 |
|||||||
 |
|
||||||
|
||||||||||||||||
 |
|
 |
|
|
Quality of integrated geospatial data depends on the quality of original geospatial data sources. However, it also depends on the quality of the semantic mappings that are established between the schemas of these sources. Since semantic mappings are used in querying multiples sources, their quality will affect decision making in domain where there is a need for high quality and accurate data such as in the health domain. Thus, being aware of the quality of the semantic mappings can help to avoid propagation of low quality data through the mapping process. However, the question that is still unanswered is how quality of semantic mapping can be assessed. In our presentation, we propose a new method for characterization and assessment of the quality of semantic mapping. We define a set mapping conflict predicates that are used to detect mappings of low quality between geospatial databases. We also propose a method that uses these predicates for the assessment of the quality of semantic mappings. We have applied this method with geospatial databases of the health domain and it shows that it prevents risks of misinterpretation of integrated geospatial data. Mohamed Bakilah, étudiant
|
