Experiences with text mining large collections of unstructured systems development artifacts at jpl
Proceedings of the 33rd International Conference on Software Engineering
NLARE, a natural language processing tool for automatic requirements evaluation
Proceedings of the CUBE International Information Technology Conference
Hi-index | 0.00 |
Anomaly data can be used to estimate baseline values for operational mission software anomaly frequencies; these estimates can be used for future missions to determine whether software reliability is improving. The accuracy of anomaly frequency estimates can be affected by characteristics of the anomaly data and the problem reporting system maintaining that data. We have been using text mining and machine learning techniques to address one of these issues, in which the number of software-related anomalies is incorrectly reported because the problem reporting system does not tag them correctly. Results to date indicate that these techniques may substantially increase the accuracy of anomaly frequency estimates.