International Journal on Advanced Science, Engineering and Information Technology, Vol. 12 (2022) No. 2, pages: 521-529, DOI:10.18517/ijaseit.12.2.15945

Optimization of Search Environments for Learning Contexts

Jaurès S. H. Kameni, Bernabé Batchakui, Roger Nkambou

Abstract

This article proposes an improvement of search engines in a learning or training context. Indeed, the learner requests resources or learning content in a training or learning situation. The same goes for the trainer, who wishes to select the appropriate resources available to his learners. Unfortunately, existing search engines produce an enormous mass of content but sometimes do not match the learning context, thus causing an enormous loss of time for the learner or the teacher to find the appropriate resources among this important batch. Therefore, we suggest associating a complementary layer with search engines to extract the most relevant information related to learning or training situations from the engine results. For this purpose, an integrated filter eliminates irrelevant results to the current learning or training situation; and performs a weighted reclassification of these results based on Bloom’s taxonomy. In terms of the HMI, this layer allows having more informative result snippets. The experimentation of this environment is based on Google APIs. According to the Bloom hierarchy, the classification of the user question and the classification of the search results are carried out from Natural Language Processing based on Logistic Regression of Machine Learning Algorithms. The result obtained presents an intuitively favorable environment for education, leading to the implementation of a specific search engine capable of collecting, storing, and indexing educational concepts in the next stage of this project. A project to empirically evaluate the results obtained is currently underway.

Keywords:

Information retrieval; search engine; search-as-learning; bloom’s taxonomy; natural language processing; question classification.

Viewed: 89 times (since abstract online)

cite this paper     download