A Local Mining and Global Learning Approach for Bridging the Vocabulary Gap

In Healthcare domain, there is a wide gap prevalent between health seekers and healthcare experts. This vocabulary gap is due to the presence of ambiguity in the natural language in which the users post their queries in online medical portals and therefore the emerging community generated health data is more colloquial, in terms of inconsistency, complexity and ambiguity. This poses challenges for data access and analytics. To bridge the vocabulary gap present, a new scheme has been introduced in this paper which combines two approaches namely local mining and global learning based on machine learning. Local mining extracts the individual medical concepts from medical records and map them to their appropriate medical terminologies. Global learning enhances the local mining database by jointly finding the missing key terms with the help of medical related resources.