The primary objective of this paper is to propose two nonlinear extensions for macroeconomic forecasting using large datasets. First, we propose an alternative technique for factor estimation, i.e., kernel principal component analysis, which allows the factors to have a nonlinear relationship to...