A Study on the Software Reliability Model Using Deep Neural Network for General Situation  
Author Youn Su Kim


Co-Author(s) Kwang Yoon Song; Da Hye Lee; In Hong Chang


Abstract As software becomes important in all fields, software reliability has become a very important discipline. The software reliability growth model, a tool for estimating the quality and reliability of software, has been proposed to solve these problems because if the software breaks down due to large or small problems such as a part of coding or system errors, social and economic damage is large. The software reliability model induces software failures through assumptions about statistical distributions. However, the proposed model through assumptions is difficult to generalize. In this study, to overcome difficulties, a model that relies on given data was proposed, not a model that depends on assumptions through deep learning. In this study, we propose a software reliability model based on data rather than a software reliability model based on assumptions by using the DNN, which is the most basic of deep learning, and the RNN. As a result, RNN showed good performance as the three criteria values were the smallest.


Keywords Software Reliability, Deep Neural Network (DNN), Recurrent Neural Network (RNN)
    Article #:  RQD27-23

Proceedings of 27th ISSAT International Conference on Reliability & Quality in Design
Virtual Event

August 4-6, 2022