Selecting the Optimal Configuration ofSwarm Algorithms for an NLP Task

George Tambouratzis

Abstract: The present article involves the generation of phrase boundaries in unconstrained free text. Particle Swarm Optimisation (PSO) is applied to determine the optimal values for a set of parameters, based on a limited amount of training data. The starting point is a detailed analysis of generated solutions, which leads to a reformulation of the phrasing task. Based on this reformulation, the optimal swarm configuration is investigated, including the interconnection between swarm particles as well as velocity initialisation. Another aspect studied is whether a homogeneous or a heterogeneous swarm provides a better optimisation, established by comparing three PSO variants. The experimental results are further analysed with statistical tests to determine their significance.