Research On Quality Factor Classification And Dynamic Prediction Of Kano Model Based On Online Reviews

With the rapid development of social economy,the market competition is becoming increasingly fierce,how to effectively tap the deep-seated needs of customers and carry out targeted optimization iteration of products / services is an important issue for enterprises to improve customer satisfaction and maintain competitive advantage.Kano model describes the nonlinear relationship between customer satisfaction and product and service performance from a two-dimensional perspective.It can effectively realize the fine classification of quality elements,and is widely used in the fields of quality management,product design,service innovation and so on.However,with the deepening of research and application,the limitations of the traditional Kano model begin to show: the quality elements of Kano model are usually summarized through literature or practice,which is more subjective;The questionnaire design of Kano model adopts the form of "positive and negative questions",which is easy to lead to problems such as excessive questions and time-consuming.At the same time,customers often have choice problems,resulting in the phenomenon that the survey results are difficult to reflect the real customer psychology;The final classification result of Kano model is based on the frequency of each classification reflected in the overall sample.Too simple statistical methods will affect the final classification accuracy of quality factors;In addition,the quality elements of Kano model change periodically,but there is a lack of relevant research to predict the dynamic changes of quality elements.Considering the limitations of data acquisition and subjective classification criteria of traditional Kano model,this thesis proposes an online reviews driven Kano model to classify quality elements.The details are as follows :(1)By obtaining the data from users’ online comments,LDA(Latent Dirichlet Allocation,LDA)model is used to analyze users’ online reviews so as to acquire the quality elements;Sentiment analysis of online reviews is carried out based on a SVM(Support Vector Machine,SVM)model to excavate customers’ perception of quality elements deeply.Emotional tendency is integrated to the Kano model for achieving further objective classification results of quality elements;(2)Combined with the residual GM(1,1)model,predict and analyze the classification results of quality factors,realize the dynamic prediction of quality factors based on online comment data,and timely understand the changes of customer demand;(3)Combined with the online medical service platform to carry out empirical research to verify the feasibility and effectiveness of this method.This research helps to improve the precision of quality elements classification and to improve the support decision-making effects of Kano model.Also,it can guide the improvement of online healthcare service quality.