Analysis of string encoding functions during de-duplication process

Abstract

The present research is aimed to help users to identify which encoding functions are more effective than others in terms of data matching quality. Therefore, we have carried out an evaluation of data matching considering Soundex Daitch-Mokotoff [1], Fuzzy Soundex [2] and Modified Soundex [3] in terms of precision-recall and f-measure. As far as we know such comparison against these encoding functions has not been presented before in such terms. The Daitch Mokotoff Soundex function has a better performance during string comparison. However, its execution time is greater than Fuzzy Soundex and Modified Soundex. This is justificable since the Daitch-Mokotoff Soundex algorithm is more complex and meticulous.