Ontology-Based Data Mining Workflow Construction

Currently, Data mining is applied in various domains. Many data science researchers are confused on which algorithms are suitable for the context. Hundreds of the operators/algorithms are combined within complex data mining workflows. A data mining assistant can significantly improve the efficiency of workflow construction. In our previous work, we constructed data mining ontologies based on “semantic meta mining” to support the selection of algorithms and operators for solving data mining tasks. But the use of such ontologies is still unfriendly. Strict query syntax still plagues many users. This paper proposes an interactive interface based on the reasoning mechanism to help users generate queries and build suitable data mining workflows.