<![CDATA[The rapid development of mobile application encourages the creation of many applications with a variety of uses to fulfill user needs. Each application allows users to post a review about the application. The aim of the review is to evaluate and improve the quality of future products. For that purpose, analysis sentiment can be used to classify the review into positive or negative sentiment. Application reviews usually have spelling errors which makes them difficult to understand. The word that have spelling error needs to be normalized so it can be transformed into standard word. Hence, words normalization is needed to solve spelling error problem. This research used word normalization based on Levenshtein distance. Based on testing, the highest accuracy is found in ratio of 70% training data and 30% testing data. The highest accuracy of this research using edit value <=2 is 100%, the second highest of edit value is obtained at edit value <=1 with accuracy of 96,4%, while edit value with the lowest accuracy is obtained at edit value <=4 and <=5 with accuracy of 66,6%. The result of using Naive Bayes-Levenshtein Distance has accuracy value of 96,9% compared to Naive Bayes without the Levenshtein Distance with accuracy value of 94,4%.]]>
Copyrights © 2017
