Stock market prediction using Google Trends data

Absztrakt (kivonat)

Since the 2000s sentiment analysis is getting a more and more popular field of research in many areas. My paper examines whether measuring investor attention has a place amongst the investment decision-supporting tools as well. Based on the research of Preis et al. (2013) I will be conducting a case study on how effectively Google search volume could be utilized for stock market predictions. Also, my intention is to create a well-documented and reliable methodology for this fairly new type of prediction technique. While developing my modeling framework, I intend to thoroughly analyze the possibly untouched limitations and overlooked biases involved in the seminal article, and shape a prudent framework for my research accordingly, supported by other best practice examples from this area. To prove the validity of my model (and its results), I will also perform a robustness testing in the end. Finally, I will be listing several points and recommendations which should be considered if one decides to utilize the model in real-time day-to-day trading.