Semi-Supervised Algorithms

Abstract

Semi-supervised learning, or SSL, falls somewhere between supervised and unsupervised learning. The algorithm is provided with some supervision data in addition to unlabeled data. There are two primary learning paradigms in it. Transductive education aims to use the trained classifier on unlabeled instances observed during training. This kind of algorithm is mainly used for node embedding on graphs, like random walks, where the goal is to label the graph's unlabeled nodes at the training time. Inductive learning aims to develop a classifier that can generalize unobserved situations during a test. This chapter details different semi-supervised algorithms in healthcare.