Predictive (Classification and Regression) Models Homologator

Methods to unify the different ways of creating predictive models and their different predictive formats for classification and regression. It includes methods such as K-Nearest Neighbors Schliep, K. P. (2004) , Decision Trees Leo Breiman, Jerome H. Friedman, Richard A. Olshen, Charles J. Stone (2017) , ADA Boosting Esteban Alfaro, Matias Gamez, Noelia García (2013) , Extreme Gradient Boosting Chen & Guestrin (2016) , Random Forest Breiman (2001) , Neural Networks Venables, W. N., & Ripley, B. D. (2002) , Support Vector Machines Bennett, K. P. & Campbell, C. (2000) , Bayesian Methods Gelman, A., Carlin, J. B., Stern, H. S., & Rubin, D. B. (1995) , Linear Discriminant Analysis Venables, W. N., & Ripley, B. D. (2002) , Quadratic Discriminant Analysis Venables, W. N., & Ripley, B. D. (2002) , Logistic Regression Dobson, A. J., & Barnett, A. G. (2018) and Penalized Logistic Regression Friedman, J. H., Hastie, T., & Tibshirani, R. (2010) .

Tidy Tools for 'Raster' Data

Facilities to work with vector and raster data in efficient repeatable and systematic work flow. Missing functionality in existing packages is included here to allow extraction from raster data with 'simple features' and 'Spatial' types and to make extraction consistent and straightforward. Extract cell numbers from raster data and return the cells as a data frame rather than as lists of matrices or vectors. The functions here allow spatial data to be used without special handling for the format currently in use.

Tools for Standardizing Variables for Regression in R

Tools which allow regression variables to be placed on similar scales, offering computational benefits as well as easing interpretation of regression output.

Quadrangle Mesh

Create surface forms from matrix or 'raster' data for flexible plotting and conversion to other mesh types. The functions 'quadmesh' or 'triangmesh' produce a continuous surface as a 'mesh3d' object as used by the 'rgl' package. This is used for plotting raster data in 3D (optionally with texture), and allows the application of a map projection without data loss and many processing applications that are restricted by inflexible regular grid rasters. There are discrete forms of these continuous surfaces available with 'dquadmesh' and 'dtriangmesh' functions.

Spatial Extent Tools

Functions to produce a fully fledged 'geo-spatial' object extent as a 'SpatialPolygonsDataFrame'. Also included are functions to generate polygons from raster data using 'quadmesh' techniques, a round number buffered extent, and general spatial-extent and 'raster-like' extent helpers missing from the originating packages. Some latitude-based tools for polar maps are included.

Raster Georeferencing, Grid Affine Transforms, Cell Abstraction

Tools for raster georeferencing, grid affine transforms, and general raster logic. These functions provide converters between raster specifications, world vector, geotransform, 'RasterIO' window, and 'RasterIO window' in 'sf' package list format. There are functions to offset a matrix by padding any of four corners (useful for vectorizing neighbourhood operations), and helper functions to harvesting user clicks on a graphics device to use for simple georeferencing of images. Methods used are available from < https://en.wikipedia.org/wiki/World_file> and < https://gdal.org/user/raster_data_model.html>.

Constrained Triangulation for Simple Features

Build a constrained high quality Delaunay triangulation from simple features objects, applying constraints based on input line segments, and triangle properties including maximum area, minimum internal angle. The triangulation code in 'RTriangle' uses the method of Cheng, Dey and Shewchuk (2012, ISBN:9781584887300). For a low-dependency alternative with low-quality path-based constrained triangulation see < https://CRAN.R-project.org/package=decido> and for high-quality configurable triangulation see < https://github.com/hypertidy/anglr>. Also consider comparison with the 'GEOS' lib which since version 3.10.0 includes a low quality polygon triangulation method that starts with ear clipping and refines to Delaunay.

Convolution-Based Nonstationary Spatial Modeling

Fits convolution-based nonstationary Gaussian process models to point-referenced spatial data. The nonstationary covariance function allows the user to specify the underlying correlation structure and which spatial dependence parameters should be allowed to vary over space: the anisotropy, nugget variance, and process variance. The parameters are estimated via maximum likelihood, using a local likelihood approach. Also provided are functions to fit stationary spatial models for comparison, calculate the Kriging predictor and standard errors, and create various plots to visualize nonstationarity.

Meridional and Parallel Lines for Maps

Create graticule lines and labels for maps. Control the creation of lines or tiles by setting their placement (at particular meridians and parallels) and extent (along parallels and meridians). Labels are created independently of lines.

Tools for 'ROMS' the Regional Ocean Modeling System

Helper functions for working with Regional Ocean Modeling System 'ROMS' output. See < https://www.myroms.org/> for more information about 'ROMS'.