Skip to main content

Table 8 Weights (probabilities) obtained for each method applying the Placket-Luce method to the ranking in Table 7, after random triebreakers, and averaging the results. We average 100 runs undoing ties randomly

From: Pattern recognition in data as a diagnosis tool

KM k = 3

KM k = 4

KM k = 5

HC 3C

HC 4C

DBSCAN MinPts = 3, ε = 3

0.148702

0.137208

0.11005

0.20365

0.213026

0.187354