DeepAI AI Chat
Log In Sign Up

The Lagrange approach in the monotone single index model

by   Piet Groeneboom, et al.
Delft University of Technology

The finite-dimensional parameters of the monotone single index model are often estimated by minimization of a least squares criterion and reparametrization to deal with the non-unicity. We avoid the reparametrization by using a Lagrange-type method and replace the minimization over the finite-dimensional parameter by a `crossing of zero' criterion at the derivative level. In particular, we consider a simple score estimator (SSE), an efficient score estimator (ESE), and a penalized least squares estimator (PLSE) for which we can apply this method. The SSE and ESE were discussed in Balabdaoui, Groeneboom and Hendrickx (2018, but the proofs still used reparametrization. Another version of the PLSE was discussed in 20 pagesKuchibhotla and Patra (2017), where also reparametrization was used. The estimators are compared with the profile least squares estimator (LSE), Han's maximum rank estimator (MRE), the effective dimension reduction estimator (EDR) and a linear least squares estimator, which can be used if the covariates have a elliptically symmetric distribution. We also investigate the effects of random starting values in the search algorithms.


page 1

page 2

page 3

page 4


Profile least squares estimators in the monotone single index model

We consider least squares estimators of the finite dimensional regressio...

Score estimation in the monotone single index model

We consider estimation of the regression parameter in the single index m...

Slice Weighted Average Regression

It has previously been shown that ordinary least squares can be used to ...

Regression estimators for the tail index

We propose a class of weighted least squares estimators for the tail ind...

Least Squares Estimation of a Monotone Quasiconvex Regression Function

We develop a new approach for the estimation of a multivariate function ...

Error Reduction from Stacked Regressions

Stacking regressions is an ensemble technique that forms linear combinat...

On Optimal and Feasible Regularization in Linear Models in Time Series

We discuss predictive linear modeling in the presence of: (i) stochastic...