Abstract:This paper proposes a novel dynamic forecasting method using a new supervised Principal Component Analysis (PCA) when a large number of predictors are available. The new supervised PCA provides an effective way to bridge the gap between predictors and the target variable of interest by scaling and combining the predictors and their lagged values, resulting in an effective dynamic forecasting. Unlike the traditional diffusion-index approach, which does not learn the relationships between the predictors and the target variable before conducting PCA, we first re-scale each predictor according to their significance in forecasting the targeted variable in a dynamic fashion, and a PCA is then applied to a re-scaled and additive panel, which establishes a connection between the predictability of the PCA factors and the target variable. Furthermore, we also propose to use penalized methods such as the LASSO approach to select the significant factors that have superior predictive power over the others. Theoretically, we show that our estimators are consistent and outperform the traditional methods in prediction under some mild conditions. We conduct extensive simulations to verify that the proposed method produces satisfactory forecasting results and outperforms most of the existing methods using the traditional PCA. A real example of predicting U.S. macroeconomic variables using a large number of predictors showcases that our method fares better than most of the existing ones in applications. The proposed method thus provides a comprehensive and effective approach for dynamic forecasting in high-dimensional data analysis.

Improvement of simultaneous prediction using principal components approach

A Two-Step PLS Inspired Method for Linear Prediction with Group Effect

Targeted principal components regression

Enhanced Model Predictions through Principal Components and Average Least Squares-Centered Penalized Regression

Principal Simple Linear Regression

Prediction of multivariate responses with a select number of principal components

Enhancing model predictions through the fusion of stein estimator and principal component regression

Supervised Dynamic PCA: Linear Dynamic Forecasting with Many Predictors

Prediction of Multivariate Time Series Based on Principal Component Analysis and Neural Networks

Comparison of prediction methods for multicollinear data

Two Classes of Almost Unbiased Type Principal Component Estimators in Linear Regression Model

Principal Components Analysis in Clinical Studies

Principal component-based hybrid model for time series forecasting

Stock Price Prediction using Principle Components

A note on the variance in principal component regression

Principal Components Regression Estimation in Semiparametric Partially Linear Additive Models

Real-estate price prediction with deep neural network and principal component analysis

Nonparametric Principal Subspace Regression

Prediction of multivariate time series based on reservoir principal component analysis

Supervised Principal Component Regression for Functional Responses with High Dimensional Predictors

Forecasting Using Principal Components From a Large Number of Predictors