2. Bias reduction in GLMs- brglm provides brglm.fit which is an alternative method to glm.fit that works under the usual glm interface.
Estimation is either by correcting the bias of the maximum likelihood estimates or using an adjusted-score approach to improving bias.
5. drawExpression- This package aims at giving easy access to the abstract syntax of R through graphical representation.
R expressions and R data structure are represented through some graphical conventions.
8. Bioclim Variables- Bioclimatic envelope variables calculation. This intends to be useful for several researchers making use of niche modeling techniques. The packages here are provided to calculate the necessary input variables from monthly datasets.
11. Gauss-Seidel Estimation of GLMM- This project implements a block-wise Gauss-Seidel (GS) algorithm to fit Generalized Linear Mixed Models (GLMMs). The GS algorithm allows for a significant performance increase and memory saving.
13. PwrGSD- Tools for the design and analysis of sequentially monitored trials on a time to event endpoint. Allows for time varying hazards, allows a flexible specification of contamination and non-compliance.
14. SPecies' LImits by Threshold Statistics- splits contains tools for delimiting species and automated taxonomy at many levels of biological organization (eg. DNA barcoding, morphometrics), top-down (merging phylogenetic and phylogeographic methods) and bottom-up (single samples into >1 groups).
15. Quantmod extension for Russian market- This is an extension for quantmod framework to simplify working with Russian stock market. Adds additional datas to work with MOEX (Moscow Exchange) using finam.ru and mfd.ru datasources.
18. simdat: data simulation- simdat-base aims to provide an easy to use package to simulate data from various experimental designs with various distributional assumptions. The package simdat-gui provides a GUI for less experienced users.
19. semPLS- This package offers an implementation to fit structural equatation models (SEM) by the partial least squares (PLS) algorithm. The PLS approach is referred to as 'soft-modeling' technique and requires no distributional assumptions on the observed data.