Toponym Disambiguation in Natural Language Processing

In recent years, geography has acquired a great importance in the context of Information Retrieval (IR) and, in general, of the automated processing of information in text. Mobile devices that are able to surf the web and at the same time inform about their position are now a common reality, together with applications that can exploit this data to provide users with locally customised information, such as directions or advertisements. Therefore, it is important to deal properly with the geographic information that is included in electronic texts. The majority of such information is contained as place names, or toponyms. The objective of this Ph.D. thesis is to study toponym ambiguity and the effects of its resolution in applications such as Geographical Information Retrieval (GIR), Question Answering (QA) and Web Retrieval. In GIR toponym ambiguity represents an important issue, due to the fact that queries are geographically constrained. There has been a struggle to find specific geographical IR methods that actually outperform traditional IR techniques. Toponym ambiguity may constitute a relevant factor in the inability of current GIR systems to take advantage from geographical knowledge. The work presented in this thesis starts with an introduction to the applications in which Toponym Disambiguation (TD) may result useful, together with an analysis of the ambiguity of toponyms in news collections. It could not be possible to study the ambiguity of toponyms without studying the resources that are used are place names repositories; these resources are the equivalent to language dictionaries, which provide the different meanings of a given word. It will be shown that the choice of a particular toponym repository should be done as a result of an analysis of the task that it is going to be carried out or depending on the specific kind of application that it is going to be developed. The choice of a proper Toponym Disambiguation method is also key. In this work two methods, a knowledge-based method and a map-based method, were developed and compared over the same test set. A case study of the application of TD methods to a corpus of Italian news is presented. The effects of the choice of a particular toponym resource and method in GIR have been studied, showing that TD may result useful if query length is short and a detailed resource is used. It has been found that the level of errors in disambiguation is not relevant, even in the case the errors represent 60% of the total number of toponyms in the collection, if the resource used has a little coverage and detail. Ranking methods that sort the results on the basis of geographical criteria were observed to be more sensitive to the use of TD or not, especially in the case of a detailed resource. It has been observed that the disambiguation of toponyms does not represent an issue in the case of Question Answering, because errors in TD are usually less important than other kind of errors in QA. In GIR, the geographical constraints contained in most queries are area constraints, such that the information need of users can be resumed to the form “X in P”, where P is a place name, and X represents the thematic part of the query. A frequent issue is when place name cannot be found in any resource because it is a fuzzy region or a vernacular name. In order to overcome this issue, Geooreka!, a prototype search engine with a map-based interface, was developed. A preliminary testing of this system is presented in this work. The work carried out on this search engine showed that Toponym Disambiguation can be particularly useful on web documents, especially for applications like Geooreka! that need to estimate the occurrence probabilities for places.