We have recently run some experiments in nominal classification using our new gold-standards in English and Spanish.

Here, we summarize the results obtained in the task of noun classification with these gold-standards. We carried out the experiments using Decision Trees (see Bel et al., 2010) and the Corpus Tècnic de l'IULA (Cabré et al., 2006) to extract noun occurrences. This corpus contains a collection of written texts from the fields of Law, Economy, Genomics, Medicine, and Environment, as well as a contrastive corpus from the newspapers. In our experiments, for Spanish we used the 21M tokens newspaper corpora and the Economy corpus (1 milion words) to study domain-specific and sparse data problems. For English, we used exts of different domains: Economy, Medicine, Computer science and Environmental issues, of about 3.2M tokens.

Results for English

Class	Acc. %	FP %	FN %
Abstract	71.36	9.86	18.78
Human	80.48	5.16	14.36
Non-deverbal eventive nouns	80.24	1.80	17.96

Results for Spanish

Class	General Corpus			Economy Corpus
	Acc. %	FP %	FN %	Acc. %	FP %	FN %
Human	78.14	9.13	12.74	71.70	5.25	23.05
Semiotics	71.46	7.94	20.60	65.54	10.46	24.00
Non-deverbal eventive nouns	86.43	3.02	10.55	68.55	18.87	12.58