Back end for FixedEffectModels.jl
Author FixedEffects
16 Stars
Updated Last
1 Year Ago
Started In
December 2018

Build status

This package solves least squares problem with high dimensional fixed effects. For a matrix D of high dimensional fixed effects, it finds b and ϵ such that y = D'b + ϵ with E[Dϵ] = 0.

It is the back end for the package FixedEffectModels.jl, that estimates linear models with high-dimensional fixed effect.

The package defines two functions solve_coefficients!, that returns the coefficients b, and solve_residuals!, that returns the residuals ϵ. See the documentation ?solve_residuals! or ?solve_coefficients!.

using FixedEffects
# define fixed effects:
p1 = FixedEffect(repeat(1:5, inner = 2))
# combine fixed effects
p2 = FixedEffect(repeat(1:2, outer = 5), repeat(1:2, inner = 5))
# define interacted fixed effects
p3 = FixedEffect(repeat(1:5, outer = 2), interaction = rand(10))

# find residuals
x = rand(10)
solve_residuals!(x, [p1, p2])

# find the fixed effect coefficients
solve_coefficients!(x, [p1, p3])


The package is registered in the General registry and so can be installed at the REPL with ] add FixedEffects.