期刊名称：Journal of Reliability and Statistical Studies
印刷版ISSN：0974-8024
电子版ISSN：2229-5666
出版年度：2009
卷号：2
期号：1
页码：52-59
出版社：Ankur Printing Palace
摘要：This paper presents the development of weighted least squares (WLS) method inregression modeling when data are characterized by a high degree of heteroscedasticity in theresponse variable. An algorithm is developed to obtain the weighting parameter in the WLSmodel. Tests for coincidence and parallelism in the WLS model are studied. Finally results aredemonstrated empirically by modeling the effect of age, body mass index and sex, on PeakExpiratory Flow Rate (PEFR).
