在「素養導向」為主軸的課綱設計下，教師為了提供跨領域且多元的課程內容，教學重心必須從過去單純的教科書使用者，轉而成為教學內容的設計者，承擔起課程設計的責任，尤其在進行跨領域教材設計時，不但考驗教師知識轉化能力，同時也將面對尋找課程主題相關參考書目挑戰。本研究建置課程相關書籍推薦模型，從TAAZE讀冊網站爬取147,990本書籍資訊，利用文字探勘技術分析11個國小及高中跨領域課程教案，提供課程關鍵字的關聯分析，以隱含狄利克雷分佈模型成功找尋隱含於課程中的主題，再以餘弦原理計算課程與書籍間的距離，並給予課程相關的書籍推薦。問卷結果顯示書籍推薦系統能夠提高課程品質及教師的工作效率，減少教師在尋找課程資訊的時間與壓力，有助於教師在備課時取得課程相關的學習資源，使教師持續朝核心素養的教學精神邁進。

Under the curriculum framework centered on "competency-based education," teachers are shifting from merely using textbooks to becoming designers of instructional content. To provide interdisciplinary and diverse course materials, they must tackle the challenges of curriculum design, particularly in developing interdisciplinary teaching materials. This process tests teachers’ ability to transform knowledge and presents difficulties in accessing relevant reference books for course topics. This study establishes a book recommendation model for curriculum development by extracting data from the TAAZE online bookstore, gathering information on 147,990 books. Using text mining techniques, we analyze lesson plans from 11 interdisciplinary courses in elementary and high school, conducting keyword association analysis to identify key themes. Through the Latent Dirichlet Allocation (LDA) model, we uncover implicit topics within the courses, and by applying cosine similarity, we calculate the distance between courses and books to generate relevant book recommendations. The results indicate that the book recommendation system enhances course quality and improves teachers’ work efficiency by reducing the time and effort required to find course-related materials. This, in turn, helps teachers access essential learning resources during lesson preparation, supporting the continued implementation of competency-based teaching practices.