Summary
In this paper we illustrate the usage of text mining workflows to automatically extract instances of microorganisms and their habitats from free text; these entries can then be curated and added to different databases. To this end, we use a Conditional Random Field (CRF) based classifier, as part of the workflows, to extract the mention of microorganisms, habitats and the inter-relation between organisms and their habitats.
Results indicate a good performance for extraction of microorganisms and the relation extraction aspects of the task (with a precision of over 80%), while habitat recognition is only moderate (a precision of about 65%). We also conjecture that pdf-to-text conversion can be quite noisy and this implicitly affects any sentence-based relation extraction algorithms.
© 2011 The Author(s). Published by Journal of Integrative Bioinformatics.
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.