Performance Evaluation of Phonetic Matching Algorithms on English Words and Street Names - Comparison and Correlation

Abstract

Researchers confront major problems while searching for various kinds of data in a large imprecise database, as they are not spelled correctly or in the way they were expected to be spelled. As a result, they cannot find the word they are looking for. Over the years of struggle, relying on pronunciation of words was considered to be one of the practices to solve the problem effectively. The technique used to acquire words based on sounds is known as "Phonetic Matching". Soundex is the first algorithm proposed and other algorithms like Metaphone, Caverphone, DMetaphone, Phonex etc., have been also used for information retrieval in different environments. This paper deals with the analysis and evaluation of different phonetic matching algorithms on several datasets comprising of street names of North Carolina and English dictionary words. The analysis clearly states that there is no clear best technique in general since Metaphone has the best performance for English dictionary words, while NYSIIS has better performance for datasets having street names. Though Soundex has high accuracy in correcting the misspelled words compared to other algorithms, it has lower precision due to more noise in the considered arena. The experimental results paved way for introducing some suggestions that would aid to make databases more concrete and achieve higher data quality.

Automated Linking of Historical Data

“…Second, in real ground truth data based on comparing two independent transcriptions of the 1940 census, we find that while automated methods miss many links because of transcription differences, the links they create are almost 100% correct. Third, Bailey et al (2019) continues to report results that mix the Abramitzky, Boustan, and Eriksson approach with an outdated name standardizing algorithm (Soundex) that is not used in contemporary linking papers (see Koneru et al (2016), which suggests that NYSIIS result in fewer false positives than Soundex). Not surprisingly, this method is reported to have the highest false positive rates (43%).…”

Implantation of indexing optimization technology for highly specialized terms based on Metaphone phonetical algorithm

“…It is possible to solve the specified issue by using phonetic algorithms that compare words such as the modified Metaphone algorithm given in Listing 1. It takes into consideration, first of all, peculiarities in the formation of Ukrainian last names (Table 1) and the titles of medicines [17], and, in contrast to those existing [4,5,11], makes it possible to form phonetic transformations for indexes of highly specialized Ukrainian words/terms.…”

“…An attempt to adapt several algorithms for indexing place names was made in [11]. Due to a small sample of the experimental data and a small divergence between the results from the proposed algorithms, it is difficult to give an objective estimation of the quality of operation of each separate one.…”

Social Networks and (Political) Assimilation in the Age of Mass Migration

“…At last, if no match is found, we look for unique matches within a two-year duration period for people whose first and last name are the same based on the New York State Identification and Intelligence System Phonetic Code. This phonetic code has been shown to result in fewer false positives than alternative algorithms (Koneru et al, 2016). In each of these steps, only unique matches are accepted.…”