“…(
2012) proposed to use MOO in a semisupervised context by generating a variety of partitions and selecting one to optimize a supervised objective (ARI; Vinh, Epps, & Bailey,
2010) in addition to the unsupervised objectives. Khorshiri, Aickelin, Haffari, and Hassani‐Mahmooei (
2019) further included the MOO inside the construction of the partitions itself by identifying Pareto optimal and nonoptimal solutions after each iteration of the
‐medians clustering, with an additional iteration of the
‐medians for the nonoptimal solutions. Pareto MOO allows, considering a finite set of objectives, selecting solutions that are better on at least one objective than each other solution.…”