Proposta de dissertação do MEI
Título: Mining Extremes through Clustering
Proponente(s): Susana Nascimento (snt@fct.unl.pt)
Créditos: 42 ECTS
Área científica: Decision Support and Artificial Intelligence
Início preferencial: Qualquer semestre
URL:
Já estão em curso trabalhos preliminares executados pelo alunos:
Gonçalo Sousa Mendes, nº 42082
Breve descrição: Archetypes are extreme points that synthesize data representing “pure” individual types, and are assigned by the most discriminating features of data points. Archetypes are almost always useful and easy to interpret as they represent extreme combinations of features.

Recent applications where this concept is being explored include talent analysis in sports and science, fraud detection, profiling of users and products in recommendation systems, as well as benchmark analysis.

Archetypal Analysis (AA) and some versions of Fuzzy Clustering (FC) propose complementary definitions to relate entity-to-feature data points to clusters’ prototypes.

The main goal of this work is to perform a computational study comparing Archetypal Analysis and Fuzzy Clustering algorithms using recent benchmark data. Unsupervised validation method will be developed for assessment of the results.
Observações: