STEMMING FOR TERM CONFLATION IN MALAY TEXTS

摘要

Stemming is an important process to improve the performance of an information retrieval. It reduces the variant word forms to common forms. Various algorithms for stemming have been developed for the English and foreign languages. Due to its essence and popularity, research for stemming Malay words has been extended for retrieval of Malay documents. The existing Malay stemming algorithm is studied and new algorithm is proposed to improve the performance of the stemming process especially for a specific domain. The modified algorithm utilizes important morphological aspects of Malay language. Experiments on the ability of the stemmer were done and the results indicating major improvement in terms of less errors and greater rate of correctness are shown.