How to help end users to get better decisions? Personalising OLAP aggregation queries through semantic recommendation of text documents
Talita dos Reis Lopes Berbel, Sahudy Montenegro González · International Journal of Business Intelligence and Data Mining · 2015
This paper describes an effective solution for recommending textual OLAP over data warehousing environments. The recommendation process is based on text semantics and query personalisation to improve the relevance of the retrieved results. In order to aggregate and recommend documents, we need a measure of semantic similarity. The first issue we addressed was the meaning of similarity between two concepts. For this, we used ontologies and the distance between terms over the ontology based on the least common ancestor. The second issue we dealt with was the meaning of similarity between two documents. For that, we calculated the statistical metric frequency. The purpose of query personalisation is to offer to the user an interactive way for obtaining relevant aggregation of documents based on adjustable parameters. We implemented the solution for multidimensional analysis over PubMed database. In the case study, we used the Medical Subject Headings provided by the US National Library of Medicine. At the end, we present the results of some experiments that show that good recommendations are possible. The results are discussed based on the evaluation metrics: precision, recall and F1-measure.