On 30th of January, our colleague Timotheos Paraskevopoulos defended his dissertation “Time-Frequency Decompositions and the Impact of Surprise” and thus successfully completed his Ph.D. During the talk, he presented his Ph.D. thesis and competently defended his research against the faculty members.
His research focuses on the development and application of actionable interdisciplinary methods stemming from signal processing and natural language processing for prominent Finance topics. The first part of his thesis is devoted to the analysis and forecasting of financial markets based on time-frequency decompositions. The developed hybrid forecasting algorithm combines two powerful methods: Support- Vector-Machines and Wavelets. The insights gained from time-frequency decompositions also shed new light on the dynamics of European stock markets and the European exchange rate.
In the second part of the dissertation, Timotheos contributes to the research community with a paradigm for quantifying information flow in unstructured data from an agnostic point of view. The proposed metric allows quantifying the aggregated surprise level on Twitter to explain the potential volatility of U.S. companies stock returns. The research journey of the proposed surprise index takes Timotheos and his Ph.D. supervisor Prof. Dr. Peter N. Posch into state-of-the-art natural language processing models such as word2vec and a refined Twitter-LDA for topic modeling before returning to the classical Finance world again.
The entire team of FiRRM congratulates Timotheos on this achievement!