This is an attractive project which comes up with an original perspective about predicting the movement of stock market. A sentiment indicator, built from mining the tweets and social media, would be embeded in trading models to improve their performance. The project will focus on sentiment from professionals, who exert major influence on the market.
Strengths:
Grounded rationale and feasible methodology. Traditional statistical models emphasize that the past would repeat in the future. However, this project selects sentiment from professionals to explain the market fluctuation. It is intuitive that those major players with abundant financial resources and sufficient information will trade to maximize their profit and lead the market. Thus, the sentiment is a reasonable indicator.
Suggestions:
Definition: "sentiment" is a psychological term. While applied in data analysis, the term needs to be defined in a straightforward way. Then, select and process data to reflect sentiment.
Dataset: The data can be obtained from mining the tweets directly or data providers. What is your choice?
This is an attractive project which comes up with an original perspective about predicting the movement of stock market. A sentiment indicator, built from mining the tweets and social media, would be embeded in trading models to improve their performance. The project will focus on sentiment from professionals, who exert major influence on the market.
Strengths: Grounded rationale and feasible methodology. Traditional statistical models emphasize that the past would repeat in the future. However, this project selects sentiment from professionals to explain the market fluctuation. It is intuitive that those major players with abundant financial resources and sufficient information will trade to maximize their profit and lead the market. Thus, the sentiment is a reasonable indicator.
Suggestions: Definition: "sentiment" is a psychological term. While applied in data analysis, the term needs to be defined in a straightforward way. Then, select and process data to reflect sentiment. Dataset: The data can be obtained from mining the tweets directly or data providers. What is your choice?