Crandore Hub

HDElliptical

High-Dimensional Methods for Elliptically Symmetric Distributions

Fast, documented implementations of robust estimation, testing, dimension reduction, classification, and clustering methods for high-dimensional elliptically symmetric data. Computational kernels use 'Rcpp' and 'RcppArmadillo'. The package follows methods reviewed in Feng (2026), "High-Dimensional Data Analysis for Elliptically Symmetric Distributions" <https://github.com/flnankai/HDElliptical/releases>.

README

# Deterministic benchmarks

Run the installed-package benchmark with:

```sh
Rscript run-benchmarks.R HDElliptical-benchmarks.csv 5 1 0.05
```

The harness times one representative callable workflow per book chapter and
records a deterministic numeric fingerprint. It uses no package beyond
HDElliptical itself. The timings are descriptive machine-local evidence, not a
paper-specific simulation, size/power experiment, or performance guarantee.

Arguments are the output CSV, the number of measured repetitions, and the
number of warm-up calls, followed by the minimum elapsed time used to calibrate
each measurement batch. The defaults are 5, 1, and 0.05 seconds. The script
reports per-call timings together with the number of calls in each batch,
avoiding zero-resolution timings for fast methods. It fails if repeated calls
produce different numeric fingerprints.

Versions across snapshots

VersionRepositoryFileSize
0.1.2 rolling linux/jammy R-4.5 HDElliptical_0.1.2.tar.gz 3.8 MiB
0.1.2 rolling linux/noble R-4.5 HDElliptical_0.1.2.tar.gz 3.9 MiB
0.1.2 rolling source/ R- HDElliptical_0.1.2.tar.gz 983.5 KiB
0.1.2 latest linux/jammy R-4.5 HDElliptical_0.1.2.tar.gz 3.8 MiB
0.1.2 latest linux/noble R-4.5 HDElliptical_0.1.2.tar.gz 3.9 MiB
0.1.2 latest source/ R- HDElliptical_0.1.2.tar.gz 983.5 KiB
0.1.2 2026-04-23 source/ R- HDElliptical_0.1.2.tar.gz 0 B

Dependencies (latest)

Imports

LinkingTo

Suggests