Russian oil and gas market index predictions based on hybrid neural network and autoregressive techniques

P. 14–15.

Kuletskaya L.

In this paper, we constructed a series of hybrid and combined techniques based on ANNs and ARIMA time series models to provide accurate predictions of Russian major stock market index MICEX daily values in the period from September 2007 to January 2015, and show that both hybrid and combination techniques outperform the individual models in terms of mean absolute percentage prediction error. However, this also suggests that the question of which of the hybrid or combination approach is the best is subject to further research.

Keywords: Russian Oil and Gas Industry

In book

Актуальные проблемы экономики и менеджмента. Избранные доклады I студенческой научно-практической конференции Санкт-Петербургской школы экономики и менеджмента НИУ ВШЭ

СПб.: Отдел оперативной полиграфии НИУ ВШЭ, 2016.