Visualizing Principal Components Analysis for Multivariate Process Data

Abstract

In this article, we suggest a simple method for visualizing the results of principal components analysis (PCA) intended to complement existing graphical methods for multivariate time series data applicable for process analysis and control. The idea is to visualize multivariate data as a surface that in turn can be decomposed with PCA. The surface plots developed in this paper are intended for statistical process analysis but may also help visualize economic data and, in particular, cointegration.

• Cointegration
• Data Mining
• Econometrics
• Multivariate Analysis
• Statistical Graphics
• Statistical Process Control

Additional information

Notes on contributors

Søren Bisgaard

Dr. Bisgaard is a Professor at University of Massachusetts Amherst and a fellow of ASQ. His email address is [email protected].

Xuan Huang

Ms. Huang is a Doctoral Student at University of Massachusetts Amherst and a student member of ASQ. Her email address is [email protected].