Title: Effort estimation in software development using story point: a machine learning approach

Authors: Beulah Moses; Shyam Singhal

Addresses: SP Jain School of Global Management, 5 Figtree Drive, Sydney Olympic Park, NSW 213, Australia ' SP Jain School of Global Management, 5 Figtree Drive, Sydney Olympic Park, NSW 213, Australia

Abstract: Agile methodologies are besieged with problems and potential solutions around predictive insights on a project. These problems range from estimation, quality, to effort and duration requirements. Despite of having innumerable predictive models not a single reliable solution is available to estimate the duration and effort required to complete an agile project on an ongoing basis. This is due subjective nature of 'story point', and progressive elaboration of 'user story'. This paper analyses the relationship between story points and effort, across a sample of software development projects, in an organisation. A novel machine learning predictive model has been developed and is implemented across agile projects that infer relationship between 'effort' and 'story points', directly in contrast with agile literature. This predictive model was tested and worked accurately, reliably and effectively across various agile projects. This research can be extended to agile projects having sprints of less than fifty story points.

Keywords: forecasting; estimation model; machine learning; regression; agile; scrum; story points; effort estimation.

DOI: 10.1504/IJKESDP.2020.112630

International Journal of Knowledge Engineering and Soft Data Paradigms, 2020 Vol.7 No.1, pp.25 - 44

Received: 14 Jan 2020
Accepted: 22 Aug 2020
Published online: 25 Jan 2021 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article