The feature extraction for classifying words on social media with the Naïve Bayes algorithm

To classify Naïve Bayes classification (NBC), however, it is necessary to have a previous pre-processing and feature extraction. Generally, eliminates unnecessary words while extraction processes these words. This paper focuses on in which calculations searches are used by applying word2vec frequency using term frequency-Inverse document (TF-IDF). The process of classifying Twitter with 1734 tweets defined as weight the calculation TF-IDF that often come out tweet, value decreases vice versa. Following achievement word carried test data, yielding an accuracy 88.8% Slack category tweet verb 78.79%. It can be concluded data form available twitter classified those refer slack verbs fairly good level accuracy. so manifests from habit social media user.