TABLE I.
Description of the Mechanisms of the 5 Employed Machine Learning Algorithms
| Machine Learning Algorithm | Mechanism* |
| Support vector machine | Uses a hyperplane to separate data in ≥2 groups and maximizes the distance between the closest points from both groups and the hyperplane |
| Linear discriminant analysis | Projects multidimensional data (many metrics) on a single dimension to maximize the distance between the means of the groups and minimize the variance within each group |
| k-nearest neighbors | Uses distance functions such as the Euclidean distance to determine the closest neighbors to a point. A parameter (k) corresponds with the number of neighbors considered. The class of a participant is determined on the basis of its relationship with the nearest participants in a multidimensional space |
| Naive Bayes | Classifies participants on the basis of probabilities that the chosen metrics belong to experts or novice surgeons. It assumes that all of the chosen metrics are independent from each other |
| Decision tree | Classifies individuals by building a series of nodes whereby subjects are divided according to the value of a certain metric. The algorithm finds the optimal values to divide subjects in classes |