After 12 months of hard work
we are proud to announce the release 3.0 of Carrot2
The 3.0 release introduces significant improvements in all areas of the framework:
Simplified Java APIDownload||
With new adopters in mind, the 3.0 release significantly simplifies
the Java API of Carrot2. Three statements are enough to get started:
Map<String, Object> attributes =
new HashMap<String, Object>();
attributes.put(AttributeNames.QUERY, "cluster me");
List<Cluster> clusters = new SimpleController()
Contrary to the 2.x line, the 3.0 release requires Java JDK 1.5 or higher.
Improved clustering quality
||Carrot2's default clustering algorithm, Lingo, has been re-implemented
for improved clustering quality and extensibility. For both Lingo and STC algorithms,
we have exposed a significant number of attributes that will allow you to fine-tune
them for your needs.
Brand new Carrot2 Document Clustering WorkbenchDownload||
The 3.0 release introduces a brand new GUI application for
experimenting and tuning Carrot2 clustering.
You can use Carrot2 Document Clustering Workbench to quickly start clustering
your own data or to tune clustering results.
Brand new Carrot2 Web ApplicationView online||Carrot2 Web Application has been reimplemented for easier customization
and top front-end performance. We also included an attractive visualization.
Improved Carrot2 Document Clustering ServerDownload||
The 3.0 release improves the Document Clustering Server by adding the ability to query external
document sources. It also adds a simple web-based quick start screen.
You can use Carrot2 Document Clustering Server to integrate Carrot2 with your non-Java software.
| User and Developer ManualView online||Carrot2 now comes with an initial version of the user and developer manual.
We will be extending the manual based on the frequency of questions
on the mailing list.
If you have concerns about choosing between the 2.x and 3.0 versions or migrating
from 2.x to 3.0, please see the release 3.0 Frequently Asked Questions.