Classification problems

I am using sen2cor to correct a Sentinel 2 image. Then i am using 4 classes with 5 polygons per class as a training dataset. I am interested in a coastal region. Then utilizing supervised classification with KNN classifier.

However the output is an image without any value (attached a screenshot)

Noticing a strange mark next to the band (aAttaching screenshot).

Any ideas?

If you use the Random Forest Classification then the following issue could be the reason:

Do you mean the image by strange mark?
This is just indicating an “index band”. Indexed bands usually are the result of a classification. Each integer value is the index into the map of meanings. If you open the Colour Manipulation you will see this map.

Thank you for the answer. When you mention 'If the values are scaled by some factor (e.g. 10000) then the RF classification works ` how you can scale the values e g multiple all the bands of the roaster by 10000?

You can use the band maths for this.
Create new bands for those you use in the classification, multiply them by a factor of 10000, or 1000 is probably already enough.