Edit on GitHub

BayesQR.jl

Bayesian quantile regression (BQR) models in Julia.

Installation

Pkg.add("BayesQR")

Function Documentation

BayesQR.bqr — Function
bqr(y::AbstractVector{<:Real}, X::AbstractMatrix{<:Real}, τ::Real, niter::Int, burn::Int)

Runs the Bayesian quantile regression with dependent variable y and covariates X for quantile τ. Priors currently implemented are the Normal and Laplace.

Arguments

  • σᵦ::Real: variance of π(β)
  • prior::String : "Normal" or "Laplace"
bqr(f::FormulaTerm, df::DataFrame, τ::Real, niter::Int, burn::Int)

Runs the Bayesian quantile regression with dependent variable y and covariates X constructed from f and df.

Fitting BayesQR models

Two methods can be used to fit a BQR: bqr(formula, data, τ, niter, burn) and bqr(y, X, τ, niter, burn). Their arguments must be: -formula: a StatsModels.jl Formula object referring to columns in data.

  • data: a table in the Tables.jl definition, e.g. a data frame; NAs are dropped
  • X a matrix holding values of the independent variable(s) in columns
  • y a vector holding values of the dependent variable

Both method returns a MCMCChains.jl Chains object