Smooth Curve Fit

Asked Aug 15 '20 at 17:55

Active Aug 15 '20 at 17:55

Viewed 49 times

I uses supsmu to fit a smooth curve to a scatter plot of data .

I am quite new to curve fitting, but my graph looks somewhat sinusoidal. What would you recommend in terms of next steps to determine a predictive model for this patterns. I have tried this simple code:

xc<-cos(2*pi*trend$x/104)
xs<-sin(2*pi*trend$x/104)
fit.lm <- lm(trend$y~xc+xs)

Any other opinions on methods?

asked Aug 15 '20 at 17:55

ceo_stackoverflow

Use the pink line itself as a table of predictions? – Alexis Aug 15 '20 at 18:00
@Alexis I did. I am asking about any other regression methods? – ceo_stackoverflow Aug 15 '20 at 18:14
Are you familiar with Fourier transforms? If not, you might want to read up on them. since this is a time series, you also might want to try a time series model. – kurtosis Aug 15 '20 at 18:16
@kurtosis is it much different from what I have tried in the code I showed? – ceo_stackoverflow Aug 15 '20 at 21:52

Smooth Curve Fit

0 Answers0