Bio Medical Named Entity Recognition Using Machine Learning Algorithms

Abstract

Named-entity recognition system (NER) [1] identifies different entities in many ways such name of person
locations, and organizations from news articles, reports, blogs, tweets. Main steps of named entity recognition are
boundary detection of entities and classification of entities into already defined classes. This results of recognition and
classification is widely used in information retrieval and extraction.
The main component of biomedical natural language processing is named entity recognition system which extracts
information from the text and finally does the knowledge discovery. As amount of health and biomedical text being
available is huge and since much of the data is recorded in non-structured text, like in clinical notes and biomedical
publications the bottleneck of biomedical information processing is how to make use of the knowledge resources and
build scalable models to process large amounts of text. Biomedical named-entity recognition (BM-NER), also known as
biomedical concept identification or concept mapping, is a key step in biomedical language processing.
In this paper, we are proposing a Biomedical Named Entity Recognition System for extracting Name, Problem and Test
from the Textual Clinical Lab Reports using two widely accepted datasets, i2b2 [2] and GENIA corpora [3]and we are
attempting to correlate the appropriate Treatment associated with it using Machine Learning Algorithms [4] and
Natural Language Processing [5].
Rest of the paper is organized as follows, section 2 gives in depth literature survey, in section 3 we discuss different
approaches used for Bio-NER. In section 4 describes proposed system architecture.