A Learning Sentiment Database for Machine Learning

Abstract Learners’ affective states play a crucial role in learning evaluation, and the external expressions that can directly reflect affect are facial expressions. However, sample size of database for process learners is limited, most existing expression databases ordinary people, which difficult to support an in-depth study emotional scenarios. In order better influence learners’ emotion on state, we built machine learning, at same time labeled data with expression-emotion finally formed 8,000 images containing 36 learners, important studying state under