The Statistical Learning Theory (SLT) provides the theoretical background to ensure that a supervised algorithm generalizes the mapping f:X -> Y given f is selected from its search space bias F. This formal result depends on the Shattering coefficient function N(F,2n) to upper bound the empirical risk minimization principle, from which one can estimate the necessary training sample size to ensure the probabilistic learning convergence and, most importantly, the characterization of the capacity of F, including its under and overfitting abilities while addressing specific target problems. In this context, we propose a new approach to estimate the maximal number of hyperplanes required to shatter a given sample, i.e., to separate every pair of points from one another, based on the recent contributions by Har-Peled and Jones in the dataset partitioning scenario, and use such foundation to analytically compute the Shattering coefficient function for both binary and multi-class problems. As main contributions, one can use our approach to study the complexity of the search space bias F, estimate training sample sizes, and parametrize the number of hyperplanes a learning algorithm needs to address some supervised task, what is specially appealing to deep neural networks. Reference: de Mello, R.F. (2019) "On the Shattering Coefficient of Supervised Learning Algorithms" <arXiv:1911.05461>; de Mello, R.F., Ponti, M.A. (2018, ISBN: 978-3319949888) "Machine Learning: A Practical Approach on the Statistical Learning Theory".

This package implements the SE-test for equivalence according to Hoffelder et al. (2015) <DOI:10.1080/10543406.2014.920344>. The SE-test for equivalence is a multivariate two-sample equivalence test. Distance measure of the test is the sum of standardized differences between the expected values or in other words: the sum of effect sizes (SE) of all components of the two multivariate samples. The test is an asymptotically valid test for normally distributed data (see Hoffelder et al.,2015). The function SE.EQ() implements the SE-test for equivalence according to Hoffelder et al. (2015). The function SE.EQ.dissolution.profiles() implements a variant of the SE-test for equivalence for similarity analyses of dissolution profiles as mentioned in Suarez-Sharp et al.(2020) <DOI:10.1208/s12248-020-00458-9>). The equivalence margin used in SE.EQ.dissolution.profiles() is analogically defined as for the T2EQ approach according to Hoffelder (2019) <DOI:10.1002/bimj.201700257>) by means of a systematic shift in location of 10 [\% of label claim] of both dissolution profile populations. SE.EQ.dissolution.profiles() checks whether the weighted mean of the differences of the expected values of both dissolution profile populations is statistically significantly smaller than 10 [\% of label claim]. The weights are built up by the inverse variances.

Calculates a Satorra-Bentler scaled chi-squared difference test between nested models that were estimated using maximum likelihood (ML) with robust standard errors, which cannot be calculated the traditional way. For details see Satorra & Bentler (2001) <doi:10.1007/bf02296192> and Satorra & Bentler (2010) <doi:10.1007/s11336-009-9135-y>. This package may be particularly helpful when used in conjunction with Mplus software, specifically when implementing the complex survey option. In such cases, the model estimator in Mplus defaults to ML with robust standard errors.

Procedure to optimally split a dataset for training and testing. SPlit is based on the method of support points, which is independent of modeling methods. Please see Joseph and Vakayil (2021) <doi:10.1080/00401706.2021.1921037> for details. This work is supported by U.S. National Science Foundation grant DMREF-1921873.

This package provides basic functions that support an implementation of multi-profile case (Case 3) best-worst scaling (BWS). Case 3 BWS is a question-based survey method to elicit people's preferences for attribute levels. Case 3 BWS constructs various combinations of attribute levels (profiles) and then asks respondents to select the best and worst profiles in each choice set. A main function creates a dataset for the analysis from the choice sets and the responses to the questions. For details on Case 3 BWS, refer to Louviere et al. (2015) <doi:10.1017/CBO9781107337855>.

The focus is on simulating and modeling families with founders drawn from a structured population (for example, with different ancestries or other potentially non-family relatedness), in contrast to traditional pedigree analysis that treats all founders as equally unrelated. Main function simulates a random pedigree for many generations, avoiding close relatives, pairing closest individuals according to a 1D geography and their randomly-drawn sex, and with variable children sizes to result in a target population size per generation. Auxiliary functions calculate kinship matrices, admixture matrices, and draw random genotypes across arbitrary pedigree structures starting from the corresponding founder values. The code is built around the plink FAM table format for pedigrees. Described in Yao and Ochoa (2022) <doi:10.1101/2022.03.25.485885>.

Specific and class specific multiple correspondence analysis on survey-like data. Soc.ca is optimized to the needs of the social scientist and presents easily interpretable results in near publication ready quality.

RNA sequencing analysis methods are often derived by relying on hypothetical parametric models for read counts that are not likely to be precisely satisfied in practice. Methods are often tested by analyzing data that have been simulated according to the assumed model. This testing strategy can result in an overly optimistic view of the performance of an RNA-seq analysis method. We develop a data-based simulation algorithm for RNA-seq data. The vector of read counts simulated for a given experimental unit has a joint distribution that closely matches the distribution of a source RNA-seq dataset provided by the user. Users control the proportion of genes simulated to be differentially expressed (DE) and can provide a vector of weights to control the distribution of effect sizes. The algorithm requires a matrix of RNA-seq read counts with large sample sizes in at least two treatment groups. Many datasets are available that fit this standard.

Visualization and analysis of spatially resolved transcriptomics data. The spatialGE R package provides methods for visualizing and analyzing spatially resolved transcriptomics data, such as 10X Visium, CosMx, or csv/tsv gene expression matrices. It includes tools for spatial interpolation, autocorrelation analysis, tissue domain detection, gene set enrichment, and differential expression analysis using spatial mixed models.

Fast and efficient sampling from general univariate probability density functions. Implements a rejection sampling approach designed to take advantage of modern CPU caches and minimise evaluation of the target density for most samples. Many standard densities are internally implemented in C for high performance, with general user defined densities also supported. A paper describing the methodology will be released soon.

Predictive margins like in the Stata procedure margins'. This package is based on the survey::svypredmean() function. Supported regression analyses are based on survey::svyglm() and svyVGAM::svy_vglm() functions (for multinomial logistic regression models).

Computes the Exposure-At-Default based on the standardized approach of CRR2 (SA-CCR). The simplified version of SA-CCR has been included, as well as the OEM methodology. Multiple trade types of all the five major asset classes are being supported including the Other Exposure and, given the inheritance- based structure of the application, the addition of further trade types is straightforward. The application returns a list of trees per Counterparty and CSA after automatically separating the trades based on the Counterparty, the CSAs, the hedging sets, the netting sets and the risk factors. The basis and volatility transactions are also identified and treated in specific hedging sets whereby the corresponding penalty factors are applied. All the examples appearing on the regulatory papers (both for the margined and the unmargined workflow) have been implemented including the latest CRR2 developments.

Spatial forecast verification refers to verifying weather forecasts when the verification set (forecast and observations) is on a spatial field, usually a high-resolution gridded spatial field. Most of the functions here require the forecast and observed fields to be gridded and on the same grid. For a thorough review of most of the methods in this package, please see Gilleland et al. (2009) <doi: 10.1175/2009WAF2222269.1> and for a tutorial on some of the main functions available here, see Gilleland (2022) <doi: 10.5065/4px3-5a05>.

This package provides a set of tools developed at Simularia for Simularia, to help preprocessing and post-processing of meteorological and air quality data.

Capture screenshots in Shiny applications. Screenshots can either be of the entire viewable page, or a specific section of the page. The captured image is automatically downloaded as a PNG image, or it can also be saved on the server. Powered by the html2canvas JavaScript library.

We introduce improved methods for statistically assessing birth seasonality and intra-annual variation. The first method we propose is a new idea that uses a nonparametric clustering procedure to group individuals with similar time series data and estimate birth seasonality based on the clusters. One can use the function SCEM() to implement this method. The second method estimates input parameters for use with a previously-developed parametric approach (Tornero et al., 2013). The relevant code for this approach is makeFits_OLS(), while makeFits_initial() is the code to implement the same method but with given initial conditions for two parameters. The latter can be used to show the disadvantage of the existing approach. One can use the function makeFits() to generate parametric birth seasonality estimates using either initialization. Detailed description can be found here: Chazin Hannah, Soudeep Deb, Joshua Falk, and Arun Srinivasan (2019) <doi:10.1111/arcm.12432> "New Statistical Approaches to Intra-Individual Isotopic Analysis and Modeling Birth Seasonality in Studies of Herd Animals".

Select the most suitable shape to describe the relationship between the exposure and the outcome among increasing, decreasing, convex, and concave shapes (Yin et al. (2021) <DOI:10.1007/s13571-020-00246-7>); estimate the direct and indirect effects with prior knowledge on the relationship between the mediator and the outcome with binary exposure (Yin et al. (2024) <DOI:10.1007/s13571-024-00336-w>); estimate the direct and indirect effects using linear regression-based approach (VanderWeele (2015, ISBN:9780199325870)).

Forms queries to submit to the Cleveland Federal Reserve Bank web site's financial stress index data site. Provides query functions for both the composite stress index and the components data. By default the download includes daily time series data starting September 25, 1991. The functions return a class of either type easing or cfsi which contain a list of items related to the query and its graphical presentation. The list includes the time series data as an xts object. The package provides four lattice time series plots to render the time series data in a manner similar to the bank's own presentation.

This package provides a comprehensive statistical toolbox for National Statistics Institutes (INS) in Africa. Provides functions for survey data import ('KoboToolbox', ODK', CSPro', Excel', Stata', SPSS'), data processing and validation, weighted statistical analysis (descriptive statistics, cross-tabulations, regression, Human Development Index (HDI), Multidimensional Poverty Index (MPI) following Alkire and Foster (2011) <doi:10.1093/oep/gpr051>, inequalities), visualization (age pyramids, thematic maps, official charts) and dissemination ('SDMX export, DDI metadata, anonymization, Word/PDF reports). Designed to work in resource-constrained environments, offline and in French.

This package provides a general-purpose implementation of synthetic control methods that accounts for potential spillover effects between units. Based on the methodology of Cao and Dowd (2019).

This package provides a system that computes metrics to assess the segmentation accuracy of geospatial data. These metrics calculate the discrepancy between segmented and reference objects, and indicate the segmentation accuracy. For more details on choosing evaluation metrics, we suggest seeing Costa et al. (2018) <doi:10.1016/j.rse.2017.11.024> and Jozdani et al. (2020) <doi:10.1016/j.isprsjprs.2020.01.002>.

Loads and manages statistical genetics data objects including reference panels, genotypes, LD matrices, annotations, and summary statistics. Follows the statgen specification for use across Python', R', and MATLAB'/'Octave runtimes.

Implementation of a shiny app to easily compare supervised machine learning model performances. You provide the data and configure each model parameter directly on the shiny app. Different supervised learning algorithms can be tested either on Spark or H2O frameworks to suit your regression and classification tasks. Implementation of available machine learning models on R has been done by Lantz (2013, ISBN:9781782162148).

Evaluating probabilistic forecasts via proper scoring rules. scoring implements the beta, power, and pseudospherical families of proper scoring rules, along with ordered versions of the latter two families. Included among these families are popular rules like the Brier (quadratic) score, logarithmic score, and spherical score. For two-alternative forecasts, also includes functionality for plotting scores that one would obtain under specific scoring rules.