I have a dataset with 20 rows and 480 columns. When I run plsr command with validation="LOO" my output shows RMSEP or CV is increasing with number of components and stabilizing after 6 components. My impression is that RMSEP should decrease. Confused! Help!
You are right, generally RMSECV is supposed to decrease first, followed by stabilization or increment or very slow decrement. If you are observing a steady increment of RMSECV starting from first component, it is likely:
- Your dependent and independent variables are not linearly related to each other(enough).
- You have insufficient number of observations to reveal the relationship between dependent and independent variables.
- You really need only 1 component to model the data.