Model-Free Time-Aggregated Predictions for Econometric Datasets

Forecasting volatility from econometric datasets is a crucial task in finance. To acquire meaningful predictions, various methods were built upon GARCH-type models, but these classical techniques suffer instability of short and volatile data. Recently, novel existing normalizing variance-stabilizing (NoVaS) method for predicting squared log-returns financial data was proposed. This model-free has been shown to possess more accurate stable prediction performance than methods. However, whether this can sustain high long-term still doubt. In article, we firstly explore the robustness NoVaS time-aggregated predictions. Then, develop parsimonious variant method. With systematic justification extensive analysis, our new shows better current standard GARCH(1,1) on both short- The success remarkable since efficient predictions with always carry great importance. Additionally, article opens potential avenues where one design structure meet specific needs.