An SPRT Procedure for an Ungrouped Data using MMLE Approach

Abstract

Sequential Analysis of statistical science could be adopted in order to decide up on the reliable/unreliable of the developed software very quickly. Sequential analysis is different from classical hypothesis testing. In classing hypothesis testing one can draw conclusion during the data collection and final conclusion can possible be reached at a later stage, whereas sequential analysis are easy to see as data collection can be terminated after fewer cases and decision taken earlier stage, thus saving in terms of reliability of software. In this paper we proposed the performance of sequential probability ratio test (SPRT) on time domain data using exponential model and analyzed the result by applying on 3 data sets. The parameters are estimated using Modified Maximum likelihood estimation.

Software Reliability Using SPRT: Burr Type III Process Model

“…The Wald"s SPRT procedure can be used to distinguish the software under test into one of the two categories like reliable/unreliable, pass/fail and certified/uncertified [14]. SPRT is the optimal statistical test that makes the correct decision in the shortest time among all tests that are subject to the same level of decision errors [15]. SPRT is used to detect the fault based on the calculated likelihood of the hypotheses.…”

An Analogy of Burr Type III and Pareto Type II using SPRT

“…The Wald's SPRT procedure is a method that can be used to distinguish the software under test into one of the two categories like reliable/unreliable, pass/fail and certified/uncertified [8]. SPRT is the optimal statistical test that makes the correct decision in the shortest time among all tests that are subject to the same level of decision errors [9]. It is used to detect the fault based on the calculated likelihood of the hypotheses.…”