A recommender system for learning objects on a learning management system: Implementation experience

Abstract

This paper aims to present the experience of implementing a recommender system for selecting learning objects (LO) for lessons through the Learning Management System (LMS) of Costa Rica Institute of Technology. Our implementation approach is made up of three components: A user model based on the descriptions of the lessons of a course given in an Instructional Design (ID), a serie of items which constitute the metadata bank of resources described using IEEE Standard for Learning Object Metadata model (IEEE-LOM) and harvested by Open Archives Initiative Protocol for Metadata Harvesting (OAI-PMH) and finally, an utility function derived from the weighting between the collaborative assessment of LO and similarity measures to compare LO's metadata with descriptions of activities given for course. As validation mechanism, ID's tests were conducted in order to determine whether or not recover resources for lessons and precision tests with a pilot plan remained pending.