Sentiment Analysis Using Word2vec And Long Short-Term Memory (LSTM) For Indonesian Hotel Reviews

• DOI: 10.1016/j.procs.2021.01.061
• Publication Date: 2021-02-21T16:04:07Z
• Authors: Putra Fissabil Mu...; Retno Kusumaningrum; Adi Wibowo
• Keywords: Word2vec; Softmax function; Dropout (neural networks); Pooling; Sentiment Analysis

Generally, Online Travel Agent (OTA) has a review element where clients can give reviews of the facilities they have used. Availability huge volume makes it troublesome for service executives to know percent that an effect on their services. Thus, is essential develop sentiment assessment technique with respect hotel reviews, particularly in Indonesian language. This research use Long-Short Term Memory (LSTM) model as well Word2Vec model. The integration and LSTM variables used this are architecture, vector dimension, evaluation method, pooling technique, dropout value, learning rate. On basis experimental performed through 2500 texts dataset, best performance was obtained had accuracy 85.96%. parameter combinations Skip-gram Hierarchical Softmax 300 dimension. Whereas value 0.2, type average pooling, rate 0.001.