Using Dependency Bigrams and Discourse Connectives for Predicting the Helpfulness of Online Reviews

Mathias Mertz, Nikolaos Korfiatis, Roberto V Zicari

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

Helpfulness prediction of online consumer reviews is an interesting research topic with immediate practical applications both from a data mining and marketing perspective. As such a set of studies have been published in the last few years to tackle this problem, targeting the reviews' textual characteristics. In this paper, we propose and evaluate two text-based features that have not been used in the context of consumer review helpfulness prediction before. The first considers a variation of the bigram feature, utilizing grammatical dependencies instead of word adjacency. The second captures the type and amount of discourse in a text by looking for discourse connectives. In our experiments, we treat the helpfulness prediction problem as a binary classification task. The results show that both features contain valuable information for evaluating review helpfulness, however they should be used with caution due to the restrictive experimental setup. The study serves as a ground for future work regarding the usefulness of the proposed features in that perspective.
Original languageEnglish
Title of host publicationE-Commerce and Web Technologies
Subtitle of host publication15th International Conference, EC-Web 2014, Munich, Germany, September 1-4, 2014. Proceedings
PublisherSpringer
Pages146-152
Number of pages6
Volume188
ISBN (Electronic)978-3-319-10491-1
ISBN (Print)978-3-319-10490-4
DOIs
Publication statusPublished - 2014

Cite this