Generalised additive P-spline regression models estimation using the separation of overlapping precision matrices (SOP) method. Estimation is based on the equivalence between P-splines and linear mixed models, and variance/smoothing parameters are estimated based on restricted maximum likelihood (REML). The package enables users to estimate P-spline models with overlapping penalties. Based on the work described in Rodriguez-Alvarez et al. (2015) <doi:10.1007/s11222-014-9464-2>; Rodriguez-Alvarez et al. (2019) <doi:10.1007/s11222-018-9818-2>, and Eilers and Marx (1996) <doi:10.1214/ss/1038425655>.

L2 penalized logistic regression for both continuous and discrete predictors, with forward stagewise/forward stepwise variable selection procedure.

This package provides methods for computing spatial, temporal, and spatiotemporal statistics as described in Gouhier and Guichard (2014) <doi:10.1111/2041-210X.12188>. These methods include empirical univariate, bivariate and multivariate variograms; fitting variogram models; phase locking and synchrony analysis; generating autocorrelated and cross-correlated matrices.

Dual interfaces, graphical and programmatic, designed for intuitive applications of Multilevel Regression and Poststratification (MRP). Users can apply the method to a variety of datasets, from electronic health records to sample survey data, through an end-to-end Bayesian data analysis workflow. The package provides robust tools for data cleaning, exploratory analysis, flexible model building, and insightful result visualization. For more details, see Si et al. (2020) <https://www150.statcan.gc.ca/n1/en/pub/12-001-x/2020002/article/00003-eng.pdf?st=iF1_Fbrh> and Si (2025) <doi:10.1214/24-STS932>.

Fitting of non-parametric production frontiers for use in efficiency analysis. Methods are provided for both a smooth analogue of Data Envelopment Analysis (DEA) and a non-parametric analogue of Stochastic Frontier Analysis (SFA). Frontiers are constructed for multiple inputs and a single output using constrained kernel smoothing as in Racine et al. (2009), which allow for the imposition of monotonicity and concavity constraints on the estimated frontier.

Fits singular linear models to longitudinal data. Singular linear models are useful when the number, or timing, of longitudinal observations may be informative about the observations themselves. They are described in Farewell (2010) <doi:10.1093/biomet/asp068>, and are extensions of the linear increments model <doi:10.1111/j.1467-9876.2007.00590.x> to general longitudinal data.

Spatiotemporal individual-level model of seasonal infectious disease transmission within the Susceptible-Exposed-Infectious-Recovered-Susceptible (SEIRS) framework are applied to model seasonal infectious disease transmission. This package employs a likelihood based Monte Carlo Expectation Conditional Maximization (MCECM) algorithm for estimating model parameters. In addition to model fitting and parameter estimation, the package offers functions for calculating AIC using real pandemic data and conducting simulation studies customized to user-specified model configurations.

This package provides a set of functions that can be used to spatially thin species occurrence data. The resulting thinned data can be used in ecological modeling, such as ecological niche modeling.

This package implements a physics-informed one-dimensional convolutional neural network (CNN1D-PINN) for estimating the complete soil water retention curve (SWRC) as a continuous function of matric potential, from soil texture, organic carbon, bulk density, and depth. The network architecture ensures strict monotonic decrease of volumetric water content with increasing suction by construction, through cumulative integration of non-negative slope outputs (monotone integral architecture). Four physics-based residual constraints adapted from Norouzi et al. (2025) <doi:10.1029/2024WR038149> are embedded in the loss function: (S1) linearity at the dry end (pF in [5, 7.6]); (S2) non-negativity at pF = 6.2; (S3) non-positivity at pF = 7.6; and (S4) a near-zero derivative in the saturated plateau region (pF in [-2, -0.3]). Includes tools for data preparation, model training, dense prediction, performance metrics, texture classification, and publication-quality visualisation.

This package provides a sparklyr extension that enables reading and writing TensorFlow TFRecord files via Apache Spark'.

This package provides a framework for modeling cellular metabolic states and continuous metabolic trajectories from single-cell RNA-seq data using pathway-level scoring. Enables lineage-restricted metabolic analysis, metabolic pseudotime inference, module-level trend analysis, and visualization of metabolic state transitions.

This package provides a coordinate descent algorithm for computing the solution paths of the sparse and coupled sparse asymmetric least squares, including the (adaptive) elastic net and Lasso penalized SALES and COSALES regressions.

Several functions and S3 methods to construct a super learner in the presence of censored times-to-event and to evaluate its prognostic capacities.

This package provides two methods for segmentation and joint segmentation/clustering of bivariate time-series. Originally intended for ecological segmentation (home-range and behavioural modes) but easily applied on other series, the package also provides tools for analysing outputs from R packages moveHMM and marcher'. The segmentation method is a bivariate extension of Lavielle's method available in adehabitatLT (Lavielle, 1999 <doi:10.1016/S0304-4149(99)00023-X> and 2005 <doi:10.1016/j.sigpro.2005.01.012>). This method rely on dynamic programming for efficient segmentation. The segmentation/clustering method alternates steps of dynamic programming with an Expectation-Maximization algorithm. This is an extension of Picard et al (2007) <doi:10.1111/j.1541-0420.2006.00729.x> method (formerly available in cghseg package) to the bivariate case. The method is fully described in Patin et al (2018) <doi:10.1101/444794>.

Population genetics package for designing diagnostic panels. Candidate markers, marker combinations, and different panel sizes are assessed for how well they can predict the source population of known samples. Requires a genotype file of candidate markers in STRUCTURE format. Methods for population cross-validation are described in Jombart (2008) <doi:10.1093/bioinformatics/btn129>.

Implementation of the SIC epsilon-telescope method, either using single or distributional (multiparameter) regression. Includes classical regression with normally distributed errors and robust regression, where the errors are from the Laplace distribution. The "smooth generalized normal distribution" is used, where the estimation of an additional shape parameter allows the user to move smoothly between both types of regression. See O'Neill and Burke (2022) "Robust Distributional Regression with Automatic Variable Selection" for more details. <doi:10.48550/arXiv.2212.07317>. This package also contains the data analyses from O'Neill and Burke (2023). "Variable selection using a smooth information criterion for distributional regression models". <doi:10.1007/s11222-023-10204-8>.

This package implements a Super Learner framework for right-censored survival data. The package fits convex combinations of parametric, semiparametric, and machine learning survival learners by minimizing cross-validated risk using inverse probability of censoring weighting (IPCW). It provides tools for automated hyperparameter grid search, high-dimensional variable screening, and evaluation of prediction performance using metrics such as the Brier score, Uno's C-index, and time-dependent area under the curve (AUC). Additional utilities support model interpretation for survival ensembles, including Shapley additive explanations (SHAP), and estimation of covariate-adjusted restricted mean survival time (RMST) contrasts. The methodology is related to treatment-specific survival curve estimation using machine learning described by Westling, Luedtke, Gilbert and Carone (2024) <doi:10.1080/01621459.2023.2205060>, and the unified ensemble framework described in Lyu et al. (2026) <doi:10.64898/2026.03.11.711010>.

Includes bases for litholog generation: graphical functions based on R base graphics, interval management functions and svg importation functions among others. Also include stereographic projection functions, and other functions made to deal with large datasets while keeping options to get into the details of the data. When using for publication please cite Sebastien Wouters, Anne-Christine Da Silva, Frederic Boulvain and Xavier Devleeschouwer, 2021. The R Journal 13:2, 153-178. The palaeomagnetism functions are based on: Tauxe, L., 2010. Essentials of Paleomagnetism. University of California Press. <https://earthref.org/MagIC/books/Tauxe/Essentials/>; Allmendinger, R. W., Cardozo, N. C., and Fisher, D., 2013, Structural Geology Algorithms: Vectors & Tensors: Cambridge, England, Cambridge University Press, 289 pp.; Cardozo, N., and Allmendinger, R. W., 2013, Spherical projections with OSXStereonet: Computers & Geosciences, v. 51, no. 0, p. 193 - 205, <doi: 10.1016/j.cageo.2012.07.021>.

RNA sequencing analysis methods are often derived by relying on hypothetical parametric models for read counts that are not likely to be precisely satisfied in practice. Methods are often tested by analyzing data that have been simulated according to the assumed model. This testing strategy can result in an overly optimistic view of the performance of an RNA-seq analysis method. We develop a data-based simulation algorithm for RNA-seq data. The vector of read counts simulated for a given experimental unit has a joint distribution that closely matches the distribution of a source RNA-seq dataset provided by the user. Users control the proportion of genes simulated to be differentially expressed (DE) and can provide a vector of weights to control the distribution of effect sizes. The algorithm requires a matrix of RNA-seq read counts with large sample sizes in at least two treatment groups. Many datasets are available that fit this standard.

An algorithm for identifying high-resolution driver elements for datasets from a high-definition reporter assay library. Xinchen Wang, Liang He, Sarah Goggin, Alham Saadat, Li Wang, Melina Claussnitzer, Manolis Kellis (2017) <doi:10.1101/193136>.

An efficient implementation of SCCI using Rcpp'. SCCI is short for the Stochastic Complexity-based Conditional Independence criterium (Marx and Vreeken, 2019). SCCI is an asymptotically unbiased and L2 consistent estimator of (conditional) mutual information for discrete data.

Manage a collection/library of R source packages. Discover, document, load, test source packages. Enable to use those packages as if they were actually installed. Quickly reload only what is needed on source code change. Run tests and checks in parallel.

Soft-margin support vector machines (SVMs) are a common class of classification models. The training of SVMs usually requires that the data be available all at once in a single batch, however the Stochastic majorization-minimization (SMM) algorithm framework allows for the training of SVMs on streamed data instead Nguyen, Jones & McLachlan(2018)<doi:10.1007/s42081-018-0001-y>. This package utilizes the SMM framework to provide functions for training SVMs with hinge loss, squared-hinge loss, and logistic loss.

This package produces LaTeX code, HTML/CSS code and ASCII text for well-formatted tables that hold regression analysis results from several models side-by-side, as well as summary statistics.