Open AccessApplying Information Measure for Predicting Release Time of Open Source Software
Author(s) -
Talat PARVEEN,
Hari Darshan ARORA
Publication year - 2017
Publication title -
walailak journal of science and technology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.146
H-Index - 15
eISSN - 2228-835X
pISSN - 1686-3933
DOI - 10.48048/wjst.2018.2631
Subject(s) - computer science , source code , software , measure (data warehouse) , software release life cycle , open source software , software metric , data mining , code (set theory) , open source , software quality , software development , operating system , programming language , set (abstract data type)
Open Source Software (OSS) is updated regularly to meet the requirements posed by the customers. The source code of OSS undergoes frequent change to diffuse new features and update existing features in the system, providing a user friendly interface. The source code changes for fixing bugs and meeting user end requirements again affects the complexity of the code change and creates bugs in the software which are accountable to the next release of software. In this paper, the complexity of code changes in various Bugzilla open source software releases, from version 2.0 on 19th Sep, 1998, to 5.0.1 on 10th Sep, 2015, bugs in each software version release, and the time of release of each software version are considered, and the data used to predict the next release time. The Shannon entropy measure is used to quantify the code change process in terms of entropy for each software release. Observed code changes are utilized to quantify them into entropy units and are further used to predict the next release time. A neural network-based regression model is used to predict the next release time. The performance is compared with the R measure calculated using the multi linear regression model, and a goodness of fit curve is produced.