Quality Driven Source Selection Using Data Envelope Analysis

Due to the development of the Internet there is an increasing number of information sources available to users. This makes it necessary to query only the most appropriate sources. The information quality ooered by these sources can and must be a criterion for source selection. However, information quality has many dimensions, both subjective and objective, and it is thus diicult to directly compare sources with one another or give a ranking of sources. We propose to use Data Envelopment Analysis (DEA) to solve these shortcomings. This method does not directly compare sources, but focuses on individual sources, determining their eeciency in terms of information quality and cost. We adapt DEA to the task of source selection and discuss its advantages over other methods. Furthermore , we expand the model to include cost criteria and suggest a solution to additionally select sources in a query-dependent way. 1 Motivation The Situation The development of the Internet and the World Wide Web during recent years has made it possible and useful to access many diierent information systems anywhere in the world to obtain information. For a certain piece of information a user can typically choose from many similar information sources. Due to time and cost considerations it must be decided which ones to query, either by the user or by an automated process.