Smoothing-based Initialization for Learning-to-Forecast Algorithms

Abstract

Under adaptive learning, recursive algorithms are proposed to represent how agents up-date their beliefs over time. For applied purposes these algorithms require initial estimatesof agents perceived law of motion. Obtaining appropriate initial estimates can become pro-hibitive within the usual data availability restrictions of macroeconomics. To circumventthis issue we propose a new smoothing-based initialization routine that optimizes the useof a training sample of data to obtain initials consistent with the statistical properties of thelearning algorithm. Our method is generically formulated to cover different specificationsof the learning mechanism, such as the Least Squares and the Stochastic Gradient algo-rithms. Using simulations we show that our method is able to speed up the convergence ofinitial estimates in exchange for a higher computational cost.