Resumen de An introduction to nonparametric multimodal regression
The mean, the median and the mode are the classical location measures introduced in any elementary course in statistics. Although mean and median based statistical methods are the usual approaches in different contexts, the mode seems to be somehow neglected. This paper gives a review on nonparametric multimodal regression, an approach for regression where, instead of seeking the mean of the conditional density, as in classical regression models, the conditional local modes are targeted. In addition to revising the existing literature on multimodal regression, the finite sample performance of the multimodal regression estimator is explored with both simulated and real data examples. © 2020 SEIO
