Desarrollo de un sistema de aprendizaje automático supervisado para la desambiguación léxica automática utilizando DAMIEN (Data Mining Encountered).

Word sense ambiguity is one of the major challenges we face when we carry out tasks related to Natural Language Processing, in particular those related to the processing of electronic language resources. In this study we address word sense disambiguation within the computing environment DAMIEN (Data Mining ENcountered). DAMIEN is an online workbench that embeds several techniques from different fields (corpus linguistics, statistics and text mining) in order to deal with text analysis to help in linguistic research tasks such as data collection, information retrieval and text classification, among others. By way of experiment, we carry out the analysis of the polysemic lexical units "cabeza", "cara" and "carta" in Spanish and present the results of the automatic disambiguation system developed with DAMIEN. Among the models that the environment offers we have deployed the supervised machine learning method with ingenious bayes algorithm because it has traditionally given the best results for automatic word sense disambiguation. It is a mathematical model that consists in extracting information from a corpus, setting from previously tagged datasets (training corpus), so that new datasets can be automatically classified by the system (trained corpus). It is important to highlight the flexibility and potentialities of DAMIEN for both the processing of electronic linguistic resources and the design of experiments in the field of natural language processing. [ABSTRACT FROM AUTHOR]