FR. Ren [1] (ren@mri.tmd.ac.jp)
H. Tanaka [1] (tanaka@mri.tmd.ac.jp)
T. Okayama [2] (tokayama@genes.nig.ac.jp)
[1] Medical Research Institute, Tokyo Medical and Dental Univ.
Yushima 1-5-45, Tokyo 113, Japan
[2] Center for Information Biology,
National Institute of Genetics
1111 Yata, Mishima, Shizuoka 411, Japan
As a novel application of MDL, the concept of model-based complexity is proposed in this study. This method have good asymptotic property and compensates the bias of maximum likelihood method. The efficiency of this method for reconstructing the correct phylogenetic tree is studied by computer simulation, and results suggest that it is superior to the traditional maximum likelihood method or its modification by Akaike's AIC.