Discover owidi-001/delidash-customer Open Source project

Soundex-Algorithm

People can have very similar sounding names that are spelled differently. When hearing a name without knowing the spelling, someone may spell the name in several different ways. We would like a tool that can roughly group similar sounding names together to allow us to search for names when an exact spelling may not be known. Different sounding names can arise for a variety of reasons, including individual preference of parents in assigning first names and differing ways to transliterate names from other languages into the Latin alphabet, for example. There are obvious limitations and biases to approaches to group similar sounding names together: many were originally designed for names that were historically common in the United States, and likely perform better on common Western European names. They will likely not sufficiently capture differences or similarities in names from other backgrounds. Phonetic algorithms are tools used to compare similar sounding names. In this assignment, we will implement a variety of the Soundex algorithm, originally developed and patented in 1918. In the version of Soundex algorithm that we are implementing, each name will be translated into a string of the form AXXX, where A is a lower-case letter and XXX is a three-digit numerical code. We call this string the Soundex encoding of the name. Two names will be judged to be similar sounding if their Soundex encodings are the same. We are choosing the Soundex algorithm for simplicity, as it relies on a relatively small number of rules. However, as an older algorithm, Soundex will likely not perform as well as more modern algorithms on diverse groups of names.

owidi-001/delidash-customer

owidi-001

Reviews

Repository Details

More Repositories