Discourse Processing for Text Mining

Discourse Processing for Text Mining

Nadine Lucas (GREYC CNRS, Université de Caen Basse-Normandie Campus 2, France)
DOI: 10.4018/978-1-60566-274-9.ch012
OnDemand PDF Download:


This chapter presents the challenge of integrating knowledge at higher levels of discourse than the sentence, to avoid “missing the forest for the trees”. Characterisation tasks aimed at filtering collections are introduced, showing use of the whole set of layout constituents from sentence to text body. Few text descriptors encapsulating knowledge on text properties are used for each granularity level. Text processing differs according to tasks, whether individual document mining or tagging small or large collections prior to information extraction. Very shallow and domain independent techniques are used to tag collections to save costs on sentence parsing and semantic manual annotation. This approach achieves satisfactory characterisation of text types, for example reviews versus clinical reports, or argumentation-type articles versus explanation-type. These collection filtering techniques are fit for a wider domain of biomedical literature than genomics.
Chapter Preview


In this chapter we address higher-levels of text processing, as related to text mining. The domain of biomedical language processing (BLP or bio-NLP) “encompasses the many computational tools and methods that take human generated texts as input, generally applied to tasks such as information retrieval, document classification, information extraction, plagiarism detection, or literature-based discovery” (Hunter & Bretonnel Cohen, 2006 p. 589).

Access to biomedical literature itself (primary sources) is provided since 2004 through PubMed Central established by the American National Library of Medicine (NLM) as a repository of free access articles. The search system Entrez PubMed offers abstracts from Medline along with on line full-text indexed by Mesh and access to databases (see NLM site). This has fostered a new circular situation where data and text bases feed literature in turn feeding databases and ontologies.

Related events are first, advances in genomics and the information deluge. A double exponential growth of published material is recorded in the biomedical field, creating in turn an increased amount of facts to be stored (Shatkay & Craven, 2007). Second, text mining techniques for specific purposes were developed in particular to help in database curation. Automats now directly fill a growing part of databases (Hunter & Bretonnel Cohen, 2006). Third, a new field called systems biology emerged at the frontier between data and text mining (Krallinger & Valencia, 2005). Text mining is used to back data interpretation. Computational processes are ubiquitous and the frontier between text and data mining is blurred as well as the frontier between human and automated processes. Integrated text mining systems inherit from expert systems (nomenclatures linked with inference rules) and from statistical data mining. They rely on what might be called tertiary sources of knowledge: unified nomenclatures, hand curated interaction databases and hand annotated corpora. These are sometimes grouped under the term “ontologies” (Ananiadou et al., 2006). Last, users now take it for granted that raw information is quickly translated into secondary and tertiary sources, and rely on computer-manageable “concepts” (Rebholz-Schumann et al., 2005). Recent developments can be watched by consulting the Biomedical Literature Mining Publication portal (Blimp) (2008).

Success in the genomics field opened the way for less specific purposes. As text mining is advertised in more publications, not only “omics” researchers, but also clinicians, general practitioners and medical librarians call for text processing (Fluck et al., 2005; Hunter & Bretonnel Cohen, 2006; Mizuta et al., 2006). One emerging trend in research is to take patients into consideration to best respond to users’ needs (Leroy et al., 2006). This implies a change in the way to produce results. While researchers can do with highly specialised words, evoking for them research trends, a wider public need full explanations, therefore lengthier passages of text. Robust text processing is needed but it is still in its infancy.

Another trend calls for semantic characterisation of texts in a collection. Most semantic oriented tasks, such as characterizing original findings on a topic, or eliciting hypotheses, require a wider context than the sentence. Valuable meta-information that could be used to qualify texts is still lacking. Some attempts at qualifying parts of them, like conclusions, e.g. tentative or definite are on the way. Yet, very few studies address text at a global level as a semantic unit. The gap between expectations and realisations is blatant.

The approach explained here relies on combining text mining characterisation techniques for collections and robust high-level discourse parsing techniques. We advocate a shift of paradigm from word-level description to text-level description. In the domain of biomedical text processing, text is characterized as “unstructured data” as compared to databases (Hunter & Bretonnel Cohen, 2006), or at best as semi-structured data (Hakenberg et al., 2005). Yet, texts are structured by layout, a feature that has been overlooked. Academic articles in particular are highly structured. Scale issues are seldom addressed, although they are important for information retrieval and knowledge integration in biomedicine.

Complete Chapter List

Search this Book:
Table of Contents
Violaine Prince, Mathieu Roche
Chapter 1
Sophia Ananiadou
Text mining provides the automated means to manage information overload and overlook. By adding meaning to text, text mining techniques produce a... Sample PDF
Text Mining for Biomedicine
Chapter 2
Dimitrios Kokkinakis
The identification and mapping of terminology from large repositories of life science data onto concept hierarchies constitute an important initial... Sample PDF
Lexical Granularity for Automatic Indexing and Means to Achieve It: The Case of Swedish MeSH®
Chapter 3
M. Teresa Martín-Valdivia, Arturo Montejo-Ráez, M. C. Díaz-Galiano, José M. Perea Ortega, L. Alfonso Ureña-López
This chapter argues for the integration of clinical knowledge extracted from medical ontologies in order to improve a Multi-Label Text... Sample PDF
Expanding Terms with Medical Ontologies to Improve a Multi-Label Text Categorization System
Chapter 4
Piotr Pezik, Antonio Jimeno Yepes, Dietrich Rebholz-Schuhmann
The present chapter discusses the use of terminological resources for Information Retrieval in the biomedical domain. The authors first introduce a... Sample PDF
Using Biomedical Terminological Resources for Information Retrieval
Chapter 5
Laura Diosan, Alexandrina Rogozan, Jean-Pierre Pécuchet
The automatic alignment between a specialized terminology used by librarians in order to index concepts and a general vocabulary employed by a... Sample PDF
Automatic Alignment of Medical Terminologies with General Dictionaries for an Efficient Information Retrieval
Chapter 6
Vincent Claveau
This chapter presents a simple yet efficient approach to translate automatically unknown biomedical terms from one language into another. This... Sample PDF
Translation of Biomedical Terms by Inferring Rewriting Rules
Chapter 7
Nils Reiter, Paul Buitelaar
This chapter is concerned with lexical enrichment of ontologies, that is how to enrich a given ontology with lexical information derived from a... Sample PDF
Lexical Enrichment of Biomedical Ontologies
Chapter 8
Torsten Schiemann, Ulf Leser, Jörg Hakenberg
Ambiguity is a common phenomenon in text, especially in the biomedical domain. For instance, it is frequently the case that a gene, a protein... Sample PDF
Word Sense Disambiguation in Biomedical Applications: A Machine Learning Approach
Chapter 9
M. Narayanaswamy, K. E. Ravikumar, Z. Z. Hu, K. Vijay-Shanker, C. H. Wu
Protein posttranslational modification (PTM) is a fundamental biological process, and currently few text mining systems focus on PTM information... Sample PDF
Information Extraction of Protein Phosphorylation from Biomedical Literature
Chapter 10
Yves Kodratoff, Jérôme Azé, Lise Fontaine
This chapter argues that in order to extract significant knowledge from masses of technical texts, it is necessary to provide the field specialists... Sample PDF
CorTag: A Language for a Contextual Tagging of the Words Within Their Sentence
Chapter 11
Yun Niu, Graeme Hirst
The task of question answering (QA) is to find an accurate and precise answer to a natural language question in some predefined text. Most existing... Sample PDF
Analyzing the Text of Clinical Literature for Question Answering
Chapter 12
Nadine Lucas
This chapter presents the challenge of integrating knowledge at higher levels of discourse than the sentence, to avoid “missing the forest for the... Sample PDF
Discourse Processing for Text Mining
Chapter 13
Dimosthenis Kyriazis, Anastasios Doulamis, Theodora Varvarigou
In this chapter, a non-linear relevance feedback mechanism is proposed for increasing the performance and the reliability of information (medical... Sample PDF
A Neural Network Approach Implementing Non-Linear Relevance Feedback to Improve the Performance of Medical Information Retrieval Systems
Chapter 14
Yitao Zhang, Jon Patrick
The fast growing content of online articles of clinical case studies provides a useful source for extracting domain-specific knowledge for improving... Sample PDF
Extracting Patient Case Profiles with Domain-Specific Semantic Categories
Chapter 15
Laura I. Furlong, Ferran Sanz
SNPs constitute key elements in genetic epidemiology and pharmacogenomics. While data about genetic variation is found at sequence databases... Sample PDF
Identification of Sequence Variants of Genes from Biomedical Literature: The OSIRIS Approach
Chapter 16
Francisco M. Couto, Mário J. Silva, Vivian Lee, Emily Dimmer, Evelyn Camon, Rolf Apweiler
Molecular Biology research projects produced vast amounts of data, part of which has been preserved in a variety of public databases. However, a... Sample PDF
Verification of Uncurated Protein Annotations
Chapter 17
Burr Settles
ABNER (A Biomedical Named Entity Recognizer) is an open-source software tool for text mining in the molecular biology literature. It processes... Sample PDF
A Software Tool for Biomedical Information Extraction (And Beyond)
Chapter 18
Asanee Kawtrakul, Chaveevarn Pechsiri, Sachit Rajbhandari, Frederic Andres
Valuable knowledge has been distributed in heterogeneous formats on many different Web sites and other sources over the Internet. However, finding... Sample PDF
Problems-Solving Map Extraction with Collective Intelligence Analysis and Language Engineering
Chapter 19
Christophe Jouis, Magali Roux-Rouquié, Jean-Gabriel Ganascia
Identical molecules could play different roles depending of the relations they may have with different partners embedded in different processes, at... Sample PDF
Seekbio: Retrieval of Spatial Relations for System Biology
Chapter 20
Jon Patrick, Pooyan Asgari
There have been few studies of large corpora of narrative notes collected from the health clinicians working at the point of care. This chapter... Sample PDF
Analysing Clinical Notes for Translation Research: Back to the Future
About the Contributors