and The strength of the relationship This indicates the entity pair in the previous state. and The strength of the relationship For time-sensitive parameters, To indicate the current time point, For reference time points, is the base of the natural logarithm.

[0012] The present invention is improved in that the step of obtaining the optimized entity relationship information is as follows: S321: Based on the entity relationship change trend information, extract the degree of change in the relationship strength of entity pairs and the current relationship strength of entity pairs to obtain updated association information; S322: Based on the updated association information, using the formula: ; Calculate the updated entity pair relationship strength and generate optimized entity relationship information, where... Is the current entity pair and The strength of the relationship Represents entity pairs and The degree of change in the strength of the relationship at the two points in time. and These are the weighting coefficients. This is the logarithmic scaling factor. For the updated entity pair and The strength of the relationship.

[0013] The present invention is improved in that the method for adjusting the addition interval of new entities is as follows: S411: Based on the optimized entity relationship information, import the relationships between entities into the health and wellness knowledge graph, continuously track new data and user feedback in the health and wellness environment, and obtain interval association information; S412: Based on the aforementioned interval association information, using the formula: ; Calculate the adjusted entity addition interval to obtain the interval adjustment result, where, It is the frequency at which new data is received. For user feedback frequency, As a balance factor, Add spacing to the base entity. Add spacing to the adjusted entity.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, AI technology is used to enhance the recognition accuracy of entity data items and the dynamic updating of relationships between entities, enabling deeper analysis and application of health and wellness-related data. The processing logic employs continuous entity recognition, entity relationship enhancement, dynamic data tracking, and adaptive graph updates, improving data utilization efficiency and information accuracy. Through continuous evaluation and optimization of entity attributes and relationships, the knowledge graph reflects the latest health and wellness data, improving the targeting and personalization of robot services, enhancing data usability, and strengthening the graph's adaptability. This allows the graph to maintain its current state in the ever-changing health and wellness environment, thereby providing more accurate decision support.

[0015] The constructed health and wellness robot knowledge graph can provide personalized service recommendations, suggesting suitable health and wellness plans and activities based on the user's health status, rehabilitation history, and personal preferences. The knowledge graph enables the robot to engage in intelligent interaction, accurately answering users' inquiries about treatment and rehabilitation, improving the relevance and practicality of the services. In addition, the knowledge graph supports symptom association assessment and early warning functions, helping the robot to promptly identify health risks corresponding to user behaviors and suggest appropriate measures. The robot can use the information in the knowledge graph to provide users with health education, rehabilitation training tutorials, and daily life guidance, enhancing users' health awareness and self-management capabilities, and optimizing users' health and wellness experience and quality of life. Attached Figure Description

[0016] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a flowchart illustrating the process of identifying key entities in a dataset according to the present invention; Figure 3 This is a flowchart illustrating how the present invention obtains entity recognition information; Figure 4 This is a flowchart illustrating the process of evaluating relationships between entities in this invention; Figure 5 This is a flowchart illustrating how the present invention obtains basic entity relationship information; Figure 6 This is a flowchart illustrating the changing trends of entity relationships in this invention. Figure 7 This is a flowchart illustrating how the optimized entity relationship information is obtained in this invention. Figure 8 This is a flowchart illustrating the process of adjusting the addition interval of new entities in this invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0018] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0019] Please see Figure 1 This invention provides a technical solution: a method for constructing a knowledge graph for elderly care robots based on AI technology, comprising the following steps: S1: Based on health and wellness related data, AI is used to analyze the data and extract entity data items, including user information, symptom descriptions and treatment history. Semantic analysis of keywords and concepts is performed to identify key entities in the dataset and extract their corresponding attributes. Entity matching technology is applied to find the same entities in the dataset, merge duplicate entity information, reduce data redundancy, and obtain entity recognition information. S2: Based on entity recognition information, using entity relationship recognition technology, analyze the interaction data between entities, evaluate the relationship between entities, including the relationship between treatment and symptoms, assess the strength of entity relationship, and obtain basic entity relationship information; S3: Based on basic entity relationship information, analyze and remove entity pairs that do not fit the health and wellness scenario, track the changing trend of entity relationships, update the relationship strength between entities, match the current actual situation, optimize the accuracy and practicality of the data, and generate optimized entity relationship information; S4: Based on the optimized entity relationship information, the relationships between entities are imported into the health and wellness knowledge graph. New data and user feedback in the health and wellness environment are continuously tracked. The interval for adding new entities is adjusted according to the frequency of new data reception and user feedback. New entities are automatically introduced, and the attributes and relationship strength of existing entities are updated to match the changes in the health and wellness environment, resulting in a dynamically updated health and wellness knowledge graph.

[0020] Entity identification information includes entity name, entity type (such as symptom, treatment, patient) and entity attributes (such as duration, treatment response). Basic entity relationship information includes direct contact points between entities, contact type (such as impact, response, belonging) and weight score for each type of contact. Optimized entity relationship information includes updated entity relationships and adjusted relationship weights. Dynamically updated health and wellness knowledge graph includes health and wellness-related entity views, dynamic relationships between entities and integration of real-time data feedback.

[0021] Please see Figure 2 The method for identifying key entities in a dataset is as follows: S111: Based on health and wellness-related data, AI is used to analyze the data and extract entity data items, including user information, symptom descriptions, and treatment history; S112: Based on entity data items, using the formula: ; Calculate word frequency parameters ,in, It is the frequency of occurrence of terms in the dataset. It is the total number of words in the dataset. It is the number of documents in which the term appears in the differentiated documents. Indicates the term frequency parameter; S113: Based on term frequency parameters The system compares the frequency parameters with a preset word frequency threshold, marks entities whose word frequency parameters exceed the preset word frequency threshold as key entities, and uses AI to analyze and extract the attributes corresponding to the key entities to obtain the key entity recognition results.

[0022] formula: ; The meaning and acquisition method of the parameters: The frequency of a term in the dataset is a direct counting operation, typically achieved by scanning each document in the dataset and counting the number of times a term appears. Total number of words in the dataset: It is the sum of all terms in the dataset, obtained by accumulating the number of terms in each document. This count reflects the size of the entire dataset and is achieved by iterating through all documents in the dataset and counting the number of terms in each document. The number of documents in the differential signature: This is a document frequency, representing the number of documents in the dataset that contain the term. This is calculated by iterating through all documents in the dataset and tagging each document containing the term. Calculation example: Given a dataset containing the following values: This indicates that the term appeared 15 times in the dataset. This indicates that there are a total of 1000 words in the dataset. This indicates that the term appears in 10 different documents; The calculation yielded: ; result The term frequency (TF) reflects the relative importance of a term within the entire dataset. This calculation method emphasizes the importance of a term in a few documents by adjusting the TF. Therefore, if a term appears in only a few documents but with a high frequency, its importance is reduced. The value will be relatively large, which helps to identify keywords and core concepts when processing text data.

[0023] Please see Figure 3 The steps for obtaining entity recognition information are as follows: S121: Based on the key entity identification results, for each entity, the following formula is used: ; Calculate each entity identifier ,in, It is a hash function. It is an abbreviation of entity name. For timestamps and source code; S122: Based on each entity identifier Through the formula: ; Calculate entity frequency, where, express The number of times it appears, It is the total number of entities in the dataset; S123: Based on the entity frequency, compare it with the preset frequency threshold, merge the entity information whose entity frequency exceeds the preset frequency threshold, reduce data redundancy, and obtain entity recognition information.

[0024] formula: and ; Meaning and acquisition method of parameters : A unique identifier used to calculate each entity ,in Represents a hash function. It is an abbreviation for entity name. This is combined with timestamps and source code. This combination ensures that each entity has a unique identifier in the dataset, facilitating subsequent processing; : Represents an entity identifier The number of times something appears in the dataset. This value is calculated by scanning the dataset and counting each... The occurrence of these values ​​directly reflects the frequency of entities in the dataset. The total number of entities in the dataset, obtained by counting the number of all entity identifiers in the dataset. This number provides a reference for the size of the dataset and is crucial for the frequency of calculations; Calculation example: The dataset contains the following entities: total number of entities. A certain entity The name is abbreviated as "PatInfo", timestamp and source code =“20231120XYZ”, therefore, ID=h(“PatInfo+20231120XYZ”). Setting The corresponding hash value is "abc123", and this It appeared in the dataset. Second-rate.

[0025] Calculation frequency: ; In the example, This represents the relative frequency of the identifier "abc123" for the "PatInfo" entity across the entire dataset. If a threshold is set... ,because Greater than Therefore, the information in "PatInfo" will be considered duplicates and merged. This process reduces data redundancy and improves the accuracy and usability of the dataset.

[0026] Please see Figure 4 The method for assessing the relationships between entities is as follows: S211: Based on entity recognition information, traverse the already recognized entity data using the formula: ; Calculate the co-occurrence frequency of entity pairs, where, It is an entity pair and The number of documents that appear together This represents the total number of documents in the dataset. For entity pairs and The co-occurrence frequency; S212: Based on the co-occurrence frequency of entity pairs, compare it with a preset co-occurrence threshold, mark entity pairs with a co-occurrence frequency higher than the preset co-occurrence threshold as related, and extract the relationship type between entities through AI to obtain the relationship information between entities.

[0027] formula: ; The meaning and acquisition method of the parameters: For entity pairs and The co-occurrence frequency represents the entity and The frequency of common occurrences within the same document; For entity pairs and The number of documents that appear together represents entities. and The number of documents that co-occur in the dataset is obtained by iterating through the dataset, checking whether each document contains both entities, and accumulating the total number of occurrences. The total number of documents in the dataset. This is calculated by counting the total number of documents in the dataset. Calculation example: Set entity and It appeared 10 times in 100 documents, that is... Total number of documents in the dataset .

[0028] Calculate co-occurrence frequency: ; result This indicates a low co-occurrence frequency, suggesting that there is no or a low degree of association between the entities.

[0029] Please see Figure 5 The steps for obtaining basic entity relationship information are as follows: S221: Based on the relationship information between entities, extract the co-occurrence frequency of entity pairs and calculate the product of the number of documents in which the entity pairs appear in the dataset to obtain the relationship strength association information; S222: Based on relational strength information, using the formula: ; Evaluate the strength of the relationship between entity pairs to obtain basic entity relationship information, where... For entity pairs and The strength of the relationship It is an entity pair and The product of the number of documents appearing in the dataset. This is the magnification factor.

[0030] formula: ; Meaning and acquisition method of parameters For entity pairs and The co-occurrence frequency was obtained through the previous steps.

[0031] Entity pairs and The product of the number of documents appearing in the dataset represents an entity. and The product of the number of documents in which each entity appears. The parameter is obtained by calculating the number of documents in which each entity appears and then multiplying them.

[0032] It is the amplification factor, used to adjust the scoring sensitivity.

[0033] Calculation Example set up ,entity It appeared in 300 documents. It appeared in 200 documents, therefore Set the magnification factor. 000.

[0034] Calculate the score: ; result Representing entities and The low scores for importance and strength of the relationship indicate that there is no strong association between them or that their occurrence is common in the environment, thus reducing the significance of the relationship.

[0035] Please see Figure 6 The method for tracking the changing trends of entity relationships is as follows: S311: Based on basic entity relationship information, AI is used to analyze the relationships between current entities, eliminate entity pairs that do not fit the health and wellness scenario, and obtain the filtered entity pairs; S312: Based on the filtered entity pairs, extract the entity pair relationship strength data over a period of time, using the formula: ; Calculate the degree of change in the strength of entity-to-entity relationships to obtain information on the trend of changes in entity-to-entity relationships. Represents entity pairs and The degree of change in the strength of the relationship at the two points in time. Indicates the current entity pair and The strength of the relationship This indicates the entity pair in the previous state. and The strength of the relationship For time-sensitive parameters, To indicate the current time point, For reference time points, is the base of the natural logarithm.

[0036] formula: ; Meaning and acquisition method of parameters Current entity pair and The strength of the relationship between entities in the latest state. and The strength of the relationship.

[0037] This indicates the entity pair in the previous state. and The strength of the relationship is obtained from the database.

[0038] The time sensitivity parameter is used to adjust the influence of the time weighting factor. The selection of this parameter is based on experimental or historical data analysis to optimize the model's response speed and sensitivity.

[0039] The current time point being considered is typically the real-time timestamp at the time of analysis.

[0040] Reference time point, used as a benchmark for time difference.

[0041] Calculation example: set up: , , Time sensitivity parameter Current month, The reference month is set to three months ago.

[0042] Calculate the time-weighted difference: ; Calculate the degree of change in relationship strength: ; The calculated degree of change , representing entities and The strength of the relationship between them has increased.

[0043] Please see Figure 7 The optimized steps for obtaining entity relationship information are as follows: S321: Based on the information on the changing trend of entity relationships, extract the degree of change in the relationship strength of entity pairs and the current relationship strength of entity pairs to obtain updated association information; S322: Based on updated association information, using the formula: ; Calculate the updated entity pair relationship strength and generate optimized entity relationship information, where... Is the current entity pair and The strength of the relationship Represents entity pairs and The degree of change in the strength of the relationship at the two points in time. and These are the weighting coefficients. This is the logarithmic scaling factor. For the updated entity pair and The strength of the relationship.

[0044] formula: ; The meaning and acquisition method of the parameters: : Current entity pair and The strength of the relationship represents the entity pair before the update. and The strength of the relationship between them.

[0045] Entity pairs and The degree of change in the strength of the relationship at the two time points was calculated using the previously mentioned method.

[0046] and Weighting coefficients These are linear weighting coefficients used for adjustment. The direct contribution to the update process. This coefficient is set based on statistical analysis of the adjustment effect of historical data. These are logarithmically adjusted weighting coefficients used to amplify or weaken the weighting effect of... The nonlinear effects of the resulting changes are considered. The coefficients are determined through optimization algorithms to ensure optimal sensitivity and adaptability of the model.

[0047] : Logarithmic scaling factor, used to adjust the sensitivity of the logarithmic part so that small changes are properly reflected in the update of relation strength. This parameter was determined experimentally to find the optimal balance.

[0048] Calculation example: set up: Indicates the strength of the relationship before the update. This indicates the change in relationship strength since the last update. Linear weighting coefficients, Logarithmically adjusted weighting coefficients. Logarithmic scaling factor.

[0049] calculate : ; Calculation results The data shows that the relationship strength between entities increased from 0.5 to 0.7046, reflecting the latest entity interaction data and adjustments.

[0050] Please see Figure 8 The method to adjust the interval for adding new entities is as follows: S411: Based on the optimized entity relationship information, import the relationships between entities into the health and wellness knowledge graph, continuously track new data and user feedback in the health and wellness environment, and obtain interval association information; S412: Based on interval correlation information, using the formula: ; Calculate the adjusted entity addition interval to obtain the interval adjustment result, where, It is the frequency at which new data is received. For user feedback frequency, As a balance factor, Add spacing to the base entity. Add spacing to the adjusted entity.

[0051] formula: ; The meaning and acquisition method of the parameters: The frequency of new data reception is obtained through regular recording and automatic aggregation by sensors and data collection systems in the health and wellness environment. This frequency reflects the average number of times health and wellness data is updated to the knowledge graph, typically calculated in terms of daily data updates.

[0052] User feedback frequency is automatically recorded and summarized by the user interaction platform through monitoring user activities (such as queries, feedback submissions, etc.). This frequency indicates how frequently users interact with the system, providing feedback or data corrections.

[0053] The preset balance factor is used to adjust the impact of the frequency of new data reception and the frequency of user feedback on the interval for adding new entities. This is preset based on historical data analysis and system optimization requirements, with the aim of balancing the impact of these two frequencies on the entity addition interval to ensure the sensitivity and rationality of the system response.

[0054] The basic entity addition interval refers to the frequency of entity additions set in the system before the update strategy is adjusted. This is usually based on early requirements analysis and system design.

[0055] Calculation example: The following actual parameter values ​​are set for calculation: On average, it receives 120 new data entries per day. We receive an average of 30 pieces of user feedback per day. The balance factor is set to 0.4 to reduce the impact of data reception frequency on interval adjustment. The original interval is to add a new entity once every 24 hours.

[0056] Calculate the new addition interval : ; Calculation results The time interval for adding new entities should be reduced from once every 24 hours to once every 2.4 hours, which significantly improves the update frequency of the knowledge graph.

[0057] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for constructing a knowledge graph for elderly care robots based on AI technology, characterized in that, Includes the following steps: Based on health and wellness-related data, AI is used to analyze the data, extract entity data items, identify key entities in the dataset, extract corresponding attributes, merge duplicate entity information, reduce data redundancy, and obtain entity identification information. Based on the entity identification information, the interaction data between entities is analyzed, the relationship between entities is evaluated, including the relationship between treatment and symptoms, the strength of entity relationship is assessed, and basic entity relationship information is obtained. Based on the aforementioned basic entity relationship information, entity pairs that do not fit the health and wellness scenario are analyzed and removed, and the changing trend of entity relationships is tracked. The relationship strength between entities is updated to match the current actual situation, optimize the accuracy and practicality of the data, and generate optimized entity relationship information. Based on the optimized entity relationship information, the relationships between entities are imported into the health and wellness knowledge graph. New data and user feedback in the health and wellness environment are continuously tracked. The interval for adding new entities is adjusted according to the frequency of receiving new data and the frequency of user feedback. New entities are automatically introduced, and the attributes and relationship strength of existing entities are updated to match the changes in the health and wellness environment, resulting in a dynamically updated health and wellness knowledge graph.

2. The method for constructing a knowledge graph for a health and wellness robot based on AI technology according to claim 1, characterized in that, The method for identifying key entities in the dataset is as follows: Based on health and wellness-related data, AI is used to analyze the data and extract entity data items, including user information, symptom descriptions, and treatment history. Based on the aforementioned entity data item, using the formula: ; Calculate word frequency parameters ,in, It is the frequency of occurrence of terms in the dataset. It is the total number of words in the dataset. It is the number of documents in which the term appears in the differentiated documents. Indicates the term frequency parameter; Based on the word frequency parameters The system compares the frequency parameters with a preset word frequency threshold, marks entities whose word frequency parameters exceed the preset word frequency threshold as key entities, and uses AI to analyze and extract the attributes corresponding to the key entities to obtain the key entity recognition results.

3. The method for constructing a knowledge graph for a health and wellness robot based on AI technology according to claim 2, characterized in that, The steps for obtaining the entity identification information are as follows: Based on the key entity identification results, for each entity, the formula is as follows: ; Calculate each entity identifier ,in, It is a hash function. It is an abbreviation of entity name. For timestamps and source code; Based on each entity identifier Through the formula: ; Calculate entity frequency, where, express The number of times it appears, It is the total number of entities in the dataset; Based on the entity frequency, it is compared with a preset frequency threshold. Entity information whose entity frequency exceeds the preset frequency threshold is merged to reduce data redundancy and obtain entity recognition information.

4. The method for constructing a knowledge graph for a health and wellness robot based on AI technology according to claim 1, characterized in that, The method for assessing the relationships between entities is as follows: Based on the entity recognition information, the already recognized entity data is traversed using the formula: ; Calculate the co-occurrence frequency of entity pairs, where, It is an entity pair and The number of documents that appear together This represents the total number of documents in the dataset. For entity pairs and The co-occurrence frequency; Based on the co-occurrence frequency of the entity pairs, a preset co-occurrence threshold is compared with the co-occurrence frequency. Entity pairs with a co-occurrence frequency higher than the preset co-occurrence threshold are marked as associated, and the relationship type between entities is extracted by AI to obtain the relationship information between entities.

5. The method for constructing a knowledge graph for a health and wellness robot based on AI technology according to claim 4, characterized in that, The steps for obtaining the basic entity relationship information are as follows: Based on the relationship information between the entities, the co-occurrence frequency of entity pairs is extracted, and the product of the number of documents in which the entity pairs appear in the dataset is calculated to obtain the relationship strength association information. Based on the aforementioned relationship strength information, using the formula: ; Evaluate the strength of the relationship between entity pairs to obtain basic entity relationship information, where... For entity pairs and The strength of the relationship It is an entity pair and The product of the number of documents appearing in the dataset. This is the magnification factor.

6. The method for constructing a knowledge graph for a health and wellness robot based on AI technology according to claim 1, characterized in that, The method for tracking the changing trends of entity relationships is as follows: Based on the aforementioned basic entity relationship information, AI is used to analyze the relationships between current entities, and entity pairs that do not fit the health and wellness scenario are eliminated to obtain filtered entity pairs. Based on the filtered entity pairs, entity pair relationship strength data over a period of time is extracted using the formula: ; Calculate the degree of change in the strength of entity-to-entity relationships to obtain information on the trend of changes in entity-to-entity relationships. Represents entity pairs and The degree of change in the strength of the relationship at the two points in time. Indicates the current entity pair and The strength of the relationship This indicates the entity pair in the previous state. and The strength of the relationship For time-sensitive parameters, To indicate the current time point, For reference time points, is the base of the natural logarithm.

7. The method for constructing a knowledge graph for a health and wellness robot based on AI technology according to claim 6, characterized in that, The steps for obtaining the optimized entity relationship information are as follows: Based on the entity relationship change trend information, the degree of change in the relationship strength of entity pairs and the current relationship strength of entity pairs are extracted to obtain updated association information; Based on the updated association information, using the formula: ; Calculate the updated entity pair relationship strength and generate optimized entity relationship information, where... Is the current entity pair and The strength of the relationship Represents entity pairs and The degree of change in the strength of the relationship at the two points in time. and These are the weighting coefficients. This is the logarithmic scaling factor. For the updated entity pair and The strength of the relationship.

8. The method for constructing a knowledge graph for a health and wellness robot based on AI technology according to claim 1, characterized in that, The method for adjusting the interval between adding new entities is as follows: Based on the optimized entity relationship information, the relationships between entities are imported into the health and wellness knowledge graph, and new data and user feedback in the health and wellness environment are continuously tracked to obtain interval association information. Based on the aforementioned interval association information, using the formula: ; Calculate the adjusted entity addition interval to obtain the interval adjustment result, where, It is the frequency at which new data is received. For user feedback frequency, As a balance factor, Add spacing to the base entity. Add spacing to the adjusted entity.

Images