Nowadays, Neural Networks are considered one of the most effective methods for various tasks such as anomaly detection, computer-aided disease detection, or natural language processing. However, these networks suffer from the ``black-box'' problem which makes it difficult to understand how they make decisions. In order to solve this issue, an R package called neuralGAM is introduced. This package implements a Neural Network topology based on Generalized Additive Models, allowing to fit an independent Neural Network to estimate the contribution of each feature to the output variable, yielding a highly accurate and interpretable Deep Learning model. The neuralGAM package provides a flexible framework for training Generalized Additive Neural Networks, which does not impose any restrictions on the Neural Network architecture. We illustrate the use of the neuralGAM package in both synthetic and real data examples.
This exposition presents nimblewomble, a software package to perform wombling, or boundary analysis, using the nimble Bayesian hierarchical modeling language in the R statistical computing environment. Wombling is used widely to track regions of rapid change within the spatial reference domain. Specific functions in the package implement Gaussian process models for point-referenced spatial data followed by predictive inference on rates of change over curves using line integrals. We demonstrate model based Bayesian inference using posterior distributions featuring simple analytic forms while offering uncertainty quantification over curves.
The BayesPPDSurv (Bayesian Power Prior Design for Survival Data) R package supports Bayesian power and type I error calculations and model fitting using the power and normalized power priors incorporating historical data with for the analysis of time-to-event outcomes. The package implements the stratified proportional hazards regression model with piecewise constant hazard within each stratum. The package allows the historical data to inform the treatment effect parameter, parameter effects for other covariates in the regression model, as well as the baseline hazard parameters. The use of multiple historical datasets is supported. A novel algorithm is developed for computationally efficient use of the normalized power prior. In addition, the package supports the use of arbitrary sampling priors for computing Bayesian power and type I error rates, and has built-in features that semi-automatically generate sampling priors from the historical data. We demonstrate the use of BayesPPDSurv in a comprehensive case study for a melanoma clinical trial design.
The woylier package implements tour interpolation paths between frames using Givens rotations. This provides an alternative to the geodesic interpolation between planes currently available in the tourr package. Tours are used to visualise high-dimensional data and models, to detect clustering, anomalies and non-linear relationships. Frame-to-frame interpolation can be useful for projection pursuit guided tours when the index is not rotationally invariant. It also provides a way to specifically reach a given target frame. We demonstrate the method for exploring non-linear relationships between currency cross-rates.
The Neyman-Scott point process is a widely used point process model which is easily interpretable and easily extendable to include various types of inhomogeneity. The inference for such complex models is then complicated and fast methods, such as minimum contrast method or composite likelihood approach do not provide accurate estimates or fail completely. Therefore, we introduce Bayesian MCMC approach for the inference of Neymann-Scott point process models with inhomogeneity in any or all of the following model components: process of cluster centers, mean number of points in a cluster, spread of the clusters. We also extend the Neyman-Scott point process to the case of overdispersed or underdispersed cluster sizes and provide a Bayesian MCMC algorithm for its inference. The R package binspp provides these estimation methods in an easy to handle implementation, with detailed graphical output including traceplots for all model parameters and further diagnostic plots. All inhomogeneities are modelled by spatial covariates and the Bayesian inference for the corresponding regression parameters is provided.
The AcceptReject package, available for the R programming language on the Comprehensive R Archive Network (CRAN), versioned and maintained on GitHub, offers a simple and efficient solution for generating pseudo-random observations of discrete or continuous random variables using the acceptance-rejection method. This method provides a viable alternative for generating pseudo-random observations in univariate distributions when the inverse of the cumulative distribution function is not in closed form or when suitable transformations involving random variables that we know how to generate are unknown, thereby facilitating the generation of observations for the variable of interest. The package is designed to be simple, intuitive, and efficient, allowing for the rapid generation of observations and supporting multicore parallelism on Unix-based operating systems. Some components are written using C++, and the package maximizes the acceptance probability of the generated observations, resulting in even more efficient execution. The package also allows users to explore the generated pseudo-random observations by comparing them with the theoretical probability mass function or probability density function and to inspect the underlying probability density functions that can be used in the method for generating observations of continuous random variables. This article explores the package in detail, discussing its functionalities, benefits, and practical applications, and provides various benchmarks in several scenarios.
The MultiATSM package provides estimation tools and a wide range of outputs academics, and policymakers. All models extend the single-country framework of Joslin et and Candelon and Moura (2024). These model extensions incorporate, respectively, the presence of a dominant (global) economy and adopt a global vector autoregressive (GVAR) setup to capture the joint dynamics of risk factors. The package generates diverse outputs for each ATSM, including graphical representations of model fit, risk premia, impulse response functions, and forecast error variance decompositions. It also implements bootstrap methods for confidence intervals and produces bond yield forecasts.
Recent work in survey domain estimation has shown that incorporating a priori assumptions about orderings of population domain means reduces the variance of the estimators. The new R package csurvey allows users to implement order constraints using a design specified in the well-known survey package. The order constraints not only give estimates that satisfy a priori assumptions, but they also provide smaller confidence intervals with good coverage, and allow for design-based estimation in small-sample or empty domains. A test for constant versus increasing domain means is available in the package, with generalizations to other one-sided tests. A cone information criterion may be used to provide evidence that the order constraints are valid. Examples with well-known survey data sets show the utility of the methods. This package is now available from the Comprehensive R Archive Network at http://CRAN.R-project.org/package=csurvey.
Multicollinearity is relevant in many different fields where linear regression models are applied since its presence may affect the analysis of ordinary least squares estimators not only numerically but also from a statistical point of view, which is the focus of this paper. Thus, it is known that collinearity can lead to incoherence in the statistical significance of the coefficients of the independent variables and in the global significance of the model. In this paper, the thresholds of the Redefined Variance Inflation Factor (RVIF) are reinterpreted and presented as a statistical test with a region of non-rejection (which depends on a significance level) to diagnose the existence of a degree of worrying multicollinearity that affects the linear regression model from a statistical point of view. The proposed methodology is implemented in the rvif package of R and its application is illustrated with different real data examples previously applied in the scientific literature.
The development of methods to study intratumoral heterogeneity and tumor phylogenies is a highly active area of research. However, the advancement of these approaches often necessitates access to substantial amounts of data, which can be challenging and expensive to acquire. Moreover, the assessment of results requires tools for visualizing and comparing tumor phylogenies. In this paper, we introduce GeRnika, an R package designed to address these needs by enabling the simulation, visualization, and comparison of tumor evolution data. In summary, GeRnika provides researchers with a user-friendly tool that facilitates the analysis of their approaches aimed at studying tumor composition and evolutionary history.
Optimal Latin hypercube designs (LHDs), including maximin distance LHDs, maximum projection LHDs and orthogonal LHDs, are widely used in computer experiments. It is challenging to construct such designs with flexible sizes, especially for large ones, for two main reasons. One reason is that theoretical results, such as algebraic constructions ensuring the maximin distance property or orthogonality, are only available for certain design sizes. For design sizes where theoretical results are unavailable, search algorithms can generate designs. However, their numerical performance is not guaranteed to be optimal. Another reason is that when design sizes increase, the number of permutations grows exponentially. Constructing optimal LHDs is a discrete optimization process, and enumeration is nearly impossible for large or moderate design sizes. Various search algorithms and algebraic constructions have been proposed to identify optimal LHDs, each having its own pros and cons. We develop the R package LHD which implements various search algorithms and algebraic constructions. We embedded different optimality criteria into each of the search algorithms, and they are capable of constructing different types of optimal LHDs even though they were originally invented to construct maximin distance LHDs only. Another input argument that controls maximum CPU time is added to each of the search algorithms to let users flexibly allocate their computational resources. We demonstrate functionalities of the package by using various examples, and we provide guidance for experimenters on finding suitable optimal designs. The LHD package is easy to use for practitioners and possibly serves as a benchmark for future developments in LHD.
