To solve the problem of sparseness in collaborative filtering algorithms, Zuech proposes a way of thinking that the clustering is based on the attributes of the project and uses the mean of the project categories to fill in the null values in the original scoring data