Overall average number of hits obtained by the classification model, represented as a percentage.

In rule A → C, A is the antecedent of the rule. In other words, it must appear in the instance so that C will also appear with a high probability.

A set of measurements recorded when evaluating body composition.

This was the first algorithm proposed in the scientific literature to obtain ARs from a data set. This algorithm uses a breadth-first generation process to extract all the frequent itemsets and then generates the RAs from them. It makes use of the anti-monotonicity property of the support metric to improve the efficiency of generating the frequent itemsets.

These are defined as expressions of type A → C, where A and C are itemsets whose intersection is empty. These rules represent the fact that when the elements of A appear in an instance of the data set, there is a high probability that the elements of C will also appear in that instance.

Association rules are used to identify and represent dependencies between the elements or values of a data set in which we do not know the class to which they belong.

A quality metric based on the output probabilities of the classifier and the balance achieved between true and false positives for each probability threshold value.

Type of ensemble that uses a different subset of the training data, in this case, M estimators are trained independently.

A supervised learning problem where the output variable has only two possible states.

An interdisciplinary field dedicated to the development of methods and software for understanding biological data. Bioinformatics combines biology, computer science, engineering, mathematics, and statistics to analyze and interpret biological data.