A Neuro-Based Software Fault Prediction with Box-Cox Power Transformation

Full-Text HTML XML Download Download as PDF (Size:1232KB) PP. 288-309
DOI: 10.4236/jsea.2017.103017    1,160 Downloads   1,659 Views Citations
Author(s)
Momotaz Begum, Tadashi Dohi

Affiliation(s)

Department of Information Engineering, Hiroshima University, Hiroshima, Japan.

ABSTRACT

Software fault prediction is one of the most fundamental but significant management techniques in software dependability assessment. In this paper we concern the software fault prediction using a multilayer-perceptron neural network, where the underlying software fault count data are transformed to the Gaussian data, by means of the well-known Box-Cox power transformation. More specially, we investigate the long-term behavior of software fault counts by the neural network, and perform the multi-stage look ahead prediction of the cumulative number of software faults detected in the future software testing. In numerical examples with two actual software fault data sets, we compare our neural network approach with the existing software reliability growth models based on nonhomogeneous Poisson process, in terms of predictive performance with average relative error, and show that the data transformation employed in this paper leads to an improvement in prediction accuracy.

KEYWORDS

Software Reliability, Artificial Neural Network, Box-Cox Power Transformation, Long-Term Prediction, Fault Count Data, Empirical Validation

Cite this paper

Begum, M. and Dohi, T. (2017) A Neuro-Based Software Fault Prediction with Box-Cox Power Transformation. Journal of Software Engineering and Applications, 10, 288-309. doi: 10.4236/jsea.2017.103017.
JSEA Subscription
E-Mail Alert
JSEA Most popular papers
Publication Ethics & OA Statement
JSEA News
Frequently Asked Questions
Recommend to Peers
Recommend to Library
Contact Us

Copyright © 2020 by authors and Scientific Research Publishing Inc.

Creative Commons License

This work and the related PDF file are licensed under a Creative Commons Attribution 4.0 International License.