This preview shows page 1. Sign up to view the full content.
Unformatted text preview: e dependent
variable gives Which models are suggested by the shape of the data? Thinking back Section 2.5, we try a
Quadratic Regression, with pretty good results.
Since there are only 94 distinct ASCII keyboard characters, to achieve this strength, the number of characters in
the password should be increased.
Derived from the Henderson-Hasselbalch Equation. See Exercise 8 in Section 6.2. Hasselbalch himself was
studying carbon dioxide dissolving in blood - a process called metabolic acidosis. 388 Exponential and Logarithmic Functions However, is there any scientiﬁc reason for the data to be quadratic? Are there other models which
ﬁt the data equally well, or better? Scientists often use logarithms in an attempt to ‘linearize’ data
sets - in other words, transform the data sets to produce ones which result in straight lines. To see
how this could work, suppose we guessed the relationship between N and t was some kind of power
function, not necessarily quadratic, say N = BtA . To try to determine the A and B , we can take
the natural log of both sides and get ln(N ) = ln B tA . Using properties of logs to expand the right
hand side of this equation, we get l...
View Full Document