Model for Improving Intelligent Interaction of Guide System Based on Deep Learning

This paper proposes an Attentive-GraphNet algorithm that integrates attention mechanism and graph neural network, which is committed to improving the intelligent interaction performance of the guide system. This algorithm takes a unique approach. It uses the attention mechanism to accurately highlight key information and effectively handles complex relationships with the help of graph neural network for complex spatial topological relationships and diverse user behavior data in the guide system. After pre-processing operations such as cleaning and encoding of spatial maps and user trajectory data collected from real scenes such as shopping malls and airports, a multi-layer graph neural network model is constructed, and a reinforcement learning strategy is adopted to train with user interaction satisfaction as the reward function. Experimental data show that in terms of accuracy, the Attentive-GraphNet algorithm reaches 95%, which is much higher than the 70% of the traditional rule-based model and the 80% of the ordinary neural network model; in terms of response time, the algorithm is only 0.5 seconds on average, while the traditional model is 1.2 seconds and the ordinary neural network model is 0.8 seconds. User satisfaction surveys also show that the algorithm scores significantly higher than other models. The innovative integration of Attentive-GraphNet algorithm effectively improves the processing capability of complex data of the guidance system and shows excellent advantages in intelligent interaction performance.