Figure 5.2 shows an example of this model. Just as in the abstract model of
ranking, various features are extracted from the document. This picture shows
only a few features, but in reality there will be many more. These correspond
to the fi(D) functions in the equation just shown. We could easily name these
ftropical(D) or ffish(D); these values will be larger for documents that contain
the words “tropical” or “fish” more often or more prominently.