The Effect of Data Centering on PCA Models

A common misconception in PCA is that PC 1 is the mean of the data when the data are not centered. It was shown in this paper that PC 1 is not the mean of the data but it can point in the direction of the mean. The extent to which PC 1 points in the direction of the mean depends on how far away the data set mean is from the origin. More details can be found in the paper [1].

References

[1] The Effect of Data Centering on PCA Models