Robust consumer preference analysis with a social network

The popularity of social media makes it possible for online consumers to seek decision-making support for product selections from their social networks. Users (platforms, manufacturers, etc.) can employ social networks in turn to identify products that consumers prefer, which is important for users to launch marketing strategies such as market segmentation and advertisements. However, there is a challenge for users with regard to knowing consumer preferences about products in a social network environment. To address this issue, we establish a robust consumer preference analysis that includes social network information. First, based on a social network analysis, we estimate a target consumer's missing preference, which is represented by pairwise comparisons between candidate products. Second, we utilize a consensus reaching process to obtain the bounds of the consumer's preferences. Finally, we apply robust optimization to obtain the priority weights of products such that the consumer's preferences regarding these products can be shown. As a tool for analyzing consumer preferences, the robust optimization method only requires the lower and upper bounds of consumer preferences, and it is robust to errors with respect to the preferences. For illustration purposes, we apply this method to analyze consumer preferences based on a rating dataset called filmtrust.