Algorithm Research Based on Labels of Scenic Spots

Abstract

In collaborative filtering recommendation algorithm based on user's social relations, sometimes the ratings of target user for the target items can't be predicted. What's more, in traditional item-based collaborative filtering, user ratings for different types of items are not comparable. To handle this problem, two new algorithms of collaborative filtering recommendation were proposed, in which the labels of scenic spot's type were introduced to compute the similarity between two scenic spots. The experimental results on the data set of scenic spot's ratings show that, the accuracy and coverage of collaborative filtering recommendation algorithm based on user's social relations and item labels are improved by 10% and 4% respectively compared with the collaborative filtering recommendation algorithm based on user's social relations, the accuracy of collaborative filtering recommendation algorithm based on items and item labels are improved by 15% compared with the collaborative filtering recommendation algorithm based on items, this indicates that introducing the labels of scenic spot's type can make the computation of the similarity between two scenic spots more accurate.