An Implementation

Abstract

Software based on the minimum message length (MML) principle can be implemented in any reasonable computer programming language. The original version of the MML mixture modelling program Snob (Wallace and Boulton, An information measure for classification. Comput J 11(2):185–194, 1968) was written in ALGOL-60 and later versions were written in FORTRAN and in C (Wallace, Statistical and inductive inference by minimum message length. Springer, Berlin, 2005. ISBN 978-0-387-23795-4. https://doi.org/10.1007/0-387-27656-4). Prototype libraries or packages of more general MML code, as opposed to individual application programs, were written in Java (Comley et al, Flexible decision trees in a general data-mining environment. In: Intelligent data engineering and automated learning. LNCS, vol 2690. Springer, Berlin, pp 761–767, 2003. https://doi.org/10.1007/978-3-540-45080-1_102; Fitzgibbon et al, Probability model type sufficiency. In: Intelligent data engineering and automated learning. LNCS, vol 2690. Springer, Berlin, pp 530–534, 2003. https://doi.org/10.1007/978-3-540-45080-1_72) and in Haskell (Allison, Models for machine learning and data mining in functional programming. J Funct Program 15(1):15–32, 2005). A Java package, mml, of classes and methods for inductive inference using MML accompanies this book.