Estimating the number of faults in a software program using an exponential model.

This memorandum takes a look at some of the problems encountered when trying to estimate the number of faults in a software program using a exponential model. The maximum likelihood estimator, a jackknife estimator, and a simple Bayesian estimator are investigated. We are drawn to one of two conclusions. First, estimator failure or bias may be a problem inherent with the exponential model. Second, we have yet to determine the proper method of estimation. We feel that the former is the proper conclusion but cannot proof it with certainty.