Use of Statistical Methodology based on Weighted Median Approach To analyze Loudness Discomfort Data
DOI:
https://doi.org/10.69844/438a7m88Keywords:
Least absolute deviations, asymptotic tests, least squares, L1 -norm, LAD, Marginal medians, Defining observationsAbstract
#A real-life application related to data for estimating the Loudness Discomfort from Acoustic Thresholds is presented to demonstrate the use of statistical methodology based on the L1 norm as a robust alternative to the usual least squares. In this paper, a robust procedure using the L1 norm estimator was considered to estimate the loudness discomfort level (LDL) from the acoustic threshold level (ART). The estimation results have shown that LDL can conveniently be predicted from ART. To have more confidence in the results and conclusions, the weighted median of L1 norm was utilized to compute intervals for the values of the response (LDL) variable. Such intervals would provide a critical examination of 1-regression fit and the data set as well.