A new approach for clustering of effects in quantile regression
- Autori: Sottile, G.; Adelfio, G.
- Anno di pubblicazione: 2017
- Tipologia: Proceedings (TIPOLOGIA NON ATTIVA)
- OA Link: http://hdl.handle.net/10447/238651
In this paper we aim at nding similarities among the coefficients from a multivariate regression. Using a quantile regression coefficients modeling, the effect of each covariate, given a response (also multivariate) is a curve in the multidimensional space of the percentiles. Collecting all the curves, describing the effects of each covariate on each response variable, we could be able to assess if only one or more covariates have same effects on different responses.