Tiếng Việt

Abstract

In the context of digital transformation in education, creating electronic tests from heterogeneous sources remains a significant challenge. This research presents a comprehensive automation solution built on the Google Apps Script platform, capable of flexibly processing data from diverse sources such as Google Docs, PDFs, and websites. By utilizing regular expressions for text analysis and a dedicated algorithm for image mapping, the system converts raw exam data into a structured question bank based on a three-layer architecture. The solution was evaluated through experiments with a dataset of 60 exams and a survey involving 50 teachers at the Ho Chi Minh City University of Education. Quantitative results indicate that the tool reduces processing time by over 90%, achieves a 91% extraction accuracy for multiple-choice questions, and received a high user satisfaction rating (4.4/5 points). In conclusion, the solution has demonstrated high effectiveness and feasibility, providing a useful tool for teachers to optimize their question bank creation and management workflow, aligning with the TPACK framework.