Suggesting Software Measurement Plans with Unsupervised Learning Data Analysis

Sarah Dahab, Stephane Maag

Abstract

Software measurement processes require to consider more and more data, measures and metrics. Measurement plans become complex, time and resource consuming, considering diverse kinds of software project phases. Experts in charge of defining the measurement plans have to deal with management and performance constraints to select the relevant metrics. They need to take into account a huge number of data though distributed processes. Formal models and standards have been standardized to facilitate some of these aspects. However, the maintainability of the measurements activities is still constituted of complex activities. In this paper, we aim at improving our previous work, which aims at reducing the number of needed software metrics when executing measurement process and reducing the expertise charge. Based on unsupervised learning algorithm, our objective is to suggest software measurement plans at runtime and to apply them iteratively. For that purpose, we propose to generate automatically analysis models using unsupervised learning approach in order to efficiently manage the efforts, time and resources of the experts. An implementation has been done and integrated on an industrial platform. Experiments are processed to show the scalability and effectiveness of our approach. Discussions about the results have been provided. Furthermore, we demonstrate that the measurement process performance could be optimized while being effective, more accurate and faster with reduced expert intervention.

Download


Paper Citation


in Harvard Style

Dahab S. and Maag S. (2019). Suggesting Software Measurement Plans with Unsupervised Learning Data Analysis.In Proceedings of the 14th International Conference on Evaluation of Novel Approaches to Software Engineering - Volume 1: ENASE, ISBN 978-989-758-375-9, pages 189-197. DOI: 10.5220/0007768101890197


in Bibtex Style

@conference{enase19,
author={Sarah Dahab and Stephane Maag},
title={Suggesting Software Measurement Plans with Unsupervised Learning Data Analysis},
booktitle={Proceedings of the 14th International Conference on Evaluation of Novel Approaches to Software Engineering - Volume 1: ENASE,},
year={2019},
pages={189-197},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0007768101890197},
isbn={978-989-758-375-9},
}


in EndNote Style

TY - CONF

JO - Proceedings of the 14th International Conference on Evaluation of Novel Approaches to Software Engineering - Volume 1: ENASE,
TI - Suggesting Software Measurement Plans with Unsupervised Learning Data Analysis
SN - 978-989-758-375-9
AU - Dahab S.
AU - Maag S.
PY - 2019
SP - 189
EP - 197
DO - 10.5220/0007768101890197