Mining Modern App Stores to Improve the Perceived Quality of Software Products

移动的应用程序（app）正在成为我们日常活动中不可或缺的一部分。由于移动的市场竞争激烈，应用程序开发人员需要避免部署错误的版本，并不断部署满足用户需求的高质量版本。多年来，不断增长的移动的应用程序市场拥有数百万个应用程序和数十亿的年下载量，产生了全面的数据。挖掘移动的应用市场数据（例如，分析用户评论）可以帮助应用程序开发人员提高其应用程序的感知质量。多年来，研究已经获得了很多关注，以提供指导方针，帮助应用程序开发人员构建高质量的应用程序。然而，现有的指南没有为应用开发者提供明确的、自动可操作的建议来增强他们的应用（例如，用于修复所报告的问题或实现所请求的功能的自动方法）。此外，现有的指引并没有考虑所开发的应用程序的性质。例如，设计电信应用程序的指引可能不适合设计游戏或金融应用程序。因此，我的研究计划旨在通过将数据挖掘技术应用于新兴的软件存储库（即，应用程序市场数据），以提供自动和可定制的指导和方法，帮助应用程序开发人员提高移动的应用程序的感知质量。我的项目有两个主要的研究方向（RD）。 RD 1：支持快速发布高质量的移动的应用。我的研究旨在帮助应用程序开发人员执行快速发布（而不影响其应用程序的感知质量），如下所示。首先，我们将通过分析竞争对手的应用评论，帮助应用开发者提取需要实现的最重要的功能。在收集需求之后，开发人员需要设计他们的应用程序，以促进应用程序对所有用户的公平访问。最近的研究表明，移动的应用程序在其设计中存在可访问性问题。因此，我的研究计划旨在帮助应用程序开发人员设计他们的应用程序，以实现可访问性，允许所有用户公平访问。接下来，我的研究计划旨在提供自动推荐，以帮助应用程序开发人员在开发阶段实现所需的功能。 RD 2：保持部署的应用程序的质量。为了保持交付的移动的应用程序的质量，在部署移动的应用程序之后，我的研究旨在构建预测任何运行时问题的方法（例如，性能异常），并自动建议修复报告的错误。四名硕士和五名UG学生将接受分析大规模软件工程数据和构建工具的培训，以促进高质量移动的应用程序的快速开发。我们将把我们所有的方法作为开源项目发布，以帮助研究和开发者社区利用我们的工作。我的工作将帮助应用程序开发人员快速产生高感知质量，使软件市场在加拿大增长。

Mobile applications (apps) are becoming an integral part of our daily activities. With the mobile market's highly competitive nature, app developers need to avoid deploying buggy releases and continuously deploy high-quality releases that satisfy user needs. Over the years, the growing mobile app market with millions of apps and billions of yearly downloads generates comprehensive data. Mining mobile app market data (e.g., analyzing user reviews) can help app developers improve their apps' perceived quality. Over the years, research has gained lots of focus to provide guidelines that help app developers build high-quality apps. However, the existing guidelines do not provide clear, automatic actionable advice for app developers to enhance their apps (e.g., an automatic approach for fixing the reported issues or implementing the requested features). Moreover, the existing guidelines do not consider the nature of the developed app. For example, the guidelines for designing a telecommunication app may not be suitable for designing games or financial apps. Therefore, my research program aims to close the existing research gap by applying data mining techniques to the emerging software repositories (i.e., app market data) to provide automatic and customizable guidance and approaches that help app developers improve the perceived quality of mobile apps. My program has two main research directions (RD). RD 1: Supporting rapid releases of high-quality mobile apps. My research aims to help app developers perform rapid releases (without impacting their apps' perceived quality) as follows. First, we will help app developers extract the most important features that need to be implemented by analyzing their competitor apps' reviews. After collecting the requirements, developers need to design their apps to facilitate the app's fairness access to all users. Recent studies show that mobile apps suffer from accessibilities issues in their design. Hence, my research program aims to help app developers design their apps for accessibility, allowing fair access to all users. Next, my research program aims to provide automatic recommendations to help app developers implement the requested features in the development phase. RD 2: Towards maintaining the quality of the deployed apps. To maintain the quality of the delivered mobile apps, after deploying the mobile app, my research aims to build approaches that predict any run-time issues (e.g., performance anomaly) and automatically recommend fixes to the reported bugs. Four MSc and five UG students will be trained on analyzing large-scale software engineering data and building tools that facilitate the rapid development of high-quality mobile apps. We will release all our approaches as open-source projects to help the research and the developer community leverage our work. My work will help app developers produce rapidly high-perceived quality, enabling the software market to grow in Canada.