| Abstract [eng] |
The paper dealt with the unemployment rate modeling opportunities, analyzing the theoretical and practical assessment. Studied not only in Lithuania, but also in Ireland, Denmark, Great Britain, Estonia, Iceland, Latvia and the U.S. unemployment rate changes in 2003 – 2013 years (132 months), including the 2008 global economic crisis. In order to obtain reliable results, the work carried out comparison of modeling techniques. Object of the research – the time series with long memory. This work examines the theoretical and empirical long memory ARFIMA models and problems. The relationship between economic theory and a long memory is not yet studied, in this thesis used several methods to determine the presence of long memory. The U.S. unemployment rate analysis (for the period from July 1968 to October 1999) using the nonlinear long memory model (Van Dijk, 2000), it was observed that the unemployment rate is growing faster than the decline in the recession year of economic growth . According to the U.S. above mentioned analysis was performed Slovak unemployment rate modeling nonlinear long memory model (Komorník, 2005). Based on these two analyzes and modeling of time series with modern principles (Philip Hans-France, 1998) modeled the unemployment rate for the eight countries mentioned above. It was found that the Irish, British, U.S. and Lithuanian unemployment rate time series has a long memory. Stated main goal - to identify changes in the unemployment rate really value the dynamics of the model, which can be assessed using the current unemployment rate in the state. This is very important because if the unemployment rate rise nationwide, so the economy shrinks. Part of the master's work was published in the conference \"Mathematics and Mathematics Teaching – 2014\". |
