Inferential Procedures For The Truncated Exponential Distribution

Abstract

Suppose that X1, X2,…, Xn are independent and identically distributed with density f(x,θ) = [θ(l-e-t/θ)]-1 e-x/θ, 0 ≤ x ≤ t and that inferences about θ are to be made. The exact distribution of U = Σi=nX1 known but is quite complicated and so an approximation to its distribution is needed. It is shown here that the beta approximation for the density of (nt)-1U obtained by equating the first two moments performs better, for moderate n, than the normal approximation given by the central limit theorem and is asymptotically equivalent to it. The use of this approximation in making inferences in some life testing situations is discussed via an example. © 1977, Taylor & Francis Group, LLC. All rights reserved.

Department(s)

Mathematics and Statistics

Second Department

Geosciences and Geological and Petroleum Engineering

Comments

National Science Foundation, Grant DAAG29-76-G-0271

Keywords and Phrases

beta approximation; censoring; reliability

International Standard Serial Number (ISSN)

1532-415X; 0361-0926

Document Type

Article - Journal

Document Version

Citation

File Type

text

Language(s)

English

Rights

© 2023 Taylor and Francis Group; Taylor and Francis, All rights reserved.

Publication Date

01 Jan 1977

Share

 
COinS