Natural Language Processing for Evaluating Customer Emotion
By Thomas Metcalfe
I recently had the opportunity to speak at Online CX Emotion, a conference organized by Seth Grimes of Alta Plana to bring together spokespersons of multiple companies operating in the emotion analysis industry.
The conference was a full day of excellent presentations from data scientists and leaders of customer experience, detailing their strategies and practices in sentiment, social, behavior, and emotion technologies. A personal highlight was the talk from Dr Danica Damljanovic on crafting emotion-centric customer interactions, where she shared some of her research on correlations between agent performance and sentiment in support queries. Alya Yacoubi, team lead at Zaion, also delivered an insightful discussion into the parallel importance of both understanding and conveying emotion accurately in voice-platform chat bots.
Working as a deep learning engineer, I thought it would be a good place to give an overview of the algorithmic strategies that are commonly employed in the emotion analytics space, evaluating each one with its pros and cons. Armed with this, I turned to discuss why techniques such as machine learning are invaluable when attempting to understand emotion at scale, and then gave a brief insight into the future of these practices.
You can catch the whole of my talk below:
If you are interested in viewing the rest of the conference lineup, the event videos are available on the CX Emotion website.