A0381
Title: Unbiased mixed variables distance
Authors: Michel van de Velden - Erasmus University Rotterdam (Netherlands) [presenting]
Carlo Cavicchia - Erasmus University Rotterdam (Netherlands)
Alfonso Iodice D Enza - Universita di Napoli Federico II (Italy)
Angelos Markos - Democritus University Of Thrace (Greece)
Abstract: Defining a distance in a mixed data setting requires the quantification of observed differences of variables of different types and of variables that are measured on different scales. There exist several proposals for mixed variable distances, however, such distances tend to be biased towards specific variable types or measurement units. That is, the variable types or scales influence the contribution of individual variables to the overall distance. Unbiased mixed variable distance is defined as a distance for which the contributions of individual variables to the overall distance, are not influenced by measurement types or scales. The relevant concepts are defined to quantify such biases, and a general formulation is provided that can be used to construct unbiased mixed variable distances.
