Longitudinal Sentiment Classification of Reddit Posts
This work addresses sentiment analysis for university student posts on Reddit, but it is incremental as it applies an existing method to new data without major innovations.
The study tackled sentiment classification of Reddit posts from students at four Canadian universities from 2020-2023, achieving consistent results across datasets by fine-tuning a sentiment threshold to [-0.075,0.075] for categorizing posts as positive or negative.
We report results of a longitudinal sentiment classification of Reddit posts written by students of four major Canadian universities. We work with the texts of the posts, concentrating on the years 2020-2023. By finely tuning a sentiment threshold to a range of [-0.075,0.075], we successfully built classifiers proficient in categorizing post sentiments into positive and negative categories. Noticeably, our sentiment classification results are consistent across the four university data sets.