Improvement on Stemmer Algorithm for Indonesian Language With Spellchecker

Stemming for Indonesian text arranged by morphology rules, however the diversity of Indonesian writing text sometimes make text not standard, for example “sholat” or “shalat” (pray) it should be “salat”. This matter has effect on the stemming process. Previous algorithm is a stemming algorithm that uses novel as its dataset. The result of its algorithm found as many as 15% words unstemmed. Most unstemmed word were found due to misspelling and use of non-standard text. This study improving previous algorithm by adding spellchecker to correct the misspelling text or non-standard text on stemming process. Dataset taken from humanities and religious novel. The result of this research can improve accuracy of stemming algorithm as much as 12%.