![IconGems-Print](images/IconGems-Print.png)
427 – Contributed Poster Presentations:Government Statistics Section
Time Series Reconciliation Through Flexible Least Squares Estimation
Luis Frank
University of Buenos Aires
The research improves the current flexible least squares procedure for time series reconciliation. This kind of estimation was already introduced by the author but assuming a prior knowledge of a key weighting parameter. The improved procedure allows the estimation of the weighting parameter from the sample data by means of the Newton-Raphson iterative method. Encouraging results arise from the reconciliation of the monthly growth rates of Argentina’s Monthly Economic Activity Estimator (EMAE) and the quarterly growth rates of the Gross Domestic Product.