On Estimation of the Spatial Clustering: Case Study of Epidemiological Data In Olomouc Region, Czech Republic

An evaluation of spatial patterns and a clustering play an important role among methods of spatial statistics. However, traditional clustering techniques are seldom suitable for analyses of spatial data and patterns because they usually do not count on spatial relations and qualities of objects. This paper aims to introduce usage of methods of spatial clustering estimation, which are based mainly on the position of events and not only on the events attribute space. Firstly, the methods of the spatial clustering or randomness estimation are introduced and applied on a real dataset, then spatial clusters are identified and the intensity of processes is quantified. Non-spatial properties and a time are considered together with the location data. Also methods of the multivariate statistics are used for the purpose of the classification of regions with similar properties. Particularly, occurrence data of selected infectious diseases in Olomouc Region in period 2004 – 2010 provided by Regional Public Health Service in Olomouc are used for the case study.