在網路發達的年代，有許多英語影片資源可用來推薦給學習者進行英語學習。然而，在茫茫的影片來源中，如何根據學習者的能力給予適當的教學影片，則是一個很重要的研究課題。本研究提出利用潛在語義分析技術，比對學生不精熟概念與影片字幕來推薦適當的學習影片，補強學生不精熟概念。利用語詞權重和潛在語義分析（Latent Semantic Analysis, LSA）影片字幕（texts），再根據學生不精熟的概念推薦學習影片。共有48位大一新生參與實驗，實驗組學生將會透過推薦系統的推薦影片進行學習，對照組則根據影片列表來進行學習。實驗結果發現，基於學習者不精熟概念推薦英語學習影片可以提升學生的英語聽力和閱讀水平以及學習成績，學習成效也優於對照組。

According to technology development, more and more English video resources can be used to recommend to learners to learn English. However, among the vast video sources, how to give appropriate teaching videos according to the learners’ abilities is a critical research topic. This study proposes to use latent semantic analysis integrated with students’ non-mastery concepts to recommend appropriate learning videos to improve students’ non-mastery conceptions. Video recommendation corpus texts were analyzed using term weights and Latent Semantic Analysis (LSA) rules. Based on learners’ error responses, enhanced English instructions were developed for the video recommendation system. A total of 48 first-year students participated in the experiment. Students in the experimental group learned English with recommended videos through the proposed recommendation system, and students in the control group learned English according to the list of videos. The experimental results found that recommended English learning videos based on learners’ non-mastery concepts can improve learners’ English listening and reading levels. The learning effect is better than that of the control group.