Your Cart

Your cart is empty

Home  >  Volume 22 (2012)

Assessing the Detection and Correction of Model Violations Using Residuals in a Linear Regression Diagnostics by Osemeke Reuben. F. and Ehiwario, J.C. Volume 22 (November, 2012), pp 249 – 256
Sale price: $5.00
Quantity:

Abstract

In ordinary linear regression, graphical diagnostics and numerical test were used to detect and correct the model violations of regression assumption. Residual plots against the predictor or predicted (ŷi) were used to show error violations of heteroscedasticity, autocorrelation, outliers, clustering of data points and nonlinearity. Non-normality was diagnosed using letter value display. The error violations were corrected by plausible trial and error transformation of the variables. Analysis of residuals after the correction were improved upon as shown in the coefficient of determination (r2),multiple r, p value < 0.05, Durbin Watson Test above 1.6, increase in T Test for the observed value, increase in F change, minimal standard error of the estimate, strong midpoint values in letter value display and  well behave scatter plots 

Keywords: Correction; Detention; Model Violations, Residuals, Linear Regression.

Close
Loading...