This is a test version of Biostars. For the public version, visit https://www.biostars.org.
PCA Plot componenets

PCA PLOT with 2 components is better over 3 components or PLOT with 3 components is better over 2 and why ?

r
• 2,239 views • link
updated by Mahdiar Sadeghi • Mahdiar Sadeghi • written by #### ####

2 components are easier to plot; 3 components might tell you more about the structure of your data. This is a bit of a subjective question for Biostars though

• 0 views • link
written by russhh russhh

This totally depends on your data, i.e., how much percentage of variation is represented in each PC. Usually you see that one factor is explained in one direction (PC), and another factor in the other direction. For example, if you compare diseased versus healthy persons, this difference might be shown in PC1 (healthy to the left, diseased to the right), and gender might be explained in the PC2 direction (women up, and men down). If there is a clear third factor which is present in PC3, 3D plot might be useful. Depends on how complex your dataset is.

• 0 views • link
written by Benn Benn

1 answer

Whatever that presents the data better. 3D plots might be confusing depending on your data. In general, 2 component plots are good for presentation and easy to digest. You can add the 3rd component by color scale.

• 0 views • link
written by Mahdiar Sadeghi • Mahdiar Sadeghi

Log in to answer this question.

More posts like this