Semantic Relation from Biomedical Text Documents using Machine Learn-ing Algorithm

Abstract

Semantic relations are underlying relation between the concepts present in sentences. Seman-tic relationplays an important role for building applications namely, Information Retrieval, Information Extraction (IE), Question Answering (QA) and chatbots. This paper proposed a model to extract entities and identifying the semantic relations exist between the entities. This paper is to extract entities using novel feature extraction techniques with naïve bayes ap-proach. Furthermore, it is used to identify the semantic relations using various machine learn-ing approaches. This paper addresses eight semantic relations namely, Prevent, No cure, Dis-ease Only, Side Effect, Vague, Treatment Only, Cure and None. The model is evaluated by standard metrics and produce the result of 68.42% F-score for this dataset.