基于图嵌入和CaGBDT的多模态出行推荐

iming at the problems of single recommended methods in transportation services and easy to ignore user travel preferences, a cascade gradient boosting decision tree is proposed for reference to the multi-grained cascade forest structure. The model uses the cascade structure to increase the depth of the model and realize the deep representation learning of features. At the same time, in order to solve the problem of sample category imbalance, an index optimization layer is added to the model, and the weight of the model\'s prediction results is corrected by searching for a threshold for each category. In addition, the CaGBDT model can construct a user travel global relationship graph based on the user\'s travel record, and use the graph embedding method to automatically extract the spatial context relationship of the user\'s travel, which improves the efficiency of feature extraction. The experimental results show that the proposed model can better mine user preferences in multi-modal travel recommendation, and the accuracy and stability of recommendation have better performance.