This document discusses classifying students' e-learning experiences shared on social media via text mining. It aims to identify problems students face in their learning by analyzing unstructured social media data like tweets, posts and comments. The proposed method introduces a new label "Good Things" to classify positive student experiences alongside existing labels for problems. A naive Bayes multi-label classifier is used to calculate the probability of words in tweets belonging to each label category. The classified tweets with the new label will then be compared to tweets with existing labels to improve understanding of student experiences and enhance e-learning quality.