Tracking English and Translated Arabic News using GHSOM

Abstract

4.1 Map quality For exploratory data analysis tasks where the SOM and GHSOM are used as tools to display similarity relationships from a high-dimensional input space on a low dimensional mapping space, the quality of this mapping is essential. The quality of data representation is measured in terms of the deviation between a units model vector Best Matching Unit (BMU) and the subset of input data vectors represented by this unit. The deviation can be calculated as either the mean quantization error (mqe) or the quantization error (qe) of the single unit. This error is used to measure the map resolution as mentioned by Liu et al. (Liu, et al., 2005)