Gaining insights from text-based data can be a daunting task, even when the data is labeled with ground truth categories and ready for usage in machine learning tasks.Researchers often rely on simple methods like the frequency of words in each category to understand the collection’s characteristics. However, this approach is not always insightful, as term…… Continue reading Log Odds Ratio: Going Beyond Simple Term Frequencies to Characterize Textual Categories