Recommender systems could be used to help users in their access processes to relevant information. Collaborative filtering is one of the most successful recommendation approaches. It typically associates a user with a group of like-minded users based on their preferences over all the items, and recommends to the user those items enjoyed by others in the group. This paper presents a new similarity measure for collaborative filtering and several schemes. As a whole similarity measures in collaborative filtering apply user ratings data to find similar users to active user. In real world, these user rating matrices are sparse. Since there are a lot of items in a system, so user can't pay attention all of them. To reduce the limitation here, we investigate different ways to fill matrix and describe the combinations of the ways. It helps to increase the quality of proposed new similarity measure.
