<![CDATA[Along with the increasing competition in hotel business, every hotel tries to improve their quality for increasing their profits. Hotel can improve their quality by understanding hotel reviews that written on the internet. However, the variety of types of review made hotels difficult to analyze the type of sentiment on review. In addition, the distribution of sentiment types in the reviews was unbalanced. Therefore, analysis sentiment is carried out to determine the sentiment of hotel reviews easily. The method that used by researcher is Boosting Weighted ELM because this method can handle unbalanced class. Sentiment analysis determine by doing some pre-processing, term weighting, normalization, and classification. Testing process were carried out using k-fold cross validation with k is 5. Data that used were 500 data consisting 343 positive class and 157 negative class. Testing result shows that the model is produced with the highest f-measure value is 0,953. Optimal value of each parameter are C =16, L = 64 and weak learner = 256.]]>
Copyrights © 2019
