Comparing the performance of index based searching and ontology based application of OLAP for information retrieval

Abstract

Analyzing the data types and structures in a business environment where data grows rapidly has become a 
serious challenge for companies. As most business data are unstructured and modelled into complex models, we 
need a solution to perform multidimensional searching for achieving sustainable results. In this paper, we present 
the idea of incorporating index searching as part of a standard information retrieval system. We test our concepts 
by using business documents from Knauf Radika AD, a leading Macedonian plasterboard manufacturer located 
in Diber. Further, we describe and propose a novel architecture which is including an ontology-based approach 
by integrating OLAP and information extraction attributes to access structured and unstructured data, mainly 
organized in form of documents. In our first demonstration, the query performance was reduced as more 
documents were added to the index, and consequently the growth factor becomes very low. While at the second 
case, when a user performs navigation throughout the OLAP report, it is possible to track the user context 
information which can be used for searching other relevant documents.