Autor/Autorin: Yousefi Djamil
Djamil Yousefi, M.Sc. University of Konstanz, Major in Econometrics and Applied Economics
This work investigates in a forecast application whether the forecast accuracy of the US unemployment growth rate can be improved when the time series model is augmented with Google Trends data. The empirical analysis is based on up to 10 distinct Google Trends search terms, which show a sufficiently high correlation with the target variable US unemployment rate. A Google index from the 10 Google search terms is derived by estimating a Factor model with the method of principal component analysis. By using a VAR model it is empirically shown that the forecast accuracy of the US unemployment growth rate can be improved by augmenting the model with the Google search terms.
Über den Autor
Djamil Yousefi, B.Sc. Economics, Major in Quantitative Economics, University of Constance.
Klappentext
This work investigates in a forecast application whether the forecast accuracy of the US unemployment growth rate can be improved when the time series model is augmented with Google Trends data. The empirical analysis is based on up to 10 distinct Google Trends search terms, which show a sufficiently high correlation with the target variable US unemployment rate. A Google index from the 10 Google search terms is derived by estimating a Factor model with the method of principal component analysis. By using a VAR model it is empirically shown that the forecast accuracy of the US unemployment growth rate can be improved by augmenting the model with the Google search terms.
