Product recommendation incorporating the consideration of product performance and customer service factors

Abstract

Purpose The purpose of this paper is to present a novel framework for recommending desirable products to active customers with the consideration of not only their preferences but also the products’ quality performances and their e-retailers’ service performances under e-commerce. Design/methodology/approach A framework in support of the product recommendation is presented. Three modules are involved in the framework, i.e. data collection and preference analysis module, hybrid recommendation module and recommendation generation module. First, preferences of different types of customers are inferred through analysis of their behavioral data and then a paradigm is adopted based on cased-based reasoning to generate candidate recommendation products. Further, recommendation lists for different types of customers are obtained through measurement of the quality and service performances concerning each candidate recommendation product. Findings To illustrate the performance of the presented framework, a simulation study comparing the approach developed based on the framework and the traditional approach is conducted. The experiment results show that the developed approach outperformed the traditional approach in term of average rank score. This means that incorporating the consideration of product performance and customer service factors can play an important role in product recommendations. Originality/value The presented framework can overcome the defect that low conversion rate of recommended products to actually purchased ones suffered by the traditional approach. In addition, the use of the presented framework can not only help customers to obtain desirable products and save searching time but also supervise and urge e-retailers to pay more attention to the quality and service performances.

Advertisement recommendation based on personal interests and ad push fairness

“…Recommendation systems involve a filtering technology based on the users' preferences or interests which filters off the information the user does not need (Alyari and Jafari Navimipour, 2018). There are several domains which apply recommender systems, such as movies (Resnick et al, 1994), knowledge recommendation for workers (Li et al, 2015), destinations and tour services (Yuan and Yang, 2017), documents and books (Liu et al, 2012;Mooney and Roy, 2000), alliance partners (Yuan et al, 2015), colleague recommendations (Hazratzadeh and Navimipour, 2016) and products (Li et al, 2017;Liu et al, 2018). In general, there are some commonly used recommendation methods, including collaborative filtering (CF) and content-based filtering (CBF).…”

Integrating collaborative topic modeling and diversity for movie recommendations during news browsing

“…Recommender systems adopt filtering techniques to solve the problem of information overload for users by analyzing their historical preferences or interests and exploring items they may like (Resnick and Varian, 1997). Nowadays, recommender systems have widely been applied in several domains, such as products (Huang et al, 2019;Li et al, 2017;Liu et al, 2018), music (Patel and Wadhvani, 2018), advertisement (Liu et al, 2019), tasks or knowledge (Li et al, 2015;Zhang and Su, 2018), and tour services (Yuan and Yang, 2017). Recommender systems can be broadly divided into three categories based on how recommendations are provided (Alyari and Jafari Navimipour, 2018).…”

Product recommendation using online reviews with emotional preferences

“…There are several methods that have been developed for facilitating the making of online purchase decisions based on the information available (Liu and Li, 2016;Kritikos, 2017;Li et al, 2017;Yue and Li, 2017;Huang et al, 2019;Mao et al, 2019). Liu and Li (2016), for example, propose a fuzzy comprehensive evaluation method for helping consumers choose mobile phones.…”

“…There are several methods that have been developed for facilitating the making of online purchase decisions based on the information available (Liu and Li, 2016; Kritikos, 2017; Li et al. , 2017; Yue and Li, 2017; Huang et al.…”