Retrieving Business Information from the WWW

Abstract

Within this work the retrieval aspects for a system to retrieve Business Information from the WWW are presented. The WWW is seen as an important source for Business Information. In this thesis a number of problems and their solutions are faced.First the results of a study conducted within this work among business decision makers, with the objective to get an insight in their external Business Information need, are discussed. The results show that external information is very valuable for them and that Business Information from the WWW is seen to be an important source.The outcomes of this study and the review of literature led to the development of a visual information seeking system for Business Information called INSYDER. The Information Retrieval aspects of this system are in the focus of this work, whereas the visualisations aspects are discussed in [Mann 2002].It has been intended to develop a system giving a big added-value to the user. For this various components have been designed. The visualisation of the query for an interactive query expansion assists the user in the first step of the information seeking process. The proposed ranking and classification components support the user when reviewing results. Hereby the ranking analyses document by document on-the-fly. This way a comparable ranking, not relying on an overall document collection, has been achieved. For a redefinition of the initial query a relevance feedback option has also been included. The evaluation of the retrieval performance using TREC data shows the system's effectiveness.As a résumé and outlook on future work the presented components and enhancements are rearranged in a sketch of a new Business Intelligence System.