Binning on Metagenomic Data for Disease Prediction Using Linear Discriminant Analysis and K-Means

Metagenomic data is one of the valuable data resources to predict human disease in personalized medicine. Metagenomic data is very potential and attracted numerous scholars to provide tools and methods to analyze and explore insights in Metagenomics. Binning techniques are promising methods to enhance disease classification on metagenomic data. This study evaluates the integration between Linear Discriminant Analysis and K-Means on preprocessing data before fetching it into prediction models. We perform our experiments on thousands of species abundance metagenomic samples of five diseases have shown that the proposed method can reach 0.913 in accuracy in disease predictions of Liver cirrhosis and obtain promising performance on other four diseases compared to other approaches.