fls: Fitting Time-Varying Linear Models via FLS

In GediminasB/fls: Time-Varying Linear Regression via Flexible Least Squares

Description

fls is used to fit Time-Varying Linear Regression via Flexible least squares (FLS) as discribed in Kalaba and Tesfatsion (1989).

Usage

fls(X, y, mu = 1, smooth = TRUE)

Arguments

X	design matrix of dimensuin n * K.
y	vector of observations of length n.
mu	parameter controling relative weight of sum of dynamic errors (r_D^2) vs sums of squared residual measurement errors (r_M^2).
smooth	logical. If TRUE, a smoothed coefficients are provided.

Value

Returns object of class "fls". An object of class "fls" is a list containing the following components:

coefficients

A n * K matrix coefficient estimates.

fitted.values

the fitted mean values.

r_D

sum of dynamic errors.

r_M

sum of squared residual measurement errors.

References

\insertRef

KALABA19891215fls

GediminasB/fls documentation built on Nov. 13, 2019, 12:43 a.m.