Comparative Analysis and Strategy Research of Enterprises Dealing with Network Public Opinion Based on Text Mining

Abstract

Enterprises need to pay attention to and track changes in online public opinion, and create a good public opinion environment for enterprise development. This study analyzes the successful experiences and deficiencies of enterprises in dealing with network public opinion, which is helpful to provide decision support for enterprises to respond to network public opinion. Two public opinion incidents of “Pinduoduo Employee's Sudden Death” and “Huolala Girl Jumps from the Car” are taken as the research objects. Based on text mining methods, the evolution of network public opinion, semantic network and sentiment trends are compared and analyzed. Through comparative analysis, it is found that Pinduoduo's comments focus on the corporate institutions and labor law, and revolve around multiple important nodes. The Huolala's comments focus on the truth and rectification of the event, and only revolve around two important nodes. The proportion of negative emotions in the early stage of the “Huolala girl jumping from the car” is much higher than that of the “Pinduoduo employee sudden death”. However, Huolala did better than Pinduoduo in terms of coping strategies, which makes the proportion of negative emotions in the later stage lower and comments tend to be more rational. This study put forward corresponding strategies and suggestions for enterprises to effectively respond to the crisis of online public opinion. Firstly, network public opinion early warning mechanism need to be established to control from the source. Secondly, companies have to understand the direction of public opinion and grasp the focus. Thirdly, companies need to strengthen responsibility awareness and have courage to take responsibility. Fourthly, follow-up management system and propose corrective measures should be carried out.