8. Conclusion
Depression is a mental condition that can result in more serious problems or even suicide if it is not properly and swiftly treated. Since a complete history of postings might provide crucial information to assist in better diagnosing patients, this type of research may be advantageous to psychiatrists. The TF-IDF characteristics and other characteristics that the authors believed might aid in a better diagnosis of depression were used to test the logistic regression model’s base model. Only the response time between posts improved the fundamental model on particular data subsets among the features the authors used, which also included post sentiment, post count, post length, and average time between posts. In comparison to the model described by Losada et al. [2], the authors’ F1 score using that model was significantly higher. As can be seen in section 5, the main goal was to identify words or groups of words that aid in the diagnosis of depression. Using deep learning models, incorporating further features not included in this paper, like the user’s gender, or experimenting with different word embeddings could all improve the work.