Re-ranking of educational materials based on topic profile for e-learning

As the Web grows, and more e-learning content becomes available, discovering content relevant to the user query and suitable specifically from the e-learning perspective becomes increasingly hard. This has resulted in the need to develop automatic methods for retrieving learning materials and ranking them corresponding to their significance to the specified query yet again from the e-learning perspective. Ranking of documents is known to be an essential process in information retrieval (IR) where retrieved documents are projected in the order of their expected level of significance to the query. Conventional document ranking methods are depends upon the essential measurements of similarity between documents and query. However from the e-learning viewpoint, the significance to the query is depends upon the relevancy of the document to the topic of the query and hence retrieval needs to be based on the query, an essence of the topic related to the query and the documents themselves. This work presents a topic profile based query expansion and re-ranking system that aims at retrieving and ranking documents from the web that would be suitable to extract learning material for an e-learning scenario. Topic profile contains chapters, topics and annotation terms represented as a hierarchical representation of keywords for a particular topic. This topic profile is designed for a particular course (subject), incorporating details from more than one curriculum and is a representation of the topic from a learning viewpoint. The topic profile is used for two purposes; one to generate queries to discover documents missed out from the topic profile and second to re-rank documents based on topic profile. Thus the topic profile is used for organizing learning materials from search engines to maximize topic coverage and minimize redundancy among retrieved results.