I recently came across Newssift: it is a semantic search engine for business. The tool is still in beta and is definitely a good start to semantic search. I read a lot of positive buzz about the engine and some calling it better than google. But the company has to fix some kinks if it has to become the destination for business search.
A search on H1N1 virus yields about 8500 results. Interestingly, about 600 (7%) of the results are positive. Intrigued, I click more. The first 2 results are the following
- Its so cool for H1N1 virus in Bangalore: the article talks about the cool city temperature and how that is a favorable factor for survival of the virus. Apparently, the algorithm picked up words like “cool” and “favorable” and assigned a positive sentiment to the article.
- Colombia president catches H1N1 flu: a news article that states that the Columbian President Alvaro Uribe has caught the swine flu. I was surprised that this is categorized as positive. The only rationale could be that further in the news the article states that “Doctors say Uribe is being treated and he is progressing satisfactorily.”

There were 2 more results on the second page (Other results on the first page were on target)
- Despite H1N1 threat, most Americans haven’t increased hand washing: the article that states that 54 percent of the people surveyed said they “wash their hands no more or less frequently” in public restrooms since the H1N1 virus emerged.
- Kids going back to school will cart H1N1 with them: you look at the heading of the news and think how can this be positive?
That said, I like Newssift for the fact that I can view results by timeframe and the firm offers a unique way to classify information and search results. Let’s give it a little more time and input to fine tune the kinks.


