New online recommendation approach based on unbalanced linguistic label with integrated cloud

Abstract

Purpose Online reviews increasingly present the characteristic of bidirectional communication with the advent of Web 2.0 era and tend to be asymmetrical and individualized in linguistic information. The authors aim to develop a new linguistic conversion model that exploits the asymmetric and personalized information from online reviews to express such linguistic information. A new online recommendation approach is provided. Design/methodology/approach The necessity of new linguistic conversation model is elucidated, and a leverage factor is incorporated into the linguistic label of negative review to handle the asymmetry problems of linguistic scale. A possible value range of the leverage factor is studied. A new linguistic conversation model is accordingly established with an unbalanced linguistic label and a cloud model. The authors develop a new online recommendation approach based on several modules, such as initialization, conversion, user-clustering and recommendation models. Findings The unbalanced effect between negative and positive reviews is verified with real data and measured using indirect methods. A new online recommendation approach of electronic products is proposed and used as an illustrative example to prove the practicality, effectiveness and feasibility of the proposed approach. Research limitations/implications Due to the unavailable transaction information of customers, the limitation of this study is the effectiveness of the authors’ established recommendation system for platform or website cannot be verified. Originality/value In most existing studies, the influence of negative review is counterbalanced by positive review, and the unbalanced effect between negative and positive reviews is ignored. The negative review receives much attention from consumers and businesses. This study thus highlights the influence of negative review.