Analysis of Stemming Influence on Indonesian Tweet Classification

Abstract

Stemming has been commonly used by some researchers in natural language processing area such as text mining, text classification, and information retrieval. In information retrieval, stemming may help to raise retrieval performance. However, there is an indication that stemming does not hand over significant influence toward the accuracy in text classification. Therefore, this paper analyzes further research about the influence of stemming on tweet classification in Bahasa Indonesia. This work examines about the accuracy result between two conditions by involving stemming and without involving stemming in pre-processing task for tweet classification. The contribution of this research is to find out a better pre-processing task in order to obtain good accuracy in text classification. According to the experiments, it is observed that all accuracy results in tweet classification tend to decrease. Stemming task does not raise the accuracy either using SVM or Naive Bayes algorithm. Therefore, this work summarized that stemming process does not affect significantly towards the accuracy performance.