I am running a principal component analysis in Orign2025, and it is only a 3D graph that gives an explanation to about 71% of the variance. Therefore, I chose to plot the variables in 3D; however, explaining it has become much of a problem. Hence, I would be glad to get a material that best explains it or any simple way to describe it.
3D scatterplot is indeed quite problematic to visualize the cluster as clearly we wanted to.
One way you can plot a PC in a really simple way is by plotting a standard deviation ellipse. However, you would plot 2D instead 3D, hence you would be using 2 PC component for each axis instead of 3. In this ellipse plot, you would plot mean at the center of the axis and multiple of SD for the radius (i.e the standard deviation ellipse). Both of this is doable with Minitab and Python.
However, I am uncertain (as I am unfamiliar with the software) whether it is doable in Origin2025. However, there's a feature called confidence ellipse plotting in the software documentation (https://www.originlab.com/fileExchange/details.aspx?fid=856). I am uncertain whether this confidence ellipse is similar to SD ellipse, but you could probably try to plot and see if it helps explaining the cluster pattern.
- Badges
- Science topic