Gaussian mixture graphical models include Bayesian networks and dynamic Bayesian networks (their temporal extension) whose local probability distributions are described by Gaussian mixture models. They are powerful tools for graphically and quantitatively representing nonlinear dependencies between continuous variables. This package provides a complete framework to create, manipulate, learn the structure and the parameters, and perform inference in these models. Most of the algorithms are described in the PhD thesis of Roos (2018) <https://tel.archives-ouvertes.fr/tel-01943718>.

Multidimensional nonparametric spatial (spatio-temporal) geostatistics. S3 classes and methods for multidimensional: linear binning, local polynomial kernel regression (spatial trend estimation), density and variogram estimation. Nonparametric methods for simultaneous inference on both spatial trend and variogram functions (for spatial processes). Nonparametric residual kriging (spatial prediction). For details on these methods see, for example, Fernandez-Casal and Francisco-Fernandez (2014) <doi:10.1007/s00477-013-0817-8> or Castillo-Paez et al. (2019) <doi:10.1016/j.csda.2019.01.017>.

Calculate the ratio of iron oxides, hematite and goethite, in soil using the diffuse reflectance technique. The Kubelka-Munk theory, second derivative analysis, and spectral region amplitudes related to hematite and goethite content are used for quantification (Torrent, J., & Barron, V. (2008) <doi:10.2136/sssabookser5.5.c13>). Additionally, the package calculates soil color in the visible spectrum using Munsell and RGB color spaces, based on color theory (Viscarra et al. (2006) <doi:10.1016/j.geoderma.2005.07.017>).

Many of the models encountered in applications of point process methods to the study of spatio-temporal phenomena are covered in stpp'. This package provides statistical tools for analyzing the global and local second-order properties of spatio-temporal point processes, including estimators of the space-time inhomogeneous K-function and pair correlation function. It also includes tools to get static and dynamic display of spatio-temporal point patterns. See Gabriel et al (2013) <doi:10.18637/jss.v053.i02>.

Computes the optimal sample size for various 2-group designs (e.g., when comparing the means of two groups assuming equal variances, unequal variances, or comparing proportions) when the aim is to maximize the rewards over the full decision procedure of a) running a trial (with the computed sample size), and b) subsequently administering the winning treatment to the remaining N-n units in the population. Sample sizes and expected rewards for standard t- and z- tests are also provided.

Stagewise techniques implemented with Generalized Estimating Equations to handle individual, group, bi-level, and interaction selection. Stagewise approaches start with an empty model and slowly build the model over several iterations, which yields a path of candidate models from which model selection can be performed. This slow brewing approach gives stagewise techniques a unique flexibility that allows simple incorporation of Generalized Estimating Equations; see Vaughan, G., Aseltine, R., Chen, K., Yan, J., (2017) <doi:10.1111/biom.12669> for details.

Simulation of reconstructed phylogenetic trees under tree-wide time-heterogeneous birth-death processes and estimation of diversification parameters under the same model. Speciation and extinction rates can be any function of time and mass-extinction events at specific times can be provided. Trees can be simulated either conditioned on the number of species, the time of the process, or both. Additionally, the likelihood equations are implemented for convenience and can be used for Maximum Likelihood (ML) estimation and Bayesian inference.

Penalized weighted least-squares estimate for variable selection on correlated multiply imputed data and penalized estimating equations for generalized linear models with multiple imputation. Reference: Li, Y., Yang, H., Yu, H., Huang, H., Shen, Y*. (2023) "Penalized estimating equations for generalized linear models with multiple imputation", <doi:10.1214/22-AOAS1721>. Li, Y., Yang, H., Yu, H., Huang, H., Shen, Y*. (2023) "Penalized weighted least-squares estimate for variable selection on correlated multiply imputed data", <doi:10.1093/jrsssc/qlad028>.

This is an implementation of design methods for binomial reliability demonstration tests (BRDTs) with failure count data. The acceptance decision uncertainty of BRDT has been quantified and the impacts of the uncertainty on related reliability assurance activities such as reliability growth (RG) and warranty services (WS) are evaluated. This package is associated with the work from the published paper "Optimal Binomial Reliability Demonstration Tests Design under Acceptance Decision Uncertainty" by Suiyao Chen et al. (2020) <doi:10.1080/08982112.2020.1757703>.

This package provides a self-contained set of methods to aid clinical trial safety investigators, statisticians and researchers, in the early detection of adverse events using groupings by body-system or system organ class. This work was supported by the Engineering and Physical Sciences Research Council (UK) (EPSRC) [award reference 1521741] and Frontier Science (Scotland) Ltd. The package title c212 is in reference to the original Engineering and Physical Sciences Research Council (UK) funded project which was named CASE 2/12.

This package provides a comprehensive framework in R for modeling and forecasting economic scenarios based on multi-level dynamic factor model. The package enables users to: (i) extract global and block-specific factors using a flexible multilevel factor structure; (ii) compute asymptotically valid confidence regions for the estimated factors, accounting for uncertainty in the factor loadings; (iii) estimate factor-augmented quantile regressions; (iv) recover full predictive densities from these quantile forecasts; and (v) estimate the density when the factors are stressed.

Neural network has potential in forestry modelling. This package is designed to create and assess Artificial Intelligence based Neural Networks with varying architectures for prediction of volume of forest trees using two input features: height and diameter at breast height, as they are the key factors in predicting volume, therefore development and validation of efficient volume prediction neural network model is necessary. This package has been developed using the algorithm of Tabassum et al. (2022) <doi:10.18805/ag.D-5555>.

Ing and Lai (2011) <doi:10.5705/ss.2010.081> proposed a high-dimensional model selection procedure that comprises three steps: orthogonal greedy algorithm (OGA), high-dimensional information criterion (HDIC), and Trim. The first two steps, OGA and HDIC, are used to sequentially select input variables and determine stopping rules, respectively. The third step, Trim, is used to delete irrelevant variables remaining in the second step. This package aims at fitting a high-dimensional linear regression model via OGA+HDIC+Trim.

This package provides a set of functions is provided for 1) the stratum lengths analysis along a chosen direction, 2) fast estimation of continuous lag spatial Markov chains model parameters and probability computing (also for large data sets), 3) transition probability maps and transiograms drawing, 4) simulation methods for categorical random fields. More details on the methodology are discussed in Sartore (2013) <doi:10.32614/RJ-2013-022> and Sartore et al. (2016) <doi:10.1016/j.cageo.2016.06.001>.

Fits the regularization path of regression models (linear and logistic) with additively combined penalty terms. All possible combinations with Least Absolute Shrinkage and Selection Operator (LASSO), Smoothly Clipped Absolute Deviation (SCAD), Minimax Concave Penalty (MCP) and Exponential Penalty (EP) are supported. This includes Sparse Group LASSO (SGL), Sparse Group SCAD (SGS), Sparse Group MCP (SGM) and Sparse Group EP (SGE). For more information, see Buch, G., Schulz, A., Schmidtmann, I., Strauch, K., & Wild, P. S. (2024) <doi:10.1002/bimj.202200334>.

Fit Bayesian hierarchical models of animal abundance and occurrence via the rstan package, the R interface to the Stan C++ library. Supported models include single-season occupancy, dynamic occupancy, and N-mixture abundance models. Covariates on model parameters are specified using a formula-based interface similar to package unmarked', while also allowing for estimation of random slope and intercept terms. References: Carpenter et al. (2017) <doi:10.18637/jss.v076.i01>; Fiske and Chandler (2011) <doi:10.18637/jss.v043.i10>.

This package provides support for transformations of numeric aggregates between statistical classifications (e.g. occupation or industry categorisations) using the Crossmaps framework. Implements classes for representing transformations between a source and target classification as graph structures, and methods for validating and applying crossmaps to transform data collected under the source classification into data indexed using the target classification codes. Documentation about the Crossmaps framework is provided in the included vignettes and in Huang (2024, <doi:10.48550/arXiv.2406.14163>).

This package provides an object-oriented modeling language for disciplined convex programming (DCP) as described in Fu, Narasimhan, and Boyd (2020, <doi:10.18637/jss.v094.i14>). It allows the user to formulate convex optimization problems in a natural way following mathematical convention and DCP rules. The system analyzes the problem, verifies its convexity, converts it into a canonical form, and hands it off to an appropriate solver to obtain the solution. Interfaces to solvers on CRAN and elsewhere are provided.

This package provides the ASUS procedure for estimating a high dimensional sparse parameter in the presence of auxiliary data that encode side information on sparsity. It is a robust data combination procedure in the sense that even when pooling non-informative auxiliary data ASUS would be at least as efficient as competing soft thresholding based methods that do not use auxiliary data. For more information, please see the paper Adaptive Sparse Estimation with Side Information by Banerjee, Mukherjee and Sun (JASA 2020).

This package provides a comprehensive approach for identifying and estimating change points in multivariate time series through various statistical methods. Implements the multiple change point detection methodology from Ryan & Killick (2023) <doi:10.1080/00401706.2023.2183261> and a novel estimation methodology from Fotopoulos et al. (2023) <doi:10.1007/s00362-023-01495-0> generalized to fit the detection methodologies. Performs both detection and estimation of change points, providing visualization and summary information of the estimation process for each detected change point.

This package provides a general, flexible framework for estimating parameters and empirical sandwich variance estimator from a set of unbiased estimating equations (i.e., M-estimation in the vein of Stefanski & Boos (2002) <doi:10.1198/000313002753631330>). All examples from Stefanski & Boos (2002) are published in the corresponding Journal of Statistical Software paper "The Calculus of M-Estimation in R with geex" by Saul & Hudgens (2020) <doi:10.18637/jss.v092.i02>. Also provides an API to compute finite-sample variance corrections.

We provide an R tool for computation and nonparametric plug-in estimation of Highest Density Regions (HDRs) and general level sets in the directional setting. Concretely, circular and spherical HDRs can be reconstructed from a data sample following Saavedra-Nieves and Crujeiras (2021) <doi:10.1007/s11634-021-00457-4>. This library also contains two real datasets in the circular and spherical settings. The first one concerns a problem from animal orientation studies and the second one is related to earthquakes occurrences.

The heterogeneous multi-task feature learning is a data integration method to conduct joint feature selection across multiple related data sets with different distributions. The algorithm can combine different types of learning tasks, including linear regression, Huber regression, adaptive Huber, and logistic regression. The modified version of Bayesian Information Criterion (BIC) is produced to measure the model performance. Package is based on Yuan Zhong, Wei Xu, and Xin Gao (2022) <https://www.fields.utoronto.ca/talk-media/1/53/65/slides.pdf>.

Rapid satellite data streams in operational applications have clear benefits for monitoring land cover, especially when information can be delivered as fast as changing surface conditions. Over the past decade, remote sensing has become a key tool for monitoring and predicting environmental variables by using satellite data. This package presents the main applications in remote sensing for land surface monitoring and land cover mapping (soil, vegetation, water...). Tomlinson, C.J., Chapman, L., Thornes, E., Baker, C (2011) <doi:10.1002/met.287>.