Large, Sparse Optimal Matching with Refined Covariate Balance

Tools for large, sparse optimal matching of treated units and control units in observational studies. Provisions are made for refined covariate balance constraints, which include fine and near-fine balance as special cases. Matches are optimal in the sense that they are computed as solutions to network optimization problems rather than greedy algorithms. See Pimentel, et al.(2015) and Pimentel (2016), Obs. Studies 2(1):4-23. The rrelaxiv package, which provides an alternative solver for the underlying network flow problems, carries an academic license and is not available on CRAN, but may be downloaded from Github at < https://github.com/josherrickson/rrelaxiv/>.

Bayesian Optimization of Hyperparameters

A Pure R implementation of Bayesian Global Optimization with Gaussian Processes.

Animal Network Toolkit Software

How animals interact and develop social relationships in face of sociodemographic and ecological pressures is of great interest. New methodologies, in particular Social Network Analysis (SNA), allow us to elucidate these types of questions. However, the different methodologies developed to that end and the speed at which they emerge make their use difficult. Moreover, the lack of communication between the different software developed to provide an answer to the same/different research questions is a source of confusion. The R package Animal Network Toolkit 'ANTs' was developed with the aim of implementing in one package the different social network analysis techniques currently used in the study of animal social networks. Hence, ANT is a toolkit for animal research allowing among other things to: 1) measure global, dyadic and nodal networks metrics; 2) perform data randomization: pre- and post-network (node and link permutations); 3) perform statistical permutation tests as correlation test (), t-test (), General Linear Model (), General Linear Mixed Model (), deletion simulation (), 'Matrix TauKr correlations' (). The package is partially coded in C++ using the R package 'Rcpp' for an optimal coding speed. The package gives researchers a workflow from the raw data to the achievement of statistical analyses, allowing for a multilevel approach (): from the individual's position and role within the network, to the identification of interaction patterns, and the study of the overall network properties. Furthermore, ANT also provides a guideline on the SNA techniques used: 1) from the appropriate randomization technique according to the data collected; 2) to the choice, the meaning, the limitations and advantages of the network metrics to apply, 3) and the type of statistical tests to run. The ANT project is multi-collaborative, aiming to provide access to advanced social network analysis techniques and to create new ones that meet researchers' needs in future versions. The ANT project is multi-collaborative, aiming to provide access to advanced social network analysis techniques and to create new ones that meet researchers' needs in future versions.

'ECOS' Plugin for the 'R' Optimization Infrastructure

Enhances the 'R' Optimization Infrastructure ('ROI') package with the Embedded Conic Solver ('ECOS') for solving conic optimization problems.

Evolutionary Learning of Globally Optimal Trees

Commonly used classification and regression tree methods like the CART algorithm are recursive partitioning methods that build the model in a forward stepwise search. Although this approach is known to be an efficient heuristic, the results of recursive tree methods are only locally optimal, as splits are chosen to maximize homogeneity at the next step only. An alternative way to search over the parameter space of trees is to use global optimization methods like evolutionary algorithms. The 'evtree' package implements an evolutionary algorithm for learning globally optimal classification and regression trees in R. CPU and memory-intensive tasks are fully computed in C++ while the 'partykit' package is leveraged to represent the resulting trees in R, providing unified infrastructure for summaries, visualizations, and predictions.

Functions for Optimal Non-Bipartite Matching

Perform non-bipartite matching and matched randomization. A "bipartite" matching utilizes two separate groups, e.g. smokers being matched to nonsmokers or cases being matched to controls. A "non-bipartite" matching creates mates from one big group, e.g. 100 hospitals being randomized for a two-arm cluster randomized trial or 5000 children who have been exposed to various levels of secondhand smoke and are being paired to form a greater exposure vs. lesser exposure comparison. At the core of a non-bipartite matching is a N x N distance matrix for N potential mates. The distance between two units expresses a measure of similarity or quality as mates (the lower the better). The 'gendistance()' and 'distancematrix()' functions assist in creating this. The 'nonbimatch()' function creates the matching that minimizes the total sum of distances between mates; hence, it is referred to as an "optimal" matching. The 'assign.grp()' function aids in performing a matched randomization. Note bipartite matching can be performed using the prevent option in 'gendistance()'.

Wrapper for the 'CVD Prevent' Application Programming Interface

Provides an R wrapper to the 'CVD Prevent' application programming interface (API). Users can make API requests through built-in R functions. The Cardiovascular Disease Prevention Audit (CVDPREVENT) is an England-wide primary care audit that automatically extracts routinely held GP health data. < https://bmchealthdocs.atlassian.net/wiki/spaces/CP/pages/317882369/CVDPREVENT+API+Documentation>.

Genetic Tools for Colony Management

Provides genetic tools for colony management and is a derivation of the work in Amanda Vinson and Michael J Raboin (2015) < https://pmc.ncbi.nlm.nih.gov/articles/PMC4671785/> "A Practical Approach for Designing Breeding Groups to Maximize Genetic Diversity in a Large Colony of Captive Rhesus Macaques ('Macaca' 'mulatto')". It provides a 'Shiny' application with an exposed API. The application supports five groups of functions: (1) Quality control of studbooks contained in text files or 'Excel' workbooks and of pedigrees within 'LabKey' Electronic Health Records (EHR); (2) Creation of pedigrees from a list of animals using the 'LabKey' EHR integration; (3) Creation and display of an age by sex pyramid plot of the living animals within the designated pedigree; (4) Generation of genetic value analysis reports; and (5) Creation of potential breeding groups with and without proscribed sex ratios and defined maximum kinships.

Fast Fixed-Effects Estimations

Fast and user-friendly estimation of econometric models with multiple fixed-effects. Includes ordinary least squares (OLS), generalized linear models (GLM) and the negative binomial. The core of the package is based on optimized parallel C++ code, scaling especially well for large data sets. The method to obtain the fixed-effects coefficients is based on Berge (2018) < https://github.com/lrberge/fixest/blob/master/_DOCS/FENmlm_paper.pdf>. Further provides tools to export and view the results of several estimations with intuitive design to cluster the standard-errors.

Model-Based Boosting

Functional gradient descent algorithm (boosting) for optimizing general risk functions utilizing component-wise (penalised) least squares estimates or regression trees as base-learners for fitting generalized linear, additive and interaction models to potentially high-dimensional data. Models and algorithms are described in , a hands-on tutorial is available from . The package allows user-specified loss functions and base-learners.