Deep Cross-Project Software Reliability Growth Model Using Project Similarity-Based Clustering

Software reliability is an essential characteristic for ensuring the qualities of software products. Predicting potential number bugs from beginning a development project allows practitioners to make appropriate decisions regarding testing activities. In initial phases, applying traditional growth models (SRGMs) with limited past data does not always provide reliable prediction result decision making. To overcome this, herein, we propose new modeling method called deep cross-project model (DC-SRGM). DC-SRGM that uses features previous projects’ through similarity. Specifically, proposed applies cluster-based selection training source and by learning method. Experiments involving 15 real datasets company 11 open show can more precisely describe ongoing projects than existing SRGMs LSTM model.