Hacker News new | past | comments | ask | show | jobs | submit login
Carrot2 Clustering Engine (carrot2.org)
17 points by dualogy on July 5, 2009 | hide | past | favorite | 11 comments



I did extensive research on this thing when I was doing my last start-up. It's a clustering engine, but it is limited in that it is specifically tuned to cluster search results. For example, it returns poor clustering results if you input a set of long text documents; it works much better when its input consists of a query string and a set of titles and short snippets.


where can I find publications, how this thing is doing the term selection for the topic description?


All of the publications related to the project are locate at

http://project.carrot2.org/publications.html#ieee-lingo

but I am no longer sure what the exact publication that would answer that question is


Am I the only person that thinks the visualization is the smart part (and wonders why it's not the front page)?


both are smart. i definitely agree that this is the first search visualization that looked like more than eye candy. i mean, it is fun to use, but it also seems to be useful.

i love the tabs, too. news v web v images etc.


Nothing to do with "Clustering"; it's a multi search-engine result filter, albeit a fancy one.


How do you think they generate the left side bar?

Maybe this page is more useful to you: http://project.carrot2.org/


grrr damn web designers .... I need to READ not to watch nice shadows everywhere!

... aw a Cmd+ resolve this, but the whole design do to hell nicely :)


huh? i was unable to parse anything after "grrr damn web designers."


This is really cool. One of the most useful search engines I've seen in a long time, not to mention open source.


impressive user iteraction, and fake queries seem to return good results. i'll use the engine throughout the day and see if the results are good 'nuff.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: