Research On Fuzzy Statistical Method And Application Based On Tourism Attraction Review Data

With the rapid development of the tourism industry and technology,tourists pay much attention to the travel experience and are strongly inclined to adopt the opinions on the travel website to determine their travel plans,hoping to have a satisfactory itinerary.However,it is difficult to make a decision only manually from the plethora of reviews of many attractions.This thesis will propose a multi-attribute decision making method to deal with the problem of tourist attractions selection based on fuzzy statistical theory.Firstly,this thesis expands and improves the information entropy and divergence measures under the background of Pythagorean fuzzy sets.It makes up for some deficiencies of existing measures and can describe the characteristics of fuzzy sets more reasonably and effectively.In addition,The VIKOR approach is improved based on the proposed measure,which includes attribute weight determination and distance calculation.By using R software to simulate the decision-making process,this study proves the effectiveness of the proposed VIKOR method after comparative analysis with the results of some representative methods,and adopts sensitivity analysis by changing the decision-making mechanism coefficients to test stability of this method.Secondly,this thesis takes four scenic spots of famous mountain peaks in our country as the research objects.After using Octopus and Python software to crawl the review data on Qunar,Ctrip and Tongcheng travel platforms,this study processes the review text by R software and constructs a scenic spot index system based on the LDA model and high-frequency phrases analysis.After sentiment analysis for the reviews under different criteria,the review information is converted into Pythagorean fuzzy numbers,which provides data support for the subsequent decision-making method.Finally,this thesis adopts two methods to construct weights.One is to construct objective weights using the proposed entropy and divergence measures,and the other is to use AHP method to determine subjective weights with review information.Then,the two weights respectively combining with the VIKOR approach are applied to solve the problem of tourist attractions selection,so as to provide tourists with objective or personalized scenic spot selection.To sum up,this thesis extends and improves Pythagorean fuzzy entropy,divergence measures and VIKOR method.Combined with the text mining technology,the decision making process of tourist attractions selection based on objective and subjective attribute weights is discussed respectively.It can provide new options when people are hesitant to travel or choose things,which greatly improves the decision making efficiency.