Patterns of influence in a recommendation network

摘要: Information cascades are phenomena in which individuals adopt a new action or idea due to influence by others. As such process spreads through an underlying social network, it can result widespread adoption overall. We consider information the context of recommendations, and particular study patterns cascading recommendations that arise large networks. investigate person-to-person recommendation consisting four million people who made sixteen on half products. Such dataset allows us pose number fundamental questions: What kinds frequently real life? features distinguish them? enumerate count cascade subgraphs directed graphs; as one component this, we develop novel efficient heuristic based graph isomorphism testing scales datasets. discover patterns: distribution sizes is approximately heavy-tailed; tend be shallow, but occasional bursts propagation occur. The relative abundance different suggests subtle properties network process.