Abstract:

The act of communication of a learner's knowledge is most of the time in the form of a paragraph, in a context. This paragraph, when evaluated, gets cognitively classified to a domains/ sub-domains in that context by the evaluator. The degree of accuracy in classification depends on the level of expertise and experience of the evaluator. The challenge is translating this process of almost automatic and natural classification into a computational model. This work proposes a method of classification of a learner generated paragraph in a technical context by assigning weighted score. It also takes into account the comprehension index of the learner as a direct measure of the number of domain relevant keywords in the generated paragraph. This work proposes a means of classification based on matching keywords extracted from the paragraph to keywords extracted from a reliable knowledge base. The extensive use and availability of knowledge sources on the persistent storage media, web or cloud makes it possible for multiple sources to be considered for classification into most appropriate domain. The classified content can then be used for purposes such as furthering learner comprehension of a sub-domain and related sub-domains and assessment of the extent learner's existing knowledge.

Keywords:

Domain, Classification, assessment, comprehension.