NAMED ENTITY RECOGNITION FROM BIOMEDICAL TEXT

Abstract

Named entity recognition is commonly considered an essential task in natural language processing. It represents a major phase in information extraction methodology to discover the named entities referenced in unstructured text and classify them into predefined class labels. Identifying biomedical entities has been recognized as a challenging task in named entity recognition. In this thesis, the applicability of using structured support vector machine to classify flat and nested biomedical entities combined with the feature selection techniques to enhance the performance of biomedical named entity recognition has been thoroughly investigated. The proposed approach used a combination of various types of features to explore the classification performance in a combination of structured support vector machine as a machine learning technique. These features include linguistic, morphological, orthographical, context, and word representation features. The experimental results showed that the performance of the proposed approach surpassed that produced from other benchmark approaches in extracting the biomedical entities such as genes, proteins, cell lines, cell types, DNAs and RNAs. Derived by these promising results, we were motivated to explore the effect of different types of features on structured support vector machine performance in extracting the biomedical entities. This was achieved by