Translate R control flow expressions into Tensorflow graphs.

Create structured, formatted HTML tables of in a flexible and convenient way.

Unit testing is a solid component of automated CI/CD pipelines. tinytest - a lightweight, zero-dependency alternative to testthat was developed. To be able to integrate tinytests results into common CI/CD systems the test results from tinytest need to be caputred and converted to JUnit XML format. tinytest2JUnit enables this conversion while staying also lightweight and only have tinytest as its dependency.

This package implements a probabilistic ensemble time-series forecaster that combines an auto-encoder with a neural decision forest whose split variables are learned through a differentiable feature-mask layer. Functions are written with torch tensors and provide CRPS (Continuous Ranked Probability Scores) training plus mixture-distribution post-processing.

Parsing (R)Markdown files with numerous regular expressions can be fraught with peril, but it does not have to be this way. Converting (R)Markdown files to XML using the commonmark package allows in-memory editing via of markdown elements via XPath through the extensible R6 class called yarn'. These modified XML representations can be written to (R)Markdown documents via an xslt stylesheet which implements an extended version of GitHub'-flavoured markdown so that you can tinker to your hearts content.

High-performance parsing of Tableau workbook files into tidy data frames and dependency graphs for other visualization tools like R Shiny or Power BI replication.

Several statistical test functions as well as a function for exploratory data analysis to investigate classifiers allocating individuals to one of three disjoint and ordered classes. In a single classifier assessment the discriminatory power is compared to classification by chance. In a comparison of two classifiers the null hypothesis corresponds to equal discriminatory power of the two classifiers. See also "ROC Analysis for Classification and Prediction in Practice" by Nakas, Bantis and Gatsonis (2023), ISBN 9781482233704.

Targets parameters that solve Ordinary Differential Equations (ODEs) driven by a vector of cumulative hazard functions. The package provides a method for estimating these parameters using an estimator defined by a corresponding Stochastic Differential Equation (SDE) system driven by cumulative hazard estimates. By providing cumulative hazard estimates as input, the package gives estimates of the parameter as output, along with pointwise (co)variances derived from an asymptotic expression. Examples of parameters that can be targeted in this way include the survival function, the restricted mean survival function, cumulative incidence functions, among others; see Ryalen, Stensrud, and Røysland (2018) <doi:10.1093/biomet/asy035>, and further applications in Stensrud, Røysland, and Ryalen (2019) <doi:10.1111/biom.13102> and Ryalen et al. (2021) <doi:10.1093/biostatistics/kxab009>.

Fits 2D and 3D geometric transformations via Stan probabilistic programming engine ( Stan Development Team (2021) <https://mc-stan.org>). Returns posterior distribution for individual parameters of the fitted distribution. Allows for computation of LOO and WAIC information criteria (Vehtari A, Gelman A, Gabry J (2017) <doi:10.1007/s11222-016-9696-4>) as well as Bayesian R-squared (Gelman A, Goodrich B, Gabry J, and Vehtari A (2018) <doi:10.1080/00031305.2018.1549100>).

This package provides tools for measuring similarity among documents and detecting passages which have been reused. Implements shingled n-gram, skip n-gram, and other tokenizers; similarity/dissimilarity functions; pairwise comparisons; minhash and locality sensitive hashing algorithms; and a version of the Smith-Waterman local alignment algorithm suitable for natural language.

Higher Criticism (HC) test between two frequency tables. Test is based on an adaptation of the Tukey-Donoho-Jin HC statistic to testing frequency tables described in Kipnis (2019) <arXiv:1911.01208>.

The best way to implement middle ware for shiny Applications. tower is designed to make implementing behavior on top of shiny easy with a layering model for incoming HTTP requests and server sessions. tower is a very minimal package with little overhead, it is mainly meant for other package developers to implement new behavior.

Built on top of the tibble package, tibbletime is an extension that allows for the creation of time aware tibbles. Some immediate advantages of this include: the ability to perform time-based subsetting on tibbles, quickly summarising and aggregating results by time periods, and creating columns that can be used as dplyr time-based groups.

Statistical extreme value modelling of threshold excesses, maxima and multivariate extremes. Univariate models for threshold excesses and maxima are the Generalised Pareto, and Generalised Extreme Value model respectively. These models may be fitted by using maximum (optionally penalised-)likelihood, or Bayesian estimation, and both classes of models may be fitted with covariates in any/all model parameters. Model diagnostics support the fitting process. Graphical output for visualising fitted models and return level estimates is provided. For serially dependent sequences, the intervals declustering algorithm of Ferro and Segers (2003) <doi:10.1111/1467-9868.00401> is provided, with diagnostic support to aid selection of threshold and declustering horizon. Multivariate modelling is performed via the conditional approach of Heffernan and Tawn (2004) <doi:10.1111/j.1467-9868.2004.02050.x>, with graphical tools for threshold selection and to diagnose estimation convergence.

Trusted Timestamps (tts) are created by incorporating a hash of a file or dataset into a transaction on the decentralized blockchain (Stellar network). The package makes use of a free service provided by <https://stellarapi.io>.

Social Relation Model (SRM) analyses for single or multiple round-robin groups are performed. These analyses are either based on one manifest variable, one latent construct measured by two manifest variables, two manifest variables and their bivariate relations, or two latent constructs each measured by two manifest variables. Within-group t-tests for variance components and covariances are provided for single groups. For multiple groups two types of significance tests are provided: between-groups t-tests (as in SOREMO) and enhanced standard errors based on Lashley and Bond (1997) <DOI:10.1037/1082-989X.2.3.278>. Handling for missing values is provided.

Visualize your Tidyverse data analysis pipelines via the Tidy Data Tutor'(<https://tidydatatutor.com/>) web application.

This package provides a simple type annotation for R that is usable in scripts, in the R console and in packages. It is intended as a convention to allow other packages to use the type information to provide error checking, automatic documentation or optimizations.

This package provides threshold sweep methods for Qualitative Comparative Analysis (QCA). Implements Condition Threshold Sweep-Single (CTS-S), Condition Threshold Sweep-Multiple (CTS-M), Outcome Threshold Sweep (OTS), and Dual Threshold Sweep (DTS) for systematic exploration of threshold calibration effects on crisp-set QCA results. These methods extend traditional robustness approaches by treating threshold variation as an exploratory tool for discovering causal structures. Built on top of the QCA package by Dusa (2019) <doi:10.1007/978-3-319-75668-4>, with function arguments following QCA conventions. Based on set-theoretic methods by Ragin (2008) <doi:10.7208/chicago/9780226702797.001.0001> and established robustness protocols by Rubinson et al. (2019) <doi:10.1177/00491241211036158>.

Two stage curvature identification with machine learning for causal inference in settings when instrumental variable regression is not suitable because of potentially invalid instrumental variables. Based on Guo and Buehlmann (2022) "Two Stage Curvature Identification with Machine Learning: Causal Inference with Possibly Invalid Instrumental Variables" <doi:10.48550/arXiv.2203.12808>. The vignette is available in Carl, Emmenegger, Bühlmann and Guo (2025) "TSCI: Two Stage Curvature Identification for Causal Inference with Invalid Instruments in R" <doi:10.18637/jss.v114.i07>.

Pure R implementation of Apache Thrift. This library doesn't require any code generation. To learn more about Thrift go to <https://thrift.apache.org>.

This package provides an extensible formula system to quickly and easily create production quality tables. The processing steps are a formula parser, statistical content generation from data as defined by formula, followed by rendering into a table. Each step of the processing is separate and user definable thus creating a set of composable building blocks for highly customizable table generation. A user is not limited by any of the choices of the package creator other than the formula grammar. For example, one could chose to add a different S3 rendering function and output a format not provided in the default package, or possibly one would rather have Gini coefficients for their statistical content in a resulting table. Routines to achieve New England Journal of Medicine style, Lancet style and Hmisc::summaryM() statistics are provided. The package contains rendering for HTML5, Rmarkdown and an indexing format for use in tracing and tracking are provided.

Fit, compare, and visualize Bayesian graphical vector autoregressive (GVAR) network models using Stan'. These models are commonly used in psychology to represent temporal and contemporaneous relationships between multiple variables in intensive longitudinal data. Fitted models can be compared with a test based on matrix norm differences of posterior point estimates to quantify the differences between two estimated networks. See also Siepe, Kloft & Heck (2024) <doi:10.31234/osf.io/uwfjc>.

Interface to TensorFlow Datasets, a high-level library for building complex input pipelines from simple, re-usable pieces. See <https://www.tensorflow.org/guide> for additional details.