Research topic recommendation service: an approach based on semantic relation in context

Research Topic Recommendation Service, which is an application of Knowledge Discovery, recommends users worthy research topics that have not conducted from public literatures such as papers or patents. Past studies had a few disadvantages such as low precisions or requirements of domain-specific knowledge because the studies used statistical method or descriptions written by human experts. This paper approaches the research topic recommendation problem with the method that parses sentences of full text and uses semantic relations of context. We will also describe the architecture of Scientific Tech Mining System designed for the research topic recommendation service.