Table 1.

Terminology of AI subsets

Machine learning (ML)	Algorithms that are able to improve the prediction accuracies by training on large data [10]
Decision tree	A model that consists of nodes and branches, representing variables and related outcomes. Various combinations of outcomes give several predictions. The end model will be the smallest tree that fits the data best [11]
Gradient boosting (GBM)	Builds models that focus on inaccuracies of preceding models and improves these parts until the most accurate model is formed [12]
Random forest	Combines multiple decision trees to build the final accurate prediction model [13]
Support vector machine (SVM)	Finds the optimal border in the dataset to classify outcomes in two groups [14]
Artificial neural networks (ANNs)	Trains by using various processing layers to automatically find relevant features for the prediction. Additionally, weights of the extracted features are adjusted to form the most accurate model [15]
Convolutional neural networks (CNNs)	Similar to ANNs, except these models use filters instead of weight for extracted features [16]
Deep learning	Deep learning algorithms function similarly to neural networks, however, deep learning models have more layers or depth than neural networks [17]
Radiomics	Extracts quantitative features of clinical images to construct predictive or prognostic associations with the predicted medical outcomes [18]

ML machine learning, SVM support vector machine, GBM gradient boosting machine, RF random forest, ANN artificial neural networks, CNN convolutional neural networks