CUSTOMIZING A SEMANTIC SEARCH ENGINE AND RESOURCE AGGREGATOR FOR DIFFERENT SCIENCE DOMAINS

The goal for search engines is to return results that are both accurate and complete. The search engines should find only what the users really want and everything they really want. Search engines (even meta search engines) lack semantics. The basis for search is simply string matching between the user's query term and the resource database, thus the semantics associated with the search string is not captured. For example, if a atmospheric scientist is searching for “pressure” related web resources, most search engines return inaccurate results such as web resources related to blood pressure. Noesis is a meta-search engine and a resource aggregator that has been designed to utilize domain specific ontologies to provide specialized scoped search capabilities. Noesis uses domain ontologies to help the user scope the search query to ensure that the search results are both accurate and complete. Semantics are captured in the domain ontologies by defining terms along with their synonyms, specializations, generalizations and related concepts. These domain ontologies are used by Noesis to guide the user to refine their search query and thereby reduce the user's burden of experimenting with different search strings. Noesis also serves as a resource aggregator as tt categorizes the search results from different online resources such as education materials, publications, datasets, web search engines that might be of interest to the user. Noesis can be customized for use in different domains by configuring it to access different ontologies or to search different online resources. Currently we have a general purpose atmospheric science version of Noesis available, and we are creating a coastal ecology version specialized for the Gulf of Mexico Research Collaborative, a NASA Applications project. In addition Noesis portlets for the ESIP Federation Environmental Information Exchange and Geospatial One Stop are planned.