Which of the following statements is/are true concerning linear regression?
I. One can take the $\log$ of the $x$ and $y$-values to transform an exponential relationship into a linear relationship.
II. If the original data is of exponential form, one can increase the correlation coefficient, $r$, by transforming the data to linear form.
III. As long as there is a strong correlation ($|r| > 0.7$), the data is most likely linear.