The existing recommendation systems typically employ explicit feedbacks

such as ratings and reviews

to achieve personalized recommendations. However

since the explicit feedbacks are hard to obtain in practice or have poor quality

their applications in the recommendation field are limited. In contrast

implicit feedbacks such as click behaviors and browsing histories are widely available in reality. A game recommendation method based on implicit feedbacks is presented

to provide the personalized game recommendation for game players. This method integrates the time-based implicit feedbacks

such as the times and duration of user operations

and constructs the player's preference model based on the pseudo rating. Meanwhile

it gives the personalized game recommendation by the improved SVD + + (Singular Value Decomposition + +) algorithm. The extensive results of the experiment based on the real data set show that the proposed method has the higher precision and recall.