A study of the grey relational model of interval numbers for panel data. Issue 1 (24th July 2020)

• Record Type: Journal Article
• Authors: Yin, Kedong; Xu, Tongtong; Li, Xuemei; Cao, Yun
• Is Part Of: Grey systems. Volume 11:Issue 1(2021)
• Page Start: 200
• Page End: 211
• Publication Date: 2020-07-24
• Subjects: Grey relational model -- Panel interval number -- Grey relational operators for interval numbers; Cybernetics -- Periodicals; Systems engineering -- Periodicals; 003.5
• DOI: 10.1108/GS-02-2020-0021
• Languages: English
• ISSNs: 2043-9377

Abstract

Abstract : Purpose: This paper aims to deal with the grey relational problem of panel data with an attribute value of interval numbers. The grey relational model of interval number for panel data is constructed in this paper. Design/methodology/approach: First, three kinds of interval grey relational operators for the behavior sequence of a dimensionless system are proposed. At the same time, the positive treatment method of interval numbers for cost-type and moderate-type indicators is put forward. On this basis, the correlation between the three-dimensional interval numbers of panel data is converted into the correlation between the two-dimensional interval numbers in time series and cross-sectional dimensions. The grey correlation coefficients of each scheme and the ideal scheme matrix are calculated in the two dimensions, respectively. Finally, the correlation degree of panel interval number and scheme ordering are obtained by arithmetic mean. Findings: This paper proves that the grey relational model of the panel interval number still has the properties of normalization, uniqueness and proximity. It also avoids the problem that the results are not unique due to the different orders of objects in the panel data. Practical implications: The effectiveness and practicability of the model is verified by taking supplier selection as an example. In fact, this model can also be widely used in agriculture, industry, society and other fields. Originality/value: The accuracy of the relational results is higher and more accurate compared with the previous studies.