The Early-Warning and Control of Service Complaint Based on Time Series Forecasting Method and SPC Model-Take Ctrip as an Example

摘要

This paper applies theories and methods of time series model, emotional classification model, statistical process control (SPC) model and natural language processing to complaint management, and gives basic framework, processes and methods of collecting, using, predicting and analyzing the complaint data in complaint management. By verifying the effect of specific forecasting model which combines the emotional classification of external big data, early-warning model and text mining based on LDA model, the result of the experiment, taking Ctrip as an example, shows that it effectively improves the capability of searching and solving the complaints resulting from service failure, optimizing corporate services, and improving the timeliness and effectiveness of industry supervision.