Hurst exponent estimation, verification, portability and parallelization
Keywords:
Hurst exponent, rescaled range, parallel computing, time series, message passing interface,neutron monitorAbstract
We present multiple software programs for the Hurst exponent calculations for a sample time series collected by a neutron monitor detectors array. The first application is carried out by the finite differences approach, using a spreadsheet-type application for a single one hour long data series; the second is a complete, one and a half week long, mathematical and graphical analysis of six acquisition channels in Matlab; the third and the fourth are the data file parser and analyzer in C/C++ compiler on Windows platform, and its modified Linux version for simultaneous, parallel computing on a virtual cluster of three machines. All applications produce the same results proving the codes’ validity and portability across the operational systems and software packages.
References
2. Hurst J., Morandi O., Manfredi G., Hervieux P.-A. SemiclassicalVlasov and fluid models for an electron gas with spin effects // The European Physical Journal D.-2014.-Vol.68.-Issue 6.-P. 11.
3. Hurst H.E. Long term storage capacity of reservoirs // Trans. Am. Soc. Civ. Eng.-1951.-Vol. 116.-P. 770-779.
4. Sankar N.P. et al. Scaling and Fractal Dimension Analysis of DailyForbush Decrease Data // International Journal of Electronic Engineering Research.-2011.- Vol. 3.-Num. 2.-P. 237-246.
5. Flynn M, Pereira W. Ecological studies from biotic data by Hurst exponent and the R/S analysis adaptation to short time series // Biomatematica.-2013.-Vol.23.-P.1-14.
6. GorskiA.Z. et al. Financial multifractality and its subtleties: an example of DAX // Physica.-2002.-Vol. 316 .-P.496 –510.
7. The Shepetov’s database in IZMIRAN at http://cr29.izmiran.ru/vardbaccess/frames-vari.htmlaccessed on August 10, 2014.
8. Pesnell W. D. Solar Cycle Predictions (Invited Review) // Sol. Phys.-2012.-Vol.281.-P.507–532.
