Showing papers in "Calcutta Statistical Association Bulletin in 1977"
TL;DR: In this article, it was shown that the p.d.f. distribution of the BVE distribution can be used to characterize the life span of a series of independent components.
Abstract: In analysing reliability characteristics of multicomponent systems it is rational to assume in many cases that component life lengths are positively or negatively interdependent. This positive interdependence may be due to common environmental and incidental stresses, functional dependence on common sources of energy, 'mechanical sticking of assemblies' etc. Multivariate parametric families of distributions which have been proposed and studied (incompletely) to describe joint random variations of component failure times are multivariate generalisations of exponential, weibull, gamma and normal distributions. Although several useful results characterising life distributions of coherent systems of independent components have already been obtained (Barlow and Proschan, 1975), very little is yet known about dependent systems even of two components. Downton (1970) has worked out probability distributions, means and variances of the life of a series and a parallel system for the bivariate exponential (BVE) distribution arising out of the 'successive damage model' proposed by him as well for the one arising out of the 'shock model' due to Marshall and Olkin (1967) . An easily tractable form of the BVE distribution, proposed by Morgenstern (1956) and studied by Gumbel (1960) has the p.d.f.
6 citations
5 citations
TL;DR: For grouped data under time-sequential studies, Majumdar and Sen as discussed by the authors studied the asymptotic properties of a general class of progressively censored rank order tests for the simple regression model.
Abstract: For grouped data under time-sequential studies, Majumdar and Sen (1977a) have studied the asymptotic properties of a general class of progressively censored rank order tests for the simple regression model. The results are extended to the multiple regression situation.
5 citations
3 citations
3 citations
TL;DR: In this paper, the authors considered a specific sampling strategy employing an unbiased estimator based on a sample chosen following the scheme of selection, where on the first occasion a sample is chosen following Midzuno's (1952) (modified) 1rPS selection scheme and on the second occasion a randomly selected portion of this (of size m = mi, 0 < i, < 1) is retained and a fresh independent sample of size u = np, (whe1'e A.
Abstract: In estimating a finite population total on a current occasion utilizing the results of two consecutive surveys several sampling strategies involving selection with varying probabilities are available in the literature. In this article we shall consider a specific one employing an unbiased estimator (vide section 2 for details) based on a sample chosen following the scheme of selection, where on the first occasion a sample is chosen following Midzuno's (1952) (modified) 1rPS selection scheme and on the second occasion a randomly selected portion of this (of size m = mi\\., 0 < i\\. < 1) is retained and a fresh independent sample of size u = np, (whe1'e A.+;t = 1) is chosen from the universe again employing Midzuno's (modified) 1rPS selection scheme. This strategy, applicable only under certain restrictive conditions (vide (2.3) below), (which, if satisfied, make this strategy particularly simple to apply) is compared in 1'espect of its relative efficieny, with those earlier p1'oposed by Des Raj (1965), Ghangurde and Rao (1969) and Ghotai (1974). It is obse1'ved that certain conditions ((3.1) in particular and others stated in sections 3 and 4), if satisfied, are particularly fa.vom'able to this particular strategy relative to its competitors considered here.
3 citations
1 citations
TL;DR: In this paper, the authors studied the properties of sample extreme values in the regular case and modified them for a modified version of the classic jackknifing technique, and some applications were also considered.
Abstract: : Sample extreme values are biased estimators of the end-points of a distribution, and hence, jackknifing is useful. However, the properties of jackknifing in such a case differ considerably from those in the regular case. These are studied here. Along with a modification of jackknifing, some applications are also considered. (Author)
1 citations
TL;DR: It appears that in India the length of PPA is longer than in Western Countries, and for over 75% of the cases, there is resumption of menstrual cycles within 6 weeks following the termination of pregnancies.
Abstract: Post partum amenorrhoea (PPA) or' the temporaiy cessation of menstrual cycle afte1' child birth is an important item of consideration in the estimation of intm'pregnancy interval. It appears that in India the length ofPPA is longer than in Western Countries. A survey in the nursing homes at Belgium reveals that for over 75% of the cases, thel't is resumption of menstrual cycles within 6 weeks following the termination of pregnancies; as against another survey at Bombay reporting only 1.5% of the women resuming the same after the fourth post partum month. Apart from the variation of PPA from community to community, the measm'ement of PPA is in general beset with many practical\" difficulties; especially, from the 1'ecords of a sample survey. For example, a conception occurringat theovaluatory period, just preceding the otherwise normal data of resumption of menstrual cycle may wrongly be ascribed as the continuation of PPA. False symptons, as occasional post natal haemorrhage may also be mistaken, possibly as the expiry of PPA .