SOCIAL RECOMMENDATION MODEL BASED ON RESTART RANDOM WALK ALGORITHM IN GLOBAL VIEW

Due to the non-connectivity and a large number of isolated nodes in real social networks, it is impossible to accurately describe the strength of social relationships between nodes, so that the problems of data sparseness and cold start of social recommendation models cannot be alleviated. In view of this situation, a social recommendation model based on the restart random walk algorithm is constructed in the global view. The model introduced super nodes to construct a directed connected network, used the restart random walk algorithm to characterize the strength of social relations between nodes, and integrated the characterised social relations strength into a social recommendation model based on probability decomposition technology. The experimental results show that compared with the traditional recommendation model, this model can effectively improve the recommendation effect. The experimental results show that compared with traditional social recommendation models, this model can effectively improve recommendation performance.