I ran the keyword and prefix search for some generic keywords like it, there, he, etc. The most amazing part about these was that it gave wrong results and took around 10 times more time to process the request than some named entities like Nokia, Samsung, McDonald's.
Can anyone explain the weird results I get for these keywords
there ====> http://dbpedia.org/resource/United_States
Why are the results wrong and why does it take so much time to process these requests?
I wonder what kind of results you were looking for with a query like "there" or "it"?
In the context of search engine terms these are often referred to as stop words and are sometimes ignored completely due to the fact they are so common that they add very little relevance to the search query or result. I think actually this is what the lookup tool does now as I do not get the same results you mentioned.
Why did the query take longer? This is likely because the words are very frequent and a query for them returns many more results. This means the search engine has more work to do in figuring out the most relevant result.
Why is United_States the top result? Probably because the wiki page for United_States is the highest ranked in terms of inbound links from other Wikipedia pages. This is the heart of the relevance algorithm used within the lookup tool. Essentially there are more links with the words "there", "it", etc pointing to United_States than any other page, so it is judged to be the most relavent for those terms.
