English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

In the past two decades, several Machine Learning (ML) libraries have becomefreely available. Many studies have used such libraries to carry out empiricalinvestigations on predictive Software Engineering (SE) tasks. However, thedifferences stemming from using one library over another have been overlooked,implicitly assuming that using any of these libraries would provide the userwith the same or very similar results. This paper aims at raising awareness ofthe differences incurred when using different ML libraries for softwaredevelopment effort estimation (SEE), one of most widely studied SE predictiontasks. To this end, we investigate 4 deterministic machine learners as providedby 3 of the most popular ML open-source libraries written in differentlanguages (namely, Scikit-Learn, Caret and Weka). We carry out a thoroughempirical study comparing the performance of the machine learners on 5 SEEdatasets in the two most common SEE scenarios (i.e., out-of-the-box-ml andtuned-ml) as well as an in-depth analysis of the documentation and code oftheir APIs. The results of our study reveal that the predictions provided bythe 3 libraries differ in 95% of the cases on average across a total of 105cases studied. These differences are significantly large in most cases andyield misestimations of up to approx. 3,000 hours per project. Moreover, ourAPI analysis reveals that these libraries provide the user with differentlevels of control on the parameters one can manipulate, and a lack of clarityand consistency, overall, which might mislead users. Our findings highlightthat the ML library is an important design choice for SEE studies, which canlead to a difference in performance. However, such a difference isunder-documented. We conclude by highlighting open-challenges with suggestionsfor the developers of libraries as well as for the researchers andpractitioners using them.

Do Not Take It for Granted: Comparing Open-Source Libraries for Software Development Effort Estimation