[edm-announce] New student modeling toolkit: FAST (Knowledge Tracing with Features)

From: Stephen Fancsali <sfancsali@xxxxxxxxx>
To: edm-announce@xxxxxxxxxxxxx
Date: Tue, 7 Oct 2014 09:11:53 -0400

[on behalf of José P. González-Brenes]

Dear colleagues,

We are happy to announce that we are releasing the Feature-Aware Student
Knowledge Tracing (FAST) student modeling toolkit:

http://ml-smores.github.io/fast/

FAST has the following advantages:
* Easy to use:  FAST toolkit is easy to use, it only requires a
comma-separated file with the features you want to use. Unlike, BNT-SM,
which requires researchers to design a different Bayes Net for each feature
set they want to prototype.
* Accurate:  25% more predictive than conventional Knowledge Tracing.
* Fast:  The FAST toolkit is up to 300x faster than BNT-SM.

We presented FAST at the 7th International Conference on Educational Data
Mining (EDM 2014), where it was selected as one the top 5 paper
submissions. In a follow-up study, we compared FAST to the best EDM paper
of that year, with favorable results. This suggests that FAST has similar
performance than custom, single-purpose models, but requires much less
engineering effort.


Thanks,
Jose, Yun (Cloud) and Peter

Links:
* Software: http://ml-smores.github.io/fast/
* Slides: josepablogonzalez.com/files/fast_presentation.pdf
* Follow up Comparison:  http://ceur-ws.org/Vol-1181/pale2014_paper_01.pdf

[edm-announce] New student modeling toolkit: FAST (Knowledge Tracing with Features)

Other related posts: