Recency, frequency, quality: novel feature from sentiment analysis for clustering and ranking in tourism big data analytics

Abstract

Understanding tourist perceptions has been a key benefit of sentiment analysis in tourism data. However, its outcomes can be further utilized to gain insights into the characteristics of tourist attractions and hotels. This study aims to develop a new feature, called recency, frequency, quality (RFQ), derived from sentiment analysis results to cluster and rank tourist attractions and hotels in Bali. RFQ consists of three components: review recency, review frequency, and review quality. These dimensions reflect the recentness of reviews, the popularity based on the number of reviews, and the review quality measured by the ratio of positive to negative sentiment polarity. Using big data analytics through clustering and ranking, the study finds that the quality of tourist attractions and hotels is primarily concentrated in Badung and Gianyar regencies. More tourist attractions are found in the silver cluster than in the gold, indicating the need to enhance quality. In the hotel sector, the diamond cluster dominates among star-rated hotels, suggesting overall high quality. Budget hotels show fairly good quality, with most falling under the gold cluster.