Interval Delimitation of Mean Line of Geographical Phenomena from a Set of Polygons: Case Study of Mental Maps for Haná Region

Abstract

The paper proposes an innovative method for estimation of mean border from a set of polygons. This method is based on statistical estimation of mean line using distribution of distances at given directions. Confidence intervals for mean line are calculated, which allows to incorporate uncertainty into the analysis. Compared with currently used methods, the proposed procedure offers an objective statistical evaluation of data, it enables analyst to include his/her expert knowledge into the estimation process and enables statistical testing of hypotheses. Confidence intervals can be used to evaluate variability in different directions. Performance of method is demonstrated on both experimental and real data representing mental maps of ethnographic region Haná. The results with experimental data are very promising, since these data are generated by a well-known process. The results with real data are comparable with those, reported by previously published studies.