Publisher: Bachudo Science Co. Ltd

Weighted Least Squares Method and Associated Time Series Data Problems

A. C. Eke, D. Eni, J. Atsu
KEYWORDS: Time series, autocorrelation, multi-collinearity heteroscodasticity transformation.

ABSTRACT:

An apparently high R2 in any set of data and especially in a time series data which may suggest a good fit may not necessarily imply that the basic assumptions of a regression model have been well met. Even in the presence of the problems of heteroscedasticity, serial correlation and multi collinearity, it is possible to have a very high R2 This paper seeks to address the problem arising from heteroscedasticity and serial correlation in time series data and to suggest how the problems can be dealt with so as to have data that will yield valid statistical inference.


Download
View Online
180
views
0
downloads
Subject
Publication Year
Place Of Publication
Contributors