Evaluation and Comparison of multilingual fusion strategies for similarity-based Word Sense Disambiguation

In this article, we investigate the effects on the quality of the disambiguation of exploiting multilingual features with a similaritybased WSD system based on an Ant Colony Algorithm. We considered features from one, two, three or four languages in order to quantify the improvement brought by using features from additional languages. Using BabelNet as a multilingual resource, we considered three data fusion strategies: an early fusion strategy, and two late fusion strategies (majority vote and weighted majority vote). We found that the early fusion approach did not produce any significant improvements while voting strategies adding features from more languages led to an increase in the quality of the disambiguation of up to 2.84%. Furthermore, a simple majority vote led to better results than the weighted variant.