Despite the success of massive open online courses (MOOCs) in attracting learners, low completion rates are a major issue. The complexity of forum information often means that urgent student problems are not addressed promptly. This paper introduces a model using deep learning, specifically an improved recurrent convolutional neural network, to identify “urgent” posts that need immediate instructor attention. The model outperforms existing methods, potentially aiding instructors in managing discussion forums more effectively, which could help reduce dropout rates and improve the learning experience.