Improving text summarization of online hotel reviews with review helpfulness and sentiment

The considerable volume of online reviews for today's hotels are is difficult for review readers to manually process. Automatic review summarizations are a promising direction for improving information processing of travelers. Studies have focused on extracting relevant text features or performing sentiment analysis to compile review summaries. However, numerous reviews contain nonspecific or nonsentimental content, hindering the ability of sentiment-based techniques' to accurately summarize useful information from hotel reviews. This paper proposes a systematic approach that first constructs classifiers to identify helpful reviews and then classifies the sentences in the helpful reviews into six hotel features. Finally, the sentiment polarities of sentences are analyzed to generate the review summaries. Experiment results indicated that the performance of the proposed approach was superior to other methods.