PBLSTM-TRS: Personalized Bi-directional LSTM Travel Recommendation System

This paper introduces the Personalized Bidirectional LSTM Travel Recommendation System, a model designed to improve personalized travel route recommendations by leveraging tourist behavior’s temporal and spatial dynamics. Utilizing a bi-directional LSTM architecture and a temporal attention mechanism, the model effectively captures complex dependencies in travel data. Tested on a dataset of over 79,000 geotagged photos from tourists in Shanghai, PBLSTM-TRS outperforms baseline models, achieving an F1-score of 0.85. The model’s ability to maintain sequential visit order and adapt to varying trajectory lengths makes it a powerful tool for real-time, context-aware travel recommendations, setting a new standard in the field.