8 Jun 2011

Biomedical Article Classification Using an Agent-Based Model of T-Cell Cross-Regulation


We propose a novel bio-inspired solution pro biomedical article classification. Our method draws from an existing develop of T-cell cross-regulation in the vertebrate immune logic (IS), which is a complicated adaptive logic of millions of cells interacting to distinguish linking harmless and detrimental intruders. Analogously, automatic biomedical article classification assumes with the intention of the interaction and co-occurrence of thousands of terms in text can be used to identify conceptually-related classes of articles-at a smallest, two classes with significant and irrelevant articles pro a agreed thought (e.G. Articles with protein-protein interaction information). Our agent-based method pro paper classification expands the existing analytical develop of Carneiro et al. [1], by allowing us to deal at once with many evident T-cell facial appearance (epitomes) and their collective dynamics using agent based modeling. We already extended this develop to develop a bio-inspired spam-detection logic [2, 3]. Here we develop our agent-base develop additional, and test it on a dataset of publicly unfilled full-text biomedical articles provided by the BioCreative challenge [4].We study several extra parameter configurations leading to cheering results comparable to state-of-the-art classifiers. These results help us understand both T-cell cross-regulation and its applicability to paper classification in all-purpose. Therefore, we trade show with the intention of our bio-inspired algorithm is a promising novel method pro biomedical article classification and pro binary paper classification in all-purpose. 

No comments:

Post a comment