Gallery examples: Multi-class AdaBoosted Decision Trees Class Likelihood Ratios to measure classification performance Post-tuning the decision threshold for cost-sensitive learning