Additional file 1 of scSemiAE: a deep model with semi-supervised learning for single-cell transcriptomics

Additional file 1. Results for the first experiment Change of ARI (Louvain & K-means) and ACC (kNN) values with the increasing labeled proportion for six　methods　(AE, netAE, PCA, scSemiAE, scANVI and scVI) and four datasets(Cortex, Heart, Limb Muscle and　Embryos). In addition, mean and std is the mean and standard deviation.

附加文件1：首次实验结果。本结果展示了六种方法（自动编码器（AutoEncoder，AE）、netAE、主成分分析（Principal Component Analysis，PCA）、scSemiAE、scANVI与scVI）在四个数据集（大脑皮层（Cortex）、心脏（Heart）、肢体肌肉（Limb Muscle）、胚胎（Embryos））中，随标记比例提升时，调整兰德指数（Adjusted Rand Index，ARI，基于Louvain与K-means算法）及准确率（Accuracy，ACC，基于k近邻（k-Nearest Neighbor，kNN）方法）的数值变化情况。此外，文中所用的mean与std分别对应平均值与标准差。