Using the Canadian census wage data from Problem 5, smooth the data with Friedman's method using the cross-validation span and bass=0. Vary the bass parameter to obtain a reasonable fit. Comment on the graphical comparison of the estimates with varying bass values. Which value of the bass seems most reasonable?
Problem 5
Pagan and Ullah (1999) reported data from the 1971 Canadian census on wages for male high school graduates. This data in the R object cps71. Use the logarithm of wages log wage as the response and age as the predictor. Smooth this data using Friedman's method with the span found by trial and error. Comment on the graphical comparison between this estimate to that obtained when using the span selected by cross-validation.