Analogue and Weighted Averaging Methods for Palaeoecology

Fits Modern Analogue Technique and Weighted Averaging transfer function models for prediction of environmental data from species data, and related methods used in palaeoecology.

analogue

What is analogue?

analogue is an R package for use with palaeoecological data. Originally, analogue was intended as an R implementation of analogue methods such as analogue matching, ROC curves, and MAT transfer function models, and the computation of dissimilarity coefficients. Since then the scope of the package has grown to include a number of other methods applicable to data routinely encountered in palaeoecology and palaeolimnology.

Features

Transfer functions
- MAT
- Weighted Averaging with monotonic, inverse, and classical deshrinking, with and without tolerance down-weighting
- Principal Component Regression (using ecologically-relevant transformations)
- Cross-validation (Bootstrapping, leave-one-out, k-fold)
- Analogue statistics
Analogue matching
Dissimilarity coefficients
- Chord, Bray-Curtis, Gower's Generalised coefficient, Manhattan, ...
Dissimilarity decisions thresholds
- ROC curves
- Monte Carlo resampling
- Logistic regression
Stratigraphic diagrams
Principal curves