A Concept of Multicriteria Stratification: A Definition and Solution
Mikhail Yu. Orlov, Boris Mirkin · Procedia Computer Science · 2014
This paper develops an approach to the problem of multicriteria ranking referredto as multicriteria stratification. The target of stratification is an ordered partition with predefined number of classes rather than a complete ranking of the set of objects.We formulate the problem of multicriteria stratification as a task of minimization of a cost function depending on criteria weights so that strata are to form compact layers on the axis of an aggregate criterion. A quadratic programming algorithm for the problem is proposed. A synthetic data generatorfor a comparative study of the stratification algorithm is developed. The novel algorithm appears to be competitive to a bunch of other approaches on synthetic data. Also, the algorithm is applied to two real-world datasets in the field of scientometrics and leads tosensible and well interpretable results.