About NOESIS

Noesis is a meta-search engine and resource aggregator that uses ontologies to help users produce intelligent searches of internet-based resources. It suggests search terms by drawing information from its underlying domain ontologies. These ontologies encode domain specific knowledge of concepts, constraints and the relationships among them. For the GoMRC project Noesis has ontologies that express the relationships among terms for Submerged Aquatic Vegetation (SAV) and Harmful Algal Blooms (HABs). Noesis helps users refine their search query and thereby achieve better precision and completeness in their results. The search results are aggregated according to filters selectable by the user. This following describes how Noesis is used to search for information and how the user can organize the results.

Searching for Information

Noesis presents a very simple search interface as shown below. The user simply enters a search term in a text box. As the user types in their term, Noesis will automatically show a list of terms that begin with the typed letters. For example, typing “Sed” as a start toward typing the word “sediment” will result in Noesis suggesting a list of terms beginning with “Sed.” The user can continue typing or double click on one of the suggested terms.

Once the search term is entered, a new page will appear containing four sections - click link below for an example. At the top of the page is information about the search term including its definition. As the user adds or deletes terms from their query the list of search terms currently active will appear there. In the middle of the page below the definition are the search results. On the left side is a list of terms with check boxes for each term. On the right is a list of filters with check boxes next to each.

Link to image of Noesis search results page

Modifying the Search

As the search results are returned and listed in the center of the page, Noesis presents a list of terms on the left side of the page - shown at the left. These terms are generated from the ontology and fall into three categories: Specializations, Synonyms and Related Terms. Users may add these terms to the current search query simply by clicking the text box next to the desired term. Noesis will immediately begin a refined search based on the combination of terms selected.

Specializations can be used to provide a more detailed search. For example a search for “Cyclone” would show specializations, “Hurricane” and “Typhoon”.

Synonyms are different terms that have the same meaning. In ontological terms these are the equivalent concepts. For example a search for “Reflectance” shows synonym, “Albedo”. Appending this term to the query expands the search, thus providing better search coverage.

Every concept has a set of related properties that are neither in the same inheritance hierarchy nor equivalent. These are called Related Terms. They are captured in the ontology through the property relationships. The user can search for resources on a concept with respect to a particular property. For example, a search for “Cyclone” shows “Rain” as a Related Term. Appending this term to the search narrows the search to resources that contain information about “Cyclone” within the context of “Rain. ”

In addition to the terms suggested by Noesis, there is a text box at the bottom left where the user can type in a free text term. User- added terms can be removed from the search easily by unchecking the associated checkbox.

Filtering the Results

On the right side is a list of filters. The search results can be managed by selecting returns from a number of resources by checking the check boxes next to their name. For example if the user only wants to see search results from Google they can check that check box only.

As shown to the right, search returns are available from several resources. The major web search engines Google and Yahoo are available as are a number of publications databases and data catalogs such as the Global Change Master Directory (GCMD). This example shows selected search results from Google, Yahoo and a publications database.

Launching the RTIV and Map Maker

In many cases it is desirable to allow Noesis to search localized catalogs. Research projects may develop a specialized database of information that relates to a special topic, geographic region or research area. In cases such as these, domain experts may have developed conceptual models, ontologies or databases that are very specific. If there exists sufficient information about a data product and there are well defined methods for importing that product into an application, then Noesis can produce specialized search returns that allow the user to launch an application directly. This is the case with imagery and map products that are viewable with the Real Time Image Viewer (RTIV) and the Interactive Maps. When Noesis encounters a search result for such a product, it drops in special icons that will launch the respective tool when clicked. Shown at left are search results from a query on SST with icons that will launch the RTIV and Map Maker.

Noesis Search Architecture

Noesis uses a three step algorithm to search resources. The three steps are Query Analysis, Semantics Presentation and Resource Search. The algorithm architecture is depicted below.

A. Query Analysis

In this step, the user-provided search query is broken down to identify concepts that are defined in the domain ontology. Once they are identified, they are annotated with the associated concepts from the ontology.

B. Semantics Presentation

The annotated concepts from the query string are used to search the Ontology Inference Service (OIS) for associated concepts (Specializations, Generalizations, Synonyms and Related Terms). The Specializations and Generalizations are shown in a tree structure to allow users to navigate through the hierarchy. Synonyms and Related Terms are shown in separate categories and a check box is provided to let the user select the term to append to the search. The user employs these terms to refine the search query.

C. Resource Search

The selected terms are then used for searching the resources. For open web resource searches the refined query is directly used to provide results since no semantic information is encoded (annotated) in these resources. For hidden web resources like data archives, an Application Ontology is added for every new vocabulary used. The concepts in the refined query are used to search the Ontology.

Helpful Hints

On occasion Noesis appears to get stuck when searching for information. This can result from several sources. If you find that the search is taking too long press the "Stop" button and restart your search.

For further information contact the Noesis Development Team.

Noesis Development Team

• Dr. Rahul Ramachandran
• Sunil Movva
• Dr. Xiang Li
• Phani Cherukuri