Show Reel - Automatic Categorisation

Automatic categories are good for organising information or linking documents to categories of banner ads.

Categories are created by drawing out the top WordRank words in all the documents that are sampled from a ready-made set of categorised documents. Now that Grapeshot knows which are the optimal words behind each category (as determined from the training set), then any new document can be parsed in real time and given a set of suggested categories that best apply - based on degrees of probability confidence.

The training set can come from any category systems: DMOZ Open Directory categories, UK Government SIC Industry Codes, MESH medical taxonomy or any other taxonomy for which you have a set of sample documents already manually categorised - to use as the training set. The training set we use here comes from Moreover News.