Machine Learning for Microbiome Statistics

Machine learning fundamentally learns from the past experiences (seen data) to make predictions about future (unseen data). Predictions in nature are often uncertain. Microbiome data have unique characteristics, including high-dimensionality, over-dispersion, sparsity and zero-inflation, and heterogeneity. Thus, machine learning involving microbiome data for predicting the outcome of phenotypes is...