This package implements a model-based clustering method for categorical life-course sequences relying on mixtures of exponential-distance models introduced by Murphy et al. (2021) <doi:10.1111/rssa.12712>. A range of flexible precision parameter settings corresponding to weighted generalisations of the Hamming distance metric are considered, along with the potential inclusion of a noise component. Gating covariates can be supplied in order to relate sequences to baseline characteristics and sampling weights are also accommodated. The models are fitted using the EM algorithm and tools for visualising the results are also provided.

Enables you to create accessible modal dialogs, with confidence and with minimal configuration.

This package provides functions that (1) fit multivariate discrete distributions, (2) generate random numbers from multivariate discrete distributions, and (3) run regression and penalized regression on the multivariate categorical response data. Implemented models include: multinomial logit model, Dirichlet multinomial model, generalized Dirichlet multinomial model, and negative multinomial model. Making the best of the minorization-maximization (MM) algorithm and Newton-Raphson method, we derive and implement stable and efficient algorithms to find the maximum likelihood estimates. On a multi-core machine, multi-threading is supported.

This package provides tools for constructing, computing, and using distance measures for numerical, categorical, and mixed-type data. The package implements a flexible framework in which continuous and categorical components can be combined under additive, commensurable, and association-aware specifications. Supported methods include classical distances such as Gower, Euclidean, Manhattan, and Mahalanobis-type distances; categorical dissimilarities such as simple matching, occurrence-frequency, and association-based measures; and mixed-type presets designed to reduce biases due to variable type, scale, distribution, redundancy, and number of categories. The package also provides scaling options, supervised and unsupervised distance constructions, leave-one-variable-out tools for distance-based variable importance, and integration with distance-based learning workflows such as nearest-neighbour prediction, partitioning around medoids, and spectral clustering. Methods are motivated by van de Velden, Iodice D'Enza, Markos, and Cavicchia (2026) <doi:10.1080/10618600.2026.2680181> and related work on categorical and mixed-type dissimilarities.

This package provides a set of utility functions for analysing and modelling data from continuous report short-term memory experiments using either the 2-component mixture model of Zhang and Luck (2008) <doi:10.1038/nature06860> or the 3-component mixture model of Bays et al. (2009) <doi:10.1167/9.10.7>. Users are also able to simulate from these models.

This package provides a comprehensive and computationally fast framework to analyze high dimensional data associated with an experimental design based on Multiple ANOVAs (MultANOVA). It includes testing the overall significance of terms in the model, post-hoc analyses of significant terms and variable selection. Details may be found in Mahieu, B., & Cariou, V. (2025). MultANOVA Followed by Post Hoc Analyses for Designed Highâ Dimensional Data: A Comprehensive Framework That Outperforms ASCA, rMANOVA, and VASCA. Journal of Chemometrics, 39(7). <doi:10.1002/cem.70039>.

Matching longitudinal methodology models with complex sampling design. It fits fixed and random effects models and covariance structured models so far. It also provides tools to perform statistical tests considering these specifications as described in : Pacheco, P. H. (2021). "Modeling complex longitudinal data in R: development of a statistical package." <https://repositorio.ufjf.br/jspui/bitstream/ufjf/13437/1/pedrohenriquedemesquitapacheco.pdf>.

Count data is prevalent and informative, with widespread application in many fields such as social psychology, personality, and public health. Classical statistical methods for the analysis of count outcomes are commonly variants of the log-linear model, including Poisson regression and Negative Binomial regression. However, a typical problem with count data modeling is inflation, in the sense that the counts are evidently accumulated on some integers. Such an inflation problem could distort the distribution of the observed counts, further bias estimation and increase error, making the classic methods infeasible. Traditional inflated value selection methods based on histogram inspection are easy to neglect true points and computationally expensive in addition. Therefore, we propose a multiple-inflated negative binomial model to handle count data modeling with multiple inflated values, achieving data-driven inflated value selection. The proposed approach provides simultaneous identification of important regression predictors on the target count response as well. More details about the proposed method are described in Li, Y., Wu, M., Wu, M., & Ma, S. (2023) <arXiv:2309.15585>.

Implementation of Matched Wake Analysis (mwa) for studying causal relationships in spatiotemporal event data, introduced by Schutte and Donnay (2014) <doi:10.1016/j.polgeo.2014.03.001>.

Machine learning algorithms have been used for performing single missing data imputation and most recently, multiple imputations. However, this is the first attempt for using automated machine learning algorithms for performing both single and multiple imputation. Automated machine learning is a procedure for fine-tuning the model automatic, performing a random search for a model that results in less error, without overfitting the data. The main idea is to allow the model to set its own parameters for imputing each variable separately instead of setting fixed predefined parameters to impute all variables of the dataset. Using automated machine learning, the package fine-tunes an Elastic Net (default) or Gradient Boosting, Random Forest, Deep Learning, Extreme Gradient Boosting, or Stacked Ensemble machine learning model (from one or a combination of other supported algorithms) for imputing the missing observations. This procedure has been implemented for the first time by this package and is expected to outperform other packages for imputing missing data that do not fine-tune their models. The multiple imputation is implemented via bootstrapping without letting the duplicated observations to harm the cross-validation procedure, which is the way imputed variables are evaluated. Most notably, the package implements automated procedure for handling imputing imbalanced data (class rarity problem), which happens when a factor variable has a level that is far more prevalent than the other(s). This is known to result in biased predictions, hence, biased imputation of missing data. However, the autobalancing procedure ensures that instead of focusing on maximizing accuracy (classification error) in imputing factor variables, a fairer procedure and imputation method is practiced.

This package provides a unified and consistent S3 interface for training and predicting with a variety of machine learning models in R. The package wraps popular algorithms (e.g., from glmnet', lightgbm', ranger', e1071', and caret') under a common workflow based on simple wrap_*() and predict() functions, allowing users to switch between models without changing their code structure. It supports both classification and regression tasks and facilitates rapid experimentation, benchmarking, and comparison of models. By abstracting away package-specific APIs while preserving flexibility in parameter specification, the package streamlines machine learning workflows and promotes reproducibility.

Offers a general framework of multivariate mixed-effects models for the joint analysis of multiple correlated outcomes with clustered data structures and potential missingness proposed by Wang et al. (2018) <doi:10.1093/biostatistics/kxy022>. The missingness of outcome values may depend on the values themselves (missing not at random and non-ignorable), or may depend on only the covariates (missing at random and ignorable), or both. This package provides functions for two models: 1) mvMISE_b() allows correlated outcome-specific random intercepts with a factor-analytic structure, and 2) mvMISE_e() allows the correlated outcome-specific error terms with a graphical lasso penalty on the error precision matrix. Both functions are motivated by the multivariate data analysis on data with clustered structures from labelling-based quantitative proteomic studies. These models and functions can also be applied to univariate and multivariate analyses of clustered data with balanced or unbalanced design and no missingness.

An implementation of metaheuristic algorithms for continuous optimization. Currently, the package contains the implementations of 21 algorithms, as follows: particle swarm optimization (Kennedy and Eberhart, 1995), ant lion optimizer (Mirjalili, 2015 <doi:10.1016/j.advengsoft.2015.01.010>), grey wolf optimizer (Mirjalili et al., 2014 <doi:10.1016/j.advengsoft.2013.12.007>), dragonfly algorithm (Mirjalili, 2015 <doi:10.1007/s00521-015-1920-1>), firefly algorithm (Yang, 2009 <doi:10.1007/978-3-642-04944-6_14>), genetic algorithm (Holland, 1992, ISBN:978-0262581110), grasshopper optimisation algorithm (Saremi et al., 2017 <doi:10.1016/j.advengsoft.2017.01.004>), harmony search algorithm (Mahdavi et al., 2007 <doi:10.1016/j.amc.2006.11.033>), moth flame optimizer (Mirjalili, 2015 <doi:10.1016/j.knosys.2015.07.006>, sine cosine algorithm (Mirjalili, 2016 <doi:10.1016/j.knosys.2015.12.022>), whale optimization algorithm (Mirjalili and Lewis, 2016 <doi:10.1016/j.advengsoft.2016.01.008>), clonal selection algorithm (Castro, 2002 <doi:10.1109/TEVC.2002.1011539>), differential evolution (Das & Suganthan, 2011), shuffled frog leaping (Eusuff, Landsey & Pasha, 2006), cat swarm optimization (Chu et al., 2006), artificial bee colony algorithm (Karaboga & Akay, 2009), krill-herd algorithm (Gandomi & Alavi, 2012), cuckoo search (Yang & Deb, 2009), bat algorithm (Yang, 2012), gravitational based search (Rashedi et al., 2009) and black hole optimization (Hatamlou, 2013).

Supplementary materials and datasets for the book "Modern Psychometrics With R" (Mair, 2018, Springer useR! series).

An implementation of the expectation conditional maximization (ECM) algorithm for matrix-variate variance gamma (MVVG) and normal-inverse Gaussian (MVNIG) linear models. These models are designed for settings of multivariate analysis with clustered non-uniform observations and correlated responses. The package includes fitting and prediction functions for both models, and an example dataset from a periodontal on Gullah-speaking African Americans, with responses in gaad_res', and covariates in gaad_cov'. For more details on the matrix-variate distributions used, see Gallaugher & McNicholas (2019) <doi:10.1016/j.spl.2018.08.012>.

Fit, plot, and predict a multivariate response, using an arbitrary univariate Bayesian regression model to independently fit basis components (e.g., principal components) of the response (Francom et al., 2025 <DOI:10.1137/24M1644092>).

Integrates large language model generated item responses into psychometric calibration studies through a mixed-subjects design for unidimensional two-parameter and one-parameter logistic item response theory models. Human pilot responses are augmented with model-generated responses using a prediction-powered inference estimator (Angelopoulos, Bates, Fannjiang, Jordan and Zrnic (2023) <doi:10.1126/science.adi6000>; Angelopoulos, Duchi and Zrnic (2023) <doi:10.48550/arXiv.2311.01453>) adapted to marginal maximum-likelihood estimation, following the mixed-subjects design of Broska, Howes and van Loon (2025) <doi:10.1177/00491241251326865>. The estimator is anchored to the human responses and is asymptotically unbiased for the human item parameters at any tuning weight; the weight on the synthetic responses is chosen to minimize propagated ability-score risk, down-weighting uninformative or biased generated responses. Louis-corrected sandwich standard errors, ability scoring, cross-fitted tuning, and scale linking are also provided.

The aim of the package is two-fold: (i) To implement the MMD method for attribution of individuals to sources using the Hamming distance between multilocus genotypes. (ii) To select informative genetic markers based on information theory concepts (entropy, mutual information and redundancy). The package implements the functions introduced by Perez-Reche, F. J., Rotariu, O., Lopes, B. S., Forbes, K. J. and Strachan, N. J. C. Mining whole genome sequence data to efficiently attribute individuals to source populations. Scientific Reports 10, 12124 (2020) <doi:10.1038/s41598-020-68740-6>. See more details and examples in the README file.

Recently, multiple marginal variable selection methods have been developed and shown to be effective in Gene-Environment interactions studies. We propose a novel marginal Bayesian variable selection method for Gene-Environment interactions studies. In particular, our marginal Bayesian method is robust to data contamination and outliers in the outcome variables. With the incorporation of spike-and-slab priors, we have implemented the Gibbs sampler based on Markov Chain Monte Carlo. The core algorithms of the package have been developed in C++'.

Alternative implementation of the beautiful MissForest algorithm used to impute mixed-type data sets by chaining random forests, introduced by Stekhoven, D.J. and Buehlmann, P. (2012) <doi:10.1093/bioinformatics/btr597>. Under the hood, it uses the lightning fast random forest package ranger'. Between the iterative model fitting, we offer the option of using predictive mean matching. This firstly avoids imputation with values not already present in the original data (like a value 0.3334 in 0-1 coded variable). Secondly, predictive mean matching tries to raise the variance in the resulting conditional distributions to a realistic level. This would allow, e.g., to do multiple imputation when repeating the call to missRanger(). Out-of-sample application is supported as well.

Deep Learning library that extends the mlr3 framework by building upon the torch package. It allows to conveniently build, train, and evaluate deep learning models without having to worry about low level details. Custom architectures can be created using the graph language defined in mlr3pipelines'.

Evaluate hypotheses concerning the distribution of multinomial proportions using bridge sampling. The bridge sampling routine is able to compute Bayes factors for hypotheses that entail inequality constraints, equality constraints, free parameters, and mixtures of all three. These hypotheses are tested against the encompassing hypothesis, that all parameters vary freely or against the null hypothesis that all category proportions are equal. For more information see Sarafoglou et al. (2020) <doi:10.31234/osf.io/bux7p>.

Data and code for the paper by Ehm, Gneiting, Jordan and Krueger ('Of Quantiles and Expectiles: Consistent Scoring Functions, Choquet Representations, and Forecast Rankings', JRSS-B, 2016 <DOI:10.1111/rssb.12154>).

This package provides a framework that boosts the imputation of missForest by Stekhoven, D.J. and Bühlmann, P. (2012) <doi:10.1093/bioinformatics/btr597> by harnessing parallel processing and through the fast Gradient Boosted Decision Trees (GBDT) implementation LightGBM by Ke, Guolin et al.(2017) <https://papers.nips.cc/paper/6907-lightgbm-a-highly-efficient-gradient-boosting-decision>. misspi has the following main advantages: 1. Allows embrassingly parallel imputation on large scale data. 2. Accepts a variety of machine learning models as methods with friendly user portal. 3. Supports multiple initializations methods. 4. Supports early stopping that prohibits unnecessary iterations.