12-28 Statistics
8312STAT.DOC TI-83 international English Bob Fedorisko Revised: 02/19/01 12:42 PM Printed: 02/19/01 1:37
PM Page 28 of 38
Compute the regression model for the number of hours of
daylight in Alaska during one year.
&
&
With noisy data, you will achieve better convergence
results when you specify an accurate estimate for period.
You can obtain a period guess in either of two ways.
Plot the data and trace to determine the x-distance
between the beginning and end of one complete period,
or cycle. The illustration above and to the right
graphically depicts a complete period, or cycle.
Plot the data and trace to determine the x-distance
between the beginning and end of N complete periods,
or cycles. Then divide the total distance by N.
After your first attempt to use SinReg and the default value
for iterations to fit the data, you may find the fit to be
approximately correct, but not optimal. For an optimal fit,
execute SinReg 16,Xlistname,Ylistname,2p / b where b is
the value obtained from the previous SinReg execution.
SinReg Example:
Daylight Hours in
Alaska for One
Year
1 period