Hi All, Mahout provides great tools for training classifiers. I am actually researching tools/products that can be used as a service in order to classify text/web-pages.
The main goals/constraints for me are: reduce the dev cycle by using classifications models that have already been trained, and, hence, avoid the cost of building the necessary infrastructure to manage data, measure classification quality… The classification should be rich enough for commercial use. For instance, the top few levels of the ODP taxonomy would be fine, another taxonomy I'd be interested in is ComScore. These are just examples and other taxonomies would probably work as well. I am open to commercial solutions. I have found some companies that provide this type of service, for instance uClassify.com However, the main short coming is, in general, that the classes provided are too coarse and I am looking for a taxonomy that provides content classification at a finer grain (number of classes in the order of ~100). I am seeking pointers to projects/companies that may provide this type of service. Any suggestion would be greatly appreciated. Thanks, S.
