In Chap. 11 we presented placement in the context of Accreditation of Prior Learning and showed that in the scenario we address we do not assume the availability of controlled vocabulary with which the contents of the learner portfolio or the learning material in the learning network is described. Our placement service is based on the assumption that similarity between material produced or studied by the learner on the one hand and the learning material of the learning network on the other, can be used as a proxy to similarity in learning outcomes. The first task of any such placement service is therefore to establish whether these similarities are present for a given learner. The technology with which this is done, latent semantic analysis, is presented here. The emphasis here is on the technical and computational aspects of data preparation and analysis.


No posts