Bayesian Ontologies in Spatial Integrating Medical Information Systems
Geographical Information Systems (GIS) play a major role in all areas of health research, especially for the understanding of spatial variations concerning disease monitoring. The information produced by the spatial analysis can be modelled and displayed using maps. Spatial analysis (as an alternative statistical technique) may be used in order to suggest health patterns for describing the spreading of various diseases. Areas where GIS can be of benefit include the point mapping of patients and aggregated analyses within different geographical areas. The incorporation of GIS sections in Healthcare Information Systems aims towards the efficient and automated follow-up of prevalence of various diseases in diverse geographic regions. A very important feature of the current system is the integration of queries for the extraction of specific information regarding the above parameters. The queries have been developed through the ontologies of the system. Each ontology refers to each of the correlations that are being explored. The appropriate ontology design techniques have been used to assure the validity of the query output. This work describes the methodological approach for the development of a real time electronic health record, for the statistical analysis of geographic information and graphical representation for disease monitoring. Uncertainty of the ontology system may be achieved by proposing Bayesian type statistical techniques like Bayesian network and Markov logic. Implementation of the proposed techniques will be illustrated considering real accident data.