Essay on A Analysis On Binary Particle Swarm Optimisation

1293 Words Jan 27th, 2015 6 Pages
Abstract—Feature selection, used as a preprocessing step, can reduce the dimensionality of data and thereby increase the efficiency, accuracy, and clarity of learning systems. However feature selection can be costly endeavour. This paper proposes two new feature selection algorithms, based on binary particle swarm optimisation, with the aim of reducing running time without affecting classification accuracy by combining filter and wrapper approaches. The first algorithm proceeds cautiously by only updating the pbest and gbest, two critical values, after the learning system has been consulted. The second algorithm performs more reckless updates with the aim of sacrificing some performance for speed. Some theoretical analysis is performed in order to establish a general approximation of the performance of the algorithms. The algorithms are benchmarked on eight datasets, using K-nearest neighbour is used as the learning system and Min-Redundancy Max-Relevancy as a filter-based evaluation function. The results show a dataset-specific improvement to running time over the standard approach, with the tradeoff of larger resultant feature subsets. I. I NTRODUCTION Datasets used in classification (the prediction of class labels from features) are often described by a large number of features, so as to represent the target concept as completely as possible. However this abundance of features is often harmful for classification, as many features are either redundant or irrelevant;…

Related Documents