Poon, Chung KeungChung KeungPoonXie, HaoranHaoranXieWang, Philips Fu LeePhilips Fu LeeWangWong, Tak LamTak LamWongYu, Y. T.Tang, C. M.2021-07-112021-07-112017https://repository.sfu.edu.hk/handle/sfu/850Designing a good curriculum or an appropriate learning path for learners is challenging because it requires a very good and clear understanding of the subjects concerned as well as many other factors. One common objective of educational data mining and learning analytics is to assist learners to enhance their learning via the discovery of interesting and useful patterns from learning data. We have recently developed a technique called <i>skill2vec</i>, which utilizes an artificial neural network to automatically identify the relationship between skills from learning data. The outcome of <i>skill2vec</i> can help instructors, course planners and learners to have a more objective and data-informed decision making. <i>Skill2vec</i> transforms a skill to a vector in a new vector space by considering the contextual skills. Such a transformation, called <i>embedding</i>, allows the discovery of relevant skills that may be implicit. We conducted experiments on two real-world datasets collected from an online intelligent tutoring system. The results show that the outcome of <i>skill2vec</i> is consistent and reliable.enconference proceedings