Abstract
This paper proposes an unsupervised classification algorithm for Lorenz curves. It extends the density based multivariate cluster approach to the functional framework. In particular, the modes of the small-ball probability are used as starting points to build the clusters. The method is applied to an economic dataset.
Original language | English |
---|---|
Title of host publication | Contributions in infinite-dimensional statistics and related topics |
Publisher | Società Editrice Esculapio |
Pages | 25-30 |
Number of pages | 6 |
ISBN (Print) | 9788874887637 |
Publication status | Published - 1 Jan 2014 |