<![CDATA[Hotel reviews not only provide useful information for business owners but also shape the image of the hotel in the eyes of customers. Reviews generally cover various aspects expressed honestly by customers. In the era of technological development, the number of hotel reviews online is increasing, making the processing of hotel assessments discussed in reviews a challenge for many parties. To overcome this, a multi-aspect sentiment analysis has been developed to help extract more specific information about hotel evaluations from each review sentence. The aspects evaluated in the reviews include price, location, service, food, facilities, and rooms. In developing the multi-aspect sentiment analysis model, a deep learning method based on LSTM is used. The LSTM model architecture is built using a sequential model with four layers: embedding, SpatialDropout1D, LSTM, and dense. The model is trained with 10 epochs and a batch size of 32. The model is evaluated through three scenarios, including testing sentences with one aspect, testing sentences with a combination of two aspects, and testing sentences with a combination of three aspects. Top-1 and Top-2 accuracy are applied to test sentences with a combination of two and three aspects. Meanwhile, F1_score is used for testing sentences with one aspect and sentiment analysis. The obtained accuracy results are 79% for Top-2 accuracy in sentences with a combination of two and three aspects, 85.7% for F1_score in sentences with one aspect, and 83% for F1_score accuracy in sentiment analysis. These results indicate that the developed model is capable of performing multi-aspect sentiment analysis on hotel reviews.]]>
Copyrights © 2023
