You are here

Using adaptive feedback and recommendation engines

15 January, 2016 - 09:48

In the last years, numerous repositories of educational digital resources have been created. These repositories are added to the unclassified resources provided by Internet itself. In this overcrowded space of online educational resources, the e-learning users feel the need of services which can help them identify the proper learning objects. Recommender systems (RS) serve this purpose (Manouselis et al., 2010). A RS guides the user to interesting objects (concepts) in a large space of possible options. In educational area, RS started to spread more and more: some assist students to plan their semester schedule, by checking courses that comply with constraint regulation and with students’ preferences (Hsu, 2009), others are used at course ranking (Farzan & Brusilovsky, 2011) or to give proper knowledge to proper members in collaborative team contexts, by respecting role, tasks, members’ level of knowledge (Zhen, Huang & Jiang, 2010).

In e-assessment, recommender engines can be used to improve the selection of the tests, taking into account the educational objectives or to optimize the feedback mechanism: each student can receive bibliographical recommendations, according to one’s mistakes. RS can be also used by teachers and trainers in selecting the proper questions from a database, like in Cadmus case (Hage & Aimeur, 2005).

The current chapter proposes the use of RS in the feedback module of an e-assessment application. After an e-test is finished, the incorrectly answered questions are analyzed and the concepts which were not understood by the students are extracted. These concepts are mapped to a domain ontology. By using that ontology, a set of lexical instances for each misunderstood concept is obtained and used for internet search. Thus, bibliographical recommendations from the web search are offered to the students.