Classification of Customer Service Tickets in Power System Based on Character and Word Level Semantic Understanding

Recently, a large amount of electric power data has been accumulated in the power system because of the development of information technology. Customer service ticket data is one of the most important types of electric power data. Classifying customer service tickets according to the text content is conducive to understanding customer demands accurately. So it is of great significance to improve the quality and efficiency of power system operation. In this paper, we propose a classification model of power system customer service tickets based on hierarchical semantic understanding. The deep learning method is used to simultaneously model the words and characters in the tickets. And then the hierarchical semantic representations can be used to classify tickets. The contrast experiments on real dataset prove that the proposed method can accurately learn the hidden semantics of tickets, and has better classification accuracy than the comparative approaches.