Combining Light-Weight Retrieval Strategies for Robust Text Categorization

We report on the development of a general purpose text categorization system designed to automatically assign biomedical categories to any input text. Unlike usual automatic text categorization systems, which rely on data-intensive models extracted from large sets of training data, our categorizer is largely dataindependent and so it can be used when training data are not available provided that a small set of instances is available for tuning the system. Like it is usual with information retrieval engines, the tool provides a ranked list of categories, which can then be interactively filtered by the user.