Hierarchical Text Categorization and Its Application to Bioinformatics

In a hierarchical categorization problem, categories are partially ordered to form a hierarchy. In this dissertation, we explore two main aspects of hierarchical categorization: learning algorithms and performance evaluation. We introduce the notion of consistent hierarchical classification that makes classification results more comprehensible and easily interpretable for end-users. Among the previously introduced hierarchical learning algorithms, only a local top-down approach produces consistent classification. The present work extends this algorithm to the general case of DAG class hierarchies and possible internal class assignments. In addition, a new global hierarchical approach aimed at performing consistent classification is proposed. This is a general framework of converting a conventional “flat” learning algorithm into a hierarchical one. An extensive set of experiments on real and synthetic data indicate that the proposed approach significantly outperforms the corresponding “flat” as well as the local top-down method. For evaluation purposes, we use a novel hierarchical evaluation measure that is superior to the existing hierarchical and non-hierarchical evaluation techniques according to a number of formal criteria. Also, this dissertation presents the first endeavor of applying the hierarchical text categorization techniques to the tasks of bioinformatics. Three bioinformatics problems are addressed. The objective of the first task, indexing biomedical articles with Medical Subject Headings (MeSH), is to associate documents with biomedical concepts from the specialized vocabulary of MeSH. In the second application, we tackle a challenging problem of gene functional annotation from biomedical literature. Our experiments demonstrate a considerable advantage of hierarchical text categorization techniques over the “flat” method on these two tasks. In the third application, our goal is to enrich the analysis of plain experimental data with biological knowledge. In particular, we incorporate the functional information on genes directly into the clustering process of microarray data with the outcome of an improved biological relevance and value of clustering results.