The LOLITA User-Definable Template Interface

Abstract

The development of user-definable templates interfaces which allow the user to design new templates definitions in a user-friendly way is a new issue in the field of information extraction. The LOLITA user-definable templates interface allows the user to define new templates using sentences in natural language text with a few restrictions and formal elements. This approach is rather different from previous approaches to information extraction which require developers to code the template definitions directly in the system. After describing LOLITA as a general purpose base NLP System and other approaches to user-definable template interfaces that could be taken, the paper describes the design and the implementation of the LOLITA userdefinable template interface. The performance of the interface is evaluated comparing the results with those produced by pre-defined financial templates produced by the LOLITA System.