A new stemmer for Farsi language

In this paper, we report on the design and implementation of a stemmer for the Farsi language, according to combination of Kazem Taghva's method and improved Krovetz's method. The first method removes the suffixes and prefixes according to the word's structure. And the second method is based on saving the information in a Database. This paper reports a kind of combination of these methods. The results of our evaluation on a small Farsi document collection show a significant improvement in precision/recall.