<![CDATA[In an effort to utilize technology in public services, the Malang Office of Communication and Information has launched the SAMBAT Online web application (Integrated Questions Society Application System) to accommodate criticism, suggestions, and complaints given by the public. To improve time efficiency and make it easier for admins to classify incoming complaints the text classification method is needed. The Naive Bayes Multinomial method is widely used because this algorithm is very simple and efficient. But the Naive Bayes Multinomial algorithm has the disadvantage of having dependence on the amount of data. To improve these deficiencies researchers used a support method as feature extraction, N-gram. The test results using the Multinomial Naive Bayes method and N-gram show that the unigram n-gram can provide the highest accuracy rate of 88.23% with an average overall accuracy of 80.88% with an f-measure value of 0,8013.]]>
Copyrights © 2020
