A Parametrizable Task-adaptive Named-Entity Recognition System

Jakub Piskorski, Tilman Jäger, Feiyu Xu

In: H.-M. Haav , A. Kalja (Hrsg.). Databases and Information Systems II, Selected Papers from the 5th International Baltic Conference DB&IS'2002. ISBN 1-4020-1038-9 Kluwer Academic Publishers 2003.


Robust Named-Entity Recognition software is an essential preprocessing tool for performing more complex text processing tasks in business information systems. In this paper we present a Framework for Domain and Task Adaptive Named-Entity Recognition. It consists of several clear-cut subcomponents which can be flexibly and variably combined together in order to construct a task-specific NE-Recognition tool. Additionally, a diagnostic tool for automatic prediction of best system configuration is provided, which speeds up the development cycle.

Deutsches Forschungszentrum für Künstliche Intelligenz
German Research Center for Artificial Intelligence