Customer Reviews

Text Mining for Biology And Biomedicine

5 star:		(0)
4 star:		(1)
3 star:		(0)
2 star:		(0)
1 star:		(0)

 

 

This book was written to help biologists manage the flood of literature in their profession. Online search tools help some, but not enough. Text analysis, "the process of discovering and extracting knowledge from unstructured data," can help with this literature tsunami. Biologists--and researchers in related disciplines--can use the tools and techniques of text analysis to identify relevant documents, extract useful information from them, and discover new relationships between these hard-won nuggets.

Sophe Ananiadou and John McNaught have assembled a collection of chapters written by experienced text analysts. They cover the primary subareas of text analysis with little overlap. These chapters are well-documented, use examples from biology, and explain them well enough to make the book accessible to non-biologists. After a concise introductory outline in the first chapter, readers become familiar with the general techniques of natural language processing, the various research communities using NLP techniques, and the tools and other resources they have created. Subsequent chapters cover terminology management, abbreviations, named entities (think "proper nouns"), information extraction, and corpus tagging. The final two chapters discuss how text analysis techniques are evaluated and how results are integrated into structured data analysis.

The clear explanations and accessible examples make this book a reasonable first read for anyone curious about text analysis. Although readers need not be biologists, some familiarity with scientific inquiry is helpful. I found the chapter on evaluation of text analysis methods most useful. It explains the role of challenge competitions in advancing the field and describes the common evaluation metrics well enough to facilitate understanding of the research literature. All chapters include numerous web links to software tools, data sources, and relevant research groups. This is a particular strength of the book, and helps readers bridge the gap between text analysis as it was in 2005, and as it has become since the book's publication.

It is recommended for scientists who want to learn about text analysis by observing it's practical applications in biology. Readers preferring a "straight" introduction without the biology might turn to The Text Mining Handbook: Advanced Approaches in Analyzing Unstructured Data or the last chapter in Data Mining Techniques: For Marketing, Sales, and Customer Relationship Management.