Predicting the Time between failures of Software Projects Using Novel Machine Intelligence Algorithms

Abstract

In this paper, machine intelligence technique is used for software effort estimation considering failure patterns and time boundaries in simulation time. Experiments are performed using publicly available software failure dataset and considering two neural machine intelligence techniques, the Hierarchical Temporal Memory (HTM) and the Auditory Machine Intelligence (AMI). The results show that the HTM is a better technique with an error improvement of over 4 times that of the AMI. However, its complexity may hamper its applicability in real time scenarios.