TACAT: tagged corpus analyser tool

Abstract

This document describes the structure, funtional capabilities as well as the user manual of TACAT (Tagged Corpus Analyser Tool), a parser and syntactic editor for tagged text. TACAT has been developed as part of ITEM (TIC96-1243-CO3-02) project for providing multilingual background for information extraction and information retrieval tasks. The main goal of TACAT is to provide a way of obtaining at a moderate human cost large amounts of brachetted and parsed corpora, both general and domain specific, for several purposes. Within ITEM the use of these corpora (tree-banks) is limited to learn probabilities for a SCFG. Other uses can, however, be considered.