# Copyright (c) 2014  Takafumi Kanamori  [kanamori@is.nagoya-u.ac.jp]
# All rights reserved.  See the file COPYING for license terms. 

This code is supplementary material for 
T. Kanamori, H, Fujisawa: "Robust Estimation under Heavy Contamination using Unnormalized Models", in Biometrika. 

Unfortunately, it is not well documented yet. If you have any questions please send an email to 
kanamori@is.nagoya-u.ac.jp


====
EXAMPLE
====
After starting R in this folder run 
> source("demo-normalDist.r")  ## estimation of multivariate normal distribution
> source("demo-regression.r")  ## regression problem


====
MAIN FUNCTION
====
The main function you can find in "./functions.r". 
"est_UnnormalizedModel" takes a (sample size x dimension) dimensional matrix as input and provides an estimate of mean
vector, covariance matrix and contamination ratio as output. 
"reg_unnormalizedLSmodel" takes a (sample size x dimension) dimensional design matrix and a (sample size) dimensional
response vector as input and provides an estimate of regression coefficient and contamination ratio as output. 


====
EXPERIMENTS of Table 2
====
After starting R, run 
> source("sim_regression_synthetic.r")
R libraries, MASS, quantreg, robustbase, and snow, are required. 


====
LICENSE
====
See COPYING for license terms.