Impact of jumps in prices and volatility on dynamic portfolio strategies — assessment on public mood-driven asset allocation

• Main Author: Bey, Traacy Jing Ling
• Other Authors: Nicolas Privault
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2021
• Subjects: Science::Mathematics::Analysis
• Online Access: https://hdl.handle.net/10356/148521
• Institution: Nanyang Technological University

As the financial market is often volatile and influenced by many factors, the analysis of different kinds of data may be looked into in determining the significance of different factors in affecting the changes in the financial market, specifically through stock movements. Instead of looking at individual stock market analysis that are generally more researched upon, the concept of dynamic asset allocation of adjusting the weights on different assets at different time periods may be looked into in terms of creating a specific strategy set that can be applied to obtain higher returns over time. This research looks into the use of sentiment analysis towards the portfolio allocation strategies in maximising returns and whether the machine learning algorithm is better able to outperform that of a benchmark allocation strategy. The main focus of the research includes the analysis during specific time periods that may depict the event risk faced in a downturn as well as the use of specific groups of the data set, involving the best and worst performing stocks, and observing if there are additional effects onto our allocation strategy in being able to obtain the best outcome. Although sentiment data ideally should impact the movements in the financial market, there are limitations towards how effective and useful the data and the advancement of the tools used as well, as observed in our research outcomes. Therefore, this research aims to discuss the procedure involved and focusing the analysis on specific cases to determine the effectiveness of sentiment data and machine learning algorithms.