QUX : 問卷基底GUI模型之軟體架構

Abstract

在現今的社會，手機應用程式已經十分普及，人們透過這些程式來協助完成他們各式各樣的任務，像是時間管理、購物、通訊等等。但跟電腦應用程式不同的是，手機應用程式的開發通常要同時面對時間上的壓力和廣大同類型應用的競爭。此時，除了效率和創新以外，另一個能令其脫穎而出的關鍵就是GUI的設計。但UI設計是個很耗時的工程，需要不斷根據反饋和測試結果進行調整，確保最終的設計能夠滿足用戶的期望。正因如此，我們希望能找出方法，為UI開發過程提供有效的自動化支持。在本論文中，我們利用問卷蒐集回答作為模型訓練用的標籤，進一步結合機器學習的技術去做出一個能夠針對給定的GUI，自動回答問卷問題的模型。我們希望能透過這個模型自動化地檢測和識別UI設計中的問題，並向開發者提供建議，進而提高整個UI開發過程的效率，使設計師和開發者能夠更好地集中精力在創造性的方面，而不是繁瑣的錯誤修復上。

In today's society, mobile applications have become ubiquitous, assisting people in various tasks such as time management, shopping, and communication. However, unlike computer applications, the development of mobile applications typically involves time constraints and fierce competition among similar apps. In addition to efficiency and innovation, another key factor that distinguishes an app is the design of its Graphical User Interface (GUI). UI design is a time-consuming process that requires continuous adjustments based on feedback and testing to ensure the final design meets user expectations. Therefore, we aim to find methods to provide effective automation support for the UI development process. In this thesis, we use questionnaires to collect responses as labels for model training. We further integrate machine learning techniques to develop a model capable of automatically answering questionnaire questions for a given GUI. Our goal is to automate the detection and identification of issues in UI design, providing suggestions to developers. This approach aims to enhance the overall efficiency of the UI development process, allowing designers and developers to focus more on creative aspects rather than tedious error fixes.