Feasibility predictability model for software test automation projects in DevOps setting. (9th October 2020)

• Record Type: Journal Article
• Authors: Angara, Jayasri; Prasad, Srinivas; Gutta, Sridevi
• Is Part Of: International journal of forensic software engineering. Volume 1:Number 2/3(2020)
• Page Start: 231
• Page End: 246
• Publication Date: 2020-10-09
• Subjects: DevOps -- agile test automation -- project feasibility prediction algorithm -- machine learning -- logistic regression
• Languages: English
• ISSNs: 1743-5099

Abstract

DevOps is an outgrowth of agile practice and evolved to manage the continuous change. The goal is to shorten the project timelines, increase productivity, without impacting business and quality. Automation has become one of the key enablers for success. However, test automation gets little time. This poses a challenge to managers whether to automate the test function or not. Managers need to take swift go/no-go decision. The objective of this paper is to develop a predictability model for test automation project feasibility. Authors conducted a literature and practitioner's survey and identified 21 key factors which determine the viability of a project. Authors surveyed 38 test automation projects and created a dataset. A custom simulation model was developed, augmenting the dataset with 23, 407 more records. Authors attempted to predict the success using machine learning algorithms. Further, factor analysis was conducted to reduce the number of factors for operational simplicity.