Physical activity and inactivity are two independent dimensions over which children aggregate into distinct behavioural profiles. Read my new article ‘Probabilistic principal component analysis to identify profiles of physical activity behaviours in the presence of non-ignorable missing data’ in the Journal of the Royal Statistical Society: Series C at http://onlinelibrary.wiley.com/doi/10.1111/rssc.12105/abstract.
Read my new article ‘Improved transformation-based quantile regression’ in the Canadian Journal of Statistics at http://onlinelibrary.wiley.com/doi/10.1002/cjs.11240/abstract!
Read my #OpenAccess article ‘A Gradient Search Maximization Algorithm for the Asymmetric Laplace Likelihood’ at http://bit.ly/1j99BKN!
Here is a talk I recently gave at the UCL Department of Statistical Science.
I’ve uploaded lqmm 1.02 into CRAN. A few bugs have been fixed. I made a substantial change to the function boot.lqmm which provides bootstrapped standard errors of the coefficients estimates. The starting values used in each bootstrap sample are by default those from a least squares fit (startQR = FALSE). In previous versions of the package, starting values were taken from the fitted object (startQR = TRUE) to speed up convergence in each bootstrap sample. However, this causes the standard errors to be under-estimated.
The lqmm package in on CRAN! This package provides commands for fitting linear quantile mixed models (LQMMs). LQMMs complement the mixed-effects models for the mean by providing estimation of conditional quantiles as function of fixed and random effects. More details can be found here and here. More posts on lqmm will follow but feel free to drop me a line if you want to know more.