A Case Study in Narne Matching

Abstract

We examined variants of a well-documented habitational surname of English origin and evaluated the performance of several common name search algorithms to determine their efficiency in identifying equivalences of many known variants of the original surname. The surname had over fifty known variants. A new algorithm to include habitational information was developed as part of the analysis of search algorithms. The new algorithm had improved performance on the data set, achieving 92% success in name matching when the best two matches were used. The new algorithm is easily automated and holds promise as a search technique for much larger data sets.