The actual relationship between users cannot be captured by the existing data imputation methods because they only consider the rating information and traditional similarity. To alleviate data sparsity and improve recommendation accuracy, a data imputation method was proposed. Firstly, the defects of traditional similarity were analyzed and a new session-based temporal similarity based on tempoaral similarity and dissimilarity was defined, which integrated time context into rating patterns to better identify neighbors for active user. Additionally, the rating sub-matrix of key item set was extracted from similar users and their consumption items which can mine the potential influence factors of users and items, and it was imputed by using matrix factorization. Then the user probabilistic topic distribution for each stage was obtained by using Latent Dirichlet Allocation (LDA) and the user dynamic profile was built with the temporal penalty weights. Finally, the items were recommended based on the correlation of probabilistic topic distribution between users and user-based collaborative filtering. Experimental results show that compared with other imputation-based methods, the proposed method can reduce the Mean Absolute Error (MAE) and improve the recommendation performance under different sparsity.

In session-based Collaborative Filtering (CF), a user interaction history is divided into sessions using fixed time window and user preference is expressed by sequences of them.But in tourism data, there is no interaction in some sessions and it is difficult to select neighbors because of high sparsity. To alleviate data sparsity and better use the characteristics of the tourism data, a new tourism route recommendation method based on dynamic clustering was proposed. Firstly, the different characteristics of tourism data and other standard data were analyzed. Secondly, a user interaction history was divided into sessions by variable time window using dynamic clustering and user preference model was built by combining probabilistic topic distribution obtained by Latent Dirichlet Allocation (LDA) from each session and time penalty weights. Then, the set of neighbors and candidate routes were obtained through the feature vector of users, which reflected the characteristics of tourist age, route season and price. Finally, routes were recommended according to the relevance of probabilistic topic distribution between candidate routes and tourists. It not only alleviates data sparsity by using variable time window, but also generates the optimal number of time windows which is automatically obtained from data. User feature vector was used instead of similarity of tourism data to select neighbors, so as to the avoid the computational difficulty caused by data sparsity. The experimental results on real tourism data indicate that the proposed method not only adapts to the characteristics of tourism data, but also improves the recommendation accuracy.