Table 1.

Glossary of commonly used terms in artificial intelligence applied to healthcare

Term	Definition
Algorithm	A set of rules that precisely defines a sequence of operations for computers
Feature	Image/instance attributes extracted by humans or machines
Feature Selection	Process of selecting relevant features for predictive model development
Instance	A single row of data is called an instance (structured data). It is a single observation from the dataset
Label	Target or reference assigned to instances*/images that machine learning algorithms aim to predict
Model	A mathematical data structure created with machine learning algorithms, which can predict and improve by transforming input data into output
Black-box	Algorithm with an unknown internal processing pattern resulting in difficulty to comprehend how the model reaches the outcome
Classification	Prediction of categorical outputs
Clustering	The task of grouping a set of objects similar to each other into clusters
Convolutional neural network	Deep learning algorithm commonly used in diagnostic imaging
Regression	Prediction of numeric outputs
Segmentation	Process of delineating the boundaries of an organ/lesion in an image
Training	The automatic process of the model after providing a machine learning algorithm (the learning algorithm) with data to learn from
Training dataset	Dataset used for model development
Testing dataset	Dataset unseen by the model during training used to evaluate the model’s performance
Overfitting	Model showing high performance with training data and poor performance with testing data
Underfitting	Model showing poor performance with both training and testing data
Neural network	A model composed of layers consisting of connected nodes inspired by neurons in the human brain