Loss reserving generally focuses on identifying a single model that can generate superior predictive performance. However, different loss reserving models specialise in capturing different aspects of loss data. This is recognised in practice in the sense that results from different models are often considered, and sometimes combined. For instance, actuaries may take a weighted average of the prediction outcomes from various loss reserving models, often based on subjective assessments. This package allows for the use of a systematic framework to objectively combine (i.e. ensemble) multiple stochastic loss reserving models such that the strengths offered by different models can be utilised effectively. Our framework is developed in Avanzi et al. (2023). Firstly, our criteria model combination considers the full distributional properties of the ensemble and not just the central estimate - which is of particular importance in the reserving context. Secondly, our framework is that it is tailored for the features inherent to reserving data. These include, for instance, accident, development, calendar, and claim maturity effects. Crucially, the relative importance and scarcity of data across accident periods renders the problem distinct from the traditional ensemble techniques in statistical learning. Our framework is illustrated with a complex synthetic dataset. In the results, the optimised ensemble outperforms both (i) traditional model selection strategies, and (ii) an equally weighted ensemble. In particular, the improvement occurs not only with central estimates but also relevant quantiles, such as the 75th percentile of reserves (typically of interest to both insurers and regulators). Reference: Avanzi B, Li Y, Wong B, Xian A (2023) "Ensemble distributional forecasting for insurance loss reserving" <doi:10.48550/arXiv.2206.08541>.

Create, upload and run Acumos R models. Acumos (<https://www.acumos.org>) is a platform and open source framework intended to make it easy to build, share, and deploy AI apps. Acumos is part of the LF AI Foundation', an umbrella organization within The Linux Foundation'. With this package, user can create a component, and push it to an Acumos platform.

Average population attributable fractions are calculated for a set of risk factors (either binary or ordinal valued) for both prospective and case- control designs. Confidence intervals are found by Monte Carlo simulation. The method can be applied to either prospective or case control designs, provided an estimate of disease prevalence is provided. In addition to an exact calculation of AF, an approximate calculation, based on randomly sampling permutations has been implemented to ensure the calculation is computationally tractable when the number of risk factors is large.

Some functions for drawing some special plots: The function bagplot plots a bagplot, faces plots chernoff faces, iconplot plots a representation of a frequency table or a data matrix, plothulls plots hulls of a bivariate data set, plotsummary plots a graphical summary of a data set, puticon adds icons to a plot, skyline.hist combines several histograms of a one dimensional data set in one plot, slider functions supports some interactive graphics, spin3R helps an inspection of a 3-dim point cloud, stem.leaf plots a stem and leaf plot, stem.leaf.backback plots back-to-back versions of stem and leaf plot.

This package provides methods (<doi:10.7717/peerj.11534>) are provided of calibrating and predicting shifts in allele frequencies through redundancy analysis ('vegan::rda()') and generalized additive models ('mgcv::gam()'). Visualization functions for predicted changes in allele frequencies include shift.dot.ggplot()', shift.pie.ggplot()', shift.moon.ggplot()', shift.waffle.ggplot() and shift.surf.ggplot() that are made with input data sets that are prepared by helper functions for each visualization method. Examples in the documentation show how to prepare animated climate change graphics through a time series with the gganimate package. Function amova.rda() shows how Analysis of Molecular Variance can be directly conducted with the results from redundancy analysis.

This package provides an efficient suite of R tools for scorecard modeling, analysis, and visualization. Including equal frequency binning, equidistant binning, K-means binning, chi-square binning, decision tree binning, data screening, manual parameter modeling, fully automatic generation of scorecards, etc. This package is designed to make scorecard development easier and faster. References include: 1. <http://shichen.name/posts/>. 2. Dong-feng Li(Peking University),Class PPT. 3. <https://zhuanlan.zhihu.com/p/389710022>. 4. <https://www.zhangshengrong.com/p/281oqR9JNw/>.

Programmatic interface to the NASA Application for Extracting and Exploring Analysis Ready Samples services (AppEEARS; <https://appeears.earthdatacloud.nasa.gov/>). The package provides easy access to analysis ready earth observation data in R.

Government Analysis Function recommended colours for use in charts on gov.uk to help meet accessibility guidance.

This package provides access to biographical and political data about Australian federal politicians who served between 1901 and 2021. This enhances how reproducible research is that uses this data.

Build and train a variational autoencoder (VAE) for mixed-type tabular data (continuous, binary, categorical). Models are implemented using TensorFlow and Keras via the reticulate interface, enabling reproducible VAE training for heterogeneous tabular datasets.

DEPRECATED. Do not start building new projects based on this package. (The (in-house) APD file format was initially developed to store Affymetrix probe-level data, e.g. normalized CEL intensities. Chip types can be added to APD file and similar to methods in the affxparser package, this package provides methods to read APDs organized by units (probesets). In addition, the probe elements can be arranged optimally such that the elements are guaranteed to be read in order when, for instance, data is read unit by unit. This speeds up the read substantially. This package is supporting the Aroma framework and should not be used elsewhere.).

Developer oriented utility functions designed to be used as the building blocks of R packages that work with ArcGIS Location Services. It provides functionality for authorization, Esri JSON construction and parsing, as well as other utilities pertaining to geometry and Esri type conversions. To support ArcGIS Pro users, authorization can be done via arcgisbinding'. Installation instructions for arcgisbinding can be found at <https://developers.arcgis.com/r-bridge/installation/>.

This package implements two complementary high-dimensional feature screening methods, Adaptive Iterative Ridge High-dimensional Ordinary Least-squares Projection (Air-HOLP, suitable when the number of predictors p is greater than or equal to the sample size n) and Adaptive Iterative Ridge Ordinary Least Squares (Air-OLS, for n greater than p). Also provides helper functions to generate compound-symmetry and AR(1) correlated data, plus a unified Air() front end and a summary method. For methodological details see Joudah, Muller and Zhu (2025) <doi:10.1007/s11222-025-10599-6>.

Automatically generate a changelog file (NEWS.md / CHANGELOG.md) from the git history using conventional commit messages (<https://www.conventionalcommits.org/en/v1.0.0/>).

Convert populations into integer number of seats for legislative bodies. Implements apportionment methods used historically and currently in the United States for reapportionment after the Census, as described in <https://www.census.gov/history/www/reference/apportionment/methods_of_apportionment.html>.

This package implements the differential equations associated to different versions of Allometric Trophic Models (ATN) to estimate the temporal dynamics of species biomasses in food webs. It offers several features to generate synthetic food webs and to parametrise models as well as a wrapper to the ODE solver deSolve.

Functionality for working with virtual machines (VMs) in Microsoft's Azure cloud: <https://azure.microsoft.com/en-us/services/virtual-machines/>. Includes facilities to deploy, startup, shutdown, and cleanly delete VMs and VM clusters. Deployment configurations can be highly customised, and can make use of existing resources as well as creating new ones. A selection of predefined configurations is provided to allow easy deployment of commonly used Linux and Windows images, including Data Science Virtual Machines. With a running VM, execute scripts and install optional extensions. Part of the AzureR family of packages.

This package provides scalable generalized linear and mixed effects models tailored for sequence count data analysis (e.g., analysis of 16S or RNA-seq data). Uses Dirichlet-multinomial sampling to quantify uncertainty in relative abundance or relative expression conditioned on observed count data. Implements scale models as a generalization of normalizations which account for uncertainty in scale (e.g., total abundances) as described in Nixon et al. (2025) <doi:10.1186/s13059-025-03609-3> and McGovern et al. (2025) <doi:10.1101/2025.08.05.668734>.

Example software for the analysis of data from designed experiments, especially agricultural crop experiments. The basics of the analysis of designed experiments are discussed using real examples from agricultural field trials. A range of statistical methods using a range of R statistical packages are exemplified . The experimental data is made available as separate data sets for each example and the R analysis code is made available as example code. The example code can be readily extended, as required.

An iterative implementation of a recursive binary partitioning algorithm to measure pairwise dependence with a modular design that allows user specification of the splitting logic and stop criteria. Helper functions provide suggested versions of both and support visualization and the computation of summary statistics on final binnings. For a thorough discussion and demonstration of the algorithm, see Salahub and Oldford (2025) <doi:10.1002/sam.70042>.

Adaptive Sparse Multi-block Partial Least Square, a supervised algorithm, is an extension of the Sparse Multi-block Partial Least Square, which allows different quantiles to be used in different blocks of different partial least square components to decide the proportion of features to be retained. The best combinations of quantiles can be chosen from a set of user-defined quantiles combinations by cross-validation. By doing this, it enables us to do the feature selection for different blocks, and the selected features can then be further used to predict the outcome. For example, in biomedical applications, clinical covariates plus different types of omics data such as microbiome, metabolome, mRNA data, methylation data, copy number variation data might be predictive for patients outcome such as survival time or response to therapy. Different types of data could be put in different blocks and along with survival time to fit the model. The fitted model can then be used to predict the survival for the new samples with the corresponding clinical covariates and omics data. In addition, Adaptive Sparse Multi-block Partial Least Square Discriminant Analysis is also included, which extends Adaptive Sparse Multi-block Partial Least Square for classifying the categorical outcome.

Which day a week starts depends heavily on the either the local or professional context. This package is designed to be a lightweight solution to easily switching between week-based date definitions.

Perform parallel factor analysis (PARAFAC: Hitchcock, 1927) <doi:10.1002/sapm192761164> on fluorescence excitation-emission matrices: handle scattering signal and inner filter effect, scale the dataset, fit the model; perform split-half validation or jack-knifing. Modified approaches such as Whittaker interpolation, randomised split-half, and fluorescence and scattering model estimation are also available. The package has a low dependency footprint and has been tested on a wide range of R versions.

Estimate diagnosis performance (Sensitivity, Specificity, Positive predictive value, Negative predicted value) of a diagnostic test where can not measure the golden standard but can estimate it using the attributable fraction.