Refactoring Android Source Code Smells From Android Applications

Result code in IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F61989100%3A27740%2F25%3A10256660" target="_blank" >RIV/61989100:27740/25:10256660 - isvavai.cz</a>

Result on the web

<a href="https://ieeexplore.ieee.org/document/10840228" target="_blank" >https://ieeexplore.ieee.org/document/10840228</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1109/ACCESS.2025.3529687" target="_blank" >10.1109/ACCESS.2025.3529687</a>

Result language

angličtina

Original language name

Refactoring Android Source Code Smells From Android Applications

Original language description

As technology advances and new features emerge, the demand for Android applications continues to grow, leading to rapid release schedules. These accelerated development timelines often push developers to make rushed changes, often resulting in suboptimal design practices, commonly known as code smells. These issues can degrade application quality, drive up maintenance costs, lead to unexpected behaviors, and complicate evolution and re-engineering efforts. While substantial research has focused on identifying Android-specific and object-oriented code smells, comparatively less attention has been devoted to their systematic refactoring and evaluation. This study introduces a web-based technique, validated through a tool specifically developed to detect 20 Android-specific code smells and automatically refactor 10 of them. Our approach surpasses traditional desktop and plugin solutions by providing easy accessibility, cross-platform compatibility, and eliminating setup requirements. When applied to six open-source and two industrial Android applications and evaluated against the ISO/IEC 25010 quality standard, our tool demonstrated considerable improvements: reducing CPU utilization by 15.39%, lowering memory consumption by 12.85%, and enhancing battery efficiency by up to 5.78%. The tool&apos;s accuracy, validated through precision, recall, and F-measure metrics, achieved averages of 91.81% precision, 97.77% recall, and a 94.67% F-measure. This study enhances the Android application development lifecycle by offering developers a feasible solution for optimizing CPU efficiency, reducing memory use, and minimizing battery consumption.

Czech name

—

Czech description

—

Type

J_imp - Article in a specialist periodical, which is included in the Web of Science database

CEP classification

—

OECD FORD branch

10700 - Other natural sciences

Project

—

Continuities

O - Projekt operacniho programu

Publication year

2025

Confidentiality

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Name of the periodical

IEEE Access

ISSN

2169-3536

e-ISSN

2169-3536

Volume of the periodical

13

Issue of the periodical within the volume

nestránkováno

Country of publishing house

US - UNITED STATES

Number of pages

29

Pages from-to

14122-14150

UT code for WoS article

001405925700028

EID of the result in the Scopus database

2-s2.0-85215225955