Research on the Improved Graph Convolutional Neural Network Algorithm Based on Historical Data Analysis in the Task of Recommending Attractions on Online Travel Platforms

In the study of scenic spot recommendation task in online tourism platform, we propose an improved graph convolutional neural network algorithm, which improves the ability of historical data analysis by applying graph attention network (GAT), heterogeneous graph attention network (HAN) and relationship aggregation modules to traditional graph convolutional neural networks. First, we introduce the graph attention network to improve the model's ability to capture the features of scenic spots by adaptively learning the importance of different neighbors. Secondly, we adopt a heterogeneous graph attention network to better handle heterogeneous information in online travel platforms. In addition, we designed a relationship aggregation module to capture complex relationships between attractions, thereby improving the predictive performance of the model. By conducting experiments on three real world online travel platforms, we demonstrate the superiority of the proposed improved graph convolutional neural network algorithm in the task of scenic spot recommendation. The experimental results show that our method achieves significant performance improvement compared with traditional graph convolutional neural networks and methods based on deep learning and reinforcement learning. Our research provides an effective solution for the task of attraction recommendation on online travel platforms, which helps to improve user experience and satisfaction.