郑隆威， 冯园园， 顾小清
【关 键 词 】：
学习分析； 学习成果； 布鲁姆分类学； 自然语言处理； 词向量
Learning outcome describes the knowledge, skills, and abilities that learners are expected to acquire at the end of the courses. Many studies have advocated that learning outcomes need to be formulated in accordance with Bloom's taxonomy so that learning outcomes are "measurable" in cognitive domain, but few studies have focused on whether the description of learning outcomes matches the measured results. This paper attempts to explore this problem by means of learning analytics. This paper uses natural language processing to test whether verbs and contextual information etc. can mark the cognitive types of learning outcomes effectively. It is found that verbs are still the most critical feature in Bloom's taxonomy. When combined with learning content and contextual information, verbs can indicate the cognitive types of learning outcomes more accurately. In addition, this paper discusses the ambiguity of learning outcomes in Bloom's taxonomy, and finds that this ambiguity can bring more contextual information, thus assisting the formulation of learning outcomes more comprehensively.