<![CDATA[Twitter is one of the social media whose users are increasing every day. With the ease of spreading information through Twitter, some people use Twitter as a place to promote their wares by writing advertisements. These ads for some Twitter users tend to be annoying, because they are irrelevant to the information they want to read. These ads can be classified so that they can separate advertising tweets from non-advertising tweets. The ad tweet classification process is carried out using the Naive Bayes method. In the classification process, tweet data is collected to be used as training data and test data, then preprocessing is carried out on the tweet data which will then be weighted using the term frequency method. The features used in the classification with Naive Bayes are the bag-of-word feature, the textual feature is time and link, and the numeral feature is money and phone number. The classification results are obtained from the comparison of the posterior results obtained from each class. The performance level of the results obtained by the Naive Bayes method using the bag-of-word feature has a precision value 0,96, recall 1, f-measure 0,98, and accuracy 0,98.]]>
Copyrights © 2022
