A Meta-Mining Ontology Framework for Data Processing

Abstract

Extracting knowledge from data streams received from observed objects through data mining is required in various domains. However, there is a lack of any kind of guidance on which techniques can or should be used in which contexts. Meta mining technology can help build processes of data processing based on knowledge models taking into account the specific features of the objects. This paper proposes a meta mining ontology framework that allows selecting algorithms for solving specific data mining tasks and build suitable processes. The proposed ontology is constructed using existing ontologies and is extended with an ontology of data characteristics and task requirements. Different from the existing ontologies, the proposed ontology describes the overall data mining process, used to build data processing processes in various domains, and has low computational complexity compared to others. The authors developed an ontology merging method and a sub-ontology extraction method, which are implemented based on OWL API via extracting and integrating the relevant axioms.

An ontology-based approach to the analysis of the acid-base state of patients at operative measures

“…We also define and use the “INPUT” sub-ontology to describe the data characteristics and task requirements to support algorithms selection. The detailed description of the DM Core ontology is presented in ( Tianxing et al, 2020 ). The main classes and properties are presented in Fig.…”