Who do trust? Combining Recommender Systems and Social Networking for Better Advice

Faced with overwhelming choice people seek advice from their peers or other trusted sources. Collaborative filter recommender systems aim to emulate this process by filtering all the options according to the user tastes expressed through prior item evaluations. Until now the recommender systems literature predominantly focused on improving the algorithms for making suitable predictions for unrated items, while usability research mainly concentrated on interface issues with existing applications. This approach ignored the social elements of decisionmaking and advice seeking. The research here aims to consider a broader range of factors that motivate people in their decision making in order to improve recommender systems. Qualitative research conducted to date has shown that the relationship between recommender and recommendee has a significant impact on decision-making. Thus it is proposed that the impact of social elements on the quality of recommendations needs to be considered for the design of effective recommender systems, which can be addressed through the integration of social networking.