Discrimination Mitigation for Machine Learning Models

Based on different statistical definitions of discrimination, several methods have been proposed to detect and mitigate social inequality in machine learning models. This package aims to provide an alternative to fairness treatment in predictive models. The ROC method implemented in this package is described by Kamiran, Karim and Zhang (2012) < https://ieeexplore.ieee.org/document/6413831/>.

Robust Nonparametric Two-Sample Tests for Location/Scale

Implementations of several robust nonparametric two-sample tests for location or scale differences. The test statistics are based on robust location and scale estimators, e.g. the sample median or the Hodges-Lehmann estimators as described in Fried & Dehling (2011) . The p-values can be computed via the permutation principle, the randomization principle, or by using the asymptotic distributions of the test statistics under the null hypothesis, which ensures (approximate) distribution independence of the test decision. To test for a difference in scale, we apply the tests for location difference to transformed observations; see Fried (2012) . Random noise on a small range can be added to the original observations in order to hold the significance level on data from discrete distributions. The location tests assume homoscedasticity and the scale tests require the location parameters to be zero.

Models for Simulating Wolf Populations

Simulate the dynamic of wolf populations using a specific Individual-Based Model (IBM) compiled in C, see Chapron et al. (2016) .

Temporal Trends in Ecological Niche Models

Computes temporal trends in environmental suitability obtained from ecological niche models, based on a set of species presence point coordinates and predictor variables.

Semi-Supervised Model for Geographical Document Classification

Semissupervised model for geographical document classification (Watanabe 2018) . This package currently contains seed dictionaries in English, German, French, Spanish, Italian, Russian, Hebrew, Arabic, Turkish, Japanese and Chinese (Simplified and Traditional).

Tools Developed by the Long Term Ecological Research Community

Set of the data science tools created by various members of the Long Term Ecological Research (LTER) community. These functions were initially written largely as standalone operations and have later been aggregated into this package.

Longitudinal Concordance Correlation

Estimates the longitudinal concordance correlation to access the longitudinal agreement profile. The estimation approach implemented is variance components approach based on polynomial mixed effects regression model, as proposed by Oliveira, Hinde and Zocchi (2018) . In addition, non-parametric confidence intervals were implemented using percentile method or normal-approximation based on Fisher Z-transformation.

Generalized Spatial-Time Sequence Miner

Implementations of the algorithms present article Generalized Spatial-Time Sequence Miner, original title (Castro, Antonio; Borges, Heraldo ; Pacitti, Esther ; Porto, Fabio ; Coutinho, Rafaelli ; Ogasawara, Eduardo . Generalização de Mineração de Sequências Restritas no Espaço e no Tempo. In: XXXVI SBBD - Simpósio Brasileiro de Banco de Dados, 2021 ).

Estimation of the Number of Fatalities from Carcass Searches

The number of bird or bat fatalities from collisions with buildings, towers or wind energy turbines can be estimated based on carcass searches and experimentally assessed carcass persistence times and searcher efficiency. Functions for estimating the probability that a bird or bat that died is found by a searcher are provided. Further functions calculate the posterior distribution of the number of fatalities based on the number of carcasses found and the estimated detection probability.

Create Network Connections Based on Chess Moves

Provides functions to work with directed (asymmetric) and undirected (symmetric) spatial networks. It makes the creation of connectivity matrices easier, i.e. a binary matrix of dimension n x n, where n is the number of nodes (sampling units) indicating the presence (1) or the absence (0) of an edge (link) between pairs of nodes. Different network objects can be produced by 'chessboard': node list, neighbor list, edge list, connectivity matrix. It can also produce objects that will be used later in Moran's Eigenvector Maps (Dray et al. (2006) ) and Asymetric Eigenvector Maps (Blanchet et al. (2008) ), methods available in the package 'adespatial' (Dray et al. (2023) < https://CRAN.R-project.org/package=adespatial>). This work is part of the FRB-CESAB working group Bridge < https://www.fondationbiodiversite.fr/en/the-frb-in-action/programs-and-projects/le-cesab/bridge/>.