BCCWords: Bayesian Text Sentiment Analysis using Crowdsourced Annotations

How do we build an automated tool for text sentiment analysis that learns from crowdsourced human annotations? This is the challenge addressed by the Bayesian Classifier Combination with Words (BCCWords) model presented in the paper:  Edwin Simpson,  Matteo Venanzi,  Steven Reece,  Pushmeet Kohli,  John Guiver,  Stephen Roberts  and  Nicholas R. Jennings, (2015)  Language Understanding in the Wild: Combining Crowdsourcing and Machine Learning. …

1