An Inquiry on using Sentiment Analysis on the Annual Reports of Selected Listed Companies in the Philippine Stock Exchange
This study explores the application of sentiment analysis, specifically utilizing the FinBERT transformer model, in assessing the sentiment expressed by company management in the Management Discussion and Analysis (MD&A) section of annual reports of Philippine listed companies. The study aims to investigate whether the sentiment scores extracted from the MD&A section are related to abnormal returns surrounding the release of the annual reports. The implementation of FinBERT on the annual reports demonstrates the effectiveness of modern NLP techniques in quantifying management sentiment. However, the findings indicate that while sentiment scores can be derived from the MD&A section, they do not provide a distinct advantage for financial decision-making or stock trading. Specifically, the sentiment derived from new information does not have a significant relationship with subsequent trading prices in the short window after the release of annual reports. This study also aims to contribute to the Philippine financial and accounting research by employing advanced Natural Language Processing (NLP) techniques for textual analysis of Philippine financial literature.
MIT - Jomar Jordas - 2023
https://github.com/teles/array-mixer for the amazing readme file that I used as a template.
https://github.com/TiesdeKok/Python_NLP_Tutorial c/o Professor de Kok as I extensively used his Youtube Channel and repositories as reference in learning NLP
https://sraf.nd.edu/textual-analysis/code/ c/o Professors Loughran and McDonald for opening to the world their research. My research is heavily inspired by their work
http://kaichen.work/?p=399 c/o Kai Chen for his extensive blog post that further brokedown the Loughran and McDonald dictionary. I referred to his code when I was working on the LM dictionary and how to apply it to my dataset.
https://github.com/jamescalam c/o James Briggs for his wonderful Medium blog and Youtube channel where I actually learned how to apply and use transformers model as well as how to make use of the GPU of my Macbook M1 in my code