With the emergence of Internet, there is more and more information disseminating all over this channel. The abundant amount of information, however, causes difficulty for users to locate desired information, which is referred to as the information overload problem due to our limited processing ability. Therefore, recommender systems arise to assist users to acquire useful information based on their past preferences or collaborative preferences from other sources. Group recommendation is the task of recommending items for a group of users who participate in a social activity intentionally or randomly. This kind of research is essential in the circumstances when a group of users participate in such activities as watching movies or TV shows with friends, finding restaurants for dinner with colleagues, and picking books to study in a book club. It notably distinguishes itself from personalized recommendation in that collective group behavior needs to be addressed by taking individuals’ behaviors into account. The objective of this research is thus to propose hybrid filtering approaches for group recommendation on documents. Particularly, latent Dirichlet allocation to uncover latent semantic structure in documents is incorporated to serve as a bridge to connect content-based filtering and collaborative filtering as a whole, and generate complementary and additive effects for better performance. Two experiments are conducted accordingly. The results show that our proposed approaches (GCBPF and GSBICF) outperform other traditional group filtering approaches on the recommendation performance, which justifies the feasibility of our proposed approaches in applications.