Prediction of a Musical Show Liking Using Bio-signals of an Audience

This paper measured the bio-signals and liking of the audience watching a musical show, and then conducted machine learning training using this data. Subsequently, the trained machine learning model was utilized to predict the audience’s liking for a musical show. As a result, it was possible to achieve a prediction accuracy of 74.38%. Through additional analysis, it was confirmed that the highest prediction accuracy could be achieved when predicting the audience’s liking for the musical show using support vector machine (SVM) and utilizing pupil and facial expression data.