At each target minimise sum K((x_i - x0) / lambda) (y_i - a - b (x_i -
x0))^2; the fit is a (ESL eqs 6.8-6.9).
Usage
morie_esl_local_linear(x0, x, y, lambda, kernel = "epanechnikov")
Arguments
- x0
Targets.
- x, y
Data.
- lambda
Bandwidth.
- kernel
"epanechnikov", "tri-cube" or "gaussian".
Value
Named list: estimate, values, slopes, n_in_window, lambda, kernel, n.
References
Hastie, Tibshirani & Friedman (2009), sec. 6.1.1.
Examples
x <- (1:30) / 30
morie_esl_local_linear(c(0.3, 0.6), x, sin(6 * x), 0.2)$values
#> [1] 0.8440317 -0.3835315
