Toward Machine-based Address Matching of Descriptive Addresses

Addresses are among the most common resources to locate geographical places. It is estimated that over 80% of data used by the governments has spatial dimensions in nature, most of which is connected to addresses in a way or other. Therefore, they are a relevant resource for spatial enablement of databases. Addresses relate to their corresponding locations on maps through the so-called address matching process. The importance of addresses and their corresponding matching process is so important that developed countries made considerable efforts in order to generate and maintain databases of their national addresses. Different countries have different conventions for addressing, which are classified to standard and non-standard (or descriptive). Although the latter class is more compatible to people’s spatial thinking, their automatic address matching is more difficult, which are discussed in this paper.