Query expansion strategies for laypeople-centred health information retrieval

Abstract

One of the most common activities on the web is the research for health information. This activity has been gaining popularity among users, but the majority of them have no training in health care, which leads to difficulties in understanding the terminology and contents of documents.In the field of health information retrieval various investigations have been carried out, which resulted in methodologies that offer solutions to improve the quality of the retrieval documents. One of the most covered techniques in this area is the query expansion, that solves one of the biggest difficulties for users in the search of health information: the limited knowledge of medical terminology. This lack of knowledge influence the formulation of queries and the expectations of the retrieval documents. The query expansion complements the original query with additional terms, making it more reliable. These new terms can be obtained through thesaurus containing several terms associated with a medical concept.The amount of research conducted on the issue of readability of the documents is greatly reduced, the most developed subject is relevance, but if a document is relevant and the user does not comprehend it's contents it ceases to be useful.In this thesis it will be proposed a methodology to improve the quality of the retrieval documents, using methods to improve the users queries, such as the query expansion, and it will be used Readability formulas to determine the level of education required to understand a document. Will be conducted several tests to determine if the source to be used in the query expansion and the readability will have an effect in the retrieval process. These tests will be evaluated with precision and NDCG in the case of relevance, and in the case of readability it will be used uRBP.