Solves quadratic programming problems where the Hessian is represented as the product of two matrices. Thanks to Greg Hunt for helping getting this version back on CRAN. The methods in this package are described in: Ormerod, Wand and Koch (2008) "Penalised spline support vector classifiers: computational issues" <doi:10.1007/s00180-007-0102-8>.

This package provides functions to fit log-multiplicative models using gnm', with support for convenient printing, plots, and jackknife/bootstrap standard errors. For complex survey data, models can be fitted from design objects from the survey package. Currently supported models include UNIDIFF (Erikson & Goldthorpe, 1992), a.k.a. log-multiplicative layer effect model (Xie, 1992) <doi:10.2307/2096242>, and several association models: Goodman (1979) <doi:10.2307/2286971> row-column association models of the RC(M) and RC(M)-L families with one or several dimensions; two skew-symmetric association models proposed by Yamaguchi (1990) <doi:10.2307/271086> and by van der Heijden & Mooijaart (1995) <doi:10.1177/0049124195024001002> Functions allow computing the intrinsic association coefficient (see Bouchet-Valat (2022) <doi:10.1177/0049124119852389>) and the Altham (1970) index <doi:10.1111/j.2517-6161.1970.tb00816.x>, including via the Bayes shrinkage estimator proposed by Zhou (2015) <doi:10.1177/0081175015570097>; and the RAS/IPF/Deming-Stephan algorithm.

This package provides a set of tools designed to enhance transparency and understanding of date-time manipulation functions from the lubridate package. It provides detailed feedback about the operations performed by lubridate functions, allowing users to better comprehend and debug their code. These insights serve as both a learning tool for newcomers and a debugging aid for programmers working with date-time data.

Conducts various numerical analyses and simulations in population genetics and evolutionary theory, primarily for the purpose of teaching (and learning about) key concepts in population & quantitative genetics, and evolutionary theory.

Lights Out is a puzzle game consisting of a grid of lights that are either on or off. Pressing any light will toggle it and its adjacent lights. The goal of the game is to switch all the lights off. This package provides an interface to play the game on different board sizes, both through the command line or with a visual application. Puzzles can also be solved using the automatic solver included. View a demo online at <https://daattali.com/shiny/lightsout/>.

Fits semi-confirmatory structural equation modeling (SEM) via penalized likelihood (PL) or penalized least squares (PLS). For details, please see Huang (2020) <doi:10.18637/jss.v093.i07>.

Gives users seeking federal lobbying disclosures an easier way to query the API maintained by the Senate federal lobbying disclosures database <https://lda.gov/api/redoc/v1/> to find out how much companies and other entities are spending to lobby Congress and the federal government. It allows for search terms such as keywords, time periods and entity names. It then attempts to clean, or at least flag, filings that could provide incorrect results when seeking to answer the question: How much is being spent on lobbying our Congress and the administration and what issues do they care about?

Analyze graph/network data using L1 centrality and prestige. Functions for deriving global, local, and group L1 centrality/prestige are provided. Routines for visual inspection of a graph/network are also provided. Details are in Kang and Oh (2026a) <doi:10.1080/01621459.2025.2520467>, Kang and Oh (2026b) <doi:10.1080/00031305.2025.2563730>, and Kang (2025) <doi:10.23170/snu.000000188358.11032.0001856>.

Allows identification of palettes derived from LTER (Long Term Ecological Research) photographs based on user criteria. Also facilitates extraction of palettes from users photos directly.

This package provides a set of all-cause and cause-specific life expectancy sensitivity and decomposition methods, including Arriaga (1984) <doi:10.2307/2061029>, others documented by Ponnapalli (2005) <doi:10.4054/DemRes.2005.12.7>, lifetable, numerical, and other analytic or algorithmic approaches such as Horiuchi et al (2008) <doi:10.1353/dem.0.0033>, or Andreev et al (2002) <doi:10.4054/DemRes.2002.7.14>.

The Gaussian location-scale regression model is a multi-predictor model with explanatory variables for the mean (= location) and the standard deviation (= scale) of a response variable. This package implements maximum likelihood and Markov chain Monte Carlo (MCMC) inference (using algorithms from Girolami and Calderhead (2011) <doi:10.1111/j.1467-9868.2010.00765.x> and Nesterov (2009) <doi:10.1007/s10107-007-0149-x>), a parametric bootstrap algorithm, and diagnostic plots for the model class.

This package provides classes and methods for objects, whose indexing naturally starts from zero. Subsetting, indexing and mathematical operations are defined naturally between lagged objects and lagged and base R objects. Recycling is not used, except for singletons. The single bracket operator doesn't drop dimensions by default.

This package creates HTML strings to embed tables, images or graphs in pop-ups of interactive maps created with packages like leaflet or mapview'. Handles local images located on the file system or via remote URL. Handles graphs created with lattice or ggplot2 as well as interactive plots created with htmlwidgets'.

This package performs Levins loop analysis of qualitatively-specified complex causal systems. Loop analysis makes qualitative predictions of variable change in a system of causally interdependent variables, where "qualitative" means direct causal relationships and indirect causal effects are coded as sign only (i.e. increases, decreases, no change, and ambiguous). This implementation includes output support for graphs in .dot file format for use with visualization software such as graphviz (<https://graphviz.org>). LoopAnalyst provides tools for the construction and output of community matrices, computation and output of community effect matrices, tables of correlations, adjoint, absolute feedback, weighted feedback and weighted prediction matrices, change in life expectancy matrices, and feedback, path and loop enumeration tools.

This package provides a collection of tools for the calculation of freewater metabolism from in situ time series of dissolved oxygen, water temperature, and, optionally, additional environmental variables. LakeMetabolizer implements 5 different metabolism models with diverse statistical underpinnings: bookkeeping, ordinary least squares, maximum likelihood, Kalman filter, and Bayesian. Each of these 5 metabolism models can be combined with 1 of 7 models for computing the coefficient of gas exchange across the airâ water interface (k). LakeMetabolizer also features a variety of supporting functions that compute conversions and implement calculations commonly applied to raw data prior to estimating metabolism (e.g., oxygen saturation and optical conversion models).

This package provides a modular end-to-end framework for dynamic risk prediction based on time-to-event and longitudinal data. This allows flexible specifications for the longitudinal and survival sub-models. The landmaRk package enables reproducible benchmarks of different model choices, including cross-validation to assess out-of-sample predictive performance. Methods are described in Velasco-Pardo, Constantine-Cooke, Lees and Vallejos (2026, manuscript under preparation) Landmarking with Latent Class Mixed Models for Dynamic Prediction of Time-to-event Data with Heterogeneous Biomarker Trajectories'.

Labels are a common construct in statistical software providing a human readable description of a variable. While variable names are succinct, quick to type, and follow a language's naming conventions, labels may be more illustrative and may use plain text and spaces. R does not provide native support for labels. Some packages, however, have made this feature available. Most notably, the Hmisc package provides labelling methods for a number of different object. Due to design decisions, these methods are not all exported, and so are unavailable for use in package development. The labelVector package supports labels for atomic vectors in a light-weight design that is suitable for use in other packages.

This package performs power and sample size calculation for non-proportional hazards model using the Fleming-Harrington family of weighted log-rank tests. The sequentially calculated log-rank test score statistics are assumed to have independent increments as characterized in Anastasios A. Tsiatis (1982) <doi:10.1080/01621459.1982.10477898>. The mean and variance of log-rank test score statistics are calculated based on Kaifeng Lu (2021) <doi:10.1002/pst.2069>. The boundary crossing probabilities are calculated using the recursive integration algorithm described in Christopher Jennison and Bruce W. Turnbull (2000, ISBN:0849303168). The package can also be used for continuous, binary, and count data. For continuous data, it can handle missing data through mixed-model for repeated measures (MMRM). In crossover designs, it can estimate direct treatment effects while accounting for carryover effects. For binary data, it can design Simon's 2-stage, modified toxicity probability-2 (mTPI-2), and Bayesian optimal interval (BOIN) trials. For count data, it can design group sequential trials for negative binomial endpoints with censoring. Additionally, it facilitates group sequential equivalence trials for all supported data types. Moreover, it can design adaptive group sequential trials for changes in sample size, error spending function, number and spacing or future looks. Finally, it offers various options for adjusted p-values, including graphical and gatekeeping procedures.

This package provides tools for performing Bayesian inference on epidemiological data to estimate the time-varying reproductive number and other related metrics. These methods were published in Li and White (2021) <doi:10.1371/journal.pcbi.1009210>. This package supports analyses based on aggregated case count data and individual line list data, facilitating enhanced surveillance and intervention planning for infectious diseases like COVID-19.

This package provides the lzrq() function for estimating logarithmic regression slopes in quantile regression models, permitting the outcome variable to take on non-positive values. lzrq() conducts regression after replacing non-positive values with a sufficiently negative value. If the fitted values of a quantile regression on this transformed outcome are all greater than the negative value, then results are displayed. The resulting coefficients can be meaningfully interpreted as logarithmic intensive-margin relationships between the outcome variable and the independent variables, even with non-positive values in the outcome variable. If the condition does not hold for the specified quantile, then the command iteratively makes the value larger and checks again. After ten iterations where the condition does not hold, the functions return an error and suppress results. This is an automated adaptation of the algorithm described by Liu & Kaplan (2025) <https://drive.google.com/file/d/1F3dnhm8MrlO5aRrGt48rBWAEaBqdCBH-/view> and implemented in the companion Stata command lzqreg', described in Fitzgerald et al. (2026) <doi:10.31222/osf.io/juda7_v1>.

Analysis of dichotomous, ordinal, and continuous response data using latent space item response model ('LSIRM'). Provides 1PL and 2PL LSIRM for binary response data as described in Jeon et al. (2021) <doi:10.1007/s11336-021-09762-5>, graded response models ('GRM') for ordinal data (De Carolis et al., 2025, <doi:10.1080/00273171.2025.2605678>), and extensions for continuous response data. Supports Bayesian model selection with spike-and-slab priors, adaptive MCMC algorithms, and methods for handling missing data under missing at random ('MAR') and missing completely at random ('MCAR') assumptions. Provides various diagnostic plots to inspect the latent space and summaries of estimated parameters.

Estimate the sufficient dimension reduction space using sparsed sliced inverse regression via Lasso (Lasso-SIR) introduced in Lin, Zhao, and Liu (2019) <doi:10.1080/01621459.2018.1520115>. The Lasso-SIR is consistent and achieve the optimal convergence rate under certain sparsity conditions for the multiple index models.

Computes comorbidity indices and combined frailty scores for multiple ICD coding systems, including ICD-10-CA, ICD-10-CM, and ICD-11. The package provides tools to preprocess episode data, map diagnosis codes to chronic categories, propagate conditions across episodes, and generate comorbidity and frailty measures. The methods implemented are original to this package and were developed by the authors for research applications; a manuscript describing the methodology is currently in preparation.

Hidden Markov Model (HMM) based on symmetric lambda distribution framework is implemented for the study of return time-series in the financial market. Major features in the S&P500 index, such as regime identification, volatility clustering, and anti-correlation between return and volatility, can be extracted from HMM cleanly. Univariate symmetric lambda distribution is essentially a location-scale family of exponential power distribution. Such distribution is suitable for describing highly leptokurtic time series obtained from the financial market. It provides a theoretically solid foundation to explore such data where the normal distribution is not adequate. The HMM implementation follows closely the book: "Hidden Markov Models for Time Series", by Zucchini, MacDonald, Langrock (2016).