The Use of Principal Component Factor Analysis to Interpret Particulate Compositional Data Sets

Abstract

The study of urban aerosols often yields large data bases consisting of elemental and meteorological data for a multitude of samples collected at various locations and times. Several statistical techniques are currently being used to identify the nature of aerosol sources using these data. However, errors present in the data base may make the interpretation difficult. In addition to its value as a receptor model, principal component factor analysis has the ability to identify the errors present in the data so that appropriate corrections may be made before detailed data interpretation is attempted. Thus the technique can save much time and effort. Principal component factor analysis can yield interpretations of aerosol data and a better understanding of the airshed being studied.