Dear Elise Kho ,

In the Predicted vs Reference plot you are able to obtain both the RMSE and R-Square values at least in the Unscrambler 10.3.

Also, you should check The Unscrambler user manual, it shows great information,

All the best

Dear Edenio,

Thank you for your suggestion. But that's the RMSE and R-square for calibration and cross-validation data sets, not the predictions. I'm actually looking for RMSEP for predictions. Any ideas? Couldn't find it on user manual...

Exactly, after that you need to predict the model using Task/ Predict/ SVR Predict...you will obtain the predicted values.

Copy them and paste in a excel sheet and with the reference values you could calculate r-square, rmse, etc. This is an issue from the unscrambler, I use for it for PLS, PCA and MLR.

From my point of view SVM in python is better. Here you have two examples:

1- https://stackabuse.com/implementing-svm-and-kernel-svm-with-pythons-scikit-learn/

2- https://pythonprogramming.net/support-vector-machine-intro-machine-learning-tutorial/

All the best

Thank you so much for your advice! Finally managed to find what I need with cut and paste into Excel!

So yea, will definitely look into Python for this!

All the best, I am moving also to Python, it looks great and powerful and everyday more tools into chemometrics, machine and deep learning analyses,

All the best