A Study On The Popularity Of Hotel Reservation Software Based On LDA And Random Forest

With the development and popularization of intelligent terminals,people's life has gradually changed from offline to online.More and more people will choose to book online when they go out to stay,so it is very important to choose a convenient and practical hotel booking app.The purpose of this paper is to analyze the evalu-ation indicators of hotel booking app,so as to get the user's concerns and play the role of recommending hotel booking app.This paper analyzes the data of hotel reservation app platform and reviews.Firstly,text mining is used to mine the potential information of user comments.Secondly,a classification model is built to identify the popularity of the software.Finally,ten software platforms are evaluated to provide users with reference.The specific research methods are as follows:firstly,through text mining of user com-ment data and sentiment analysis of the data after word segmentation,the user comment data is divided into positive evaluation set and negative evaluation set,and then LDA topic model fitting is carried out to mine the potential topics of user comments,and analyze the concerns of users on the hotel reservation platform and the advantages and disadvantages of the platform;secondly,feature selection is carried out on the software platform data.In this paper,four characteristics of user reviews,software size,number of iterations in recent year and number of key-word coverage are selected to establish three classification models:Random Forest,K-Nearest Neighbor and Support Vector Machine.The Random Forest classifica-tion model is selected as the optimal classification model through model evaluation;thirdly,objective evaluation of ten hotel reservation platform data and establishmen-t of objective evaluation system.The standard deviation method,entropy weight method and critical method are used to calculate the comprehensive weight,and the comprehensive evaluation scores of ten platforms are obtained,and then the ten platforms are ranked.Through the analysis of the full text,the following conclusions are drawn.First-ly,most users have a positive emotion towards the hotel booking app.The advantages of hotel reservation software lie in hotel hygiene,hotel service and low price.The problems of software experience,customer service and need to be improved.Sec-ondly,the prediction accuracy of the random forest classification model has reached 92.15%,which can well precognize the popularity of hotel booking software and ho-tel check-in.Thirdly,through objective evaluation,the top three best evaluations of hotel booking apps are Huazhu Club,Qunar Travel,and Elong Travel.The entire analysis process can provide developers with suggestions for improving the software,and provide users with a reference for the choice of booking hotel apps.