Optimally Robust Estimation

R infrastructure for optimally robust estimation in general smoothly parameterized models using S4 classes and methods as described Kohl, M., Ruckdeschel, P., and Rieder, H. (2010), , and in Rieder, H., Kohl, M., and Ruckdeschel, P. (2008), .

Trust Region Optimization for Nonlinear Functions with Sparse Hessians

Trust region algorithm for nonlinear optimization. Efficient when the Hessian of the objective function is sparse (i.e., relatively few nonzero cross-partial derivatives). See Braun, M. (2014) .

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.

Tools for Modeling Bumblebee Colony Growth and Decline

Bumblebee colonies grow during worker production, then decline after switching to production of reproductive individuals (drones and gynes). This package provides tools for modeling and visualizing this pattern by identifying a switchpoint with a growth rate before and a decline rate after the switchpoint. The mathematical models fit by bumbl are described in Crone and Williams (2016) .

Determining the Best Number of Clusters in a Data Set

It provides 30 indexes for determining the optimal number of clusters in a data set and offers the best clustering scheme from different results to the user.

Machine Learning in R

Interface to a large number of classification and regression techniques, including machine-readable parameter descriptions. There is also an experimental extension for survival analysis, clustering and general, example-specific cost-sensitive learning. Generic resampling, including cross-validation, bootstrapping and subsampling. Hyperparameter tuning with modern optimization techniques, for single- and multi-objective problems. Filter and wrapper methods for feature selection. Extension of basic learners with additional operations common in machine learning, also allowing for easy nested resampling. Most operations can be parallelized.

Highly Optimized Protocol Buffer Serializers

Pure C++ implementations for reading and writing several common data formats based on Google protocol-buffers. Currently supports 'rexp.proto' for serialized R objects, 'geobuf.proto' for binary geojson, and 'mvt.proto' for vector tiles. This package uses the auto-generated C++ code by protobuf-compiler, hence the entire serialization is optimized at compile time. The 'RProtoBuf' package on the other hand uses the protobuf runtime library to provide a general- purpose toolkit for reading and writing arbitrary protocol-buffer data in R.

Tools for Matrix Algebra, Optimization and Inference

Matrix is an universal and sometimes primary object/unit in applied mathematics and statistics. We provide a number of algorithms for selected problems in optimization and statistical inference. For general exposition to the topic with focus on statistical context, see the book by Banerjee and Roy (2014, ISBN:9781420095388).

Declare and Diagnose Research Designs

Researchers can characterize and learn about the properties of research designs before implementation using `DeclareDesign`. Ex ante declaration and diagnosis of designs can help researchers clarify the strengths and limitations of their designs and to improve their properties, and can help readers evaluate a research strategy prior to implementation and without access to results. It can also make it easier for designs to be shared, replicated, and critiqued.

'lp_solve' Plugin for the 'R' Optimization Infrastructure

Enhances the 'R' Optimization Infrastructure ('ROI') package with the 'lp_solve' solver.