Principal Components for Stepwise Multiple Linear Regression

<p>To ensure that multivariate covariates are independent of each other, Gobin (2000) used principal components instead of the original environmental covariates as predictors to improve on the prediction for soil-landscape modelling. Therefore, all the original environment covariates for digital soil mapping were subjected to a standardized principal component analysis (PCA) to generate a smaller number of linear combinations that capture most of the variation within the raster stack as a whole (Crawley, 2012). RStudio version 3.5.1 was used to conduct a standardized PCA using the <em>RStoolbox</em> package (Leutner and Horning, 2017). These principal components were uses as the predictors for stepwise multiple linear regression.</p>

<p>Crawley, M.J. (2012). The R book. John Wiley & Sons, pp. 1051.</p>

<p>Gobin, A. (2000). Participatory and spatial-modelling methods for land resources analysis. Doctoral dissertation, Katholieke Universiteit Leuven, pp. 344.</p>

<p>Leutner, B., & Horning, N. (2017). RStoolbox: tools for remote sensing data analysis. R package version 0.1,7.</p>