MOCO-GCN

The datasets includes 107 metagenomic samples from Spanish and 76 samples from German based on fecal microbial species. Missing values in the German metadata were imputed using the missForest algorithm, a random forest-based method for missing data imputation. The imputation process involved the construction of 100 trees to accurately estimate and fill in the missing values. Of these, every metadata was able to be defined as a binary variable with a positive and negative class, for example, alcohol was considered positive for drinkers and negative for non-drinkers.本数据集包含107份源自西班牙的粪便宏基因组样本，以及76份基于粪便微生物物种研究的德国粪便宏基因组样本。德国元数据（metadata）中的缺失值采用missForest算法进行填充，该算法是一种基于随机森林的缺失数据填充方法。本次填充流程共构建100棵决策树，以精准估算并补全所有缺失值。其中所有元数据字段均可被定义为包含正、负两类的二分类变量，例如饮酒相关元数据以饮酒者为阳性类别，非饮酒者为阴性类别。