Forecasting Youth Unemployment in Korea with Web Search Queries

Governments increase budget expenditure for youth job creation, but youth job markets tightened by prolonged recession are not improved as expected. To ease the problem of youth unemployment, developing relevant policies is important but more accurate and rapid prediction is also critical. This research develops a prediction model additionally utilizing web query information in classical statistical prediction model. Often ARIMA model is applied to estimate unemployment rate. For identified ARIMA model for Korean youth unemployment rate, we apply web query information to improve the accuracy of prediction. Our suggested model shows better performance than ARIMA model with respect to mean squared errors of estimate and prediction. We hope this research will be useful in developing a more improved model to estimate variable of interest.