Fusing hotel ratings and reviews with hesitant terms and consensus measures

Abstract

People have come to refer to reviews for valuable information on products before making a purchase. Digesting relevant opinions regarding a product by reading all the reviews is challenging. An automated methodology which aggregates opinions across all the reviews for a single product to help differentiate any two products having the same overall rating is defined. In order to facilitate this process, rating values, which capture the overall satisfaction, and written reviews, which contain the sentiment of the experience with a product, are fused together. In this manner, each reviewer's opinion is expressed as an interval rating by means of hesitant fuzzy linguistic term sets. These new expressions of opinion are then aggregated and expressed in terms of a central opinion and degree of consensus representing the agreement among the sentiment of the reviewers for an individual product. A real case example based on 2,506 TripAdvisor reviews of hotels in Rome during 2017 is provided. The efficiency of the proposed methodology when discriminating between two hotels is compared with the TripAdvisor rating and median of reviews. The proposed methodology obtains significant differentiation between product rankings.

A fuzzy aspect‐based approach for recommending hospitals

“…Using the data from Table 5, the IFNs and IVPFNs representing the sentiments towards each aspect can be calculated. Table 6 shows the IFNs applying Equations ( 2)-( 4), whereas in Table 7, the IVPFNs can be seen applying Equations ( 5)- (8).…”

“…In the literature, many aspect-based sentiment analysis applications can be found applied to rank, select, filter, recommend products, or services, such as hotels, books, DVDs, and so forth [6][7][8][9] ; nonetheless, to the best of our knowledge, little work has been done with respect to the field of hospitals, in which it is difficult to find a well-structured data set. Therefore, the main contribution of this study is the implementation of an application for recommending/ranking hospitals based on previous user reviews considering the different aspects/services of a hospital.…”

“…In Reference [8], a hybrid mechanism for ranking products based on hesitant fuzzy linguistic term sets, an associated consensus measure, and the use of the AFINN dictionary 27 is described. Multiple classifiers can simulate various experts conveying their opinions towards a product.…”

Applicability of artificial intelligence models

“…At the same time, also other consolidated AI concepts, such as ontologies, lexicons, consensus and trust, are being used in innovative ways to create practical systems. In this special issue, [6] provides some interenting examples. More in general, for many complex tasks, software systems engineering princicples can be applied effectively, for realizing structured architectures [2,5] and hierarchical systems [1,4].…”

“…In particular, the first included article [6] tackles the important task of automatic recommendations, through a structured process. The work is interesting since it avoids machine learning altogether.…”