Exploring the diverse interests of tourism participants: an analysis based on Topic model and BiLSTM-CRF

The rapid development of social network and new technology has led to great changes in tourism. Natural language processing (NLP) has advantages in obtaining information about the vast amount of content generated by online users about travel services and products. In this paper, Gansu province is taken as a sample to establish a corpus knowledge base combining user generated content (UGC) and literature specialty. And a variety of knowledge aggregation methods, such as unsupervised machine learning and knowledge graph methods, are used to effectively organize the massive data in the corpus. Finally, this paper analyzes destination image, tourist preference and tourist behavior patterns. This method effectively addresses the needs of tourism visitors and organization information retrieval. Make some contributions to destination management and tourist behavior recognition.