Fig. 1
From: scFSNN: a feature selection method based on neural network for single-cell RNA-seq data
The flow chart of scFSNN. The scFSNN model consists of two parts: model initialization and feature selection. The model initialization process begins with data augmentation and generating surrogate features. It then initializes the model to obtain the estimated value of \(p_0\). The second part is the feature selection process. This process initially calculates the importance scores of features and eliminates one or some least important features based on the importance scores. Subsequently, it estimates the False Discovery Rate (FDR) of the remaining features. If the estimated FDR is greater than the given cutoff, the feature selection process continues; otherwise, the feature selection process stops, and the remaining original variables are used to train the final model