Biomedical text mining for concept identification from traditional medicine literature

In recent years, vast amount of biomedical literature is produced and published. Recent developments in biomedical text mining shows potential for supporting scientists in understanding new information from the existing biomedical literature because volume of electronically available biomedical literature are increasing massively. Automated literature mining offers one opportunity to discover different entities from literature. Web Technologies allow these entities to be stores and publish in the form to the further reuse by the researchers. The approach presented here includes text mining methodologies to automatically extract different entities from biomedical text. For this purpose biomedical articles based on Traditional Chinese medicine are extracted from Bio Med Central and Pub Med Central and used as corpus. Using text mining techniques of tokenization, splitting, stemming, lemmatization, parsing, named entity recognition are used for preprocessing of corpus. Candidate terms are identified by applying C-Value algorithm. These candidate terms and existing Seed/Ontological Terms are tagged in corpus. Using lexical and contextual profiles comparison between candidate terms and already existed Seed/Ontological Terms, we have identified new concepts. Identified concepts are evaluated.

I. Introduction

In recent decades, data sharing over the internet has been increased rapidly. There are numbers of textual knowledge resources are available publicly. These textual sources can be journal Iconference papers, technical reports, user manuals, blogs, messages on social media and news etc. The amount of data in these resources is huge and growing quickly. An efficient methodology to access and extract information from these resources is required. Text mining is a methodology to automatically process and extract useful information from these resources. This information can be easily used by humans or further processed by other tools and web services. Using text mining approach not only knowledge given in texts could be identified and extracted but also new or implicit knowledge may also be exposed. In recent years, text mining is applied in many areas, such as finding concepts from text and their relations. End users do not have to read huge amount of text and depends only on text mining systems to extract important information from these textual resources.