Extensive hotel reviews classification using long short term memory
Abid Ishaq,Muhammad Umer,Muhammad Faheem Mushtaq,Carlo Medaglia,Hafeez Ur Rehman Siddiqui,Arif Mehmood,Gyu Sang Choi
DOI: https://doi.org/10.1007/s12652-020-02654-z
IF: 3.662
2020-11-08
Journal of Ambient Intelligence and Humanized Computing
Abstract:Reviews of users on social networks have been gaining rapidly interest on the usage of sentiment analysis which serve as feedback to the government, public and private companies. Text Mining has a wide variety of applications such as sentiment analysis, spam detection, sarcasm detection, and news classification. Reviews classification using user sentiments is an important and collaborative task for many organizations. During recent years, text classification is mostly studied with machine learning models and hand–crafted features which are not able to give promising results on short text classification. In this research, a deep neural network–based model Long Short Term Memory (LSTM) with word embedding features is proposed. The proposed model has been evaluated on the large dataset of Hotel reviews based on accuracy, precision, recall, and F1-score. This research is a classification study on the hotel review sentiments given by guests of the hotel. The results reveal that the proposed model performs better as compared to the existing state-of-the-art models when combined word embedding with LSTM and shows an accuracy of 97%, precision 83%, recall 71%, and F1-score 76.53%. These promising results reveal the effectiveness of the proposed model on any type of review classification tasks.
computer science, information systems,telecommunications, artificial intelligence