Ani Budi Astuti, Nur Iriawan, Irhamah, Heri Kuswanto
The Bayesian Model Averaging (BMA) required the validation step to determine the accuracy of BMA model. Kolmogorov-Smirnov (KS) and Continuous Ranked Probability Score (CRPS) are used to validate the BMA model. The absolute difference between the empirical cumulative distribution and the hypothesis cumulative distribution were the basic idea of these methods. The KS method uses the distance concept and CRPS method uses the area concept. The validation of BMA model on microarray data by KS and CRPS methods would be identified in this paper. The results have succeed to indentify the performance of KS and CRPS in the validation to BMA model on microarray data with an average value of KS=0.469 and CRPS=0.211 for n=10 and then the value of KS=0.403 and CRPS=0.11 for n=12. © 2014 Ani Budi Astuti, Nur Iriawan, Irhamah and Heri Kuswanto.
Department of Mathematics, Faculty of Mathematics and Natural Sciences, University of Brawijaya Malang Jl, Veteran Malang, 65145, Indonesia; Department of Statistics, Faculty of Mathematics and Natural Sciences, Institut Teknologi Sepuluh Nopember Surabaya, Kampus ITS Sukolilo, Surabaya, 60111, Indonesia
