Modelling Slovak unemployment data by a nonlinear long memory model

Conference Paper (international conference)

Komorník J., Komorníková Magda

serial: ALGORITMY 2005. Proceedings of the 17th Conference on Scientific Computing, p. 334-340 , Eds: Handlovičová A., Krivá Z., Mikula K.

action: ALGORITMY 2005, (Podbanské, SK, 13.03.2005-18.03.2005)

research: CEZ:AV0Z10750506

project(s): GA402/04/1026, GA ČR

keywords: time series modelling, smooth regime switching, spectral analysis, fractional integration, ARFIMA modelling

abstract (eng):

The main visual feature of Slovak unemployment data is a dramatic increase after the elections in October 1998 which can be expressed by a step in the deterministic level function. In the residuals from this level function we can identify a significant cyclical component and a long memory structure (with the estimated value of the Hurst parameter close to 0.8) as well as an autoregressive short memory behaviour.

abstract (cze):

Dramatický růst po volbách v roce 1998 je nápadným rysem údajů o nezaměstnanosti na Slovensku. V reziduu deterministické úrovňové funkce, která tento růst modeluje, je možné zjistit významnou cyklickou složku a dlouhodobou paměťovou strukturu, jakož i krátkodobou regresi.

Cosati: 12A

RIV: BA