|
Methods for electroencephalogram records comparison
Kliment, Juraj ; Ronzhina, Marina (oponent) ; Janoušek, Oto (vedoucí práce)
The bachelor thesis deals with the comparison of EEG signals from the human brain. The aim of this work is to find suitable nonlinear parameters, based on which it is possible to compare EEG records created in different conditions. The selected parameters are then tested on data from a publicly available database using Matlab software, the results are statistically processed and compared with the results of existing scientific studies. The comparison was based on the parameters Approximate entropy, Correlation dimension, Hurst exponent and Lyapunov exponent.
|
|
Methods for electroencephalogram records comparison
Kliment, Juraj ; Ronzhina, Marina (oponent) ; Janoušek, Oto (vedoucí práce)
The bachelor thesis deals with the comparison of EEG signals from the human brain. The aim of this work is to find suitable nonlinear parameters, based on which it is possible to compare EEG records created in different conditions. The selected parameters are then tested on data from a publicly available database using Matlab software, the results are statistically processed and compared with the results of existing scientific studies. The comparison was based on the parameters Approximate entropy, Correlation dimension, Hurst exponent and Lyapunov exponent.
|
|
Methods for electroencephalogram records comparison
Kliment, Juraj ; Ronzhina, Marina (oponent) ; Janoušek, Oto (vedoucí práce)
The bachelor thesis deals with the comparison of EEG signals from the human brain. The aim of this work is to find suitable nonlinear parameters, based on which it is possible to compare EEG records created in different conditions. The selected parameters are then tested on data from a publicly available database using Matlab software, the results are statistically processed and compared with the results of existing scientific studies. The comparison was based on the parameters Approximate entropy, Correlation dimension, Hurst exponent and Lyapunov exponent.
|
|
Methods for electroencephalogram records comparison
Kliment, Juraj ; Ronzhina, Marina (oponent) ; Janoušek, Oto (vedoucí práce)
The bachelor thesis deals with the comparison of EEG signals from the human brain. The aim of this work is to find suitable nonlinear parameters, based on which it is possible to compare EEG records created in different conditions. The selected parameters are then tested on data from a publicly available database using Matlab software, the results are statistically processed and compared with the results of existing scientific studies. The comparison was based on the parameters Approximate entropy, Correlation dimension, Hurst exponent and Lyapunov exponent.
|
|
Entropy as a Measure of Predictability in Financial Time Series
Nahodil, Vladimír ; Krištoufek, Ladislav (vedoucí práce) ; Wang, Yao (oponent)
Tato práce studuje efektivnost akciových trhů a jejich prediktabilitu pomocí konceptů z teorie informace, approximate entropy (ApEn) a sample entropy (SampEn), a porovnává jejich vlastnosti s odhadem Hurstova exponentu. U těchto měřítek je také porovnávána jejich schopnost rozlišovat rozvíjející se a rozvinuté trhy. Na závěr je testována investiční strategie postavená na hodnotě sample entropy. ApEn ukazuje velmi slabý vztah jak se SampEn, tak s Hurstovým exponentem a zároveň slabý výkon jako měřítko efektivnosti trhů. Sample entropy a Hurstův exponent jasně rozlišují nížší celkovou efektivnost rozvojových trhů. SampEn také utváří poměrně silný klesající vztah s hit-rates predikčních modelů. ARMA má nejvyšší hit-rate v obdobích, kdy je SampEn v rozmezí 1.6 - 1.7. Toto může být potenciálně využito v investičních strategiích za účelem nižšího risku; nicméně, s tím souvisí i možné nižší celkové zisky z důvodu menších investičních oken.
|
| |