基于用户属性聚类与矩阵填充的景点推荐算法

Tourist Spot Recommendation Algorithm Based on User Attribute Clustering and Matrix Filling

随着互联网和旅游业的发展,可以选择的旅游景点越来越多。在海量的景点信息中,景点的选择成为旅客出行的一个重要问题。该文采用改进的协同过滤算法,给每个旅客推荐合适的旅游景点,以解决他们出行难的问题。首先对传统的协同过滤算法进行改进,即对用户属性进行二分聚类;再利用奇异值分解算法填充稀疏的用户评分矩阵,得到多个聚类类别的中心和一个填充完整用户评分矩阵;然后计算出目标用户各属性到各个聚类中心的欧氏距离,将其分到距离最小的类别;再利用Pearson相似度方法和填充完整的用户评分矩阵计算出目标用户与同一类别中其他用户的相似度;最后结合相似度,用Top-N推荐方法将预测景点评分进行降序排序,并推荐给目标用户,从而提高推荐算法的精准度。实验结果表明,该算法比传统协同过滤算法的推荐质量有显著提高。

With the development of the Internet and tourism,more and more tourist attractions are available.Among the massive information,the choice of tourist attractions has become an important issue for travelers.We adopt an improved collaborative filtering algorithm to recommend suitable tourist attractions for each traveler to solve their travel difficulties.We firstly improve the traditional collaborative filtering algorithm,which is binary clustering of user attributes,and fill the sparse user rating matrix by the SVD algorithm to obtain the centers of multiple clustering categories and a complete user rating matrix.Then we calculate the Euclidean distance from each attribute of the target user to each cluster center which is divided into the category with the smallest distance,and then calculate the similarity between the target user and other users in the same category by using Pearson similarity method and filling the complete user rating matrix.Finally combining similarity,we use the Top-N recommendation method to sort the predicted attraction scores in descending order and recommend them to the target users for improving the accuracy of the recommendation algorithm.Experiment shows that the proposed algorithm has significantly improved the recommendation quality compared with traditional collaborative filtering algorithms.