Personalized recommendations are generated by considering the preferences of a target user and similar users. Although explanations of recommendations affect the evaluations of personalized recommender systems (PRS), PRS evaluations have focused primarily on the perceived accuracy and novelty of the recommending algorithms. The goal of this study is to examine the effectiveness of using social interaction factors (self-referencing and social presence) to explain PRS. We developed six PRS for applications (apps) on smartphones by varying the level of social presence and self-referencing. We conducted Web-based experiments using these six types of PRS, and we then obtained participant evaluations of their social interactions and PRS. Our research model is designed to determine how social interactions, such as social presence and self-referencing, affect perceived accuracy and novelty, and in turn, how these affect satisfaction and intent to purchase. The results obtained demonstrate that the social context significantly increases the perceived accuracy and novelty of PRS. The results explain that perceived accuracy and novelty positively influence user satisfaction, and how satisfaction and perceived novelty affect purchase intention. In addition, we verify the effect of mediation on perceived accuracy, perceived novelty, and satisfaction. Thus, by integrating PRS performance and social interaction, this research contributes to improving our understanding of the social cognitive process related to user evaluation of PRS.
|Number of pages||30|
|Journal||Journal of Electronic Commerce Research|
|Publication status||Published - 2017|
All Science Journal Classification (ASJC) codes
- Economics, Econometrics and Finance(all)
- Computer Science Applications