Research On Mining Potential Upgrade Complaint Customers Of Operators Based On Text Analysis

Since the 21st century,the proportion of the service economy in the national economy has been increasing,reflecting that people’s demand for services has increased,and the huge demand has brought about the problem of substandard service quality.There are more or less complaints about customer dissatisfaction in all walks of life,and the operator industry in the service industry has very typical complaints,and the number of complaints has always been high.At the same time,the emergence of the number portability policy has made customers’complaints to no avail.and then lost to other operators.The research on complaints has also increased year by year in recent years.Most of the research is on the resolution of specific complaints,as well as the indicator prediction system established based on customer historical behavior data(when the indicator is higher than the threshold,it is considered that customers will complain),escalating complaints among complaints The proportion of research is small:including the characteristics of escalated complaints,the causes of escalated complaints,and how ordinary complaints can be transformed into escalated complaints.To enrich the research field,this study uses Weibo text and black cat text to predict potential escalation complaints from the text level.The research uses sentiment analysis and topic analysis to analyze the characteristics of ordinary complaints and escalated complaints,and then uses a predictive model to predict the texts of ordinary complaints that may escalate complaints.The model evaluation effect is better.The study finally predicted 30 common complaint texts that might escalate complaints,and the prediction results also matched the thematic and emotional characteristics of escalated complaints.It also provides operators with some advice on complaint management.Based on text analysis,operators’ potential escalation complaint customer mining research can help operators obtain customers who may make escalation complaints,and then save relevant customers,which will greatly have less escalation complaints.First,it can help enterprises to improve their complaint management system;second,fewer complaints complaints can give enterprises a better service image.At the same time,the method of using text analysis is proposed in the study,which can play an enlightening role for operators,and operators can learn from relevant methods to analyze customer complaint data.