engleski

Guiding Students towards Frequent High-Utility Paths in an Ill-Defined Domain

This paper presents an exploratory data mining methodology for discovering frequent high- utility learning paths from a database of student interactions with an adaptable tutoring system. The discovered paths are used to present recommendations to students in order to make the learning process more efficient. The novelty of our approach is twofold: a) the process of data preparation, path evaluation and path discovery is completely autonomous ; and b) the process is executed on a growing dataset of learning traces while the students are advancing through the knowledge domain. We present the system overview and the obtained results.

Sequential pattern mining; computer-based learning environment; high-utility patterns; recommendations

nije evidentirano