Builds and optimizes Hopfield artificial neural networks (Hopfield, 1982, <doi:10.1073/pnas.79.8.2554>). One-layer and three-layer models are implemented. The energy of the Hopfield network is minimized with formula from Krotov and Hopfield (2016, <doi:10.48550/ARXIV.1606.01164>). Optimization (supervised learning) is done through a gradient-based method. Classification is done with S3 methods predict(). Parallelization with OpenMP is used if available during compilation.

Mediation analysis is used to identify and quantify intermediate effects from factors that intervene the observed relationship between an exposure/predicting variable and an outcome. We use a Bayesian adaptive lasso method to take care of the hierarchical structures and high dimensional exposures or mediators.

We provide functions for identifying the core community phylogeny in any microbiome, drawing phylogenetic Venn diagrams, calculating the core Faithâ s PD for a set of communities, and calculating the core UniFrac distance between two sets of communities. All functions rely on construction of a core community phylogeny, which is a phylogeny where branches are defined based on their presence in multiple samples from a single type of habitat. Our package provides two options for constructing the core community phylogeny, a tip-based approach, where the core community phylogeny is identified based on incidence of leaf nodes and a branch-based approach, where the core community phylogeny is identified based on incidence of individual branches. We suggest use of the microViz package.

Implementation of selected high-dimensional statistical and econometric methods for estimation and inference. Efficient estimators and uniformly valid confidence intervals for various low-dimensional causal/ structural parameters are provided which appear in high-dimensional approximately sparse models. Including functions for fitting heteroscedastic robust Lasso regressions with non-Gaussian errors and for instrumental variable (IV) and treatment effect estimation in a high-dimensional setting. Moreover, the methods enable valid post-selection inference and rely on a theoretically grounded, data-driven choice of the penalty. Chernozhukov, Hansen, Spindler (2016) <arXiv:1603.01700>.

This package implements an efficient algorithm for fitting the entire regularization path of quantile regression models with elastic-net penalties using a generalized coordinate descent scheme. The framework also supports SCAD and MCP penalties. It is designed for high-dimensional datasets and emphasizes numerical accuracy and computational efficiency. This package implements the algorithms proposed in Tang, Q., Zhang, Y., & Wang, B. (2022) <https://openreview.net/pdf?id=RvwMTDYTOb>.

Homomorphic encryption (Brakerski and Vaikuntanathan (2014) <doi:10.1137/120868669>) using Ring Learning with Errors (Lyubashevsky et al. (2012) <https://eprint.iacr.org/2012/230>) is a form of Learning with Errors (Regev (2005) <doi:10.1145/1060590.1060603>) using polynomial rings over finite fields. Functions to generate the required polynomials (using polynom'), with various distributions of coefficients are provided. Additionally, functions to generate and take coefficient modulo are provided.

Tests for a treatment effect using surrogate marker information accounting for heterogeneity in the utility of the surrogate. Details are described in Parast et al (2022) <arXiv:2209.08315>.

This package provides functions to access data from the US Department of Housing and Urban Development <https://www.huduser.gov/portal/dataset/fmr-api.html>.

Penalized regression for generalized linear models for measurement error problems (aka. errors-in-variables). The package contains a version of the lasso (L1-penalization) which corrects for measurement error (Sorensen et al. (2015) <doi:10.5705/ss.2013.180>). It also contains an implementation of the Generalized Matrix Uncertainty Selector, which is a version the (Generalized) Dantzig Selector for the case of measurement error (Sorensen et al. (2018) <doi:10.1080/10618600.2018.1425626>).

Processing, analysis and visualization of Hydrogen Deuterium eXchange monitored by Mass Spectrometry experiments (HDX-MS). HaDeX2 introduces a new standardized and reproducible workflow for the analysis of the HDX-MS data, including uncertainty propagation, data aggregation and visualization on 3D structure. Additionally, it covers data exploration, quality control and generation of publication-quality figures. All functionalities are also available in the accompanying shiny app.

This package provides a dependency free interface to the H3 geospatial indexing system utilizing the Rust library h3o <https://github.com/HydroniumLabs/h3o> via the extendr library <https://github.com/extendr/extendr>.

R interface for H2O', the scalable open source machine learning platform that offers parallelized implementations of many supervised and unsupervised machine learning algorithms such as Generalized Linear Models (GLM), Gradient Boosting Machines (including XGBoost), Random Forests, Deep Neural Networks (Deep Learning), Stacked Ensembles, Naive Bayes, Generalized Additive Models (GAM), ANOVA GLM, Cox Proportional Hazards, K-Means, PCA, ModelSelection, Word2Vec, as well as a fully automatic machine learning algorithm (H2O AutoML).

Collection of functions to help retrieving data from Hub'Eau the free and public French National APIs on water <https://hubeau.eaufrance.fr/>.

This package provides a stand-alone function that generates a user specified number of random datasets and computes eigenvalues using the random datasets (i.e., implements Horn's [1965, Psychometrika] parallel analysis <doi:10.1007/BF02289447>). Users then compare the resulting eigenvalues (the mean or the specified percentile) from the random datasets (i.e., eigenvalues resulting from noise) to the eigenvalues generated with the user's data. Can be used for both principal components analysis (PCA) and common/exploratory factor analysis (EFA). The output table shows how large eigenvalues can be as a result of merely using randomly generated datasets. If the user's own dataset has actual eigenvalues greater than the corresponding eigenvalues, that lends support to retain that factor/component. In other words, if the i(th) eigenvalue from the actual data was larger than the percentile of the (i)th eigenvalue generated using randomly generated data, empirical support is provided to retain that factor/component. Horn, J. (1965). A rationale and test for the number of factors in factor analysis. Psychometrika, 32, 179-185.

Software for performing the reduction, exploratory and model selection phases of the procedure proposed by Cox, D.R. and Battey, H.S. (2017) <doi:10.1073/pnas.1703764114> for sparse regression when the number of potential explanatory variables far exceeds the sample size. The software supports linear regression, likelihood-based fitting of generalized linear regression models and the proportional hazards model fitted by partial likelihood.

Using Dirichlet-Multinomial distribution to provide several functions for formal hypothesis testing, power and sample size calculations for human microbiome experiments.

This package provides a collection of utilities that support creation of network attributes for hydrologic networks. Methods and algorithms implemented are documented in Moore et al. (2019) <doi:10.3133/ofr20191096>), Cormen and Leiserson (2022) <ISBN:9780262046305> and Verdin and Verdin (1999) <doi:10.1016/S0022-1694(99)00011-6>.

An R port of the hashids library. hashids generates YouTube-like hashes from integers or vector of integers. Hashes generated from integers are relatively short, unique and non-seqential. hashids can be used to generate unique ids for URLs and hide database row numbers from the user. By default hashids will avoid generating common English cursewords by preventing certain letters being next to each other. hashids are not one-way: it is easy to encode an integer to a hashid and decode a hashid back into an integer.

The hydReng package provides a set of functions for hydraulic engineering tasks and natural hazard assessments. It includes basic hydraulics (wetted area, wetted perimeter, flow, flow velocity, flow depth, and maximum flow) for open channels with arbitrary geometry under uniform flow conditions. For structures such as circular pipes, weirs, and gates, the package includes calculations for pressure flow, backwater depth, and overflow over a weir crest. Additionally, it provides formulas for calculating bedload transport. The formulas used can be found in standard literature on hydraulics, such as Bollrich (2019, ISBN:978-3-410-29169-5) or Hager (2011, ISBN:978-3-642-77430-0).

Predict hatch and emergence timing for a wide range of wild fishes using the effective value framework (Sparks et al., (2019) <DOI:10.1139/cjfas-2017-0468>). hatchR offers users access to established phenological models and the flexibility to incorporate custom parameterizations using external datasets.

This package implements the Hierarchical Incremental GRAdient Descent (HiGrad) algorithm, a first-order algorithm for finding the minimizer of a function in online learning just like stochastic gradient descent (SGD). In addition, this method attaches a confidence interval to assess the uncertainty of its predictions. See Su and Zhu (2018) <arXiv:1802.04876> for details.

An implementation of Random Forest-based two-sample tests as introduced in Hediger & Michel & Naef (2022).

Add, share and manage annotations for Shiny applications and R Markdown documents via hypothes.is'.

Functions, Shiny apps and data for the book "Introduction to Statistics" by Wolfgang Karl Härdle, Sigbert Klinke, and Bernd Rönz (2015) <doi:10.1007/978-3-319-17704-5>.