Regensburg 2010 – scientific programme

Parts | Days | Selection | Search | Downloads | Help

SOE: Fachverband Physik sozio-ökonomischer Systeme

SOE 21: Financial Markets and Risk Management II

SOE 21.5: Talk

Thursday, March 25, 2010, 12:30–13:00, H46

Estimation of the volatility of finance data by multiscale reconstruction — •Arnold Gräbeldinger and Joachim Peinke — Workgroup turbulence, wind energy and stochastic (TWiSt), University of Oldenburg, Department of physics, 26111 Oldenburg, Germany

With the recently developed method for multiscale reconstruction according to [1] it is possible to reproduce correct statistic properties of processes on different scales and to create a decent prediction of the current standard deviation. Based on the multiscale reconstruction, predictions for quasi-equidistant finance data are performed. Comparing to standard methods it shall be tried to obtain an improved measure for the volatility and in particular volatility clusters of finance data. In depth the requirements for this ansatz are reviewed, with a focus on the Markov properties of the underlying data.

[1] A. P. Nawroth. Stochastische Analyse und Modellierung von Finanz- und Turbulenzzeitreihen und ihren Ähnlichkeiten. Dissertation, Oldenburg, 2007.