Skip to Main Content
As people and their acquaintances are connected with one another to form a social network, social network-based recommendation systems (SNRS) enable people to make decisions based on previous experiences and knowledge of acquaintances. SNRS are implemented to take advantage of the acquaintance relationships, which form the essential links within social networks, to enable people in need of item purchasing recommendations to send requests for recommendations and receive recommendations from people throughout the network. Each person in the social network maintains a pair of trust and distrust ratings for each of their acquaintances and updates the ratings based on usefulness of received from the corresponding acquaintance. A pair of trust and distrust rating is used instead of just a single trust rating where the assumed distrust rating can be derived from the trust rating. In the proposed trust and distrust mechanism, trust represents the perceived risk of choosing a recommendation from the corresponding acquaintance while distrust is a protective measure indicating the level of doubt on the level of trust that is assigned to the same acquaintance. The proposed mechanism corresponds to studies indicating that the trust and distrust are determined separately, with the level of trust being based on a relatively rational evaluation while the level of distrust is based more on fear and other irrational emotions. When the level of distrust rises over a certain threshold for an acquaintance, recommendations received from the acquaintance will not be considered during the decision making process until the level of distrust falls below the threshold. The simulation used for this study utilizes software agents that are connected with one another via a social network to model the behavior of people when soliciting recommendations and providing recommendations.