Gini-Simpson index based generalised grey target decision method for mixed attributes

Article type: Research Article

Authors: Ma, Jinshan^{; *}

Affiliations: School of Energy Science and Engineering, Henan Polytechnic University, Jiaozuo, Henan, China

Correspondence: [*] Corresponding author. Jinshan Ma, School of Energy Science and Engineering, Henan Polytechnic University, Jiaozuo, Henan, China. E-mail: majshan@hpu.edu.cn.

Abstract: A novel Gini-Simpson (G-S) index based generalised grey target decision method for mixed attributes is put forward. The proposed method improves the G-S index and adopts it as the mixed attribute-based target centre distance from the viewpoint of measuring the difference of alternative index and target centre index. The core of this algorithm is to obtain the ratio of each alternative index to the target centre index. And the ratio is regarded as the pseudo probability, as makes a bridge between the alternative index and the target centre index. This method first transforms all indices into binary connection numbers and divides them into the deterministic terms and uncertain terms to constitute the two-tuple (determinacy, uncertainty) numbers. Then the target centre indices of two-tuple (determinacy, uncertainty) number are determined. Following this, the ratios of deterministic terms and uncertain terms of alternative indices to those of the target centre indices are calculated. Finally, the comprehensive weighted Gini-Simpson indices (CWGSIs) of all alternatives are obtained for decision-making with which the smaller value being the better. A case study illustrates the proposed approach.

Keywords: Gini-Simpson index, generalised grey target decision method, mixed attributes, binary connection number

DOI: 10.3233/JIFS-182821

Journal: Journal of Intelligent & Fuzzy Systems, vol. 37, no. 2, pp. 2597-2608, 2019