Int J Performability Eng ›› 2019, Vol. 15 ›› Issue (5): 1289-1296.doi: 10.23940/ijpe.19.05.p4.12891296

Previous Articles     Next Articles

Fault Big Data Analysis Tool based on Deep Learning

Yoshinobu Tamuraa,* and Shigeru Yamadab   

  1. a Tokyo City University, Tamazutsumi 1-28-1, Setagaya-ku, Tokyo, 158-8557, Japan
    b Tottori University, Minami 4-101, Koyama, Tottori-shi, 680-8552, Japan
  • Submitted on ;
  • Contact: * E-mail address:
  • Supported by:
    This work was supported in part by the JSPS KAKENHI (No. 16K01242) in Japan.

Abstract: Software managers can obtain useful information from many fault data sets recorded on bug tracking systems (BTS). However, it is difficult to find helpful measures for software reliability, maintainability, and performability, because the data collected on the BTS are mixed with qualitative and quantitative ones. This paper discusses the methods of reliability, maintainability, and performability assessment by deep learning for big data in terms of software faults. Specifically, we implement the reliability, maintainability, and performability analysis tool discussed in our method by using the latest programing technology. Moreover, we show several performance examples of the implemented application software by using the fault big data observed in the practical projects.

Key words: open source software, software tool, reliability, maintainability, performability, fault big data, deep learning