DigitalPebble is currently developping a solution for Opinion Mining and Sentiment Analysis.

Sentiment Analysis consists in categorizing some text as being subjective or not and possibly as having a positive or negative connotation. This can be used for instance to determine whether a newspaper advert is mentioned in a negative context, e.g surrounded by bad news on the same page, which could affect the price of the advertising.

Opinion Mining provides a slightly more refined information as it aims at extracting occurences of concepts and the sentiments associated with them. These concepts can be for example products and their associated features, for instance specific brands of mobile phones and their attributes (reception, keyboard, music, camera, etc...).

The OpinionMiner can be used for analysing blogs, forums or any web page and provides a rich information about what users and consumers think about a given product. It is available as a web service and can be easily integrated into an existing workflow or application.