Srovnání booleovského modelu a modelu založeného na signaturních souborech

Abstract

The aim of the study was to compare boolean model and model based on signature files. The study presents the theoretical foundations of both models, so here is described how they represent documents and how they approach to the evaluation of the queries entered by the user. Created applications were used for comparison, where the index creation was compared in terms of required time for the creation and the memory that index occupies on the disk. After that the evaluation results were compared for different queries and time, which was needed for evaluation in both models. For the reason, that signature files select for the query even those documents, that are not relevant, so a comparison of set values was made to determine the best settings, which showed the best results.