Information Retrieval Properties of Teoma

I investigated two information retrieval behaviors of the search engine, Teoma. I first investigated whether capitalization affects the retrieval of results by varying capitalization using my name as a search term. Next, I investigated the role of stemming by looking at the number of results retrieved when searching for variations on the word "revolve."

CAPITALIZATION

I searched my name as a phrase (in quotation marks). When I varied the patterns of capitalization of my name, I obtained the same results each time. There were only 15 hits for each query. The results listed were the same in each case, and they were in the same order. On the basis of this information, I concluded that capitalization plays no role in search outcomes in Teoma.

STEMMING

When I searched for the variations on the word "revolve" shown in the table below, I obtained a different number of hits for each search. A search for "revolv*" only brought up results which included that term somewhere in the page. This is also indicated by the small number of hits for "revolv*" in comparison with all the variations on the term. Pages that included any other word that might include "revolv" such as "revolve" or "revolving" but not "revolv*" were not included in the hit list. Also, the fact that the number of hits for "revolve" is smaller than the number of hits for "revolves" indicates that the list of results is comprised only of pages that specifically include the term "revolve." On the basis of this information, I concluded that Teoma does not stem terms in indexing web documents.

From my investigations, I concluded that searches in Teoma are not affected by capitalization of search term, and that the search engine does not make use of stemming in indexing results.

Return to my portfolio.

E-mail questions to Meghan Lafferty at melaffer@email.unc.edu
Last updated October 20, 2002.