Development of Soundex Algorithm for IsiXhosa Language

In multilingual environments, users have the preference of searching using their preferred language. Information retrieval systems have a to cater for this different languages. Searching for a document or name of person involves knowing the exact spelling of that particular document or name. If the system has been developed under few typographical errors and the the user supplied an incorrect term, it is difficult to retrieve the information under those circumstance. It would be good if we search names and words using the sound rather their spelling. This book presents the development of soundex algorithm for isiXhosa language, which can be used in information retrieval applications to search words based on their sound rather than their spelling. Soundex is a phonetic algorithm which indexes words using the sound. This algorithm is therefore language dependent. It requires an in depth linguistic knowledge to make suitable for a particular language. Specific rules have been made to cater for the Xhosa language and the results are quite satisfactory.