Hello everyone,
Most of you might have heard about ALEKS. This organisation has this recommendation system which help students to know about, what he/she knows? What he/she needs to learn? and what he/she is ready to learn? They do it by forming a knowledge structure graph (sort of tree diagram )which constitutes the concepts represented by circles and the relationship among them, represented by arrows. I tried to search how do they do it, but found very little.
One can imagine forming knowledge structure graph for subjects like Maths and Mathematical Physics where the concepts have hierarchy. That means to know some concepts you have to have knowledge of some concepts in prior. I am wondering how do they do it for subject like history.
Can anyone please suggest or give me some papers/material explaining about the knowledge structure and how to create one?
Thank you.
This article provide basics for building knowledge structures, I hope you find it useful.
ALEKS (Assessment and LEarning in Knowledge Spaces) is based on theory of Knowledge Spaces. The Science behind ALEKS is described in detail in this paper: http://www.latinoamerica.aleks.com/about_aleks/Science_Behind_ALEKS.pdf
This requires educational experts to build knowledge structure. An alternative approach without any requirement of experts is using Bayesian Networks or Partial Order Knowledge Structures (POKS) (described here: http://www.professeurs.polymtl.ca/michel.desmarais/Publications-Michel/jumuai.pdf)
In my knowledge, ALEKS is not used for subjects such as History (see http://www.aleks.com/about_aleks/course_products)
A list of publications relevant to ALEKS can be found at http://www.aleks.com/about_aleks/publications_kst
Carnegie Learning is another famous Intelligent Tutoring System (https://www.carnegielearning.com/) based on theory of Knowledge Tracing described in detail at: http://act-r.psy.cmu.edu/wordpress/wp-content/uploads/2012/12/893CorbettAnderson1995.pdf
Hope this helps
http://www.latinoamerica.aleks.com/about_aleks/Science_Behind_ALEKS.pdf
http://www.professeurs.polymtl.ca/michel.desmarais/Publications-Michel/jumuai.pdf
http://act-r.psy.cmu.edu/wordpress/wp-content/uploads/2012/12/893CorbettAnderson1995.pdf
Devendra has given you a more complete answer than I will, but I just wanted to point you to the book linked below which discusses the math and computational strategies.
Along the lines of the Bayesian Network answer, one possible construction is to use right/wrong responses to a representative set of questions to derive attribute rules from a formal concept lattice. The issue is whether your questions are truly representative, it would more likely tell you how questions relate to one another based on respondents answers.
[Mathematically, having a complete set of rules and having the concept lattice are the same - you can construct one from the other.]
http://www.springer.com/us/book/9783642010385
Thank you everyone for your responses, It will surely help me.
But how about a subject like history; does there exist any theory for constructing the knowledge structure for subject like history?
Also, the organisation likes Knewton and Pearson are using the concept maps theory to extend the subjects' range from mathematics and science. What is your say on this?
http://www.knewton.com/wp-content/uploads/knewton-technical-white-paper-201501.pdf
Regarding history, I don't know the answer. I can see that there might be strong relationships with disciplines that deal with social and economic concepts, but it seems like it is possible to know history without having historical context. And, this context would be represented by knowledge structures within history. I think the assumption is that any dependencies are weak for the reasons I mentioned, but I would guess that if you talked with historians you would get a different story.
For your second question, I think it is a good idea to look at knowledge relationships across disciplines. History may be a good example.
- Badges
- Science topic
- Similar topics
- Correction