Clusty and Clustering
Clusty is a Meta Web Search Engine offered by Vivísimo [Search Done Right]. Vivísimo’s technology has received a tremendous amount of industry recognition and Vivísimo Velocity has recently won the ‘Best Enterprise Search Solution’ in InfoWorld’s 2007 Technology of the Year Awards Two Years in a Row and has been chosen by U.S. News and World Report to improve and optimize search engine resources for its editorial staff. On the other hand Firstgov.gov, the U.S. government’s official web portal uses Clusty (combined with MSN Search) as Firstgov search engine.
I have heard about Vivísimo and Clusty during a university lesson about Information Retrieval and Machine Learning.
They used a mathematical algorithm and deep linguistic knowledge to find relationships between search terms and bring them to light.
But I want to focus on Clusty. What makes Clusty unique?
Clusty is a whole new way to search the web.
Clusty queries several top search engines, combines the results, and generates an ordered list based on comparative ranking. This “metasearch” approach helps raise the best results to the top and push search engine spam to the bottom.
But what really makes Clusty unique is what happens after you search. Instead of delivering millions of search results in one long list, our search engine groups similar results together into clusters. Clusters help you see your search results by topic so you can hone in on exactly what you’re looking for or discover unexpected relationships between items. When was the last time you went to the third or fourth page of the search results? Rather than scrolling through page after page, the clusters help you find results you may have missed or that were buried deep in the ranked list.
Here is how Clusty results appear. The query is green apple.

Automated grouping of search results by topic is called Search Clustering. It allows the user to further refine the results by clicking on cluster or sub-cluster. «Are you looking for a kind of fruit or for one of the Greenpeace campaigns against Apple? Or for something else?». But there are not only clusters! Clusty Web Search allows you to see which source or sources the results are coming from (Ask, Gigablast, MSN, Wikipedia,…) and select what you prefer. You can do the same thing with the URLs.
Google has recently added clustering to the Google Search Appliance, a box that indexes documents from a company’s intranet and web sites.
It will be interesting to see if Google decides to add clustering to the web search. Other search engines that have this features (like Clusty) don’t have extraordinary results, but may help you view the big picture and guide you in refining your query.
I think clustering could help people to start obtaining information from web search engines, and not just webpages.
Clusty Search Engine is available also for News, Images, Wikipedia, Blogs, Jobs, Shopping,… For more details I refer to this Overview of Clustering and Clusty Search Engine on Read/Write Web.
Tell me what you think about Clusty in a comment.
Popularity: 65% [?]
Related Posts:











Franciov WebStore

Email me
Comments feed
January 7th, 2007 at 3:20 pm
Thanks for the link Franciov, I like Clusty.
The idea is great - I don’t think the algorithms are 100% right yet, but I’ll definitely keep my eye on this one.
F
January 7th, 2007 at 3:39 pm
One of the issues I have with clustering is that some of the categories returned for a result can be very generic and require the user to dig into each one to find “more refined / better” search results, which is problematic. 1) Overdependence on categorization / meaning of categories 2) Doesn’t give users the “instant gratification” they are looking for. Granted, this is only if the initial search came up with unsatisfying results. Just a thought.
January 8th, 2007 at 3:08 pm
I’ve seen Clusty before, but not look into it in much detail .. but I will now!
As an aside, what Clusty seems to be offering is what the Semantic Web (go to: http://www.w3.org/2001/sw/) proposes, but I suspect we’re some way off a fully Semantic Web at this stage…
January 8th, 2007 at 11:34 pm
Well, the algorithms are not 100% right, of course. As Wayne has just mentioned, we are in the beginning stages of this kind of evolution. Semantic Web is the Web 3.0 purpose (?), but now we are still in the middle of Web 2.0, right? I mean.. Clustering could be a good start, but the age of RDF/OWL framework is far away.
If the initial search, as Jing said, comes up with unsatisfying results, Clusty proposes the clustering solution I’ve explained in this post. You can refine your query as you wish. Yes, clustering algorithms have to be improved, but what about Google trends? Do you think Google will decide to add clustering to the web search?
In my opinion, it would be a useful feature for users, albeit might be expensive to realize.
January 9th, 2007 at 12:25 pm
“Do you think Google will decide to add clustering to the web search?”
Possibly, but then it becomes a question of page layout. On the one hand, it’s about search, while the commercial PPC stuff comes second.
So, if Google decide to enhance their search with clustering, something has to go, or be made smaller, right?
Dilemma…