CoCo: Coding Cost for Parameter-free Outlier Detection

How can we automatically spot all outstanding observations in a data set? This question arises in a large variety of applications, e.g. in economy, biology and medicine. Existing approaches to outlier