ReChan
Daniel Domínguez-Álvarez, Daniel Toniuc, Alessandra Gorla
University of Verona IMDEA Software Institute
内容与影响
"What's new?" This is what users wonder when they see the notification that a mobile app has just been updated on their device. New releases may involve simple bug fixes, or may include new features that users are eager to try. Regardless of the change, users do want to know what are the differences with respect to the release that have been using so far. The Google Play store has a visible section for each Android app that clearly describes the changes that affect the latest release. This description, however, is curated by developers, and may not match the actual changes in the binary code. This paper presents ReChan, a novel technique aiming to automatically detect mismatches between release notes of Android applications and the actual changes in the code. We define a taxonomy of 9 release categories by manually tagging 1,200 real samples, and we present our solution to automatically classify release notes written in English. ReChan then implements specific analyses to detect such changes in the code, and compares the analyses outcome to detect mismatches. ReChan achieves a precision, recall and f-score of 84.9% on the manually crafted ground truth of three open source apps. Experiments on a dataset of 12,706 closed source Android apps show that developers tend to correctly report changes due to bug fixes and new features, but omit changes that affect the list of requested permissions, the UI and other content that the app uses.
逐年被引趋势
关键指标
同类平均 = 1
同领域 · 同年份 · 同类型
Google Scholar 与 OpenAlex 的被引统计范围不同,数值存在差异属正常。
AI 辅助阅读
依据:摘要
回答优先基于摘要、文献信息与可获取全文;依据不足时会明确说明。
学术脉络
学科主题
计算机 / AIAdvanced Malware Detection Techniques
Software Engineering Research · Web Data Mining and Analysis
参考文献 35
此处列出前 3 条
施引文献 4
按被引量排序,此处列出前 3 条