| The "Opinions of the Central Committee of the Communist Party of China and the State Council on Doing a Good Job in Comprehensively Promoting the Key Work of Rural Revitalization in 2022",namely the No.1 Central Document in 2022,proposes to increase the development of rural tourism,guide and encourage farmers to participate in the operation of rural homestays.The outline of China’s 14 th Five-Year Plan proposes to vigorously develop the homestay economy,rural tourism and other characteristic leisure industries.As a new business form of tourism,homestays can not only promote the overall development of tourism,but also an important starting point for the rural revitalization strategy.Therefore,the state strongly encourages the in-depth development of the homestay industry.Among them,the "Opinions of the General Office of the People’s Government of Jiangxi Province on Promoting the Healthy Development of Homestays" released in August 2020 stated that Jiangxi Province is expected to launch a batch of high-quality,high-quality products with stories,experience,taste and nostalgia by 2023.For quality homestays,the number of employees,business performance,service levels,and reception levels in the province’s homestay industry should double,and gradually form a large-scale homestay industry with diverse formats,first-class services,distinctive features.Therefore,it is very meaningful to explore and study the homestay industry in my country.This thesis takes the homestays in Jiangxi Province as the research object,integrates the review data of three homestay platforms,Ctrip,Qunar and Airbnb,and crawled 56,080 valid homestays in 10 prefecture-level cities with 5A-level scenic spots in Jiangxi Province.Comment.Based on the existing LDA-AHP model for tourist attraction satisfaction research(a method that uses the LDA topic model to extract the word frequency weight value of text subject words to replace the index weight that relies on the subjective scoring of experts in the original AHP analytic hierarchy process),this thesis The machine learning classification algorithm and FCE fuzzy comprehensive evaluation method are introduced to improve the original model.On the basis of predecessors,emotional factors are introduced,and LDA topic model is used to extract topic words,and then NB,SVM,LR and other methods are used for sentiment classification.Comprehensive comparison,the SVM+NB+LR classification method proposed in this thesis has higher accuracy and better effect,and the F1 value is improved by 4.85%.Finally,the word frequency weight value of the LDA topic model and the sentiment value of the machine learning classification algorithm are weighted and normalized to obtain the comprehensive weight of homestay satisfaction in Jiangxi Province.This thesis draws and measures 9 satisfaction indicators of homestays in Jiangxi Province,which are room conditions,service attitude,geographic location,characteristic design,unique food,surrounding scenery,price perception,living experience,and sanitary conditions.The results show that:(1)The overall satisfaction of homestays in Jiangxi Province is 91.67%,with hygiene conditions,service attitude and unique food satisfaction ranking the top three,and room conditions,surrounding scenery and living experience ranking behind.(2)There is a certain gap in the satisfaction level of each city.Shangrao,Yingtan and Jingdezhen ranked the top three,while Fuzhou,Pingxiang and Yichun ranked the bottom.(3)The development level of homestays is closely related to the level of tourism development in local cities.About 68.05% of the homestays are concentrated in Nanchang,Shangrao,Jiujiang and Jingdezhen.Among them,Shangrao has the highest level of homestay development.This thesis combines qualitative and quantitative methods to construct a homestay satisfaction evaluation system in Jiangxi Province based on consumer online texts.The method has certain feasibility,and the research results of this thesis have certain theoretical implications for the service optimization of homestay industry in Jiangxi Province.meaning and reference value. |