Experimental Design

Question 1

Error metrics

Answer 1

The simplest is MAD, mean absolute deviation

given a gold standard vector G and your vector Y calculate the mean of G[i]- Y[i]

Question 2

one tailed hypo.

Answer 2

statistical significance means expected result was found

e.g. phenomenon exists if A is larger than B

Question 3

Two-tailed hypothesis:

Answer 3

only need to show that the different
results are statistically significant
 phenomenon exists if A and B differs