GMDH-based Models for Mid-term Forecast of Cryptocurrencies (on example of Waves)

Authors

Pavel Mogilev, Technological University DublinFollow
Anna Boldyreva, Moscow Institute of Physics and TechnologyFollow
Mikhail Alexandrov, Autonomous University of BarcelonaFollow
John Cardiff, Technological University DublinFollow

Document Type

Conference Paper

Rights

Available under a Creative Commons Attribution Non-Commercial Share Alike 4.0 International Licence

Disciplines

Computer Sciences, Information Science

Publication Details

2020 IEEE 15th International Conference on Computer Sciences and Information Technologies (CSIT)

Abstract

Cryptocurrencies became one of the main trends in modern economy. However by the moment the forecast of cryptocurrencies values is an open problem, which is almost non-reflected in publications related to finance market. Reasons consist in its novelty, large volatility and its strong dependence on subjective factors. In this experimental research we show possibilities of GMDH-technology to give weekly and monthly forecast for values of cryptocurrency 'Waves' (waves/euro rate). The source information is week data covering the period 2017-2019. We tests 4 algorithms from the GMDH Shell platform on the whole period and on the crisis period 4-th quarter 2017 - 2nd quarter 2018. Baseline is provided by the popular statistical method of double exponential smoothing. The results of Pilot study can be considered as the very promising ones having in view the large variability of data.

DOI

https://doi.org/10.1109/CSIT49958.2020.9321873

Recommended Citation

P. Mogilev, A. Boldyreva, M. Alexandrov and J. Cardiff, "GMDH-based Models for Mid-term Forecast of Cryptocurrencies (on example of Waves)," 2020 IEEE 15th International Conference on Computer Sciences and Information Technologies (CSIT), 2020, pp. 13-16, doi: 10.1109/CSIT49958.2020.9321873.