How to Model (Personalised) Landmarks?

摘要

There are a number of models available identifying landmarks based on what we call the landmark dimensions. However, we may assume that the selection of landmarks by a human from a pool of potential landmarks additionally depends on a number of personal dimensions. In this context, we test whether a model incorporating personal interest in a specific topic as well as prior spatial knowledge or a conventional model without personal dimensions is better able to identify landmark selections by humans. Results show that none of the modelling approaches produces the kind of recall that had been expected.