A proposal to construct biomedical triple collections for the dementia research

This paper describes a fundamental research for constructing biomedical triple collections from the biomedical literatures in order to find candidate biomedical interactions such as Protein-Protein or Gene-Gene or Cell-Cell interactions which might give a positive contribution in the dementia research. Due to the large amounts of biomedical publications such as PubMed, hundreds of thousands of valuable information is generated in human written language. Extracting a core biomedical interaction is quite difficult from the huge amount of textual data by human hands. Therefore, this paper introduces a simple and fundamental method to extract biomedical triple collections automatically from the human written literatures.