Mining Big Data Using Parsimonious Factor and Shrinkage Methods

Abstract

A number of recent studies have focused on the usefulness of factor models in the context of prediction using "big data" (see e.g., Bai and Ng (2008), Dufour and Stevanovic (2010), Forni et al. (2000, Kim and Swanson (2014), Stock and Watson (2002b, and the references cited therein). We add to this literature by analyzing the predictive bene…ts associated with the use of independent component analysis (ICA) and sparse principal component analysis (SPCA), coupled with a variety of other factor estimation and data shrinkage methods, including bagging, boosting, and the elastic net, among others. We carry out a forecasting "horse-race", involving the estimation of 28 di¤erent baseline model types, each constructed using a variety of speci…cation approaches, estimation approaches, and benchmark econometric models; and all used in the prediction of 11 key macroeconomic variables relevant for monetary policy assessment. In numerous instances we …nd that a variety of benchmark autoregressive models and model averaging methods are mean square forecast error (MSFE) dominated by more complicated nonlinear methods. For example, simple averaging methods are MSFE "best" in only 9 of 33 key cases considered. However, in order to "beat" model averaging methods, we must combine new factor estimation methods with interesting new forms of shrinkage. For example, SPCA yields MSFE-best prediction models in many cases, particularly when coupled with shrinkage. In summary, we present empirical results that provide strong new evidence of the usefulness of sophisticated factor based forecasting methods, and as a corollary, of the use of "big data" in macroeconometric forecasting.

Methods for backcasting, nowcasting and forecasting using factor‐MIDAS: With an application to Korean GDP

“…A detailed explanation of the above two methods is given in a supplemental Appendix available as Supporting Information. Recursive and standard principal component analysis (RPCA and OPCA). PCA is widely used to estimate factors or diffusion indices in large data environments (see Kim & Swanson, (), and references cited therein). In this paper, we utilize PCA, called OPCA in our later discussion of our empirical findings to differentiate it from recursive PCA (discussed below).…”

“…In the context of high‐frequency data, are measures of risk such as so‐called realized volatility useful as predictors? Finally, are alternative “sparse” diffusion index methodologies, such as sparse principal components analysis and independent component analysis, useful in real‐time prediction (see, e.g., Kim & Swanson, )?…”

In-Sample Inference and Forecasting in Misspecified Factor Models

“…where y t is a scalar, x t is a (N 1) vector of predictors, and is (N 1) vector of unknown parameters. In matrix notation, let y be a (T 1) vector, X be a (T N ) matrix and " be a (T 1) component analysis and sparse PCA (Kim and Swanson, 2014b), combining forecast PC (Huang and Lee, 2010), principal covariate regression (e.g. Tu and Lee, 2013).…”

Forecasting the Estonian rate of inflation using factor models

“…Banerjee et al (2014) present forecasts using a factor-augmented error correction model. Comparisons and reviews of various factor forecasting models can be found in Eickmeier & Ziegler (2008) and Kim & Swanson (2013).…”