A feasibility study of two-step software effort estimation

Naoki Kinoshita, Akito Monden

Research output: Contribution to journalArticlepeer-review

Abstract

In this paper, focusing on the problem that estimation accuracy of software development effort greatly varies among software projects, we propose a two-step effort estimation method, and conduct a feasibility study of the method. In the proposed method, given a project to be estimated, we evaluate the dependability of effort estimation for the project, that is, we evaluate whether the effort can be accurately estimated or not for the project. In case that the dependability is high, we conduct the estimation. Otherwise, estimation is avoided. As a result of an experiment to assess the feasibility of the proposed method, the residual square mean turned out to be a suitable measure of dependability; and, by avoiding the estimation when a project belongs to the high residual square mean group, the estimation error was reduced, which indicates the effectiveness of the two-step estimation. Also, by relaxing the conditions to avoid the estimation, we found that it is possible to increase the number of project to be estimated, while decreasing the estimation error.

Original languageEnglish
Pages (from-to)194-200
Number of pages7
JournalComputer Software
Volume35
Issue number1
DOIs
Publication statusPublished - Feb 1 2018

ASJC Scopus subject areas

  • Software

Fingerprint Dive into the research topics of 'A feasibility study of two-step software effort estimation'. Together they form a unique fingerprint.

Cite this