Analysis of Multivariate Associations with Qualitative and Quantitative Variables using Hybrid of Burt Multiple Correspondence Analysis and Cosine Association Matrices (A Case Study: The high school’s accreditation in West Java)

Irlandia Ginanjar, Ade Irma Nurwahidah, Jadi Suprijadi, Toni Toharudin, Sukono

The aim of this paper is to introduce data analysis methods to identify associations of more than two qualitative variables and several quantitative variables. In addition, the method introduced can be used for quantitative variables with any distribution (generalized). The method is hybrid of Burt multiple correspondence analysis and cosine association matrices. The Identification of information about the similarity between objects, and the associations between object characteristics, is used the modified Mahalanobis distance. The use of this method is clarified by a case study, concerning the accreditation of high schools in West Java. The result is the prioritised improvements for Aranked school are mostly on the facilities and infrastructure. The prioritised improvements for B-ranked school are mostly on the management. For C-ranked school and not accredited school (TT) cannot recommend anything because it is not associated with any characteristics.

Volume 12 | Issue 6

Pages: 826-832

DOI: 10.5373/JARDCS/V12I6/S20201098