Knowledge discovery in oceanographic databases: issues of complications in data sources

Abstract

Data mining or knowledge discovery refers to a variety of techniques having the intent of uncovering useful patterns and association from large databases. We have been working with data mining techniques for a variety of oceanographic data and have encountered a number of troublesome issues relative to available data. We describe the steps preparatory to data mining and three data mining techniques that we have applied to spatio-temporal data. We include a detailed review of various sources of geospatial, oceanographic and meteorological data and associated issues inherent in their use in knowledge discovery. We also provide issues relevant to the difficulties in providing an overall integration of this heterogeneous data for knowledge discovery.