Utilizando uma Abordagem Híbrida para Interpretação do Clustering Hierárquico

Jean MetzMaria Carolina Monard

In general, data mining tasks can be classified into two categories: predictive, which performs inference on class-labeled data objects in order to make predictions, and descriptive, which characterizes the general properties of the data by grouping unlabeled data objects into clusters. The organization of clusters into a hierarchy can facilitate its interpretation. In this work, we propose a methodology to help the domain specialist to interpret hierarchical clustering results by allowing the user/specialist to carry out an analysis and exploration of the agglomeration hierarchy at different levels of the hierarchy in order to discover symbolic concepts described by this structure. The methodology viability is illustrated on a real case, although more facilities to interact with the user/specialist are needed in the system that implements the methodology in order to improve its applicability.

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: