Abstract:

The use of discussion forums in online teaching environments has grown significantly in recent years. They have become an essential tool in any course that is taught online through an LMS (Learning Manegement System) such as Moodle. They generate so much information that often finding it, is difficult.
This paper describes an intelligent system based on text mining that finds a set of concepts which identify the semantic contents of a text that receives as input. These concepts can be found directly in the text, being a synonymous term, being an education category to which belongs the term or a related word based on experiences with previous occurrence.
To achieve this, it has been necessary to develop a glossary of terms from the educational environment. This glossary provides terms along with the category to which they belong to. The categories are: competencies, content, activities, evaluation system, timing, tutoring, student activities, program and platform. These categories have been obtained after qualitative analysis of several forums, select terms and, with the help of experts from the educational domain, the categories have been selected and allocated concepts.
The system uses a data mining process to:
(i) extract terms that appear in the text,
(ii) find concepts and associated categories,
(iii) calculate statistical measures based on appearance of terms in the current document and in documents previously processed, and
(iv) in accordance with all of them, assign a weight to each extracted term that represents how important it is the term for the current document.
The system is evaluated based on quantitative measurements of text mining field and qualitative measures that reveal users experience with the system.

Keywords:

e-learning, discussion forums, text mining.