A Listwise Collaborative Filtering baesd on Temporal

The memory-based listwise collaborative filtering (ListCF) algorithm only generates the recommendation result by calculating the similarity between users, ignoring the influence of user interest changes over time on the recommendation results. To solve this problem, temporal weighting function defined by logistic function that can reflect the interest changes over time is introduced to user interest, then get the probability distribution of top-k permutation between users, finally calculating the similarity to find neighborhood users. Experiments on Movielens proved that the temporal characteristics can further improve the accuracy of recommendation.