Using sine and cosine terms as predictors in modeling periodic time series and other kinds of periodic responses is a long-established technique, but it is often overlooked in many courses or textbooks. Such trigonometric regression is straightforward in Stata through applications of existing commands. I give various examples using classic periodic datasets on the motion of the asteroid Pallas and the daily rhythm of birth numbers. I make a brief connection to polynomial-trigonometric regression.
BlissC. I.1970. Statistics in Biology: Volume II.New York: McGraw–Hill.
2.
BloomfieldP.2000. Fourier Analysis of Time Series: An Introduction.New York: Wiley.
3.
BracewellR. N.2000. The Fourier Transform and Its Applications.New York: McGraw–Hill.
4.
CoxN. J.2002. Speaking Stata: How to face lists with fortitude. Stata Journal2: 202–222.
5.
CoxN. J.2004a. Speaking Stata: Graphing model diagnostics. Stata Journal4: 449–475.
6.
CoxN. J.2004b. Stata tip 15: Function graphs on the fly. Stata Journal4: 488–489.
7.
CoxN. J.2006a. Speaking Stata: Time of day. Stata Journal6: 124–137.
8.
CoxN. J.2006b. Speaking Stata: Graphs for all seasons. Stata Journal6: 397–419.
9.
DunningtonG. W.1955. Gauss: Titan of Science.New York: Hafner.
10.
EubankR. L., and SpeckmanP.1990. Curve fitting by polynomial-trigonometric regression. Biometrika77: 1–9.
11.
EvansI. S., and CoxN. J.2005. Global variations of local asymmetry in glacier altitude: Separation of north-south and east-west components. Journal of Glaciology51: 469–482.
12.
GullbergJ.1997. Mathematics: From the Birth of Numbers.New York: W. W. Norton.
