Transcript 下載/瀏覽
Towards a More Accurate
Knowledge Level Estimation
Provenance IEEE 2009 Sixth International Conference on
Information Technology: New Generations.
Author Samad Kardan , Ahmad Kardan
1
Outline
Introduction
The proposed method
Credit-Blame problem
The skill hierarchies
Guess-slip problem
Implementation and results
Summary
2
Introduction
Computer Adaptive testing (CAT)
Test items are selected according to the learner’s
level of ability.
Fast、Accuracy.
Need a huge test items pool.
Possibly trying to guess.
Not fairness.
3
Computer Based Testing (CBT)
No need for a large test item pool.
Questions are fixed for all the students.
4
The proposed method
Credit-Blame problem.
When the user answers a question correctly, this
means that more credit is granted to the skill
nodes with higher level of mastery.
5
The skill hierarchies
A question may require up to three skills.
6
Knowledge model for a learning objective.
Skills
Questions
7
Guess-slip problem.
Question model
8
A possible guess.
9
A possible slip.
10
Implementation and results
The actual and predicted scores.
11
Comparison of the actual and predicted
scores.
12
Summary
In this paper, proposed a method for
estimation of the knowledge level of the
students in an e-learning system.
Question model that handles the guess-slip
problem.
knowledge model based on the learning
objects hierarchy.
13