HDShOP - High-Dimensional Shrinkage Optimal Portfolios

Constructs shrinkage estimators of high-dimensional mean-variance portfolios and performs high-dimensional tests on optimality of a given portfolio. The techniques developed in Bodnar et al. (2018 <doi:10.1016/j.ejor.2017.09.028>, 2019 <doi:10.1109/TSP.2019.2929964>, 2020 <doi:10.1109/TSP.2020.3037369>, 2021 <doi:10.1080/07350015.2021.2004897>) are central to the package. They provide simple and feasible estimators and tests for optimal portfolio weights, which are applicable for 'large p and large n' situations where p is the portfolio dimension (number of stocks) and n is the sample size. The package also includes tools for constructing portfolios based on shrinkage estimators of the mean vector and covariance matrix as well as a new Bayesian estimator for the Markowitz efficient frontier recently developed by Bauder et al. (2021) <doi:10.1080/14697688.2020.1748214>.

Last updated

financial-mathematicshigh-dimensional-dataportfolio-managementshrinkage-estimators

3.95 score 6 stars 8 scripts 249 downloads

EstimDiagnostics - Diagnostic Tools and Unit Tests for Statistical Estimators

Extension of 'testthat' package to make unit tests on empirical distributions of estimators and functions for diagnostics of their finite-sample performance.

Last updated

3.70 score 4 scripts 169 downloads

rlfsm - Simulations and Statistical Inference for Linear Fractional Stable Motions

Contains functions for simulating the linear fractional stable motion according to the algorithm developed by Mazur and Otryakhin <doi:10.32614/RJ-2020-008> based on the method from Stoev and Taqqu (2004) <doi:10.1142/S0218348X04002379>, as well as functions for estimation of parameters of these processes introduced by Mazur, Otryakhin and Podolskij (2018) <arXiv:1802.06373>, and also different related quantities.

Last updated

cpp

3.00 score 20 scripts 177 downloads

deforestable - Classify RGB Images into Forest or Non-Forest

Implements two out-of box classifiers presented in <doi:10.1002/env.2848> for distinguishing forest and non-forest terrain images. Under these algorithms, there are frequentist approaches: one parametric, using stable distributions, and another one- non-parametric, using the squared Mahalanobis distance. The package also contains functions for data handling and building of new classifiers as well as some test data set.

Last updated

openblascpp

1.26 score 18 scripts 236 downloads