Research On Ontology Construction In The Field Of Aging Care

At present,there are still some problems such as unstandardized and redundant eldercare names and measures in the field of aging care.The National Bureau of Statistics released the "Statistical Classification of Pension Industry(2020)" indicating that the current standardization of the elderly industry needs to be resolved urgently.Domain ontology is a tool that can describe concepts and relationships between concepts at the level of semantics and knowledge.It has the characteristics of conceptualization,formalization and sharing,and can be used to solve the chaotic phenomenon.However,ontology construction needs a lot of domain experts,consumes a lot of resources and may be affected by the subjective consciousness of the builder.In order to solve this problem,combined with clustering research in data mining and natural language processing technology,the thesis realizes the automatic construction of ontology in aging care domain according to the following two steps.1)eldercare name fusion based on hierarchical clustering: Firstly,crawling the corpus of care and measure names.Next,longitudinal denoising is carried out on the corpus by setting the threshold of filtering through correlation calculation,construct the word bag model and convert the text vector into the word vector.Finally,using the hierarchical clustering for cluster analysis.The verification of clustering results and the manual labeling results shows that 83.3% of the classes keep their corresponding F-Measure at 0.75 or above.At last,calculate the similarity between the service names of each category,and merge the service names with the help of domain experts.2)Ontology construction in the field of aging care: it includes the steps of limiting the scope of relevant knowledge in the field,listing important terms for the domain ontology,defining the hierarchy of the ontology library and adding attributes and examples of related concepts.In the step of defining concept attributes,the optimal results of word segmentation are selected as attributes of each instance through part-of-speech tagging and correlation calculation.The experimental results show that 75.9% of the original data calculated attributes have at least one attribute that matches the manual annotation result.Finally,Use the ontology development tool Protege to display and query the results of the ontology library.The thesis proposed a method for the automated construction of ontology in the field of aged care based on hierarchical clustering,which reduces the cost of human intervention during the construction process,and can update the ontology database conveniently and quickly as the information changes.Furthermore,it provides a certain reference for the standardization of terms in the field of care,the formulation of relevant policies by the government and the service standard of the company.