Modely založené na skóre pro analýzu teploty vzduchu

This thesis examines historical temperature trends using the Generalized Autoregressive Score (GAS) and Seasonal Autoregressive Integrated Moving Average (SARIMA) models, with particular emphasis on the former because of its ability to analyze the time-varying nature of characteristics. This work not only tracks average temperatures, but also looks at maximum and minimum values, revealing an upward trend across all metrics, most notably after the year 1950. While variability changes in monthly data were not consistently evident, specific subsets like January averages and winter minimums displayed decreases in variability patterns. The SARIMA model showed superior performance in monthly data analysis, and its forecasts for 2063 predict significant temperature rises, thus affirming the increasing impact of climate change. GAS was chosen for forecasting purposes based on annual data and predicted higher temperature increases for winter than for summer months. These insights improve understanding of temperature dynamics and may assist in further research on climate modeling.