A Assignment unit 03-1

1. Do the exercise a randomly good model with the Marburg DOP and all classes of your digitized training areas. If the amount of data is too large to process or the modelling process takes painfully long, you can reduce the amount of pixels you use to train the model. Just be careful to include a somewhat balanced amount of data from each class.

2. Create a map of the spatial prediction with a legend of your classes.

3. Try to interpret the performance values of your model and your external validation in connection with the confusion matrix. Write 5 sentences about how you would interpret your results (map and performance values) (e.g. is there a problem? What could cause a proplem? Is the spatial prediction sufficient?).

Put your code and results in a PDF file and upload it to ILIAS before the start of the next session.

Optional: If you are already experienced in Rmarkdown you can put your results – both the classified images and your code – in a Rmarkdown file and convert it to a PDF document. Hint: If you need help with Rmarkdown, have a look at R Markdown Quick Tour

Comments?

You can leave comments under this gist if you have questions or comments about any of the code chunks that are not included as gist. Please copy the corresponding line into your comment to make it easier to answer the question.