A comparative study of the current technologies and approaches of relation extraction in biomedical literature using text mining

Many techniques in Relation Extraction (RE) require an understanding of the concept and association of relations between entities. This paper presents advanced approaches that combined many methods to extract simple and complex relations between different entities from the biomedical literature. The literature comparative study includes a wide variety of techniques for RE all of which fall into one of the following approaches such as supervised, semi-supervised, higher order relation and self-supervised approaches. Supervised approach needs very detailed data, while in semi-supervised approach only a small knowledge base to automatically annotates features. This approach does not need an initial set of labeled pages for extraction rules. Self-supervised is an integration of these two approaches. On the other hand, the term higher order relation indicates a relation that connects different biomedical relations together. The F-score performance differs from relation to relation when it is extracted. Finally, an extensive performance evaluations with real datasets for both supervised, semi-supervised, higher order relation and self-supervised approaches are described for entity — RE tasks.