Biomedical Research

Correspondence analysis is an analysis method that facilitates interpretation of categorical variables in the cross tables (correspondence table, contingency table) as well as the similarities, divergences and associations between the row and column variables, and represents these associations graphically in a lower-dimensional space. It is a method quite popular especially in the fields that require analysis of categorical data such as medicine, health sciences, biometrics, economics, marketing and social sciences. Correspondence analysis has some similarities to the other multivariate methods such as the principal component analysis, log-linear analysis and multi-dimensional scaling. Most purchases by public institutions are determined first 5 items to demonstrate the purchase profile of product in "Hospital Equipment, Furnishings and Equipment" product line in Turkey. These are wheelchairs, companion seats, emergency stretchers, wheeled patient nightstands and examination tables. The compatibility analysis was performed according to purchase frequency and the preferred status in the order cycle of these determined products. In result of multiple correspondence analyses, the explanation ratio of the first two dimensions of the corresponding variable was found to be 97.2%. It is observed that the preferred state of the product type is primary with a ratio of 87% and the number of product order is secondary with a ratio of 10.2%.

Author(s): Mustafa Agah Tekindal1*, Ümit Cengiz Uysal2, M.Özgür Dolgun3
Abstract | Full-Text | PDF

Share this