Ever used a voice of customer or other marketing tool that claims to tell you how your customers really feel? If so, wondered how the heck these systems can figure that out? Or maybe why they are failing to correctly figure it out? It’s not magic, it’s computational linguistics.
In working with sentiment analysis, it’s important to understand what it can and cannot do so you have realistic expectations for how to actually plan to use it. Digging into sentiment analysis is also a good way to get an introduction into some of the ins and outs of text mining in general, as this data is still a largely untapped resource for marketers.
For our February event at Rev1 we will have Dr. Marie-Catherine de Marneffe from OSU crack that black box open a little to show us what’s inside. She will give a quick overview of the techniques used for sentiment analysis as well as point out some current limitations. We will also explore the databases (corpora) developed for sentiment analysis research, highlighting some of the issues in how these are used, as well as inherent bias in the data that the community ought to be aware of.
All of our Columbus WAWs are free of charge thanks to the generosity of our sponsors: Clarivoy, TeamBuilder Search, the MS in Analytics Program at Wittenberg University, and Search Discovery.