隨著科技技術的發展,數位學習越來越普及,開放教育資源(Open Educational Resources)也是其中重要的一環,只要網路觸及的地方,使用者都能根據需求輕易地取得公開分享的教學資源,許多開放教育資源平台也就隨之發展起來。開放教育資源平台提供了豐富的教學資源,然而過多的教學資源反而導致學習者不知該從何處學習、如何學習。因此本研究應用了目前熱門的大數據分析技術開發了適性化學習路徑推薦系統,使用Spark on Yarn的運算框架來建置分析環境,透過分析使用紀錄、學習行為的資料,達到適性化學習路徑的推薦。系統會將使用者自動分群,再利用序列樣式探勘演算法找出適合使用者學習的路徑。;Following the development of technology, we can see the rise in E-Learning’s popularity; Open Educational Resources (OER) is also an important part of it. User can easily get open learning materials according to their requirements, and many OER platforms been built. OER platforms provide a variety of learning materials but also let the learners wonder which to learn and how to learn. As a result, this study develops an adaptive learning path recommendation system by applying Big Data analytics technologies. The proposed system uses Spark on Yarn framework, analyzing user logs and video information to find out suitable learning path list for each user by the combination of Clustering and Sequential Pattern Mining.