Abstract

GUERRERO ENAMORADO, Alain  and  CEBALLOS GASTELL, Daimerys. An evaluation of LVQ algorithm in a textual collection. Rev cuba cienc informat [online]. 2016, vol.10, n.4, pp.154-170. ISSN 2227-1899.

ABSTRACT The aim of this work was to obtain empirical information on the behavior of the LVQ (Learning Vector Quantization) algorithm in the task of information classification. We developed an algorithm evaluation on 24 versions of a news-collection previously created. In this collection, we varied the number of terms using the method, Information Gain, to select the attributes. The behavior of algorithms against the effects of dimensionality reduction is determined. We found a point, where the compromise between loss of information and incorporation of irrelevant information, has achieved good results.

Keywords : learning vector quantification; classification; neural networks.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )