PewePro 2

Limited Time Recommending

A method for personalized recommendation of learning objects with considering the limited time for learning before an exam.

Overview
Inputs: student's knowledge model, learning goals (concept importance), learning content metadata
Outputs: list of recommended learning objects

Addressed Problems

Our goal is to help the student achieve the best exam result possible. A common tactic used by students, which are under time pressure, is going through all topics in the curriculum quickly rather than learning some of the topics in detail and leaving out the remaining ones completely. Using such strategy, they often end up with none of the topics learnt at least to some minimal level to be able to answer at least some of the tasks given during the exam. Our method for personalized recommendation is designed to help the students to prepare for the exam by covering as many topics as possible in limited time.

Description

To achieve proper learning time distribution between all required concepts, we attempt to determine optimal knowledge levels of all concepts at the end of learning time, which are achievable at the current learning speed. Using the sum of all concept knowledge level from the beginning of learning to present time, we estimate the knowledge level increase from present time to the end of learning (the given milestone, i.e. exam time). The expected overall increase is then divided between all concepts in such way, that the final estimated knowledge levels of concepts correspond with the concept importance given by the teacher.

At the same time, we try to prevent gaining very little knowledge of many concepts - this would occur, if the student's knowledge is low and the learning time is extremly limited. The student naturally cannot pass the exam with such knowledge. Therefore, we set a minimal concept knowledge level – the estimated knowledge level for every concept is never lower than this limit.

In the recommendation phase, three orthogonal criteria are used to evaluate each learning object. The learning objects, which best match all three criteria, are recommended to the student.

Publications

  1. Michlík, P., Bieliková, M.: Exercises Recommending for Limited Time Learning. In: Proceedings of the 1st Workshop on Recommender Systems for Technology Enhanced Learning (RecSysTEL 2010), Procedia Computer Science, Vol. 1, Issue 2, pp. 2821-2828.