Merging classifiers for improved information retrieval
Anette Hulth, Lars Asker · DSpace repository (University of Tartu) · 1999
One prospective way to improve information retrieval is to use several indexing methods to retrieve different sets of documents, and then to merge (or combine) these results into one single result.The merging should be done in a way th at produces a final result that is more accurate than the output of any of the individual classifiers.A merging algorithm called S E Q U E L has been applied for this task to data in the field of information retrieval.This article describes the results of these experiments, as well as conceivable future directions.