Title: Local scaling in marked point process: Does the health of company influence another company?
Authors: Ilona Berkova - University of South Bohemia in Ceske Budejovice (Czech Republic) [presenting]
Tomas Mrkvicka - University of South Bohemia (Czech Republic)
Abstract: The localization of corporate activities is the most important decision for new company and is the result of many factors. Due to this fact, the location of firms has been a part of economics for many years. The aim was to estimate an econometric model to recognize the spatial relationships of the firms according to their health in inhomogeneous space. The data set consists of firms position in 3 regions in the Czech Republic whose health was determined according to the Neumeiers index IN05. The dependence of marks in a point pattern is usually expressed through the marked L-function. The function can be extended for inhomogeneous processes with second-order reweighted stationarity assumption through weighting by intensity. Since the behavior of firms in a distance r depends also on population density or density of firms, the appropriate inhomogeneous tool for the firms environment is local scaling. The local scaling assumes larger distances between firms in a less populated area and transforms the space according to a function. Therefore, the local scaled marked L-function that expresses dependence between the firms' health in transformed distance r is defined. There was applied a global envelope test with the null hypothesis of no dependence between the firm health. It allows deciding if the near firm with good health influences the health of the firm and if the near firm with whatever health influences the health of the firm.