In order to ensure the satisfaction of tourists with the recommended attractions, a text mining based tourist attraction recommendation algorithm is proposed. The topic model is applied to mine the contents, topics, and keywords of tourist attractions. Meanwhile, the LDA algorithm combining time factor is applied to analyze the personalized needs of users and construct the user model. The cosine similarity is calculated according to the models of tourist attractions and users to help users quickly and accurately find the tourist attractions suitable for their needs from the huge amount of tourist information. The test results show that the algorithm of this paper has a smaller average absolute error than the CF algorithm for different attractions with different numbers of near-neighbor users. When the number of nearest neighbors is 10 and 20, the two algorithms are within 0.6 of each other in prediction. It shows that the algorithm in this paper has high accuracy in recommending tourist attractions, which provides technical support for the subsequent optimization of tourism service management countermeasures.
