FIGURE 8.

Embedded approaches performs feature selection and extraction. (A) LASSO and Ridge are regularized versions of multiple linear regression used for feature selection. (B) K-means clustering is an unsupervised method for dimensionality reduction that selects feature genes allocated to the nearest centroid. (C) Random Forest (RF) is an ensemble of decision trees. (D) Convolutional Neural network (CNN) and (E) Autoencoders (AE) are deep learning-based methods of feature reduction. Hollow circles in (C) represent samples, and solid triangles in (B,C) represent genes.