Fig. 2

Epsilon results from Naïve Bayes for irregularity. N = 7,066 from 2011 to 2017 cohorts, UNAM Faculty of Medicine, Mexico. The figure is divided in two parts: on the right are the results for regularity and on the left those for irregularity. The variables are depicted along the vertical axis, ordered according to their relevance based on epsilon: those with the largest epsilon values appear at the top, which reflects their greater statistical significance. The two vertical lines represent the points where epsilon is equal to 2 and − 2, respectively. Positive epsilon values indicate the categories which favour belonging to the target group (right side), while the negatives ones the categories which do not (left side). The colour intensity of the dots represents the probability of belonging to the target group, with lighter dots representing those categories with a probability closer to 1, and darker ones with a probability closer to 0