Multi-objective Feature Selection Based on Artificial Bee Colony for Hyperspectral Images

Most of existing band selection methods based on evolutionary optimization emphasis on single indicator, ignoring the whole characteristic of hyperspectral data. This paper establishes a new multi-objective unsupervised band selection model by considering both the information contained in bands and the correlation between bands. And, an improved multi-objective feature selection based on artificial bee colony (MABCBS) is proposed by combining two new operators, i.e., multi-strategy hybrid search mechanism for employ bees and variable space crowing degree based search mechanism for onlooker bees. Appling in two typical data sets, experiment results verify the superiority of the proposed model and the proposed algorithm.