Tag: regression

Network trees: networktree 1.0.0, web page, and Psychometrika paper

news

Version 1.0.0 (and actually 1.0.1) of the R package 'networktree' with tools for recursively partitioning covariance structures is now available from CRAN, accompanied by a paper in Psychometrika, and a dedicated software web page. Read more ›

Bias reduction in Poisson and Tobit regression

news

While it is well-known that data separation can cause infinite estimates in binary regression models, the same is also true for other models with a point mass at the bounday (typically at zero) such as Poisson and Tobit. It is shown why this happens and how it can be remedied with bias-reduced estimation, along with implementations in R. Read more ›

Robust covariance matrix estimation: sandwich 3.0-0, web page, JSS paper

news

Version 3.0-0 of the R package 'sandwich' for robust covariance matrix estimation (HC, HAC, clustered, panel, and bootstrap) is now available from CRAN, accompanied by a new web page and a paper in the Journal of Statistical Software (JSS). Read more ›

Structural equation model trees with partykit and lavaan

news

To capture heterogeneity in structural equation models (SEMs), the model-based recursive partitioning (MOB) algorithm from partykit can be coupled with SEM estimation from lavaan. Read more ›

lmSubsets: Exact variable-subset selection in linear regression

news

The R package lmSubsets for flexible and fast exact variable-subset selection is introduced and illustrated in a weather forecasting case study. Read more ›

Circular regression trees and forests

news

A flexible framework for probabilistic forecasting of circular data is introduced, using distributional regression trees and random forests based on the von Mises distribution. Read more ›

bamlss: A Lego toolbox for flexible Bayesian regression

news

Modular R tools for Bayesian regression are provided by bamlss: From classic MCMC-based GLMs and GAMs to distributional models using the lasso or gradient boosting. Read more ›

Network model trees

news

The effect of covariates on correlations in psychometric networks is assessed with either model-based recursive partitioning (MOB) or conditional inference trees (CTree). Read more ›

The power of unbiased recursive partitioning

news

The significance tests underlying the unbiased tree algorithms CTree, MOB, and GUIDE are embedded into a unifying framework. This allows to assess relative strengths and weaknesses in a variety of setups, highlighting the advantages of score-based tests (as in CTree/MOB) vs. residual-based tests (as in GUIDE). Read more ›

Personalized treatment effects with model4you

news

Personalized treatment effects can be estimated easily with model-based trees and model-based random forests using the R package model4you. Read more ›