On Some Aspects of Residual Analysis in Regression Models: Examination of Residuals

In this book an attempt has been made by proposing some inferential techniques for linear regression models by using various types of residuals other than OLS residuals. Chapter I contains the general introduction about the concept of residuals in regression analysis. Various types of residuals and their properties have been described in chapter II. The literature on residual analysis and regression diagnostics have been reviewed in chapter III. In particular, it is discussed about examination of residual plots diagnostics for identifying unusual data points, detection of influential subsets, until applications of residuals. Chapter IV contains the proposed methods under is the contributory one in which outliers are detected; Leverage and influential points are identified; Goldfeld-Quandt test statistics computed to know the heteroscodastocity; Chow test statistic is computed to know the structural change and Durbin-Waston test statistic is competed to know the autocorrelation using different types of residuals.