wilcox_quantiles
Compare multiple quantiles from 2 distributions simultaneously using Wilcox et al. 2013 method.
This is a Python translation of one of the Wilcox' Robust Statistics (WRS) functions from an R package.
Dependencies
- numpy
- scipy
- (optional, for plotting only) matplotlib
Usage
Example:
import numpy as np
from wilcox_quantile import compare_quantiles_vec
rt1 = np.random.randn(100) * 52 + 350
rt2 = np.r_[np.random.randn(85) * 55 + 375, np.random.randn(15) * 25 + 220]
comp = compare_quantiles_vec(rt1, rt2, quantiles=np.arange(0.1, 1.0, 0.1))See the full example in example.py.
Output figure:
Reference
-
The paper: Rand R. Wilcox , David M. Erceg-Hurn, Florence Clark & Michael Carlson, Journal of Statistical Computation and Simulation (2013): Comparing two independent groups via the lower and upper quantiles, Journal of Statistical Computation and Simulation, DOI: 10.1080/00949655.2012.754026
-
The R code: https://raw.githubusercontent.com/nicebread/WRS/master/pkg/R/Rallfun-v39.R
-
An blog post on the subject: http://www.nicebread.de/comparing-all-quantiles-of-two-distributions-simultaneously/