Extraction of Semantic Biomedical Relations from Medline Abstracts using Machine Learning Approach

Machine Learning (ML) is a natural outgrowth of the intersection of Computer Science and Statistics. Machine Learning has now become a reliable tool in the medical domain. ML which act as a tool by which computer-based systems could be integrated in the healthcare field in order to get a better and efficient health care. This methodology for building an application is capable of identifying and extracting healthcare information. The proposed system focuses on two main tasks. The first task identifies the sentences which are published in Medline abstracts. This task is similar to the task of sentence scanning contained in the medical abstract of an article in order to present to the user-only sentences that are identified as containing relevant information. The second task has a deeper semantic dimension and it focus on identifying semantic relations exists between disease-treatment. It focuses on three relations: Cure, Prevent, Side Effect and also focuses a subset of the eight relations that the corpus is annotated with. The proposed methodology obtains reliable outcomes and that could be integrated in an application to be used in the medical care domain. The framework’s capabilities can be used in a commercial recommender system and it is integrated in a new Electronic Health Record system.