Towards Parameter-Free Data Mining: Mining Educational Data
Marta Zorrilla, Diego García-Saiz and José Balcázar
In: The 4th International Conference on Educational Data Mining, 6-8 July 2011, Eindhoven, The Neederlands.
The need of parameter-free alternatives for data mining algorithms is widely recognized, and is particularly acute in the webbased educational field, where instructors involved in the teaching process are interested in improving their virtual courses and adapting these to the learners’ behavior: most often, these instructors are not expected to know about data mining technologies. We report on a quantitative comparison of several algorithms for association rules on educational datasets, including in the comparison both well-established implementations and a recent parameter-free association miner; our purpose is to clarify whether this newer approach is actually useful for the educational data mining field. Our results indicate that it has indeed a high potential, and allows us to identify some important aspects that must be improved.
|EPrint Type:||Conference or Workshop Item (Paper)|
|Project Keyword:||Project Keyword UNSPECIFIED|
|Subjects:||Theory & Algorithms|
|Deposited By:||Diego García-Saiz|
|Deposited On:||21 February 2012|