Understanding your customer’s context – who they are – is key to offering a targeted service and personalised communication. There are currently solution providers who offer limited demographically labelled data which comes from known sources. We intend to deal with data coming from unknown sources by analysing the clues
which a customer leaves online.
From the language they use and their behaviour, to their friends and likes, XRCE’s research is focused on using a customer’s digital footprint to automatically determination their traits, beyond straightforward demographics.
When it comes to understanding what your customers think about your brand, it’s key to understand the sentiment expressed in every comment they make.
Not only can we detect the overall sentiment of a comment, but our technology now enables a very fine grained sentiment modelling in order to connect individual expressions of opinion to different aspects of a product or service: for example “food” vs “décor” in restaurants, or “price” vs “image quality” of digital cameras.
XRCE’s work on Sentiment Analysis has demonstrated very competitive results on the Aspect Based Sentiment Analysis task at the SemEval 2014 international evaluation campaign (links to the system description and the overall results ).