Multilingual Entity Task a Descriptive Analysis of Enamex in Spanish.

摘要

The task involved identifying and typing all named entity expressions (ENAMEX). numerical entity expressions (NUMEX), and temporal entity expressions (fiMEX) in Spanish news articles. The analysis of the data suggests that focusing on the high frequency expressions results in a higher payoff. This report looks primarily at ENAMEX expressions because they accounted for nearly three-quarters of the taggable data, and because they involved some of the more difficult tagging decisions.