Multicenter randomized trials involve the collection and analysis of data from numerous study participants across multiple sites. Outliers may be present. To identify outliers, this package examines data at the individual level (univariate and multivariate) and site-level (with and without covariate adjustment). Methods are outlined in further detail in Rigdon et al (to appear).

This package provides a computational toolbox designed for handling missing values in large datasets with the Multiple Imputation by Chained Equations (MICE) by using Apache Spark'. The methodology is described in Morvan et al. (2026) <doi:10.48550/arXiv.2601.21613>.

This package provides functions to perform Bayesian nonparametric univariate and multivariate density estimation and clustering, by means of Pitman-Yor mixtures, and dependent Dirichlet process mixtures for partially exchangeable data. See Corradin et al. (2021) <doi:10.18637/jss.v100.i15> for more details.

An implementation of best subset selection in generalized linear model and Cox proportional hazard model via the primal dual active set algorithm proposed by Wen, C., Zhang, A., Quan, S. and Wang, X. (2020) <doi:10.18637/jss.v094.i04>. The algorithm formulates coefficient parameters and residuals as primal and dual variables and utilizes efficient active set selection strategies based on the complementarity of the primal and dual variables.

Bisulfite-treated RNA non-conversion in a set of samples is analysed as follows : each sample's non-conversion distribution is identified to a Poisson distribution. P-values adjusted for multiple testing are calculated in each sample. Combined non-conversion P-values and standard errors are calculated on the intersection of the set of samples. For further details, see C Legrand, F Tuorto, M Hartmann, R Liebers, D Jakob, M Helm and F Lyko (2017) <doi:10.1101/gr.210666.116>.

Different adjustment methods for batch effects in biomarker data, such as from tissue microarrays. Some methods attempt to retain differences between batches that may be due to between-batch differences in "biological" factors that influence biomarker values.

Bagged OutlierTrees is an explainable unsupervised outlier detection method based on an ensemble implementation of the existing OutlierTree procedure (Cortes, 2020). This implementation takes advantage of bootstrap aggregating (bagging) to improve robustness by reducing the possible masking effect and subsequent high variance (similarly to Isolation Forest), hence the name "Bagged OutlierTrees". To learn more about the base procedure OutlierTree (Cortes, 2020), please refer to <arXiv:2001.00636>.

This package provides a matrix of agreement patterns and counts for record pairs is the input for the procedure. An EM algorithm is used to impute plausible values for missing record pairs. A second EM algorithm, incorporating possible correlations between per-field agreement, is used to estimate posterior probabilities that each pair is a true match - i.e. constitutes the same individual.

Render tables in text format in the terminal using ANSI strings thanks to the cli and crayon packages.

The congeneric normal-ogive model is a popular model for psychometric data (McDonald, R. P. (1997) <doi:10.1007/978-1-4757-2691-6_15>). This model estimates the model, calculates theoretical and concrete reliability coefficients, and predicts the latent variable of the model. This is the companion package to Moss (2020) <doi:10.31234/osf.io/nvg5d>.

Composite likelihood approach is implemented to estimating statistical models for spatial ordinal and proportional data based on Feng et al. (2014) <doi:10.1002/env.2306>. Parameter estimates are identified by maximizing composite log-likelihood functions using the limited memory BFGS optimization algorithm with bounding constraints, while standard errors are obtained by estimating the Godambe information matrix.

The user first provides design vectors n, a and b as well as null (p0) and alternative (p1) benchmark values for the probability of success. The key function "mv.plots.SM()" calculates mean values of exact upper and lower limits based on four different rank ordering methods. These plots form the basis of selecting a rank ordering. The function "inference()" calculates exact limits from a provided realisation and ordering choice. For more information, see "Exact confidence limits after a group sequential single arm binary trial" by Lloyd, C.J. (2020), Statistics in Medicine, Volume 38, 2389-2399, <doi:10.1002/sim.8909>.

Probability mass function, distribution function, quantile function and random generation for the Complex Triparametric Pearson (CTP) and Complex Biparametric Pearson (CBP) distributions developed by Rodriguez-Avi et al (2003) <doi:10.1007/s00362-002-0134-7>, Rodriguez-Avi et al (2004) <doi:10.1007/BF02778271> and Olmo-Jimenez et al (2018) <doi:10.1080/00949655.2018.1482897>. The package also contains maximum-likelihood fitting functions for these models.

This package implements the iterated RMCD method of Cerioli (2010) for multivariate outlier detection via robust Mahalanobis distances. Also provides the finite-sample RMCD method discussed in the paper, as well as the methods provided in Hardin and Rocke (2005) <doi:10.1198/106186005X77685> and Green and Martin (2017) <https://christopherggreen.github.io/papers/hr05_extension.pdf>. See also Chapter 2 of Green (2017) <https://digital.lib.washington.edu/researchworks/handle/1773/40304>.

This package implements Markowitz Critical Line Algorithm ('CLA') for classical mean-variance portfolio optimization, see Markowitz (1952) <doi:10.2307/2975974>. Care has been taken for correctness in light of previous buggy implementations.

Building on top of the RcppArmadillo linear algebra functionalities to do fast spatial interaction models in the context of urban analytics, geography, transport modelling. It uses the Newton root search algorithm to determine the optimal cost exponent and can run country level models with thousands of origins and destinations. It aims at implementing an easy approach based on matrices, that can originate from various routing and processing steps earlier in an workflow. Currently, the simplest form of production, destination and doubly constrained models are implemented. Schlosser et al. (2023) <doi:10.48550/arXiv.2309.02112>.

This package provides a tiny package to generate CRediT author statements (<https://credit.niso.org/>). It provides three functions: create a template, read it back and generate the CRediT author statement in a text file.

Computes classification accuracy and consistency indices under Item Response Theory. Implements the total score IRT-based methods in Lee, Hanson & Brennen (2002) and Lee (2010), the IRT-based methods in Rudner (2001, 2005), and the total score nonparametric methods in Lathrop & Cheng (2014). For dichotomous and polytomous tests.

This package provides some simple functions for printing text in color in markdown or Quarto documents, to be rendered as HTML or LaTeX. This is useful when writing about the use of colors in graphs or tables, where you want to print their names in their actual color to give a direct impression of the color, like â redâ shown in red, or â blueâ shown in blue.

The vctrs package provides a concept of vector prototype that can be especially useful when deploying models and code. Serialize these object prototypes to JSON so they can be used to check and coerce data in production systems, and deserialize JSON back to the correct object prototypes.

This package provides similar functionality to Microsoft Excel CUMPRINC function <https://support.microsoft.com/en-us/office/cumprinc-function-94a4516d-bd65-41a1-bc16-053a6af4c04d>. Returns principal remaining at a given month, principal paid in a month, and accumulated principal paid at a given month based on original loan amount, monthly interest rate, and term of loan.

This package creates an HTML vertical timeline from a data frame as an input for rmarkdown documents and shiny applications.

This package provides robust and efficient methods for estimating causal effects in a target population using a multi-source dataset, including those of Dahabreh et al. (2019) <doi:10.1111/biom.13716>, Robertson et al. (2021) <doi:10.48550/arXiv.2104.05905>, and Wang et al. (2024) <doi:10.48550/arXiv.2402.02684>. The multi-source data can be a collection of trials, observational studies, or a combination of both, which have the same data structure (outcome, treatment, and covariates). The target population can be based on an internal dataset or an external dataset where only covariate information is available. The causal estimands available are average treatment effects and subgroup treatment effects. See Wang et al. (2025) <doi:10.1017/rsm.2025.5> for a detailed guide on using the package.

Concatenation of multiple sequence alignments based on a correspondence table that can be edited in Excel <doi:10.5281/zenodo.5130603>.