Incorporating Domain Knowledge into Text Classification Diagnosis in Customer Service Dialogue Field

Abstract The customer service dialogue process is an important way for consumers to communicate with manufacturers. In order to enhance the consumer experience as well as to assist the staff, we build a knowledge base that can categorize consumer questions and provide suitable answers. However, due to labeling deviations, there are some errors in the knowledge base. So we propose a domain knowledge-based text classification diagnosis method, which innovatively transforms the question and answer task into the text classification task. We use an ERNIE-based structure to match consumer questions with multivariate groups of answers from the knowledge base, judged by similarity. Also for incorrectly matched pairs, our method provides a list of suitable candidates for selection. Compared with other baselines, our model achieves competitive results. At the same time, good results are obtained on cross-province data, proving that our method has good scalability.