Evaluation of Different Similarity Measures for the Extraction of Multiword Units in a Reinforcement Learning Environment

In this paper, we present an evaluation of four different similarity measures for the extraction of multiword units in the context of the GALEMU software that proposes an innovative architecture based on a floating point representation genetic algorithm. In particular, we will show that the Bray and Curtis measure leads to improved extraction results both in precision and recall
Published in 2004