ABSTRACT
It is rare for a data producer to revise a dataset from scratch. Partial revision of a dataset using various sources is usual in order to save time and cost. On the other hand, data accuracy and currency become more complex to report as multiple data capture methods are used. According to the Geospatial Positioning Accuracy Standards (FGDC, 1998a), accuracy of each method is separately recorded, but even if this guideline is applied on metadata, spatial variation of data quality across dataset coverage cannot be identified graphically. Further, currency and level of generalization are still revealed on a holistic basis. In this respect, metadata should be captured in feature level (Gan, 1999) to provide information on quality, temporality, data source, and processing step as well as generalization effects for identifying the suitability of a dataset in an area of interest. These pieces of information facilitate data users in data mining and knowledge discovery processes as described by Buttenfield et al. (2000).
