| Users find information on the web via evaluating alternatives based on a set of preferences, and selecting the most suitable ones. However, not all of these preferences are equally important some of them are more significant than the others. If a given alternative does not satisfy high priority preferences, it is discarded without checking against low priority criteria. The lexicographic approach allows for mimicking user's attitude that some criteria should be satisfied before other criteria are considered.The thesis describes two different versions of web Personal Evaluation Tool (PET) utilizing a novel approach for selecting the most suitable alternatives regarding information that fit user's needs. In both versions, the approach follows the concept of lexicographical preferences and combines it with a simple mechanism of representing user's criterion satisfaction levels. Additionally, the second version incorporates customer reviews. The performed experiments show that the proposed approach is consistent with human-like selection processes. |
