Not as an extra step: PCA is itself an orthogonal rotation of the axes onto the directions of maximum variance. A further rotation such as varimax is optional and only aids interpretation; it keeps the total explained variance but loses the variance-maximizing order.
Card 1 of 259. Answer: Not as an extra step: PCA is itself an orthogonal rotation of the axes onto the directions of maximum variance. A further rotation such as varimax is optional and only aids interpretation; it keeps the total explained variance but loses the variance-maximizing order.
About · Quizzes · History of AI · To suggest new questions or report an error send me a dm.
Not as an extra step. PCA is itself an orthogonal rotation: after centering, it rotates the axes so the first points along the direction of greatest variance, the second along the greatest remaining variance, and so on. That is why R's prcomp() returns the eigenvector matrix under the name rotation.
A further rotation such as varimax is optional, borrowed from factor analysis. It turns the k retained components within the space they span so each variable loads strongly on as few components as possible, making them easier to name. The total explained variance is unchanged but spread out: components explaining 60% and 25% might become 45% and 40%, still 85%.
After varimax the components are no longer ordered by variance and the first no longer captures the maximum, so the property that defines PCA is lost. Rotation never reduces how many components you need, and the rotated scores span the same space, so a linear model fit on them gives identical predictions. Its only purpose is interpretation.
