Summary of Enhancing Depressive Post Detection in Bangla: a Comparative Study Of Tf-idf, Bert and Fasttext Embeddings, by Saad Ahmed Sazan et al.
Enhancing Depressive Post Detection in Bangla: A Comparative Study of TF-IDF, BERT and FastText Embeddings
by Saad Ahmed Sazan, Mahdi H. Miraz, A B M Muntasir Rahman
First submitted to arxiv on: 12 Jul 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper introduces a well-grounded approach to identify depressive social media posts in Bangla by employing advanced natural language processing techniques. The study uses a dataset annotated by domain experts, including both depressive and non-depressive posts, to train and evaluate models. To address class imbalance, random oversampling is used for the minority class. Various numerical representation techniques are explored, including TF-IDF, BERT embedding, and FastText embedding, integrated with a deep learning-based CNN-BiLSTM model. The results show that the BERT approach performs better, achieving an F1-score of 84%. This highlights the efficacy of different embedding techniques and deep learning models for detecting depressive posts in Bangla. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This study helps detect depression on social media in the Bangla language. It uses special computer programs to analyze social media posts and identify those that might be from people who are depressed. The researchers used a big collection of Bangla-language posts, both happy and sad ones, to train their program. They also tried different ways of turning words into numbers to help their program understand the posts better. In the end, they found that using a special kind of computer program called BERT worked best at finding depressive posts. This is important because it can help us create tools to monitor people’s mental health on social media. |
Keywords
» Artificial intelligence » Bert » Cnn » Deep learning » Embedding » F1 score » Fasttext » Natural language processing » Tf idf
