For both tasks and all error metrics, the SVM is the best choice. The differences are higher for small tolerances and in particular for the white wine (e.g. for T=0.25, the SVM accuracy is almost two times better when compared to other methods). This effect is clearly visible when plotting the full REC curves (Fig. 3). The Kappa statistic [33] measures the accuracy when compared with a random classifier (which presents a Kappa value of 0%). The higher the statistics, the more accurate the result. The most practical tolerance values are T = 0.5 and T = 1.0. The former tolerance rounds the regression response into the nearest class, while the latter accepts a response that is correct within one of the two closest classes (e.g. a 3.1 value can be interpreted as grade 3 or 4 but not 2 or 5). For T=0.5, the SVM accuracy improvement is 3.3 pp for red wine (6.2 pp for Kappa), a value that increases to 12.0 pp for the white task (20.4 pp for Kappa).