bisquare_function: Bisquare function

In adamethods: Archetypoid Algorithms and Anomaly Detection

Description

This function belongs to the bisquare family of loss functions. The bisquare family can better cope with extreme outliers.

Usage

bisquare_function(resid, prob, ...)

Arguments

resid	Vector of residuals, computed from the m \times n residuals data matrix.
prob	Probability with values in [0,1].
...	Additional possible arguments.

Value

Vector of real numbers.

Author(s)

Irene Epifanio

References

Moliner, J. and Epifanio, I., Robust multivariate and functional archetypal analysis with application to financial time series analysis, 2019. Physica A: Statistical Mechanics and its Applications 519, 195-208. https://doi.org/10.1016/j.physa.2018.12.036

Examples

resid <- c(2.47, 11.85)  
bisquare_function(resid, 0.8)

adamethods documentation built on Aug. 4, 2020, 5:08 p.m.