3.6.1. Confusion Matrix

Try these first

1. 1.

   What is a confusion matrix?

   Hint

   Think of it as a report card for classifications.

2. 2.

   Define Overall Accuracy.

   Hint

   Consider the ratio of correct predictions to total predictions.

3. 3.

   What does a confusion matrix help assess?

   • The clarity of satellite imagery
   • The accuracy of classifications
   • The type of satellite used

   Hint

   Think about what kind of assessments we want to make about our classifications.

4. 4.

   True or False: User's Accuracy measures the probability of reference pixels being classified correctly.

   • True
   • False

   Hint

   Remember whose perspective it matters from.

5. 5.

   Given a confusion matrix for a land cover classification with the following data: True Positive (TP) = 30, False Positive (FP) = 10, True Negative (TN) = 40, False Negative (FN) = 5. Compute Overall Accuracy, User's Accuracy, Producer's Accuracy, and Kappa Coefficient.

   Hint

   Make sure to follow each step carefully.

6. 6.

   Discuss the implications of a low Kappa Coefficient value in terms of model reliability and classification performance.

   Hint

   Reflect on the importance of trust in the results attained from satellite imagery.