EPIA'03 - 11th Portuguese Conference on Artificial Intelligence

	Session: December 5, 18:15-19:15, Room A
Title:	Automatic selection of table areas in documents for information extraction
	Ana Costa e Silva, Alípio Jorge, Luís Torgo
Abstract:	information contained in documents such as the financial statements of companies is valuable for decision making at various levels. Much of the relevant information in such documents is contained in tables and is currently extracted mainly by hand. We propose a method that accomplishes a preliminary step of the task of automatically extracting information from tables in documents: selecting the lines in the document which are likely to belong to a table. Our method has been developed by empirically analyzing a set of Portuguese companies’ financial statements, using statistical and data mining techniques. Empirical evaluation indicates that more than 99% of the table lines are selected after discarding at least 50% of them. The method can cope with the complexity of styles used in assembling information on paper and adapt its performance accordingly, thus maximizing its results.

Back to schedule.