Sindhi Stemmer using Affix Removal Method

Stemming is the process of mapping various inflections of a word to its base form. Stemmer is an essential component of Information Retrieval (IR) systems and different Natural Language Processing pipelines. This research reports the development and evaluation of stemmer for a resource poor language Sindhi. The stemmer is using a lexicon-based affix removal technique for stemming. The developed lexicon represents the base forms and the algorithm uses this lexicon during the affix removal process. The overall performance accuracy is evaluated, and stemmed error rate is calculated. The results show 89.57% overall performance accuracy.