There are times when this is the appropriate choice, but it’s rare. If all of your variables are in the same units, then you may want to perform PCA on centered data, also called PCA on the covariance matrix. Where X̄ is the mean and sx is the standard deviation of the variable values. This puts all the variables on the same scale, so that when the PCs are found, each variable is weighted equally. How it works: Before performing PCA, variables are transformed so that each variable has a mean of 0 and standard deviation of 1. You would almost always choose this approach if your variables were measured using different units. It’s sometimes referred to as performing PCA on the correlation matrix. Unless you have a specific reason to do otherwise, this is the recommended approach. The most important decision is whether to perform PCA on standardized or centered data. Unless you understand why you need to do otherwise, we recommend performing PCA on standardized data and using Parallel Analysis to select the number of components.
In the options tab of the dialogue, you need to make two major decisions that can heavily influence the results and conclusions of PCA.
0 kommentar(er)
