stat_hotelling_impl: Compute Hotelling's T^2 statistic

In astamm/fdahotelling: Inference for Functional Data Analysis in R

Description

stat_hotelling_impl computes the generalized Hotelling's T^2 statistic using the RcppArmadillo interface for fast pseudoinverse calculations via divide and conquer method.

Usage

stat_hotelling_impl(x, y, mu, paired = FALSE, step_size = 0,
  use_correction = FALSE, tolerance = 0)

Arguments

x	Dataframe or matrix containing the data collected from 1st population.
y	Dataframe or matrix containing the data collected from 2nd population (default: NULL).
mu	True mean value (or difference between mean values) under the null hypothesis (default = 0).
paired	Is the input data paired? (default: FALSE).
step_size	The step size used to perform integral approximation via the method of rectangles (default: 0). When set to 0, it assumes that we are dealing with multivariate data rather than functional data and thus no integration is necessary.
use_correction	Flag to turn on and off a statistic adjustment factor that makes it Fisher-distributed under the normality assumption (default: FALSE).
tolerance	Positive value τ such that positive eigenvalues of the sample covariance matrix that are below τ trΣ are discarded for inverse computation (default: 0).

Value

A named numeric vector of size 1 storing the Hotelling's T^2 statistic defined as

\int_D

.

astamm/fdahotelling documentation built on May 10, 2019, 2:05 p.m.