Research On Large-scale Group Decision-Making Methods For Disaster Response Based On Social Media Data Mining

Once natural disasters or other emergencies occur,not only huge losses of people’s lives and properties will be caused,but also social unrest of secondary disasters will emerge.Rapid and effective emergency response and management can effectively minimize incident losses.Therefore,emergency decision-making has become an important research topic.Restricted by technology,cost and time,traditional emergency decision-making can only be completed by a small number of representative experts,which is the small group decision-making.In recent years,with the rapid development of information technology,especially media platforms,the space,time and technology for decision-making have expanded,and the cost of collecting decision-making information and preferences has decreased.Emergency decision-making can be made on the basis of directly considering ordinary individuals’ preferences,that is,large-scale group emergency decision-making methods.Social media data is the information that the public spontaneously publishes on social media platforms in real time.It is a typical type of big data and is the main source of public preferences.The use of social media data in emergency decision-making can have a positive impact on the quality of decision-making,ensuring better democratic,scientific and professional decision-making.Therefore,this paper studies the large-scale group emergency decision-making method based on social media data mining.A large-scale group emergency decision-making framework based on social media data mining is built.Under this framework,decisionmaking information such as decision-making agents,decision-making criteria and weights,decision-making preferences,decision-maker weights,and decision-making quality assessments are mined from social media data,and the maximization of player satisfaction and the minimization of largegroup conflicts are proposed.Several large-scale group decision-making methods are proposed to improve the quality,efficiency and democracy of emergency decision-making.The main contents are as follows.1.Aiming at the problem of low efficiency caused by too many decision-makers in the process of large-scale group decision-making for disaster response,a dimensionality reduction method for decision-makers based on TF-IDF and information loss entropy is proposed.On the basis of dimensionality reduction for decision makers,a two-stage large-scale group decision-making method based on emergency decision-making agent mining is proposed.The result of the case application and comparative analysis shows that the group consensus reach time of the proposed method will not increase with the increasing of decision-makers.2.Aiming at the problem of insufficient public participation in largescale group decision-making for disaster response,resulting in weak decision-making democracy,a method for extracting public decisionmaking preferences based on LDA topic model and sentiment analysis is proposed.On this basis,a large-scale group decision-making method for disaster response based on public emergency decision-making preferences is proposed.The case application and comparative analysis show that it is more refined and comprehensive in objectively depicting public preference information compared with using intuitionistic fuzzy numbers to express preferences.3.Aiming at the problem that the high time pressure of large-scale group decision-making for disaster response will affect the quality of decision-making,a two-stage large-scale group decision-making for disaster response based on the excavation of decision-making criteria,which minimizes group conflicts and maximizes the satisfaction of players.The case application and comparative analysis show that the proposed method saves decision-making time than the large-group decision-making method with feedback mechanism,and has higher decision satisfaction than the large-group decision-making method without feedback mechanism.4.Aiming at the problem that the objective weight determination of decision makers is less considered in the emergency decision-making process of large groups,a multi-stage large-scale group emergency decision-making method based on decision maker weight mining is proposed.Different from the previous method of determining the decision makers’ weights based on the preference information,the proposed method dynamically adjusts the objective weights of decision makers according to the historical decision quality of decision makers from a statistical point of view.The decision-makers’ historical decision quality is comprehensive judged based on emotional changes contained in social media data.The results of case application and comparative analysis shows that the information quality in social media data greatly affects the accuracy rate of decision makers’ weights.When the accuracy rate of decision quality assessment based on social media data reaches a certain level,the accuracy rate of decision makers’ objective weights better than other methods.