Unified coding of qualitative and quantitative variables and their analysis with ascendant hierarchical classification

O. Moschidis
DOI: https://doi.org/10.1504/IJDATS.2015.068745
2015-04-01
International Journal of Data Analysis Techniques and Strategies
Abstract:The present research deals with the grouping of individuals, who are determined by a number of categorical, ordinal and quantitative variables. This research applies the method of ascendant hierarchical classification CAH, using as criterion, the inertia and the distance metrics of X2. The usual dealing with this problem refers to the coding of the certain categorical and ordered data in a logical table 0-1 under the form: individuals x variables. However, the sub table of the table 0-1, which concerns the ordered variables, deducts information in advance as it deals with the ordered variables as being simple categorical variables. The research suggest a new encoding of the quantitative variables and of the ordinal variables that leads to a new more detailed encoding of categorical and ordinal variables in a new general table of percentages, the analysis CAH of which provides more detailed conclusion.
What problem does this paper attempt to address?