K-means clustering algorithm is employed in this paper to group preprocessed user behavior data and identify appropriate clustering centers. To enhance the performance and precision of the clustering findings, a clustering algorithm that uses chaotic ant colony optimization approach (CAS-C) is proposed. The association rule algorithm is applied to discover the association rules among various clusters to study the correlation between user behavior and attraction interests. Subsequently, the pattern recommendation algorithm is applied to compute the association weight of every attraction and suggest available attractions to target users. Six classes are used to cluster the users and the lowest support and the lowest confidence of mined association rules are greater than 65%. The calculation results of core attractions and routes with user interest show that the best recommendation effect is obtained at the point where the number of single recommendations and is 30.
