Analýza a predikce časových řad v R

The goal of this thesis is to introduce basic methods of prediction of time series and to compare suitability and precision of these methods for prediction of one particular time series in computing environment of statistical programming language R. After theoretical explanation of time series themselves, the methods of modeling time series and predictions using these are introduced. The method of trend modeling using trend (regression) functions is presented, next, the models of exponential smoothing are presented and last but not least the models of Autoregressive Integrated Moving Averages (ARIMA) are introduced as well. All of these models are applied onto three time series of author’s choice to better illustrate their differences and suitability for particular type of time series.