Assessing Review Recommendation Techniques under a Ranking Perspective

Maroun, Luciana B.; Moro, Mirella M.; Almeida, Jussara M.; Silva, Ana Paula C.

doi:10.1145/2914586.2914598

Cited by 5 publications

(5 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For each domain, the constituent reviews in a reviewed item are first partitioned into three sets, using 80%, 10%, and 10% of the data respectively for training, validation, and testing. In particular, chronological split [44,54,49,58,92] is applied over randomization to preserve the review order information.…”

Section: Categoriesmentioning

confidence: 99%

Exploiting Review Neighbors for Contextualized Helpfulness Prediction

Du,

Rong,

Wang

et al. 2020

Preprint

View full text Add to dashboard Cite

Helpfulness prediction techniques have been widely used to identify and recommend high-quality online reviews to customers. Currently, the vast majority of studies assume that a review's helpfulness is self-contained. In practice, however, customers hardly process reviews independently given the sequential nature. The perceived helpfulness of a review is likely to be affected by its sequential neighbors (i.e., context), which has been largely ignored. This paper proposes a new methodology to capture the missing interaction between reviews and their neighbors. The first end-to-end neural architecture is developed for neighbor-aware helpfulness prediction (NAP). For each review, NAP allows for three types of neighbor selection: its preceding, following, and surrounding neighbors. Four weighting schemes are designed to learn context clues from the selected neighbors. A review is then contextualized into the learned clues for neighbor-aware helpfulness prediction. NAP is evaluated on six domains of real-world online reviews against a series of state-of-the-art baselines. Extensive experiments confirm the effectiveness of NAP and the influence of sequential neighbors on a current reviews. Further hyperparameter analysis reveals three main findings. (1) On average, eight neighbors treated with uneven importance are engaged for context construction. (2) The benefit of neighboraware prediction mainly results from closer neighbors. (3) Equally considering up to five closest neighbors of a review can usually produce a weaker but tolerable prediction result.

show abstract

Section: Categoriesmentioning

confidence: 99%

Exploiting Review Neighbors for Contextualized Helpfulness Prediction

Du,

Rong,

Wang

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…To the best of our knowledge, none of the existing methods estimates the helpfulness of a review for a business administrator or POI owner. A recent survey shows that all current works on review quality prediction identify the relevance of a review for clients (users) [Maroun et al, 2016]. However, the relevance of a review to an establishment differs from the relevance to the customer because now the goal is not to help decide which product to buy or which place to visit.…”

Section: Our Problem Definitionmentioning

confidence: 99%

“…Regarding each POI review, several aspects may affect the users while writing them, including: noise level, quality of products or services, weather, season and existing expectations. In this context, identifying and managing these factors can provide customers and owners with valuable information through the interpretation of large amounts of reviews [Maroun et al, 2016]. Nonetheless, how to manage is different for each role.…”

Section: Chapter 1 Introductionmentioning

confidence: 99%

Review Recommendation for Points of Interest's Owners

Prado

Moro

2017

Proceedings of the 28th ACM Conference on Hypertext and Social Media

Self Cite

View full text Add to dashboard Cite

Online reviews have become a powerful way for users to make their opinions available to everyone. Such reviews are extremely valuable for consumers when they are looking for information before acquiring a product or service. In an attempt to help consumers identify helpful reviews, many sites allow users to vote if a review is useful. Then, recommendation algorithms come to the rescue in matching reviews to the consumers who are reading them. Such online review applications usually recommend the most useful reviews for consumers to read. Regarding reviews of points of interest, for the establishments owner (or administrator, manager, etc), it is important to have a fast and reliable way to identify the reviews with relevant information for improving the services provided, once they deal with potentially big volumes of data (many users writing many reviews for many items). In this work, we introduce a new problem: identifying the helpfulness of a review for the owner of an establishment. Therefore, we propose creating a ranking of reviews according to their relevance for decision making, i.e. targeting the owners and not clients. The proposed ranking considers aspects described and sentiments present in the reviews. Finally, our experimental evaluation considers a ground truth (constructed by experts opinion) and a baseline (considering the similarity between review and its respective answer provided by the establishment owner) also proposed in this dissertation and show that our solution is very close to the ideal ranking.

show abstract

“…O problema de automaticamente identificar a relevância de reviews sobre POIs para o dono de um estabelecimento é muito importante para que se possa entender melhor a dinâmica de suas ac ¸ões de negócio. Embora existam vários trabalhos que tentam prever a utilidade de uma revisão para um usuário (cliente) e outros que abordam os aspectos e sentimentos atrelados a uma revisão [Maroun et al 2016], até onde sabemos, nenhum abordou o problema de criar um ranking com as avaliac ¸ões mais relevantes para o dono em vez do cliente. A relevância de uma review para o estabelecimento difere da relevância para o usuário (cliente) pois agora o objetivo não é ajudar a decidir entre qual produto comprar ou qual lugar visitar.…”

Section: Soluc ¸ãO Propostaunclassified

“…Focando na perspectiva de reviews realizadas, por ser um servic ¸o baseado em consumidores, há vários fatores que podem afetar o sentimento desses, por exemplo, nível de ruído, produtos de baixa qualidade, clima, estac ¸ão do ano e até mesmo as suas expectativas. Nesse contexto, identificar e gerenciar esses fatores pode fornecer aos clientes e proprietários informac ¸ões valiosas através da interpretac ¸ão de grandes quantidades de dados [Maroun et al 2016]. Para o proprietário do estabelecimento, o mais importante é ter uma forma rápida e confiável para identificar as reviews que contêm informac ¸ões relevantes para a melhoria dos servic ¸os prestados, enquanto que para o cliente, o ideal é identificar reviews que possuam informac ¸ões com detalhes sobre o lugar que o ajudem a tomar uma decisão.…”

Section: Introduc ¸ãOunclassified

Recomendação de reviews personalizada para donos de estabelecimentos

Prado

Moro

2016

Anais Do XXXI Simpósio Brasileiro De Banco De Dados (SBBD 2016)

View full text Add to dashboard Cite

Aplicativos online de avaliação geralmente recomendam as revisões (reviews) mais úteis para os usuários leitores de avaliações. Aqui, introduzimos um novo problema: avaliar a utilidade de uma revisão para o dono de um estabelecimento. Especificamente, propomos o uso de aspectos e sentimentos das revisões, e a geração de um ranking ordenado a partir das mais úteis para o gerenciamento e desenvolvimento do estabelecimento.

show abstract

Assessing Review Recommendation Techniques under a Ranking Perspective

Cited by 5 publications

References 30 publications

Exploiting Review Neighbors for Contextualized Helpfulness Prediction

Exploiting Review Neighbors for Contextualized Helpfulness Prediction

Review Recommendation for Points of Interest's Owners

Recomendação de reviews personalizada para donos de estabelecimentos

Contact Info

Product

Resources

About