Automatic Standardization System for Free Text Addresses

Abstract

In cases where addresses entry should be entered as free text, it is necessary for the delivery service quality to be improved by correcting the errors and deficiencies of the addresses and standardizing them to geographic coordinate information. In our study, it is aimed to develop a system using Support Vector Machines algorithm which is used on matching free text address data with standard address. The model trained with using classified data serves to express the similarity between a free text address and a standard address as a numerical value. It is confirmed that using the developed system, queries made with free text addresses over a database created from 250.000 addresses, the system has achieved a matching accuracy exceeding 81%.