Poster Review - Shuting Chen

[Shuting05]

Hi Mengchen, this is your friend Shuting! Your poster looks pretty good to me, especially for the fantastic graphs! The overall structure of the poster is very concise and coherent, with a great balance between words and visualization. The idea of using news information to predict the expected volatility in the stock market is really interesting and it’s good to know that you use NLP as the main computational method. I really like the graphs that you used to display the results, which make it clear to compare the in-sample estimation performance with out-of-sample performance. Besides, the flow of the poster is really good, for example, the estimation model is presented with precise equations and helpful explanations.
Potentially, I wonder if it would be better to scrape more new articles over a longer time period. Due to the time constraint in this quarter, it might be impossible to scrape more data; but it is worth trying to collect more data if you would like to extend the project for future study. As for the problem of overfitting, you propose a possible way by adjusting parameters like number of dimensions. I wonder if this could be alleviated by trying another model such as bag-of-words model?
Overall, it’s an excellent poster and I really like the layout!