Chapter 15. Uncertainty, Error, and Sensitivity

1. What is a confusion matrix?
2. What is the principle diagonal of a confusion matrix?
3. What is producer's accuracy?
4. What is user's accuracy?
5. What is PCC and how is it calculated from a confusion matrix?
6. What does the kappa index tell you that PCC doesn't?
7. When sending observers to the field to collect the column data for a confusion matrix a stratified sampling procedure based on class area is sometimes used.  Why?
8. Errors in land cover maps can occur in the locations of boundaries of areas, as well as in the classification of areas.  Explain the difference.
9. What is accuracy?
10. What are two different definitions of precision?
11. In the interval/ratio case, the magnitude of errors is described by the root mean square error (RMSE).  What is RMSE?
12. The Gaussian distribution predicts the relative abundances of different magnitudes of error.  Explain.
13. What are the FGDC's five components of the quality of a spatial dataset?
14. What is a bivariate Gaussian error distribution?
15. What are National Map Accuracy Standards?  Give an example.
16. How is scale related to accuracy?
17. When measuring the distance between two points, why does strong positive correlation between errors lead to no error in the calculation?
18. Slope maps that are calculated from DEMs can be highly influenced by error of the input databases.  The spatial autocorrelation between errors in geographic databases helps to minimize their impacts on many GIS operations.  Why does spatial autocorrelation of error have a good impact?
19. What is error propagation?
20. What is fuzzy membership?
21. Why are soil class definitions considered fuzzy?
22. Our authors suggest several ways of dealing with uncertainty.  Explain them.
23. What is sensitivity analysis?