New model detects shifts in time series data for improved forecasting.
The article introduces a new model that can analyze time series data with long-term patterns, volatility, and sudden shifts in levels. This model helps identify changes in both the average and volatility of the data over time. By using specific statistical tests, researchers can determine if these shifts are present in the data. The model is called LS-ARFIMA for the time series analysis and LS-GARCH for the volatility analysis. The researchers also discuss how to estimate this model accurately.