Discriminant Non-Negative Matrix Factorization aims to extend the Non-negative Matrix Factorization algorithm in order to extract features that enforce not only the spatial locality, but also the separability between classes in a discriminant manner. It refers to three article, Zafeiriou, Stefanos, et al. "Exploiting discriminant information in nonnegative matrix factorization with application to frontal face verification." Neural Networks, IEEE Transactions on 17.3 (2006): 683-695. Kim, Bo-Kyeong, and Soo-Young Lee. "Spectral Feature Extraction Using dNMF for Emotion Recognition in Vowel Sounds." Neural Information Processing. Springer Berlin Heidelberg, 2013. and Lee, Soo-Young, Hyun-Ah Song, and Shun-ichi Amari. "A new discriminant NMF algorithm and its application to the extraction of subtle emotional differences in speech." Cognitive neurodynamics 6.6 (2012): 525-535.

Interactions between different biological entities are crucial for the function of biological systems. In such networks, nodes represent biological elements, such as genes, proteins and microbes, and their interactions can be defined by edges, which can be either binary or weighted. The dysregulation of these networks can be associated with different clinical conditions such as diseases and response to treatments. However, such variations often occur locally and do not concern the whole network. To capture local variations of such networks, we propose multiplex network differential analysis (MNDA). MNDA allows to quantify the variations in the local neighborhood of each node (e.g. gene) between the two given clinical states, and to test for statistical significance of such variation. Yousefi et al. (2023) <doi:10.1101/2023.01.22.525058>.

Bayesian Additive Regression Kernels (BARK) provides an implementation for non-parametric function estimation using Levy Random Field priors for functions that may be represented as a sum of additive multivariate kernels. Kernels are located at every data point as in Support Vector Machines, however, coefficients may be heavily shrunk to zero under the Cauchy process prior, or even, set to zero. The number of active features is controlled by priors on precision parameters within the kernels, permitting feature selection. For more details see Ouyang, Z (2008) "Bayesian Additive Regression Kernels", Duke University. PhD dissertation, Chapter 3 and Wolpert, R. L, Clyde, M.A, and Tu, C. (2011) "Stochastic Expansions with Continuous Dictionaries Levy Adaptive Regression Kernels, Annals of Statistics Vol (39) pages 1916-1962 <doi:10.1214/11-AOS889>.

Comprehensive toolkit for valid spatial, temporal, and grouped model evaluation. Automatically detects data dependencies (spatial autocorrelation, temporal structure, clustered observations), generates appropriate cross-validation schemes (spatial blocking, checkerboard, hexagonal, KNNDM, environmental blocking, leave-location-out, purged CV), and validates evaluation pipelines for leakage. Includes area of applicability (AOA) assessment following Meyer & Pebesma (2021) <doi:10.1111/2041-210X.13650>, forward feature selection with blocked CV, spatial thinning, block-permutation variable importance, extrapolation detection, and interactive visualizations. Integrates with tidymodels', caret', mlr3', ENMeval', and biomod2'. Based on evaluation principles described in Roberts et al. (2017) <doi:10.1111/ecog.02881>, Kaufman et al. (2012) <doi:10.1145/2382577.2382579>, Kapoor & Narayanan (2023) <doi:10.1016/j.patter.2023.100804>, and Linnenbrink et al. (2024) <doi:10.5194/gmd-17-5897-2024>.

Platform dedicated to the Global Modelling technique. Its aim is to obtain ordinary differential equations of polynomial form directly from time series. It can be applied to single or multiple time series under various conditions of noise, time series lengths, sampling, etc. This platform is developped at the Centre d'Etudes Spatiales de la Biosphere (CESBIO), UMR 5126 UPS/CNRS/CNES/IRD, 18 av. Edouard Belin, 31401 TOULOUSE, FRANCE. The developments were funded by the French program Les Enveloppes Fluides et l'Environnement (LEFE, MANU, projets GloMo, SpatioGloMo and MoMu). The French program Defi InFiNiTi (CNRS) and PNTS are also acknowledged (projects Crops'IChaos and Musc & SlowFast). The method is described in the article : Mangiarotti S. and Huc M. (2019) <doi:10.1063/1.5081448>.

This package provides functions to download, parse, and tidy statistical data published by HM Revenue and Customs ('HMRC') on GOV.UK'. Returns annotated hmrc_tbl data frames with provenance metadata (source URL, publication date, fetch time, cell methods) for reproducible fiscal research. Covers monthly tax receipts (42 series), VAT (from 1973), fuel duties (from 1990), tobacco duties (from 1991), Corporation Tax receipts, stamp duty, research and development tax credits (from 2000), tax gap estimates (from 2005), Income Tax liabilities by income range, Capital Gains Tax, Inheritance Tax, the Patent Box, creative industries reliefs, and monthly property transaction counts. File URLs are resolved at runtime via the GOV.UK Content API <https://www.gov.uk/api/content>, so data is always current without hardcoded URLs. Files are cached locally between sessions.

This package provides tools for estimating causal effects in panel data using counterfactual methods, as well as other modern DID estimators. It is designed for causal panel analysis with binary treatments under the parallel trends assumption. The package supports scenarios where treatments can switch on and off and allows for limited carryover effects. It includes several imputation estimators, such as Gsynth (Xu 2017), linear factor models, and the matrix completion method. Detailed methodology is described in Liu, Wang, and Xu (2024) <doi:10.48550/arXiv.2107.00856> and Chiu et al. (2025) <doi:10.48550/arXiv.2309.15983>. Optionally integrates with the "HonestDiDFEct" package for sensitivity analyses compatible with imputation estimators. "HonestDiDFEct" is not on CRAN but can be obtained from <https://github.com/lzy318/HonestDiDFEct>.

This package implements the area-based nonparametric spatial point pattern test of Andresen (2009) <doi:10.1016/j.apgeog.2008.12.004> for comparing two spatial distributions of events aggregated to areal units, with the applications and extensions reviewed in Andresen (2016) <doi:10.1177/2059799116630659>. Unit-level inference uses bounded Monte Carlo bootstrap p-values, with marginal confidence interval overlap retained as an option. Global similarity is summarised by the robust S index, accompanied by a test of whether two patterns are more different than two realisations of a single process, and a test of the spatial arrangement of the differences. A longitudinal extension compares each period of a time series against a base period and classifies unit-level trajectories. Publication-ready maps can be redrawn with user-chosen layouts.

An algorithm that trains a meta-learning procedure that combines screening and wrapper methods to find a set of extremely low-dimensional attribute combinations. This package works on top of the caret package and proceeds in a forward-step manner. More specifically, it builds and tests learners starting from very few attributes until it includes a maximal number of attributes by increasing the number of attributes at each step. Hence, for each fixed number of attributes, the algorithm tests various (randomly selected) learners and picks those with the best performance in terms of training error. Throughout, the algorithm uses the information coming from the best learners at the previous step to build and test learners in the following step. In the end, it outputs a set of strong low-dimensional learners.

Time Series Qn is a package with applications of the Qn estimator of Rousseeuw and Croux (1993) <doi:10.1080/01621459.1993.10476408> to univariate and multivariate Time Series in time and frequency domains. More specifically, the robust estimation of autocorrelation or autocovariance matrix functions from Ma and Genton (2000, 2001) <doi:10.1111/1467-9892.00203>, <doi:10.1006/jmva.2000.1942> and Cotta (2017) <doi:10.13140/RG.2.2.14092.10883> are provided. The robust pseudo-periodogram of Molinares et. al. (2009) <doi:10.1016/j.jspi.2008.12.014> is also given. This packages also provides the M-estimator of the long-memory parameter d based on the robustification of the GPH estimator proposed by Reisen et al. (2017) <doi:10.1016/j.jspi.2017.02.008>.

This package provides functions developed within Breeding Insight to analyze diploid and polyploid breeding and genetic data. BIGr provides the ability to filter variant call format (VCF) files, extract single nucleotide polymorphisms (SNPs) from diversity arrays technology missing allele discovery count (DArT MADC) files, and manipulate genotype data for both diploid and polyploid species. It also serves as the core dependency for the BIGapp Shiny app, which provides a user-friendly interface for performing routine genotype analysis tasks such as dosage calling, filtering, principal component analysis (PCA), genome-wide association studies (GWAS), and genomic prediction. For more details about the included breedTools functions, see Funkhouser et al. (2017) <doi:10.2527/tas2016.0003>, and the updog output format, see Gerard et al. (2018) <doi:10.1534/genetics.118.301468>.

Sequential Monte Carlo (SMC) inference for fully Bayesian Gaussian process (GP) regression and classification models by particle learning (PL) following Gramacy & Polson (2011) <doi:10.48550/arXiv.0909.5262>. The sequential nature of inference and the active learning (AL) hooks provided facilitate thrifty sequential design (by entropy) and optimization (by improvement) for classification and regression models, respectively. This package essentially provides a generic PL interface, and functions (arguments to the interface) which implement the GP models and AL heuristics. Functions for a special, linked, regression/classification GP model and an integrated expected conditional improvement (IECI) statistic provide for optimization in the presence of unknown constraints. Separable and isotropic Gaussian, and single-index correlation functions are supported. See the examples section of ?plgp and demo(package="plgp") for an index of demos.

This package provides tools for estimating length-based indicators (LBIs) from length-frequency data to assess fish stock status and evaluate growth and recruitment overfishing in data-limited fisheries. Implements the sustainability indicators of Froese (2004) <doi:10.1111/j.1467-2979.2004.00144.x>, empirical biological reference points from Froese and Binohlan (2000) <doi:10.1111/j.1095-8649.2000.tb00870.x>, and the decision framework of Cope and Punt (2009) <doi:10.1577/C08-025.1>. Incorporates a three-tier Monte Carlo and bootstrap uncertainty propagation framework for sustainability indicators, optimum bin size calculations following Wang et al. (2020) <doi:10.1016/j.fishres.2019.105474>, multi-month length-frequency harmonization, and length-weight relationship fitting. Methodology is detailed in Ali et al. (2025) <doi:10.1016/j.fishres.2025.107467>.

To help French students from scientific preparatory classes for the Grandes Ecoles (CPGE) in their choice of field of study and career options, this package provides an interactive tool and data visualization of a graph clustered by different competitive exams and sectors of activity for French selective engineering schools and selective higher education institutions like Ecoles Normales Superieures (ENS) or specialized university programs (magisteres). Besides, there are two drop-down menus to select on the graph many fields or more than 200 engineering schools or ENS or magisteres. It gives the opportunity to expand, collapse clusters of selective exams interactively too. For more information, see the demonstration video: <https://valerierobert-maths.re/index.php/maths-en-cpge/>. The data was collected via the official French website: <https://www.scei-concours.fr/statistiques.html>.

It contains six common multi-category classification accuracy evaluation measures. All of these measures could be found in Li and Ming (2019) <doi:10.1002/sim.8103>. Specifically, Hypervolume Under Manifold (HUM), described in Li and Fine (2008) <doi:10.1093/biostatistics/kxm050>. Correct Classification Percentage (CCP), Integrated Discrimination Improvement (IDI), Net Reclassification Improvement (NRI), R-Squared Value (RSQ), described in Li, Jiang and Fine (2013) <doi:10.1093/biostatistics/kxs047>. Polytomous Discrimination Index (PDI), described in Van Calster et al. (2012) <doi:10.1007/s10654-012-9733-3>. Li et al. (2018) <doi:10.1177/0962280217692830>. PDI with variance estimation using Dover et al. (2021) <doi:10.1002/sim.9187>. We described all these above measures and our mcca package in Li, Gao and D'Agostino (2019) <doi:10.1002/sim.8103>.

Setup, run and analyze NetLogo (<https://www.netlogo.org>) model simulations in R'. nlrx experiments use a similar structure as NetLogos Behavior Space experiments. However, nlrx offers more flexibility and additional tools for running and analyzing complex simulation designs and sensitivity analyses. The user defines all information that is needed in an intuitive framework, using class objects. Experiments are submitted from R to NetLogo via XML files that are dynamically written, based on specifications defined by the user. By nesting model calls in future environments, large simulation design with many runs can be executed in parallel. This also enables simulating NetLogo experiments on remote high performance computing machines. In order to use this package, Java and NetLogo (>= 5.3.1) need to be available on the executing system.

Most price indexes are made with a two-step procedure, where period-over-period elementary indexes are first calculated for a collection of elementary aggregates at each point in time, and then aggregated according to a price index aggregation structure. These indexes can then be chained together to form a time series that gives the evolution of prices with respect to a fixed base period. This package contains a collection of functions that revolve around this work flow, making it easy to build standard price indexes, and implement the methods described by Balk (2008, <doi:10.1017/CBO9780511720758>), von der Lippe (2007, <doi:10.3726/978-3-653-01120-3>), and the CPI manual (2020, <doi:10.5089/9781484354841.069>) and PPI manual (2004, <doi:10.5089/9781589063044.069>) for bilateral price indexes.

An increasing number of microbiome datasets have been generated and analyzed with the help of rapidly developing sequencing technologies. At present, analysis of taxonomic profiling data is mainly conducted using composition-based methods, which ignores interactions between community members. Besides this, a lack of efficient ways to compare microbial interaction networks limited the study of community dynamics. To better understand how community diversity is affected by complex interactions between its members, we developed a framework (Microbial community dIversity and Network Analysis, mina), a comprehensive framework for microbial community diversity analysis and network comparison. By defining and integrating network-derived community features, we greatly reduce noise-to-signal ratio for diversity analyses. A bootstrap and permutation-based method was implemented to assess community network dissimilarities and extract discriminative features in a statistically principled way.

This package provides a suite of tools to mitigate sampling bias in species occurrence records by thinning data in the environmental space (E-space). This process can improve the accuracy and precision of species distribution models (SDM, also known as ecological niche models, ENM). The package offers a data-driven protocol to determine thinning parameters using kernel-density bandwidth selection. Two thinning methods are provided (stochastic and deterministic) to reduce over-sampled environmental conditions and down-weight outlier observations. The name bean reflects the core principle of the method: each pod (a grid cell in E-space) is allowed to contain only a limited number of beans (occurrence points). See Silverman (1986, ISBN:978-0-412-24620-3) and Rousseeuw and Leroy (2003, ISBN:978-0-471-48855-2) for the underlying statistical methods.

Feasible generalized least squares (FGLS) estimators for linear regressions whose errors are serially correlated, heteroskedastic of unknown form, or both. Serial correlation is handled by the autoregressive filter of Perron and González-Coya (2026) <doi:10.1016/j.jeconom.2026.106346>, estimated from a Durbin regression whose order is chosen by an information criterion, with a variance estimator that stays valid when the regressors are only pre-determined rather than strictly exogenous. Heteroskedasticity is handled by the adaptive Lasso estimate of the skedastic function of González-Coya and Perron (2024) <doi:10.1515/jem-2023-0007>, with a bias-corrected sandwich variance. Both corrections can be combined. Benchmark estimators (ordinary least squares with heteroskedasticity and autocorrelation consistent standard errors, Cochrane-Orcutt, infeasible GLS) and the simulation designs of both papers are included.

Converts alphanumeric codes such as postcodes to whole numbers and back again, following the naxp standard <https://naxp.org/spec/>. A naxp ('encoded ASCII expression') is a pattern, defined using a similar syntax to regular expressions, that maps each text it accepts to a number from 1 to a fixed maximum, and any other text to 0. The same pattern maps text to numbers identically across all naxp libraries in all languages on all platforms (subject only to the availability of types able to store large enough integers). Standard naxps are defined for common code systems and are built into all naxp libraries. Two naxps can be compared to determine, among other things, whether stored values keep their meaning when one replaces the other. Wraps the naxp C++ library <https://naxp.org/api/#cpp>.

Reads the versioned exchange files written by the Zotero plugins zotQDA and qdaZ -- coded fragments, code systems, coding histories and team-consensus results -- validates them against the shipped contract, and reproduces the plugin's graphics with ggplot2'. Adds what those plugins deliberately leave out: six agreement coefficients with bootstrap confidence intervals, the reliability of the segmentation itself, chi-squared tests of code by group tables with effect sizes, correspondence analysis, multidimensional scaling and hierarchical clustering of codes. Projects from other programs can be read through the REFI-QDA interchange standard <https://www.qdasoftware.org/>, which makes those analyses available to users of established software that does not offer them; the subset a .qdpx supports is reported on import. Reference files are included, so every function can be tried without a Zotero installation.

An end-to-end implementation of the semi-structural quarterly projection models used in central-bank forecasting and policy analysis systems: model declaration with model-consistent expectations, a generalized Schur solver with Blanchard-Kahn diagnostics following Klein (2000) <doi:10.1016/S0165-1889(99)00045-7>, Kalman filtering and smoothing for latent states such as the output gap and the neutral rate, historical shock decompositions, conditional forecasts that distinguish announced from unanticipated policy paths, an auditable judgment ledger, forecast rounds with revision decompositions, Bayesian estimation with identification diagnostics following Iskrev (2010) <doi:10.1016/j.jmoneco.2009.12.007>, and reporting. The canonical small open economy model of Berg, Karam and Laxton (2006) <doi:10.5089/9781451863413.001> ships as a calibrated template, with extension blocks for disaggregated food inflation and managed exchange rates.

Subgroup classification is a basic task in genomic data analysis, especially for gene expression and DNA methylation data analysis. It can also be used to test the agreement to known clinical annotations, or to test whether there exist significant batch effects. The cola package provides a general framework for subgroup classification by consensus partitioning. It has the following features: 1. It modularizes the consensus partitioning processes that various methods can be easily integrated. 2. It provides rich visualizations for interpreting the results. 3. It allows running multiple methods at the same time and provides functionalities to straightforward compare results. 4. It provides a new method to extract features which are more efficient to separate subgroups. 5. It automatically generates detailed reports for the complete analysis. 6. It allows applying consensus partitioning in a hierarchical manner.