Towards a Generic Framework for Semantic Registration of Scientific Data
Shawn Bowers, Bertram Lud · 2003
Ontologies provide standard sets of terms and formal definitions for concepts in a domain, and are increasingly used to uniformly access information. For example, using concepts from an ontology to annotate disparate information (such as Web-page content) allows ontology-aware applications to retrieve and navigate otherwise heterogeneous sources. In this paper, we consider the specific problem of registering scientific data (as opposed to arbitrary Web content) with ontologies. We propose a generic framework to support semantic registration of scientific datasets, which we intend to deploy in the SEEK1 project—a multidisciplinary effort to help scientists discover, access, integrate, and analyze distributed ecological information. Our goal is to develop a framework that can support data providers by allowing them to easily enrich their datasets using ontological concepts, while at the same time providing end users semantic-discovery and data-mediation services. To illustrate the goal of semantic registration, Figure 1 gives an example ecological ontology that includes concepts for datasets, measurements, and locations. The ontology is expressed using standard description-logic syntax (Sattler 2003), but could just as easily be expressed using OWL (McGuinness & van Harmelen 2003). Figure 2 shows part of a dataset,2 which includes concepts from Figure 1. In particular, the dataset has abundance data for plant and animal species located in giant-kelp forests along specific regions of the Santa Barbara Long-Term Ecological Research (LTER) site. The dataset gives the date and location of samples, species codes, and the number of species observed. Through semantic registration it should be possible to connect the dataset of Figure 2 to the appropriate portions of the ontology of Figure 1. The rest of this paper describes our proposed framework for semantic registration. The framework lets a data provider choose the appropriate concepts within an ontology that best