Learning to Rank Items of Minimal Reviews Using Weak Supervision

Shaalan, Yassien; Zhang, Xiuzhen; Chan, Jeffrey

doi:10.1007/978-3-319-93034-3_50

Cited by 1 publication

(1 citation statement)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition to semantic information, scholars have constructed statistical indicators to measure the helpfulness of reviews. Singh(2017) [11] and Shaalan(2018) [12] built a review ranking model by using information entropy and score distribution. The above studies effectively constructed evaluation indexes of the helpfulness of reviews from semantic and statistical aspects.…”

Section: Introductionmentioning

confidence: 99%

Method for Ranking the Helpfulness of Online Reviews Based on SO-ILES TODIM

et al. 2021

View full text Add to dashboard Cite

Online shopping has become a habit for consumers, who often make purchase decisions based on online reviews. However, the gradual accumulation of reviews has caused an issue associated with information redundancy. Therefore, recommending helpful reviews for consumers has become an urgent problem. Current research on the helpfulness of reviews is mostly at the level of analyzing influencing factors, and few studies have focused on the problem of ranking the helpfulness of reviews. Taking film reviews as the research object, we proposed a SO-ILES TODIM method (a TODIM method based on the intuitive language evaluation set of emotional and ontological features). This method takes into account both semantic indicators (emotional factors and ontological features) and statistical indicators (review length), considers comprehensive information in the review text and has better domain adaptability. First, an intuitive language evaluation set that considers emotional and ontological features was constructed based on statistical rules. Second, a quantitative calculation method that includes an index weight value based on the logit regression model was designed, and it can effectively avoid the subjectivity of the manual assignment method. Finally, based on the degree of membership deviation, the score function and the exact function were designed to realize a ranking of the helpfulness of reviews. Through a case simulation, we show that this method can prioritize reviews that directly evaluate the product. Through a comparative analysis and parameter sensitivity analysis, the stability and scientificity of the SO-ILES TODIM method was demonstrated. This paper broadens the research scope of reviews, enriches the research method of review helpfulness ranking and provides insights for merchants or third-party platforms to manage online reviews.

show abstract