Brief report of the visit of Jan Choloniewski at the Netherlands Institute for Advanced Studies in the Humanities and Social Sciences (NIAS) (March 17-30, 2014) based on a STSM grant.
I consider my STSM at NIAS a considerable success. Not a spectacular one but of a vital importance for a young researcher like me. I got plenty of time and opportunities to make very first steps in the world of science.
At the beginning I would like to deeply appreciate Peter Sloot‘s and the fellowship of NIAS’ hospitality and to thank everyone who made my mission possible (especially Dr. Andrea Scharnhorst – my STSM coordinator).
The three main aspects of my STSM were development of new data mining methods, consultation with experts in data mining and finding new direction of studies. The data mining method I’m helping to develop is called Intrinsic Time Framework – a novel event-based approach to a time series multiscale analisis developend in frame of SOPHOCLES project. The method is especially suited for financial data analysis but can be applied for any real numbers series acquired from complex systems. I analyzed theoretical stationary distributions of possible system states for various values of parameters and, later, compared them with distributions calculated for a few realizations of stochastic processess.
Last two points can be viewed as one because meetings with both George Kampis and Alexander V. Boukhanovsky not only provided me with new perspectives of perceiving data mining techniques (also these I used in my theses) as well as with interesting datasets (Hungarian scientific papers database; Russian social media data of drug-promoting posts and users) on which one can experiment with data mining techniques that I learned from the experts.
For me, the mission was carried out just in time and opened new directions of development and studies. I had chance to discuss with and learn from experts in my field (and, hopefully, to start long-term collaborations) and to make a significant progress in testing and visualising Intrinsic Time Framework limits, scopes and performance. I think an experience gained in the mission would help me in my further career.
Warsaw, Jan Choloniewski
References:
[1] Anton Golub, Gregor Chliamovitch, Alexandre Dupuis, and Bastien Chopard (2014) Multiscale representation of high-frequency market liquidity. http://arxiv.org/abs/1402.2198
