基于BPSO的棉花异性纤维目标特征快速选择方法

A Fast Feature Selection for Cotton Foreign Fiber Objects Based on BPSO

摘要:

针对现有棉花异性纤维目标特征选择方法迭代次数多、速度慢等问题，提出了一种基于改进粒子群优化算法的棉花异性纤维目标特征快速选择方法。使用离散型粒子群优化算法作为特征选择算法，利用支持向量机算法作为分类器对最优特征集进行验证。实验结果表明，在分类准确率与蚁群算法相当的情况下，能减少26%的运行时间。

关键词:棉花;异性纤维;支持向量机;离散型粒子群优化算法;特征选择

Abstract:

A fast feature selection for cotton foreign fiber objects based on binary particle swarm optimization was presented, for the current feature selection of cotton foreign fiber having more iteration times and slow speed. Binary particle swarm optimization (BPSO) was used to select feature in the method, and the support vector machine algorithm was used to verify the optimal feature set. Experimental results showed that the running time could reduce by 26%, when the classification accuracy was almost with other algorithms.

Key words:Cotton;Foreign fiber;SVM;Binary particle swarm optimization;Feature selection