Financial Information Extraction Using Pre-defined and User-definable Templates in the LOLITA System

Abstract

This paper addresses the issue of information extraction in the financial domain within the framework of a large Natural Language Processing system: LOLITA. The LOLITA system, Large-scale Object-based Linguistic Interactor Translator and Analyser, is a general purpose natural language processing system. Different kinds of applications have been built around the system's core. One of these is the financial information extraction application, which has been designed in close contact with experts from the financial market in order to overcome the lack of usefulness of many other systems.

Three predefined groups of templates have been designed according to the "financial activities approach": company related templates, company restructuring templates and general macroeconomic templates. In addition, the user-definable template interface allows the user to define new templates using natural language sentences.

After describing LOLITA as a general purpose base NLP system, the paper addresses the issue of how information extraction is performed within the system an'd how the user-definable template interface has been designed.