Online recommendation systems are becoming more and more popular with the development of web. Data sparseness is a major problem for collaborative filtering (CF) techniques in recommender systems, especially for new users and items. In this paper, we try to reduce the data sparseness in the collaborative filtering problem by involving Wikipedia as an auxiliary information source. In this paper, we attempt to improve the recommendation accuracy by extracting collaborative social behavior information embedded in Wikipedia coediting history, and use this knowledge to help alleviate the data-sparsity problem in CF. In addition, we introduce a parallel computing algorithm to help scale up the transfer learning process. Our experimental results on two real world recommendation datasets show that the social co-editing knowledge in Wikipedia can be effectively transferred for CF problems.