03367ntm a22005417i 4500 000727621 CZ-PrVSE 20260201122909.0 m d cr n|||||||||| 260201s2025 xr fsbm 000 0 eng d NEZPRACOVANÝ IMPORT ABA006 cze ABA006 ABA006 rda Čelková, Barbora ISIS:147477 dis Predikční schopnost dat z online vyhledávání při předpovědi počtu případů COVID-19 eng Predictive Power of Online Search Data for COVID-19 Forecasting / Barbora Čelková 2025 ?? stran : digital, PDF soubor Vedoucí práce: Vladimír Holý Diplomová práce (Ing.)—Vysoká škola ekonomická v Praze. Fakulta informatiky a statistiky, 2026 Obsahuje bibliografii Textový (vysokoškolská kvalifikační práce) Rok obhajoby 2026 This thesis investigates whether online search data can enhance short-term forecasts of COVID-19 cases in the Czech Republic, utilizing weekly counts of new cases from March 2020 to April 2023 and search activity from Google Trends. The analyzed keywords comprise a set of general, symptom, and prevention-related queries, as well as their principal components. We first estimate an ARIMA model based solely on reported case numbers, which serves as a benchmark for all subsequent Google-augmented specifications. Afterwards, a set of models incorporating Google Trends is estimated, namely a regression with ARMA errors, a vector autoregression, and a vector error correction models. Using expanding-window forecast, we compare their predictive performance to the benchmark ARIMA model with standard forecast accuracy metrics in combination with Diebold–Mariano tests. The results show that while some Google Trends series can provide statistically significant gains in forecast accuracy, these improvements are heavily dependent on the chosen model and analyzed time period. Overall, the findings suggest that incorporating Google Trends into traditional statistical models could offer some predictive value for COVID-19 surveillance in the Czech Republic, but it does not deliver uniformly superior forecasts. Způsob přístupu: Internet ekonometrie a operační výzkum [obor dipl. práce] diplomové práce fd132022 czenas master's theses eczenas google trends COVID-19 ARIMA VECM forecasting time series VAR Holý, Vladimír ISIS:116093 ths Jindrová, Karolína ISIS:149274 opn Vysoká škola ekonomická v Praze. Fakulta informatiky a statistiky kn20010709399 dgg https://insis.vse.cz/zp/88290/podrobnosti VŠKP v InSIS https://insis.vse.cz/zp/88290 Hlavní práce https://insis.vse.cz/zp/88290/posudek/vedouci Hodnocení vedoucího https://insis.vse.cz/zp/88290/posudek/oponent/89077 Oponentura https://insis.vse.cz/zp/88290/podrobnosti dc:identifier NEPOSILAT VSKP vse88290 260129 88290