CNN LSTM Network Architecture for Modeling Software Reliability

Abstract

In this work, Convolutional Neural Network Long Short-Term Memory (CNN LSTM) architecture is proposed for modelling software reliability with time-series data. Evaluation of the model coming from 2 open source datasets that describe the development and testing of modern mobile operating systems - “Tizen” and “CyanogenMod”. The results of the proposed model are compared with four parametric Software Reliability Growth Models and simple Convolutional Neural Network model using the Root Mean Squared Error (RMSE) metric.