<![CDATA[Internet is a huge virtual space for people to share everything to others effectively including reviews. Reviews provided by someone on the internet have a big impact on other users and company. One of the most frequent reviews in internet is restaurant reviews. One restaurant review can contain several different aspects, to find out the aspects and sentiments contained in a review, an aspect-based sentiment analysis is needed. The data used in this study is restaurant review data obtained from SemEval-2016 Task 5 with 300 training data and 100 test data. To find out what aspects are contained in a review, opinion extraction is needed by doing POS tagging and extract the document into several opinions according to the basic grammar, then to classifying aspect and sentiment contained in a review, Support Vector Machine with the One-Against-All strategy is used in this research. The results of the evaluation using confusion matrix on aspect classification and sentiment classification produce precision of 0,94 and 0,86, recall of 0,6 and 0,98, accuracy of 0,88 and 0,86, and f-measure of 0,73 and 0,92.]]>
Copyrights © 2021
