A Conceptual Framework in Developing a New Location Model

Abstract

The original purpose of the location model is to deal with mixed variables discrimination for classification purposes. Due to the problem of empty cells, smoothed location model is introduced. However, the smoothed location model had smoothed all the cells either empty or not, where the smoothing process causing changesto the original information of the non-empty cells. As it is well known that those original informationis a valuable source and important in any study that should be maintained. To address the aforementioned issues, an amalgamationof maximum likelihood and smoothing estimations is introduced to construct a new location model. The amalgamation of both estimations is expected could handle all situations whether the cells are empty or not based on several settings of sample size and number of variables.