Artificial intelligence and kidney transplantation

Table 1 Different machine learning categories

	Supervised learning	Unsupervised learning	Reinforcement learning
Dataset	Labeled (input and output are known)	Unlabeled (output is not known)	No predefined data
Method	Analyze the relation between input and output. The output is predicted based on this relation	Analyze the input parameters to uncover hidden patterns. Output is predicted based on those patterns	Randomly trialing a vast number of possible inputs, then comparing and grading their performance
Example	Decision trees, support vector machines, neutral networks, k-nearest neighbors	k-means clustering, archetype analysis	Q-learning

Table 2 Machine learning applications used in different kidney transplantation areas

Kidney transplantation category	Machine learning methods used	Ref.
Radiological evaluation	Neural network, convolutional neural network, stacked autoencoders, Bayesian supervised classifier	[4-9]
Pathological evaluation	Neural network, Bayesian network, convolutional neural network, linear discriminant analysis, support vector machines, random forest, archetypal analysis	[10-23]
Prediction of graft survival	Neural network, logistic regression, decision tree, random forest, support vector machines, LASSO, gradient boosting	[24-39]
Optimizing the dose of immunosuppression	Neural network (multilayer perceptron, finite impulse response network, and the Elman recurrent network), adaptive-network-based fuzzy inference system, conditional inference trees, multiple linear regression, regression tree, multivariate adaptive regression splines, boosted regression tree, support vector regression, random forest regression, LASSO regression and Bayesian additive regression trees	[40-46]
Diagnosis of rejection	Neural network, support vector machines, Bayesian interference	[47-52]
Prediction of early graft function	Neural network, logistic regression, linear discriminant analysis, quadratic discriminant analysis, support vector machines, decision tree, random forest, gradient boosting, elastic net	[3,53-57]

LASSO: Least absolute shrinkage and selection operator.