Estimating the number of undetected software errors via the correlated capture–recapture model

Sometimes a complex software system fails because of errors undiscovered in the design stage of the development process. Detecting these errors early in the process would eliminate many downstream problems. The so-called ‘capture–recapture’ model, initially used by biologists to estimate the size of wildlife populations, has been widely used to estimate the number of software design errors. However, one simplifying assumption in capture–recapture models is that the inspections performed by various inspectors are statistically independent from each other. In the paper, we propose a novel method that is based on the correlation matrix of multiple inspectors. In a numerical analysis, we show that our method outperforms other traditional models that are based on the independence assumption.