Travel Intention Prediction Method Based on User Needs

Lingyu Zhang, Wenjie Bian, Jianxing Li, Ying Zhang, Jian Liang, Guobin Wu, Liang Zhang

Abstract: Taxi travel is playing an increasingly important role in daily travel. Accurately predicting the user's travel intention can not only cut off the time spending on inputting a place, but also can recommend travel and entertainment venues for the user according to his/her travel intention, thereby improving the user experience. In this paper, we first marked the label of travel intention on the historical taxi orders of 100 users in a city. By analyzing the user's intention to take a taxi, we found that the user's travel intention can be influenced by various factors. In addition, we propose an ensemble learning method based on Embedding DNN working on user travel data characteristics to predict the user's travel intention. When users have recommendation needs and authorize prediction, we provide travel intention prediction and make subsequent destination recommendation according to the intention. By taking comparing experiments on real-world data sets, the experimental results show that our model is better than other models in both accuracy and stability.