This package provides R bindings to the GGML tensor library for efficient machine learning computation. Implements core tensor operations including element-wise arithmetic, reshaping, and matrix multiplication. Supports neural network layers (attention, convolutions, normalization), activation functions, and quantization. Features optimization/training API with AdamW (Adam with Weight decay) and SGD (Stochastic Gradient Descent) optimizers, MSE (Mean Squared Error) and cross-entropy losses. Multi-backend support with CPU and optional Vulkan GPU (Graphics Processing Unit) acceleration. See <https://github.com/ggml-org/ggml> for more information about the underlying library.

This package provides tools for fitting sparse generalised linear mixed models with l0 regularisation. Selects fixed and random effects under the hierarchy constraint that fixed effects must precede random effects. Uses coordinate descent and local search algorithms to rapidly deliver near-optimal estimates. Gaussian and binomial response families are currently supported. For more details see Thompson, Wand, and Wang (2025) <doi:10.48550/arXiv.2506.20425>.

This package provides a native R implementation of grammatical evolution (GE). GE facilitates the discovery of programs that can achieve a desired goal. This is done by performing an evolutionary optimisation over a population of R expressions generated via a user-defined context-free grammar (CFG) and cost function.

This package provides a collection of datasets and simplified functions for an introductory (geo)statistics module at University College London. Provides functionality for compositional, directional and spatial data, including ternary diagrams, Wulff and Schmidt stereonets, and ordinary kriging interpolation. Implements logistic and (additive and centred) logratio transformations. Computes vector averages and concentration parameters for the von-Mises distribution. Includes a collection of natural and synthetic fractals, and a simulator for deterministic chaos using a magnetic pendulum example. The main purpose of these functions is pedagogical. Researchers can find more complete alternatives for these tools in other packages such as compositions', robCompositions', sp', gstat and RFOC'. All the functions are written in plain R, with no compiled code and a minimal number of dependencies. Theoretical background and worked examples are available at <https://tinyurl.com/UCLgeostats/>.

This package implements three nonparametric two-sample tests for multivariate paired data and pair matching. Methods are described in the associated preprint: <doi:10.48550/arXiv.2007.01497>.

Generating, evaluating, and selecting initialization strategies for Gaussian Mixture Models (GMMs), along with functions to run the Expectation-Maximization (EM) algorithm. Initialization methods are compared using log-likelihood, and the best-fitting model can be selected using BIC. Methods build on initialization strategies for finite mixture models described in Michael and Melnykov (2016) <doi:10.1007/s11634-016-0264-8> and Biernacki et al. (2003) <doi:10.1016/S0167-9473(02)00163-9>, and on the EM algorithm of Dempster et al. (1977) <doi:10.1111/j.2517-6161.1977.tb01600.x>. Background on model-based clustering includes Fraley and Raftery (2002) <doi:10.1198/016214502760047131> and McLachlan and Peel (2000, ISBN:9780471006268).

It provides functions to generate operating characteristics and to calculate Sequential Conditional Probability Ratio Tests(SCPRT) efficacy and futility boundary values along with sample/event size of Multi-Arm Multi-Stage(MAMS) trials for different outcomes. The package is based on Jianrong Wu, Yimei Li, Liang Zhu (2023) <doi:10.1002/sim.9682>, Jianrong Wu, Yimei Li (2023) "Group Sequential Multi-Arm Multi-Stage Survival Trial Design with Treatment Selection"(Manuscript accepted for publication) and Jianrong Wu, Yimei Li, Shengping Yang (2023) "Group Sequential Multi-Arm Multi-Stage Trial Design with Ordinal Endpoints"(In preparation).

Add trendline and confidence interval of linear or nonlinear regression model and show equation to ggplot as simple as possible. For a general overview of the methods used in this package, see Ritz and Streibig (2008) <doi:10.1007/978-0-387-09616-2> and Greenwell and Schubert Kabban (2014) <doi:10.32614/RJ-2014-009>.

The Genie algorithm (Gagolewski, 2021 <DOI:10.1016/j.softx.2021.100722>) is a robust and outlier-resistant hierarchical clustering method (Gagolewski, Bartoszuk, Cena, 2016 <DOI:10.1016/j.ins.2016.05.003>). This package features its faster and more powerful version. It allows clustering with respect to mutual reachability distances, enabling it to act as a noise point detector or a version of HDBSCAN* that can identify a predefined number of clusters. The package also features an implementation of the Gini and Bonferroni inequality indices, external cluster validity measures (e.g., the normalised clustering accuracy, the adjusted Rand index, the Fowlkes-Mallows index, and normalised mutual information), and internal cluster validity indices (e.g., the Calinski-Harabasz, Davies-Bouldin, Ball-Hall, Silhouette, and generalised Dunn indices). The Python version of genieclust is available via PyPI'.

This package implements the GAMbag, GAMrsm and GAMens ensemble classifiers for binary classification (De Bock et al., 2010) <doi:10.1016/j.csda.2009.12.013>. The ensembles implement Bagging (Breiman, 1996) <doi:10.1023/A:1010933404324>, the Random Subspace Method (Ho, 1998) <doi:10.1109/34.709601> , or both, and use Hastie and Tibshirani's (1990, ISBN:978-0412343902) generalized additive models (GAMs) as base classifiers. Once an ensemble classifier has been trained, it can be used for predictions on new data. A function for cross validation is also included.

Designed to facilitate the preprocessing and linking of GIS (Geographic Information System) databases <https://www.sciencedirect.com/topics/computer-science/gis-database>, the R package GISINTEGRATION offers a robust solution for efficiently preparing GIS data for advanced spatial analyses. This package excels in simplifying intrica procedures like data cleaning, normalization, and format conversion, ensuring that the data are optimally primed for precise and thorough analysis.

Analysis of multivariate data using generalized linear latent variable models (gllvm). Estimation is performed using either the Laplace method, variational approximations, or extended variational approximations, implemented via TMB (Kristensen et al. (2016), <doi:10.18637/jss.v070.i05>).

The function combines a scatter plot with ridgelines to better visualise the distribution between sample groups. The plot is created with ggplot2'.

This package provides functions to develop simulated continuous data (e.g., gene expression) from a sigma covariance matrix derived from a graph structure in igraph objects. Intended to extend mvtnorm to take igraph structures rather than sigma matrices as input. This allows the use of simulated data that correctly accounts for pathway relationships and correlations. This allows the use of simulated data that correctly accounts for pathway relationships and correlations. Here we present a versatile statistical framework to simulate correlated gene expression data from biological pathways, by sampling from a multivariate normal distribution derived from a graph structure. This package allows the simulation of biological pathways from a graph structure based on a statistical model of gene expression. For example methods to infer biological pathways and gene regulatory networks from gene expression data can be tested on simulated datasets using this framework. This also allows for pathway structures to be considered as a confounding variable when simulating gene expression data to test the performance of genomic analyses.

Calculate different glucose variability measures, including average measures of glycemia, measures of glycemic variability and measures of glycemic risk, from continuous glucose monitoring data. Boris P. Kovatchev, Erik Otto, Daniel Cox, Linda Gonder-Frederick, and William Clarke (2006) <doi:10.2337/dc06-1085>. Jean-Pierre Le Floch, Philippe Escuyer, Eric Baudin, Dominique Baudon, and Leon Perlemuter (1990) <doi:10.2337/diacare.13.2.172>. C.M. McDonnell, S.M. Donath, S.I. Vidmar, G.A. Werther, and F.J. Cameron (2005) <doi:10.1089/dia.2005.7.253>. Everitt, Brian (1998) <doi:10.1111/j.1751-5823.2011.00149_2.x>. Becker, R. A., Chambers, J. M. and Wilks, A. R. (1988) <doi:10.2307/2234167>. Dougherty, R. L., Edelman, A. and Hyman, J. M. (1989) <doi:10.1090/S0025-5718-1989-0962209-1>. Tukey, J. W. (1977) <doi:10.1016/0377-2217(86)90209-2>. F. John Service (2013) <doi:10.2337/db12-1396>. Edmond A. Ryan, Tami Shandro, Kristy Green, Breay W. Paty, Peter A. Senior, David Bigam, A.M. James Shapiro, and Marie-Christine Vantyghem (2004) <doi:10.2337/diabetes.53.4.955>. F. John Service, George D. Molnar, John W. Rosevear, Eugene Ackerman, Leal C. Gatewood, William F. Taylor (1970) <doi:10.2337/diab.19.9.644>. Sarah E. Siegelaar, Frits Holleman, Joost B. L. Hoekstra, and J. Hans DeVries (2010) <doi:10.1210/er.2009-0021>. Gabor Marics, Zsofia Lendvai, Csaba Lodi, Levente Koncz, David Zakarias, Gyorgy Schuster, Borbala Mikos, Csaba Hermann, Attila J. Szabo, and Peter Toth-Heyn (2015) <doi:10.1186/s12938-015-0035-3>. Thomas Danne, Revital Nimri, Tadej Battelino, Richard M. Bergenstal, Kelly L. Close, J. Hans DeVries, SatishGarg, Lutz Heinemann, Irl Hirsch, Stephanie A. Amiel, Roy Beck, Emanuele Bosi, Bruce Buckingham, ClaudioCobelli, Eyal Dassau, Francis J. Doyle, Simon Heller, Roman Hovorka, Weiping Jia, Tim Jones, Olga Kordonouri,Boris Kovatchev, Aaron Kowalski, Lori Laffel, David Maahs, Helen R. Murphy, Kirsten Nørgaard, Christopher G.Parkin, Eric Renard, Banshi Saboo, Mauro Scharf, William V. Tamborlane, Stuart A. Weinzimer, and Moshe Phillip.International consensus on use of continuous glucose monitoring.Diabetes Care, 2017 <doi:10.2337/dc17-1600>.

Individual-based simulations forward in time, simulating how patterns in ancestry along the genome change after admixture. Full description can be found in Janzen (2021) <doi:10.1111/2041-210X.13612>.

Derivative Free Gradient Projection Algorithms for Factor Rotation. For more details see ?GPArotateDF. Theory for these functions can be found in the following publications: Jennrich (2004) <doi:10.1007/BF02295647>. Bernaards and Jennrich (2005) <doi:10.1177/0013164404272507>.

In practical applications, the assumptions underlying generalized linear models frequently face violations, including incorrect specifications of the outcome variable's distribution or omitted predictors. These deviations can render the results of standard generalized linear models unreliable. As the sample size increases, what might initially appear as minor issues can escalate to critical concerns. To address these challenges, we adopt a permutation-based inference method tailored for generalized linear models. This approach offers robust estimations that effectively counteract the mentioned problems, and its effectiveness remains consistent regardless of the sample size.

An interface for fitting generalized additive models (GAMs) and generalized additive mixed models (GAMMs) using the lme4 package as the computational engine, as described in Helwig (2024) <doi:10.3390/stats7010003>. Supports default and formula methods for model specification, additive and tensor product splines for capturing nonlinear effects, and automatic determination of spline type based on the class of each predictor. Includes an S3 plot method for visualizing the (nonlinear) model terms, an S3 predict method for forming predictions from a fit model, and an S3 summary method for conducting significance testing using the Bayesian interpretation of a smoothing spline.

This package provides tools for sparse regression modelling with grouped predictors using the group subset selection penalty. Uses coordinate descent and local search algorithms to rapidly deliver near optimal estimates. The group subset penalty can be combined with a group lasso or ridge penalty for added shrinkage. Linear and logistic regression are supported, as are overlapping groups.

Optimal design analysis algorithms for any study design that can be represented or modelled as a generalised linear mixed model including cluster randomised trials, cohort studies, spatial and temporal epidemiological studies, and split-plot designs. See <https://github.com/samuel-watson/glmmrBase/blob/master/README.md> for a detailed manual on model specification. A detailed discussion of the methods in this package can be found in Watson, Hemming, and Girling (2023) <doi:10.1177/09622802231202379>.

Simulating single cell RNA-seq data with complicated structure. This package is developed based on the Splat method (Zappia, Phipson and Oshlack (2017) <doi:10.1186/s13059-017-1305-0>). GeneScape incorporates additional features to simulate single cell RNA-seq data with complicated differential expression and correlation structures, such as sub-cell-types, correlated genes (pathway genes) and hub genes.

Easily explore data by plotting graphs with a few lines of code. Use these ggplot() wrappers to quickly draw graphs of scatter/dots with box-whiskers, violins or SD error bars, data distributions, before-after graphs, factorial ANOVA and more. Customise graphs in many ways, for example, by choosing from colour blind-friendly palettes (12 discreet, 3 continuous and 2 divergent palettes). Use the simple code for ANOVA as ordinary (lm()) or mixed-effects linear models (lmer()), including randomised-block or repeated-measures designs, and fit non-linear outcomes as a generalised additive model (gam) using mgcv(). Obtain estimated marginal means and perform post-hoc comparisons on fitted models (via emmeans()). Also includes small datasets for practising code and teaching basics before users move on to more complex designs. See vignettes for details on usage <https://grafify.shenoylab.com/>. Citation: <doi:10.5281/zenodo.5136508>.

API bindings to the Geospatial Data Abstraction Library ('GDAL', <https://gdal.org>). Implements the GDAL Raster and Vector Data Models. Bindings are implemented with Rcpp modules. Exposed C++ classes and stand-alone functions wrap much of the GDAL API and provide additional functionality. Calling signatures resemble the native C, C++ and Python APIs provided by the GDAL project. Class GDALRaster encapsulates a GDALDataset and its raster band objects. Class GDALVector encapsulates an OGRLayer and the GDALDataset that contains it. Initial bindings are provided to the unified gdal command line interface added in GDAL 3.11. C++ stand-alone functions provide bindings to most GDAL "traditional" raster and vector utilities, including OGR facilities for vector geoprocessing, several algorithms, as well as the Geometry API ('GEOS via GDAL headers), the Spatial Reference Systems API, and methods for coordinate transformation. Bindings to the Virtual Systems Interface ('VSI') API implement standard file system operations abstracted for URLs, cloud storage services, Zip'/'GZip'/'7z'/'RAR', in-memory files, as well as regular local file systems. This provides a single interface for operating on file system objects that works the same for any storage backend. A custom raster calculator evaluates a user-defined R expression on a layer or stack of layers, with pixel x/y available as variables in the expression. Raster combine() identifies and counts unique pixel combinations across multiple input layers, with optional raster output of the pixel-level combination IDs. Basic plotting capability is provided for raster and vector display. gdalraster leans toward minimalism and the use of simple, lightweight objects for holding raw data. Currently, only minimal S3 class interfaces have been implemented for selected R objects that contain spatial data. gdalraster may be useful in applications that need scalable, low-level I/O, or prefer a direct GDAL API.