An Ontology-Based Framework for Analysis Recommendation

One of the challenges with data analysis revolves around selecting the best analysis method for a data set that will provide appropriate and meaningful results. This paper presents an ontology-based framework to address challenges around selecting an analysis method that can best represent a data set and the information you want to get out of it. Two ontologies were developed, one to capture semantic and syntactic descriptions on a data source, and likewise one to capture the description of analysis methods. Ontologies were selected for their flexibility in providing a description between a set of concepts and relationships along with their ability to reason between these descriptions.