In dealing with big data,the basic method to compute the similarity has many problems such as data sparsity and the lack of obvious recommend effects,a research for the advantage of the basic similarity calculation was done,which integrated timing model and trust model,then used HGA (hierarchical genetic algorithm)and PSO (particle swarm optimization)for hy-brid optimization.After that,based on the nearest neighbor users,comparative experiments were carried out using the algorithm in MovieLens date sets,results show that comparing with pure similarity calculation,the algorithm adopting PSO com...