There are two phases in our planned work in place name
reference (PNR) analysis. Phase I, which we present here,
involves the manual annotation of PNRs in natural
language text, and quantitative and qualitative analysis of
the results. In Phase II we expect to invest a significant
portion of our effort on incorporating standard machine
learning techniques to perform automatic annotation.
The Phase I goals are to develop software and procedural
resources to aid in the manual annotation task, and to
analyze annotations to determine the nature and extent of
the PNR ambiguity problem.