Student guidance is an always desired characteristic in any educational system, but/nit represents special difficulty if it has to be deployed in an automated way to fulfil/nsuch needs in a computer supported educational tool. In this paper we explore/npossible avenues relying on machine learning techniques, to be included in a near/nfuture -in the form of a tutoring navigational tool- in a teleeducation platform -/nInterMediActor- currently under development. Since no data from that platform is/navailable ...
Student guidance is an always desired characteristic in any educational system, but/nit represents special difficulty if it has to be deployed in an automated way to fulfil/nsuch needs in a computer supported educational tool. In this paper we explore/npossible avenues relying on machine learning techniques, to be included in a near/nfuture -in the form of a tutoring navigational tool- in a teleeducation platform -/nInterMediActor- currently under development. Since no data from that platform is/navailable yet, the preliminary experiments presented in this paper are built/ninterpreting every subject in the Telecommunications Degree at Universidad Carlos/nIII de Madrid as an aggregated macro-competence (following the methodological/nconsiderations in InterMediActor), such that marks achieved by students can be/nused as data for the models, to be replaced in a near future by real data directly/nmeasured inside InterMediActor. We evaluate the predictability of students /nqualifications, and we deploy a preventive early detection system -failure alert-, to/nidentify those students more prone to fail a certain subject such that corrective/nmeans can be deployed with sufficient anticipation.
+