<![CDATA[Google Play Store has become the largest digital market place with more than 10 million products in it. Application developers make the product review column available on the Google Play Store one of the ways to find out user satisfaction. But not all reviews on the app have an alignment between ratings and comments, there are ambiguous reviews, that are reviews marked by ratings and sentiment of comments are not the same. Machine Learning (ML) has been very useful in the field of sentiment analysis. One method that is reliable and easy to use is Naive Bayes. C4.5 method is also very popular in solving the decesion tree problem which will be used for the sentiment classification process. While the Levenshtein Disance method is used to compare two strings for the word normalization process. The method flow start with text preprocessing dataset with Levenstein Distance, then the dataset will be divided into two for the Naive Bayes and C4.5 classification process. The sentiment text and text review will be processed by the Naive Bayes method while the rating and sentiment text will be processed by C4.5.The test results using the 10-Fold evaluation method are 85.3%. While the sentiment classification without using Levenshtein Distance is 85.6%, the difference is 0,3%, making the Levenshtein Distance method not significantly affect the classification results. Other test results with the application of the Edit Distance 1, 2, 3 and 4 limits were 86.9%, 85.9%, 87.1% and 86.1%, respectively. Testing Naive Bayes algorithm without C4.5 in classifying review texts has an 85.3% same result with previous test. The results of this test illustrate the effectiveness of this program in the classification of mobile application]]>
Copyrights © 2020
