Transliterated arabic name search

We address name search for transliterated Arabic given names. In previous work, we addressed similar problems with English and Arabic surnames. In each previous case, we used a variant of Soundex and n-grams to improve precision and recall of name matching compared against well known approaches such as the Russell Soundex algorithm. Unlike prior work, the proposed approach does not rely upon Soundex algorithms. We experiment with combinations of n-grams of varying lengths. Our previous work focused on two character n-grams. As with our prior work, this approach uses standard SQL and remains portable to different relational database engines, demonstrated by implementing test in SQLAnywhere and Teradata environments.