The classic use of GIS for spatial analysis is a well-tested and at first glance seems to be a lot easier then Semantic Web approach. However, traditional analysis using GIS technologies are difficult to define in full, finite form, joining domain model, the analytical process and evaluation criteria. The definition of the process of analysis, does not come directly from the description of events contained in the logical domain model. Data application schema derived from the logic model, a description of the process of analysis and definitions of decision criteria exist as independent descriptions. This requires manual planning of processing analysis. Building of Decision-making model requires constant human intervention. Requires the participation of an expert to determine how to use available data and create analytical procedures.
GIS systems use data models, which are not self-describing. Relational database schema describing structure of data storage, refecting mutual dependences which is important to providing data consistency. But this model does not explain the method of classification. A description of the analysis process is the “external” in relation to this model, too. Therefore the process of GIS analysis is the result of arbitrarily defined sequence of analytical steps, detached from the model description.
Inference processes in Semantic Web technologies, performed by reasoning engines, benefit directly from the models stored in formal logic systems. Description Logic provides self-describing model of reality, together with the classification criteria.
Ontologies contain the recipe how to interpret the datarecorded in the defined model and how to enrich data description with new facts (how to make the reclassification of phenomena). However, support for spatial analysis by a clean DL system, now facing serious performance issues. This is mainly a problem in geometry and topology analysis.
Performance of geometric data processing and spatial analysis
Currently GIS huge advantage over the DL in the processing performance of spatial data. GIS uses powerful mathematical algorithms with well-developed analysis of topological features. It gives you a huge advantage in terms of efficiency GIS analysis, over the analysis DL. However, here we can talk about some artificial junction of the relational model with specialized spatial analysis functions too. Relational model is addressed to store and retrieve data and spatial analysis functions are added as if from the outside. Theoretically, in the future, DL model could be enriched in a similar manner.
Description Logic – integration of data from different models
The powerful advantage of Description Logic is the ability to integrate data from different databases. Description Logic, allows you to merge ontologies, mapping entities from different ontologies and ontology alignment. Such prepared resources can then be seen as a unified model describing the reality. The data stored in OWL DL ontologies and described by SWRL rules could be a subject of inference process.