A comparison of optimal map classification methods incorporating uncertainty information

Uncertainty in spatial data attributes can produce unreliable spatial patterns in choropleth maps, but only a few studies have considered uncertainty in map classification processes. Unfortunately, a less desirable classification result often is generated by existing methods. For example, most observations are assigned to a single class while the remaining classes have a very small number of observations allocated to them. Also, selection of proper criteria for an optimal map classification is difficult. The purpose of this paper is to expand the discussion about incorporating data uncertainty for map classification by extending optimal map classification strategies with Bhattacharyya distance. The proposed method is illustrated with an application of soil lead contamination measurements in the City of Syracuse.