Volatility Forecasting with Wavelet Neural Networks

APLIMAT - Journal of Applied Mathematics 2015
Andrejs Pučkovs, Andrejs Matvejevs

This paper describes volatility forecasting by using wavelet neural networks. Volatility forecasting is very important aspect of financial time series analysis which is widely used by financial institutions such as banks, pension funds, insurance companies and other stock market players. This paper describes short-term and mid-term volatility indicator forecasting such as implied volatility and expected volatility. Forecasting involves wavelet theory (which is a part of signal theory), mathematical statistics and neural networks theory. Research algorithm is related to so called ’North-East Volatility Wind’ Effect which is described in previous papers.

Atslēgas vārdi
Time-series analysis, mid-term forecasting, short-term forecasting, volatility, volatility evolution, historical volatility, implied volatility, Discrete Wavelet Transform, Wavelet decomposition, Daubechies mother wavelet function, Haar mother wavelet func- tion, Neural Networks, Wavelet Neural Networks, Nonlinear Autoregressive Neural Net- works with Exogenous Input (NARX), Mean-Squared Error (MSE), Pearson Correlation (R), ’North-East Volatility Wind’ Effect, Volatility Wave, Fourier analysis, Stock indexes, VIX, SP 500 index Daubechies mother wavelet function

Pučkovs, A., Matvejevs, A. Volatility Forecasting with Wavelet Neural Networks. APLIMAT - Journal of Applied Mathematics, 2015, No.7, 143.-150.lpp. ISSN 1337-6365.

Publikācijas valoda
English (en)