June 2021 Issue of the Statistics Digest Now Available!

Published 1
This issue includes the following:
  • Mini Paper by Rothy Chhim and Bingbing Weitner - A Performance Comparison between SAS Macro and DS2 for Deming Regression Analysis
  • Feature by Neelam Nakadi: Simple Data Summaries Using R: Beginning With Baby Steps Of Exploratory Data Analysis (EDA)
  • Columns by Donald J. Wheeler, Grant Reinman and Jim Frost.
  • Mini-Paper: A Performance Comparison between SAS Macro and DS2 for Deming Regression Analysis
*Click attachment icon in upper right corner to access the June 2021 Statistics Digest*

Mini-Paper: A Performance Comparison between SAS Macro and DS2 for Deming Regression Analysis


Key Words:
Deming regression, SAS, Jackknife estimation

Abstract:
A SAS macro to perform the Deming regression was developed by Deal, Pate and Rouby (2009) and is widely available for use. However, the macro’s performance slows down as the number of observations or number of parameters increases. Two solutions using the THREAD and Matrix package of the SAS DS2 programming language (SAS Institute Inc.) were developed as an alternative to improve Deming regression run time. For a single parameter with 100 and 5000 observations, run time was 52 seconds and 1.93 hours respectively for the Deming macro (2009). For the DS2 THREAD solution, run times were respectively 4 seconds and 0.03 hours (106 seconds). For the DS2 Matrix solution, run times were respectively 4 seconds and 0.01 hours (26 seconds). For 15 parameters with 250 observations per parameter (N=3750 total), run time was 1037 seconds for the Deming macro (2009), 73 seconds for DS2 THREAD, and 27 seconds for DS2 Matrix. Both DS2 solutions demonstrate a 13-fold reduction in run time beginning with 100 observations for one or more parameters.

Feature: Simple Data Summaries Using R: Beginning With Baby Steps Of Exploratory Data Analysis (EDA)

Key Words:
Exploratory Data Analysis, R Environment,

Abstract:
Exploratory data analysis (EDA) is a very insightful technique to understand the data. It helps in getting a look and feel of the data. Numerous simple yet powerful tools are used during EDA. Many software packages incorporating these tools are currently available. This article provides a basic look at EDA using the R environment.

News Statistics Division 06/29/2021 10:16am CDT

Comments

No Data Available

Share: